Chapter 10: Oncology

1. Modern cancer therapy is multidisciplinary, involving coordinated care by surgeons, medical oncologists, radiation oncologists, reconstructive surgeons, pathologists, radiologists, and primary care physicians.

2. Understanding cancer biology is essential to successfully implement personalized cancer therapy.

3. The following alterations are critical for malignant cancer growth: self-sufficiency of growth signals, insensitivity to growth-inhibitory signals, evasion of apoptosis, potential for limitless replication, angiogenesis, invasion and metastasis. Reprogramming of energy metabolism and evading immune destruction.

As the population ages, oncology is becoming a larger portion of surgical practice. The surgeon often is responsible for the initial diagnosis and management of solid tumors. Knowledge of cancer epidemiology, etiology, staging, and natural history is required for initial patient assessment, as well as to determination of the optimal surgical therapy.

Modern cancer therapy is multidisciplinary, involving the coordinated care of patients by surgeons, medical oncologists, radiation oncologists, reconstructive surgeons, pathologists,radiologists, and primary care physicians. Primary (or definitive) surgical therapy refers to en bloc resection of tumor with adequate margins of normal tissues and regional lymph nodes as necessary. Adjuvant therapy refers to radiation therapy and systemic therapies, including chemotherapy, ­immunotherapy, hormonal therapy, and, increasingly, biologic therapy. The primary goal of surgical and radiation therapy is local and regional control. On the other hand, the primary goal of systemic therapy is systemic control by treatment of distant foci of subclinical disease to prevent distant recurrence. Surgeons must be familiar with adjuvant therapies to coordinate multidisciplinary care and to determine the best sequence of therapy.

Recent advances in molecular biology are revolutionizing medicine. New information is being translated rapidly into clinical use, with the development of new prognostic and predictive markers and new biologic therapies. Increasingly cancer therapy is getting personalized, incorporating information about each patient’s tumor characteristics, patient’s own genome, as well as host immune responses and tumor microenvironment, into clinical decision-making. It is therefore essential that surgeons understand the principles of molecular oncology to appropriately interpret these new contributions and incorporate them into practice.

Basic Principles of Cancer Epidemiology

The term incidence refers to the number of new cases occurring. Incidence is usually expressed as the number of new cases per 100,000 persons per year. Mortality refers to the number of deaths occurring and is expressed as the number of deaths per 100,000 persons per year. Incidence and mortality data are usually available through cancer registries. Mortality data are also available as public records in many countries where deaths are registered as vital statistics, often with the cause of death. In areas where cancer registries do not exist, mortality data are used to extrapolate incidence rates. These numbers are likely to be less accurate than registry data, as the relationship between incidence and cause-specific death is likely to vary significantly among countries owing to the variation in health care delivery.

The incidence of cancer varies by geography. This is due in part to genetic differences and in part to differences in environmental and dietary exposures. Epidemiologic studies that monitor trends in cancer incidence and mortality have tremendously enhanced our understanding of the etiology of cancer. Furthermore, analysis of trends in cancer incidence and mortality allows us to monitor the effects of different preventive and screening measures, as well as the evolution of therapies for specific cancers.

The two types of epidemiologic studies that are conducted most often to investigate the etiology of cancer and the effect of prevention modalities are cohort studies and case-control studies. Cohort studies follow a group of people who initially do not have a disease over time and measure the rate of development of a disease. In cohort studies, a group that is exposed to a certain environmental factor or intervention usually is compared to a group that has not been exposed (e.g., smokers vs. nonsmokers). Case-control studies compare a group of patients affected with a disease to a group of individuals without the disease for a given exposure. The results are expressed in terms of an odds ratio, or relative risk. A relative risk <1 indicates a protective effect of the exposure, whereas a relative risk >1 indicates an increased risk of developing the disease with exposure.

Cancer Incidence and Mortality in the United States

In the year 2013, it is estimated that 1.6 million new cancer cases will be diagnosed in the United States, excluding carcinoma in situ of any site except bladder, and excluding basal cell and squamous cell carcinomas of the skin.¹ In addition, 64,640 cases of carcinoma in situ of the breast, and 61,300 of melanoma in situ are expected.¹

It is estimated that in 2013 estimated 580,350 people will die of cancer in the United States, corresponding to about 1600 deaths per day.¹ The estimated new cancer cases and deaths by cancer type are shown in Table 10-1.¹ The most common causes of cancer death in men are cancers of the lung and bronchus, prostate, and colon and rectum; in women, cancers are of the lung and bronchus, breast, and colon and rectum.¹ These four cancers account for almost half (48%) of total cancer deaths among men and women.

The annual age-adjusted cancer incidence rates among males and females for selected cancer types are shown in Fig. 10-1.¹ Incidence rates are declining for most cancer sites, but they are increasing among both men and women for melanoma of the skin, cancers of the liver and thyroid (Fig. 10-2).¹ Incidence rates are decreasing for all four major cancer sites except for breast cancer in women. Age-adjusted incidence rate of breast cancer started to decrease from 2001 to 2004.² This decrease in breast cancer incidence has at least temporally been associated with the first report of the Women’s Health Initiative, which documented an increased risk of coronary artery disease and breast cancer with the use of hormone replacement therapy; this was followed by a drop in the use of hormone replacement therapy by postmenopausal women in the United States.² Unfortunately after this initial drop, breast cancer incidence has remained relatively stable from 2005 to 2009.

Declines in colorectal cancer incidence have been mainly attributed to increased screening that allows for removal of precancerous polyps. Prostate cancer rates rapidly increased and decreased between 1995 and 1998. These trends are thought to be attributable to increased use of prostate-specific antigen (PSA) screening.³ Although analysis now suggest prostate cancer incidence has declined steadily by 1.9% per year from 2000 to 2009, annual rates fluctuate likely reflecting variations in screening.

Differences in lung cancer incidence patterns between women and men are thought to reflect historical differences in tobacco use. Differences in smoking prevalence is also thought to contribute to regional differences in lung cancer incidence. Lung cancer incidence is fourfold higher in Kentucky which has the highest smoking prevalence, compared with Utah, that has the lowest smoking prevalence (128 vs. 34 lung cancer cases per 100,000 men).¹

The 5-year survival rates for selected cancers are listed in Table 10-2. From 2005 to 2009, cancer death rates decreased by 1.8% per year in males and by 1.5% per year in females.¹ These declines in mortality have been consistent in the past decade, and larger than what was observed in the previous decade.³ Over the past two decades, death rates have decreased from their peak by more than 30% for colorectal cancer, female breast cancer, male lung cancer and more than 40% for prostate cancer. The decrease in lung cancer death rates in men is thought to be due to a decrease in tobacco use, whereas the decreases in death rates from breast, colorectal cancer, and prostate cancer reflect advances in early detection and treatment.

Global Statistics on Cancer Incidence and Mortality

The five most common cancers for men worldwide are lung, prostate, colorectal cancer, stomach, liver, and for women are breast, colorectal, cervix, lung, and stomach.⁴ Notably, for several cancer types there is wide geographical variability in cancer incidence (Fig. 10-3). The mortality rates for different cancers also vary significantly among countries. This is attributable not only to variations in incidence but also to variations in survival after a cancer diagnosis. The survival rates are influenced by treatment patterns as well as by variations in cancer screening practices, which affect the stage of cancer at diagnosis. For example, the 5-year survival rate for stomach cancer is much higher in Japan, where the cancer incidence is high enough to warrant mass screening, which is presumed to lead to earlier diagnosis. In the case of prostate cancer, on the other hand, the mortality rates diverge much less than the incidence rates among countries. Survival rates for prostate cancer are much higher in North America than in developing countries.⁵ It is possible that the extensive screening practices in the United States allow discovery of cancers at an earlier, more curable stage; however, it is also possible that this screening leads to discovery of more latent, less biologically aggressive cancers, which may not have caused death even if they had not been identified.

About one million new cases of stomach cancer were estimated to have occurred in 2008 (988,000 cases, 7.8% of the total), making it the fourth most common malignancy in the world, behind cancers of the lung, breast, and colorectal cancer. The incidence of stomach cancer varies significantly among different regions of the world. The difference in risk by country is presumed to be primarily due to differences in dietary factors. The risk is increased by high consumption of preserved salted foods such as meats and pickles, and decreased by high intake of fruits and vegetables.⁵ There also is some international variation in the incidence of infection with Helicobacter pylori, which is known to play a major role in gastric cancer development.⁵ Fortunately, a steady decline is being observed in the incidence and mortality rates of gastric cancer. This may be related to improvements in preservation and storage of foods as well as due to changes in the prevalence of H. pylori.⁵ More than 70% of cases (713,000 cases) occur in developing countries, and half the cases in the world occur in Eastern Asia (mainly in China).⁴ Age-standardized incidence rates are about twice as high for men as for women, ranging from 3.9 in Northern Africa to 42.4 in Eastern Asia for men, and from 2.2 in Southern Africa to 18.3 in Eastern Asia for women. Stomach cancer is the second leading cause of cancer death in both sexes worldwide.

Overall, the incidence of breast cancer is rising in most countries. Incidence varies from 19.3 per 100,000 women in Eastern Africa to 89.7 per 100,000 women in Western Europe, and are high in developed regions of the world (except Japan) and low in most of the developing regions.⁴ Although breast cancer has been linked to cancer susceptibility genes, mutations in these genes account for only 5% to 10% of breast tumors, which suggests that the wide geographic variations in breast cancer incidence are not due to geographic variations in the prevalence of these genes. Most of the differences, therefore, are attributed to differences in reproductive factors, diet, alcohol, obesity, physical activity, and other environmental differences. Indeed, breast cancer risk increases significantly in females who have migrated from Asia to America.⁵ The range of breast cancer mortality rates is much less (approximately 6 to 19 per 100,000) because of the more favorable survival of breast cancer in developed regions. As a result, breast cancer ranks as the fifth cause of death from cancer overall (458,000 deaths), but it is still the most frequent cause of cancer death in women in both developing (269,000 deaths, 12.7% of total) and developed regions (estimated 189,000 deaths). ⁴

There is a 25-fold variation in colon cancer incidence worldwide.⁵ The incidence of colon and rectal cancer is higher in developed countries than in developing countries. The incidence rates are highest in North America, Australia and New Zealand, and Western Europe, and especially in Japanese men.⁵ In contrast, the incidence is relatively low in North Africa, South America, and eastern, Southeastern, and Western Asia. These geographic differences are thought to reflect environmental exposures and are presumed to be related mainly to dietary differences in consumption of animal fat, meat, and fiber.⁵

Worldwide liver cancer is the fifth most common cancer in men (523,000 cases, 7.9% of the total) and the seventh in women (226,000 cases, 6.5% of the total). Almost 85% of liver cancer cases occur in developing countries, and particularly in men.⁴ The overall sex ratio male:female is 2:4. The regions of high incidence are Eastern and Southeastern Asia, Middle and Western Africa, as well as Melanesia and Micronesia/Polynesia (particularly in men). Low rates are estimated in developed regions, with the exception of Southern Europe. There were an estimated 694,000 deaths from liver cancer in 2008 (477,000 in men, 217,000 in women), and because of its high fatality (overall ratio of mortality to incidence of 0.93), liver cancer is the third most common cause of death from cancer worldwide. The geographical distribution of the mortality rates is similar to that observed for incidence. Worldwide, the major risk factors for liver cancer are infection with hepatitis B and C viruses and consumption of foods contaminated with aflatoxin. Hepatitis B immunization in children has recently been shown to reduce the incidence of liver cancer.⁵

In summary, the incidence rates of many common cancers vary widely by geography. This is due in part to genetic differences, including racial and ethnic differences. It is due also in part to differences in environmental and dietary exposures, factors that can potentially be altered. Therefore, establishment of regional and international databases is critical to improving our understanding of the etiology of cancer and will ultimately assist in the initiation of targeted strategies for global cancer prevention. Furthermore, the monitoring of cancer mortality rates and 5-year cancer-specific survival rates will identify regions where there are inequities of health care, so that access to health care can be facilitated and guidelines for treatment can be established.

Hallmarks of Cancer

Although there are >100 types of cancer, it has been proposed that there are six essential alterations in cell physiology that dictate malignant growth: self-sufficiency of growth signals, insensitivity to growth-inhibitory signals, evasion of apoptosis (programmed cell death), potential for limitless replication, angiogenesis, and invasion and metastasis.⁶ Recently two additional hallmarks have emerged—reprogramming of energy metabolism and evading immune destruction.⁷ These hallmarks of cancer are being pursued as targets for cancer therapy (Figure 10-4).

Cell Proliferation and Transformation

In normal cells, cell growth and proliferation are under strict control. In cancer cells, cells become unresponsive to normal growth controls, which leads to uncontrolled growth and proliferation. Human cells require several genetic changes for neoplastic transformation. Cell type–specific differences also exist for tumorigenic transformation. Abnormally proliferating, transformed cells outgrow normal cells in the culture dish (i.e., in vitro) and commonly display several abnormal characteristics.⁸ These include loss of contact inhibition (i.e., cells continue to proliferate after a confluent monolayer is formed); an altered appearance and poor adherence to other cells or to the substratum; loss of anchorage dependence for growth; immortalization; and gain of tumorigenicity (i.e., the ability to give rise to tumors when injected into an appropriate host).

Cancer Initiation

Tumorigenesis is proposed to have three steps: initiation, promotion, and progression. Initiating events such as gain of function of genes known as oncogenes or loss of function of genes known as tumor-suppressor genes may lead a single cell to acquire a distinct growth advantage. Although tumors usually arise from a single cell or clone, it is thought that sometimes not a single cell but rather a large number of cells in a target organ may have undergone the initiating genetic event. Thus, many normal-appearing cells may have an increased malignant potential. This is referred to as a field effect. The initiating events are usually genetic and occur as deletions of tumor-suppressor genes or amplification or mutation of oncogenes. Subsequent events can lead to accumulations of additional deleterious mutations in the clone.

Cancer is thought to be a disease of clonal progression as tumors arise from a single cell and accumulate mutations that confer on the tumor an increasingly aggressive behavior. Most tumors go through a progression from benign lesions to in situ tumors to invasive cancers (e.g., atypical ductal hyperplasia to ductal carcinoma in situ to invasive ductal carcinoma of the breast). Fearon and Vogelstein proposed the model for colorectal tumorigenesis presented in Fig. 10-5.⁹ Colorectal tumors arise from the mutational activation of oncogenes coupled with mutational inactivation of tumor-suppressor genes, the latter being the predominant change.⁹ Mutations in at least four or five genes are required for formation of a malignant tumor, while fewer changes suffice for a benign tumor. Although genetic mutations often occur in a preferred sequence, a tumor’s biologic properties are determined by the total accumulation of its genetic changes.

Gene expression is a multistep process that starts from transcription of a gene into messenger ribonucleic acid (mRNA) and then translation of this sequence into the functional protein. There are several controls at each level. In addition to alterations at the genome level (e.g., amplifications of a gene), alterations at the transcription level (e.g., methylation of the DNA leading to transcriptional silencing) or at the level of mRNA processing, mRNA stability, mRNA translation, or protein stability, all can alter the levels of critical proteins and thus contribute to tumorigenesis. Alternatively, changes in the genomic sequence can lead to a mutated product with altered function.

Cell-Cycle Dysregulation in Cancer

The proliferative advantage of tumor cells is a result of their ability to bypass quiescence. Cancer cells often show alterations in signal transduction pathways that lead to proliferation in response to external signals. Mutations or alterations in the expression of cell-cycle proteins, growth factors, growth factor receptors, intracellular signal transduction proteins, and nuclear transcription factors all can lead to disturbance of the basic regulatory mechanisms that control the cell cycle, allowing unregulated cell growth and proliferation.

The cell cycle is divided into four phases (Fig. 10-6).¹⁰ During the synthetic or S phase, the cell generates a single copy of its genetic material, whereas in the mitotic or M phase, the cellular components are partitioned between two daughter cells. The G₁ and G₂ phases represent gap phases during which the cells prepare themselves for completion of the S and M phases, respectively. When cells cease proliferation, they exit the cell cycle and enter the quiescent state referred to as G₀. In human tumor cell-cycle regulators like INK4A, INK4B, and KIP1 are frequently mutated or altered in expression. These alterations underscore the importance of cell-cycle regulation in the prevention of human cancers.

Oncogenes

Normal cellular genes that contribute to cancer when abnormal are called oncogenes. The normal counterpart of such a gene is referred to as a proto-oncogene. Oncogenes are usually designated by three-letter abbreviations, such as myc or ras. Oncogenes are further designated by the prefix “v-” for virus or “c-” for cell or chromosome, corresponding to the origin of the oncogene when it was first detected. Proto-oncogenes can be activated (show increased activity) or overexpressed (expressed at increased protein levels) by translocation (e.g., abl), promoter insertion (e.g., c-myc), mutation (e.g., ras), or amplification (e.g., HER2/neu). More than 100 oncogenes have been identified.

Oncogenes may be growth factors (e.g., platelet-derived growth factor), growth factor receptors (e.g., HER2), intracellular signal transduction molecules (e.g., ras), nuclear transcription factors (e.g., c-myc), or other molecules involved in the regulation of cell growth and proliferation. Growth factors are ubiquitous proteins that are produced and secreted by cells locally and that stimulate cell proliferation by binding specific cell-surface receptors on the same cells (autocrine stimulation) or on neighboring cells (paracrine stimulation). Persistent overexpression of growth factors can lead to uncontrolled autostimulation and neoplastic transformation. Alternatively, growth factor receptors can be aberrantly activated (turned on) through mutations or overexpressed (continually presenting cells with growth-stimulatory signals, even in the absence of growth factors), which leads cells to respond as if growth factor levels are altered. The growth-stimulating effect of growth factors and other mitogens is mediated through postreceptor signal transduction molecules. These molecules mediate the passage of growth signals from the outside to the inside of the cell and then to the cell nucleus, initiating the cell cycle and DNA transcription. Aberrant activation or expression of cell-signaling molecules, cell-cycle molecules, or transcription factors may play an important role in neoplastic transformation. Protein tyrosine kinases account for a large portion of known oncogenes. One of the best-studied oncogenes, HER2 is discussed as an example later.

HER2, also known as neu or c-erbB-2, is a member of the epidermal growth factor receptor (EGFR) family and is one of the best-characterized tyrosine kinases. Unlike other receptor tyrosine kinases, HER2/neu does not have a direct soluble ligand. It plays a key role in signaling, however, because it is the preferred partner in heterodimer formation with all the other EGFR family members (EGFR/c-erbB-1, HER2/c-erbB-3, and HER3/c-erbB-4), which bind at least 30 ligands, including epidermal growth factor (EGF), transforming growth factor α (TGFα), heparin-binding EGF-like growth factor, amphiregulin, and heregulin.¹¹ Heterodimerization with HER2 potentiates recycling of receptors rather than degradation, enhances signal potency and duration, increases affinity for ligands, and increases catalytic activity.¹¹

HER2 can interact with different members of the HER family and activate mitogenic and antiapoptotic pathways (Fig. 10-7). The specificity and potency of the intracellular signals are affected by the identity of the ligand, the composition of the receptors, and the phosphotyrosine-binding proteins associated with the erbB molecules. The Ras- and Shc-activated mitogen-activated protein kinase (MAPK) pathway is a target of all erbB ligands, which increase the transcriptional activity of early response genes such as c-myc, c-fos, and c-jun.¹² MAPK-independent pathways such as the phosphoinositide-3 kinase (PI3K) pathway also are activated by most erbB dimers, although the potency and kinetics of activation may differ. Stimulation of the PI3K pathway through HER2 signaling also can lead to activation of survival molecule Akt, which suppresses apoptosis through multiple mechanisms. The critical role of HER2 in cancer biology has been leveraged for therapeutics, leading to several HER2- targeted drugs with different mechanism of action approved by the Food and Drug Administration (FDA): monoclonal antibodies trastuzumab and pertuzumab, small molecule inhibitor lapatinib, and antibody-drug conjugate ado-trastuzumab emtansine.

The mutant rat neu gene was first recognized as an oncogene in neuroblastomas from carcinogen-treated rats.¹³ The HER2 gene is frequently amplified and the protein overexpressed in many cancers, including breast, ovarian, lung, gastric, and oral cancers. Overexpression of HER2 results in ligand-independent activation of HER2 kinase, which leads to mitogenic signaling. HER2 overexpression is associated with increased cell proliferation and anchorage-independent growth as well as resistance to proapoptotic stimuli. Further, overexpression of HER2 increases cell migration and upregulates the activities of matrix metalloproteinases (MMPs) and in vitro invasiveness. In animal models, HER2 increases tumorigenicity, angiogenesis, and metastasis. These results all suggest that HER2 plays a key role in cancer biology. More recently HER2 mutations have also been reported in human cancer. HER2 mutations have been detected in 2% to 4% of nonsmall cell lung cancer.^14,15,16,17 In frame insertions within exon 20 has been the most commonly reported mutation. HER2 mutations are more common in nonsmokers and are nonoverlapping with other oncogenic mutations in lung cancer (e.g., EGFR and Ras). Data from 8 breast cancer genome-sequencing projects identified 25 patients with HER2 somatic mutations in cancers lacking HER2 gene amplification.¹⁸ Seven of 13 mutations were functionally characterized and found to be activating mutations. All of these mutations were sensitive to the irreversible kinase inhibitor, neratinib. A prospective, multi-institutional clinical trial has been launched to screen patients with stage IV breast cancer for HER2 somatic mutations and determine the clinical outcome of treating them with HER2-targeted therapy.

Alterations in Apoptosis in Cancer Cells

Apoptosis is a genetically regulated program to dispose of cells. Cancer cells must avoid apoptosis if tumors are to arise. The growth of a tumor mass is dependent not only on an increase in proliferation of tumor cells but also on a decrease in their apoptotic rate. Apoptosis is distinguished from necrosis because it leads to several characteristic changes. In early apoptosis, the changes in membrane composition lead to extracellular exposure of phosphatidylserine residues, which avidly bind annexin, a characteristic that is used to discriminate apoptotic cells in laboratory studies. Late in apoptosis there are characteristic changes in nuclear morphology, such as chromatin condensation, nuclear fragmentation, and DNA laddering, as well as membrane blebbing. Apoptotic cells are then engulfed and degraded by phagocytic cells. The effectors of apoptosis are a family of proteases called caspases (cysteine-dependent and aspartate-directed proteases). The initiator caspases (e.g., 8, 9, and 10), which are upstream, cleave the downstream executioner caspases (e.g., 3, 6, and 7) that carry out the destructive functions of apoptosis.

Two principal molecular pathways signal apoptosis by cleaving the initiator caspases with the potential for crosstalk: the mitochondrial pathway and the death receptor pathway. In the mitochondrial (or intrinsic) pathway, death results from the release of cytochrome c from the mitochondria. Cytochrome c, procaspase 9, and apoptotic protease activating factor 1 (Apaf-1) form an enzyme complex, referred to as the apoptosome, that activates the effector caspases. In addition to these proteins, the mitochondria contain other proapoptotic proteins such as second mitochondria-derived activator of caspase/direct inhibitor of apoptosis-binding protein with low pI (Smac/DIABLO. The mitochondrial pathway can be stimulated by many factors, including DNA damage, reactive oxygen species, or the withdrawal of survival factors. The permeability of the mitochondrial membrane determines whether the apoptotic pathway will proceed. The Bcl-2 family of regulatory proteins includes both proapoptotic proteins (e.g., Bax, BAD, and Bak) and antiapoptotic proteins (e.g., Bcl-2 and Bcl-xL). The activity of the Bcl-2 proteins is centered on the mitochondria, where they regulate membrane permeability. Growth factors promote survival signaling through the PI3K/Akt pathway, which phosphorylates and inactivates proapoptotic BAD. In contrast, growth factor withdrawal may promote apoptosis through signaling by unphosphorylated BAD. The heat shock proteins, including Hsp70 and Hsp27, are also involved in inhibition of downstream apoptotic pathways by blocking formation of the apoptosome complex and inhibiting release of cytochrome c from the mitochondria.¹⁹

The second principal apoptotic pathway is the death receptor pathway, sometimes referred to as the extrinsic pathway. Cell-surface death receptors include Fas/APO1/CD95, tumor necrosis factor receptor 1, and KILL-ER/DR5, which bind their ligands Fas-L, tumor necrosis factor (TNF), and TNF-related apoptosis-inducing ligand (TRAIL), respectively. When the receptors are bound by their ligands, they form a death-inducing signaling complex (DISC). At the DISC, procaspase 8 and procaspase 10 are cleaved, yielding active initiator caspases.²⁰ The death receptor pathway may be regulated at the cell surface by the expression of “decoy” receptors for Fas (DcR3) and TRAIL (TRID and TRUNDD). The decoy receptors are closely related to the death receptors but lack a functional death domain; therefore, they bind death ligands but do not transmit a death signal. Another regulatory group is the FADD-like interleukin-1 protease-inhibitory proteins (FLIPs). FLIPs have homology to caspase 8; they bind to the DISC and inhibit the activation of caspase 8. Finally, inhibitors of apoptosis proteins (IAPs) block caspase 3 activation and have the ability to regulate both the death receptor and the mitochondrial pathway.

In human cancers, aberrations in the apoptotic program include increased expression of Fas and TRAIL decoy receptors; increased expression of antiapoptotic Bcl-2; increased expression of the IAP-related protein survivin; increased expression of c-FLIP; mutations or downregulation of proapoptotic Bax, caspase 8, APAF1, XAF1, and death receptors CD95, TRAIL-R1, and TRAIL-R2; alterations of the p53 pathway; overexpression of growth factors and growth factor receptors; and activation of the PI3K/Akt survival pathway.²⁰

Autophagy in Cancer Cells

Autophagy (self-eating) is a major cellular pathway for protein and organelle turnover. This process helps maintain a balance between anabolism and catabolism for normal cell growth and development. Inability to activate autophagy in response to nutrient deprivation, or constitutive activation of autophagy in response to stress, can lead to cell death; thus autophagy is sometimes referred to as a second form of programmed cell death. Autophagy plays an essential role during starvation, cellular differentiation, cell death, and aging. Autophagy is also involved in the elimination of cancer cells by triggering a nonapoptotic cell death program, which suggests a negative role in tumor development. Mouse models that are heterozygotes for the beclin 1 gene, an important gene for autophagy, have altered autophagic response and show a high incidence of spontaneous tumors, which establishes a role for autophagy in tumor suppression.²¹ This also suggests that mutations in other genes operating in this pathway may contribute to tumor formation through deregulation of autophagy. However, autophagy also acts as a stress response mechanism to protect cancer cells from low nutrient supply or therapeutic insults. Studies on the molecular determinants of autophagy are ongoing to determine whether autophagy can be modulated for therapeutic purposes.

Cancer Invasion

A feature of malignant cells is their ability to invade the surrounding normal tissue. Tumors in which the malignant cells appear to lie exclusively above the basement membrane are referred to as in situ cancer, whereas tumors in which the malignant cells are demonstrated to breach the basement membrane, penetrating into surrounding stroma, are termed invasive cancer. The ability to invade involves changes in adhesion, initiation of motility, and proteolysis of the extracellular matrix (ECM).

Cell-to-cell adhesion in normal cells involves interactions between cell-surface proteins. Calcium adhesion molecules of the cadherin family (E-cadherin, P-cadherin, and N-cadherin) are thought to enhance the cells’ ability to bind to one another and suppress invasion. Migration occurs when cancer cells penetrate and attach to the basal matrix of the tissue being invaded; this allows the cancer cell to pull itself forward within the tissue. Attachment to glycoproteins of the ECM such as fibronectin, laminin, and collagen is mediated by tumor cell integrin receptors. Integrins are a family of glycoproteins that form heterodimeric receptors for ECM molecules. The integrins can form at least 25 distinct pairings of their α and β subunits, and each pairing is specific for a unique set of ligands. In addition to regulating cell adhesion to the ECM, integrins relay molecular signals regarding the cellular environment that influence shape, survival, proliferation, gene transcription, and migration.

Factors that are thought to play a role in cancer cell motility include autocrine motility factor, autotaxin, scatter factor (also known as hepatocyte growth factor), TGFα, EGF, and insulin-like growth factors.

Serine, cysteine, and aspartic proteinases and MMPs have all been implicated in cancer invasion. Urokinase and ­tissue plasminogen activators (uPA and tPA) are serine proteases that convert plasminogen into plasmin. Plasmin, in return, can degrade several ECM components. Plasmin also may activate MMPs. uPA has been more closely correlated with tissue invasion and metastasis than tPA. Plasminogen activator inhibitors 1 and 2 (PAI-1 and PAI-2) are produced in tissues and counteract the activity of plasminogen activators.

MMPs comprise a family of metal-dependent endopeptidases. Upon activation, MMPs degrade a variety of ECM components. Although MMPs often are referred to by their common names, which reflect the ECM component for which they have specificity, a sequential numbering system has been adopted for standardization. For example, collagenase-1 is now referred to as MMP-1. The MMPs are further classified as secreted and membrane-type MMPs. Most of the MMPs are synthesized as inactive zymogens (pro-MMP) and are activated by proteolytic removal of the propeptide domain outside the cell by other active MMPs or serine proteinases.

MMPs are upregulated in almost every type of cancer. Some of the MMPs are expressed by cancer cells, whereas others are expressed by the tumor stromal cells. Experimental models have demonstrated that MMPs promote cancer progression by increasing cancer cell growth, migration, invasion, angiogenesis, and metastasis. MMPs exert these effects by cleaving not only structural components of the ECM but also growth factor–binding proteins, growth factor precursors, cell adhesion molecules, and other proteinases. The activity of MMPs is regulated by their endogenous inhibitors and tissue inhibitors of MMPs (TIMP-1, TIMP-2, TIMP-3, and TIMP-4).

Angiogenesis

Angiogenesis is the establishment of new blood vessels from a pre-existing vascular bed. This neovascularization is essential for tumor growth and metastasis. Tumors develop an angiogenic phenotype as a result of accumulated genetic alterations and in response to local selection pressures such as hypoxia. Many of the common oncogenes and tumor-suppressor genes have been shown to play a role in inducing angiogenesis.

In response to the angiogenic switch, pericytes retract and the endothelium secretes several growth factors such as basic fibroblast growth factor, platelet-derived growth factor (PDGF), and insulin-like growth factor. The basement membrane and stroma around the capillary are proteolytically degraded, a process that is mediated in most part by uPA. The endothelium then migrates through the degraded matrix, initially as a solid cord and later forming lumina. Finally, sprouting tips anastomose to form a vascular network surrounded by a basement membrane.

Angiogenesis is mediated by factors produced by various cells, including tumor cells, endothelial cells, stromal cells, and inflammatory cells. The first proangiogenic factor was identified by Folkman and colleagues in 1971.²² Since then, several other factors have been shown to be proangiogenic or antiangiogenic. Of the angiogenic stimulators, the best studied are the vascular endothelial growth factors (VEGFs). The VEGF family consists of six growth factors (VEGF-A, VEGF-B, VEGF-C, VEGF-D, VEGF-E, and placental growth factor) and three receptors (VEGFR1 or Flt-1, VEGFR2 or KDR/FLK-1, and VEGFR3 or Flt-4).²³ Neuropilin 1 and 2 also may act as receptors for VEGF.²⁴ VEGF is induced by hypoxia and by different growth factors and cytokines, including EGF, PDGF, TNF-α, TGFβ, and interleukin-1β. VEGF has various functions, including increasing vascular permeability, inducing endothelial cell proliferation and tube formation, and inducing endothelial cell synthesis of proteolytic enzymes such as uPA, PAI-1, urokinase plasminogen activator receptor, and MMP-1. Furthermore, VEGF may mediate blood flow by its effects on the vasodilator nitric oxide and act as an endothelial survival factor, thus protecting the integrity of the vasculature. The proliferation of new lymphatic vessels, lymphangiogenesis, is also thought to be controlled by the VEGF family. Signaling in lymphatic cells is thought to be modulated by VEGFR3.²⁵ Experimental studies with VEGF-C and VEGF-D have shown that they can induce tumor lymphangiogenesis and direct metastasis via the lymphatic vessels and lymph nodes.^25,26

PDGFs A, B, C, and D also play important roles in angiogenesis. PDGFs cannot only enhance endothelial cell proliferation directly but also upregulate VEGF expression in vascular smooth muscle cells, promoting endothelial cell survival via a paracrine effect.²³ The angiopoietins angiopoietin-1 and angiopoietin-2 (Ang-1 and Ang-2), in return, are thought to regulate blood vessel maturation. Ang-1 and Ang-2 both bind angiopoietin-1 receptor (also known as tyrosine-protein kinase receptor TIE-2), but only the binding of Ang-1 activates signal transduction; thus Ang-2 is an Ang-1 antagonist. Ang-1, via the Tie-2 receptor, induces remodeling and stabilization of blood vessels. Upregulation of Ang-2 by hypoxic induction of VEGF inhibits Ang-1–induced Tie-2 signaling, which results in destabilization of vessels and makes endothelial cells responsive to angiogenic signals, thus promoting angiogenesis in the presence of VEGF. Therefore the balance between these factors determines the angiogenetic capacity of a tumor.

Tumor angiogenesis is regulated by several factors in a coordinated fashion. In addition to upregulation of proangiogenic molecules, angiogenesis also can be encouraged by suppression of naturally occurring inhibitors. Such inhibitors of angiogenesis include thrombospondin 1 and angiostatin. Angiogenesis is a prerequisite not only for primary tumor growth but also for metastasis. Angiogenesis in the primary tumor, as determined by microvessel density, has been demonstrated to be an independent predictor of distant metastatic disease and survival in several cancers. Expression of angiogenic factors such as VEGFs has had prognostic value in many studies. These findings further emphasize the importance of angiogenesis in cancer biology.

Metastasis

Metastases arise from the spread of cancer cells from the primary site and the formation of new tumors in distant sites. The metastatic process consists of a series of steps that need to be completed successfully (Fig. 10-8).²⁷ First, the primary cancer must develop access to the circulation through either the blood circulatory system or the lymphatic system. After the cancer cells are shed into the circulation, they must survive. Next, the circulating cells lodge in a new organ and extravasate into the new tissue. Next, the cells need to initiate growth in the new tissue and eventually establish vascularization to sustain the new tumor. Overall, metastasis is an inefficient process, although the initial steps of hematogenous metastasis (the arrest of tumor cells in the organ and extravasation) are believed to be performed efficiently. Only a small subset of cancer cells is then able to initiate micrometastases, and an even smaller portion goes on to grow into macrometastases.

Metastases can sometimes arise several years after the treatment of primary tumors. For example, although most breast cancer recurrences occur within the first 10 years after the initial treatment and recurrences are rare after 20 years, breast cancer recurrences have been reported decades after the original tumor. This phenomenon is referred to as dormancy, and it remains one of the biggest challenges in cancer biology. Persistence of solitary cancer cells in a secondary site such as the liver or bone marrow is one possible contributor to dormancy.²⁸ Another explanation of dormancy is that cells remain viable in a quiescent state and then become reactivated by a physiologically perturbing event. Interestingly, primary tumor removal has been proposed to be a potentially perturbing factor.²⁹ An alternate explanation is that cells establish preangiogenic metastases in which they continue to proliferate but that the proliferative rate is balanced by the apoptotic rate. Therefore, when these small metastases acquire the ability to become vascularized, substantial tumor growth can be achieved at the metastatic site, leading to clinical detection.

Several types of tumors metastasize in an organ-specific pattern. One explanation for this is mechanical and is based on the different circulatory drainage patterns of the tumors. When different tumor types and their preferred metastasis sites were compared, 66% of organ-specific metastases were explained on the basis of blood flow alone. The other explanation for preferential metastasis is what is referred to as the “seed and soil” theory, the dependence of the seed (the cancer cell) on the soil (the secondary organ). According to this theory, once cells have reached a secondary organ, their growth efficiency in that organ is based on the compatibility of the cancer cell’s biology with its new microenvironment. For example, breast cancer cells may grow more efficiently in bone than in some other organs because of favorable molecular interactions that occur in the bone microenvironment. The ability of cancer cells to grow in a specific site likely depends on features inherent to the cancer cell, features inherent to the organ, and the interplay between the cancer cell and its microenvironment.³⁰

Many of the oncogenes discovered to date, such as HER2 and ras, are thought to potentiate not only malignant transformation but also one or more of the steps required in the metastatic process. Experimental models have suggested a role for several molecules, including RhoC, osteopontin and interleukin-11, and Twist, in tumor metastasis. Metastasis also may involve the loss of metastasis-suppressor genes. Laboratory work involving cancer cell lines that have been selected to have a higher metastatic potential have led to the realization that these more highly metastatic cells have a different gene expression profile than their less metastatic parental counterparts. This in turn has led to the currently held belief that the ability of a primary tumor to metastasize may be predictable by analysis of its gene expression profile. Indeed, several studies have recently focused on identifying a gene expression profile or a molecular signature that is associated with metastasis. It has been shown that such a gene expression profile can be used to predict the probability that the patient will remain free of distant metastasis.³¹ This suggests that the metastatic potential of a tumor is already predetermined by the genetic alterations that the cancer cells acquire early in tumorigenesis. Notably, this hypothesis differs from the multistep tumorigenesis theory in that the ability to metastasize is considered an inherent quality of the tumor from the beginning. It is assumed that metastasis develops not from a few rare cells in the primary tumor that acquire the ability to metastasize but that all cells in tumors with such molecular signatures develop the ability to metastasize. The reality probably lies in between since some early genetic changes detectable in the entire tumor can give tumors an advantage in the metastatic process, whereas additional genetic changes can give a clone of cells additional advantages, thus allowing them to succeed in metastasis.

Epithelial-Mesenchymal Transition

A regulatory program referred to as epithelial-mesenchymal transition (EMT) is a fundamental event in morphogenesis. During EMT epithelial cells are converted to migratory and invasive cells.³² EMT, has also been implicated as the mechanism through which epithelial cells acquire the ability to migrate, invade, resist apoptosis and metastasize. EMT is a developmental process, and a set of pleiotropically acting transcriptional factors, including Snail, Twist, Slug, and Zeb1/2orchestrateEMT. Several of these transcription factors can directly repress E-cadherin gene expression, depriving ­cancer cells of this key suppressor of motility and invasiveness. It has been proposed that the process of invasion and metastases requires significant plasticity, suggesting that EMT is required for invasion, intravasation and extravasation, and suppression of EMT regulators (and consequently EMT reversion, or MET) is required for metastatic outgrowth.^33,34,35

Cancer Stem Cells

Stem cells are cells that have the ability to perpetuate themselves through self-renewal and to generate mature cells of a particular tissue through differentiation.³⁶ It has recently been proposed that stem cells themselves may be the target of transformation. It was first documented for leukemia and multiple myeloma that only a small subset of cancer cells is capable of extensive proliferation. It has subsequently also been shown for many solid cancers that only a small proportion of cells is clonogenic in culture and in vivo. In leukemia and multiple myeloma only a small subset of cancer cells is capable of extensive proliferation. Similarly, in many solid tumor types only a small proportion of cells is clonogenic in culture and in vivo. If indeed tumor growth and metastasis are driven by a small population of cancer stem cells, this may alter our current approaches to cancer therapy. Currently available drugs can shrink metastatic tumors but often cannot eradicate them. The failure of these treatments usually is attributed to the acquisition of drug resistance by the cancer cells; however, the cancer stem cell hypothesis raises the possibility that existing therapies may simply fail to kill cancer stem cells effectively. Therapeutic approaches targeting stem cells specifically are under study.

Cancer Genomics

One widely held opinion is that cancer is a genetic disease that arises from an accumulation of genomic alterations that leads to the selection of cells with increasingly aggressive behavior. These alterations may lead either to a gain of function by oncogenes or to a loss of function by tumor-suppressor genes. These acquired gene alterations are termed somatic mutations to distinguish them from germline mutations that are inherited from parents and transmitted to offspring. Somatic mutations in a cancer genome may consist of several classes of DNA sequence changes. These include substitutions of one base by another; insertions or deletions of small or large segments of DNA; rearrangements, in which the DNA sequence has been broken and then rejoined to another DNA segment; copy number losses that may result in complete absence of a DNA sequence and copy number gains from the two copies present in the normal diploid genome.

Somatic mutations in a cancer cell genome have accumulated over the lifetime of the patient (Fig. 10-9).³⁷ DNA in normal cells is continuously damaged by internal and external mutagens. Most of this damage is repaired; however, a small fraction may remain as fixed mutations. Mutation rates increase in the presence of substantial exogenous mutagenic exposures, such as tobacco carcinogens or various forms of radiation, including ultraviolet light. These exposures are associated with increased rates of lung and skin cancer, respectively, and somatic mutations within such cancers often exhibit the distinctive mutational signatures known to be associated with the mutagen.³⁸ The rates of somatic mutations are also increased in several rare inherited diseases, such as Fanconi anemia, ataxia telangiectasia, and xeroderma pigmentosum, which are associated with increased risks of cancer.^39,40 The rest of the somatic mutations in a cancer cell have been acquired after the cancer cell already shows phenotypic evidence of neoplastic change. Whether the somatic mutation rate is always higher during this part of the lineage is controversial. This is clearly the case for some cancers. For instance, colorectal and endometrial cancers with defective DNA mismatch repair due to abnormalities in genes such as MLH1 and MSH2, exhibit increased rates of single nucleotide changes and small insertions/deletions atpolynucleotide tract.⁴¹ These tumor types are often referred to as “mutator phenotypes.”

To date about 300 genes that have been reported to be mutated and causally implicated in cancer development.⁴² Ninety percent of cancer genes show somatic mutations in cancer, 20% show germline mutations, and 10% show both. The most common class of genomic alterations among the known cancer genes is a chromosomaltranslocation that creates a chimeric gene. Many more cancer genes have been found in leukemias, lymphomas, and sarcomas than in other types of cancer; and these genes are usually altered by chromosomal translocation. The most common cancer genes are protein kinases. Several domains that are involved in DNA binding and transcriptional regulation are also common in proteins encoded by cancer genes. Somatic mutations in a cancer genome may be classified according to its consequences for cancer development. “Driver” mutations confer a growth advantage to the cells carrying them and have been positively selected during the evolution of the cancer. The remainder of mutations are “bystanders” or “passengers” that do not confer growth advantage. It is likely that most somatic mutations are passenger mutations. Each tumor may have dozens to hundreds of genomic alterations, making it critical to determine which alterations are indeed drivers, and potentially better therapeutic targets.

There are several ongoing large scale studies to characterize and catalogue genomic alterations in different cancer types, including the Cancer Genome Project at the Sanger Institute, United Kingdom, and The Cancer Genome Atlas project (TCGA). There are also increasing number of publically accessible resources, including COSMIC (http://www.sanger.ac.uk/cosmic), which curates comprehensive information on somatic mutations in human cancer.⁴³ These resources are being utilized to determine the most common genomic alterations in common tumor types. This information is being integrated into clinical practice in many tumor types, such as lung cancer, where molecular drivers are being chosen taking into consideration in systemic therapy selection (Fig. 10-10).

Tumor Heterogeneity and Molecular Evolution

There is increasing recognition that tumors are heterogeneous; this represents an important challenge to utilizing genomic alterations to personalize cancer therapy (Fig. 10-11).⁴⁴ First, there is significant intertumoral heterogeneity, such that patients with tumors that seem similar histologically, may differ in genomic alterations and in malignant potential.^45,46,47 Second, during cancer progression, subclones frequently arise, resulting in differences in the proportion and pattern of genomic alterations between the primary tumor and the metastases or local-regional recurrences.⁴⁴ Third, there may also be significant intratumoral heterogeneity, with spatially separated heterogeneous somatic mutations and chromosomal imbalances.⁴⁸ Such spatial heterogeneity of subclones within the primary tumor or metastases provides an additional challenge, as it has been proposed that sequencing of a biopsy specimen or only a portion of the tumor could miss therapeutically relevant genomic alterations. The genomic alterations found in a tumor can also change under the selective pressure of a targeted therapy, adding to the challenge of implementing genomically-informed personalized therapy.

Genes Associated with Hereditary Cancer Risk

Most of our information on human cancer genes has been gained from hereditary cancers. In the case of hereditary cancers, the individual carries a particular germline mutation in every cell. To date, over 70 genes have been associated with hereditary cancers (Table 10-3).⁴² A few of these hereditary cancer genes are oncogenes, but most are tumor-suppressor genes. Although hereditary cancer syndromes are rare, somatic mutations that occur in sporadic cancer have been found to disrupt the cellular pathways altered in hereditary cancer syndromes, which suggests that these pathways are critical to normal cell growth, cell cycle, and proliferation.

The following factors may suggest the presence of a hereditary cancer⁴⁹:

1.  Tumor development at a much younger age than usual

2.  Presence of bilateral disease

3.  Presence of multiple primary malignancies

4.  Presentation of a cancer in the less affected sex (e.g., male breast cancer)

5.  Clustering of the same cancer type in relatives

6.  Occurrence of cancer in association with other conditions such as mental retardation or pathognomonic skin lesions

It is crucial that all surgeons caring for cancer patients be aware of hereditary cancer syndromes, because a patient’s genetic background has significant implications for patient counseling, planning of surgical therapy, and cancer screening and prevention. Some of the more commonly encountered hereditary cancer syndromes are discussed here.

rb1Gene

The retinoblastoma gene rb1 was the first tumor suppressor to be cloned. The rb1 gene product, the Rb protein, is a regulator of transcription that controls the cell cycle, differentiation, and apoptosis in normal development.⁵⁰ ­Retinoblastoma has long been known to occur in hereditary and nonhereditary forms. Interestingly, although most children with an affected parent develop bilateral retinoblastoma, some develop unilateral retinoblastoma. Furthermore, some children with an affected parent are not affected themselves but then have an affected child, which indicates that they are rb1 mutation carriers. These findings led to the theory that a single mutation is not sufficient for tumorigenesis. Alfred Knudson hypothesized that hereditary retinoblastoma involves two mutations, of which one is germline and one somatic, whereas nonhereditary retinoblastoma is due to two somatic mutations (Fig. 10-12).⁵¹ Thus, both hereditary and nonhereditary forms of retinoblastoma involve the same number of mutations, a hypothesis known as Knudson’s “two-hit” hypothesis. A “hit” may be a point mutation, a chromosomal deletion referred to as allelic loss, or a loss of heterozygosity, or silencing of an existing gene.

p53 and Li-Fraumeni Syndrome

Li-Fraumeni syndrome (LFS) was first defined on the basis of observed clustering of malignancies, including early-onset breast cancer, soft tissue sarcomas, brain tumors, adrenocortical tumors, and leukemia.⁵² Criteria for classic LFS in an individual (the proband) include: (a) a bone or soft tissue sarcoma when younger than 45 years, (b) a first-degree relative with cancer before age 45 years, and (c) another first- or second-degree relative with either a sarcoma diagnosed at any age or any cancer diagnosed before age 45 years.⁵³ Approximately 70% of LFS families have been shown to have germline mutations in the tumor-suppressor gene p53.⁵⁴ Breast carcinoma, soft tissue sarcoma, osteosarcoma, brain tumors, adrenocortical carcinoma, Wilms’ tumor, and phyllodes tumor of the breast are strongly associated; pancreatic cancer is moderately associated; and leukemia and neuroblastoma are weakly associated with germline p53 mutations.⁵⁵ Mutations of p53 have not been detected in approximately 30% of LFS families, and it is hypothesized that genetic alterations in other proteins interacting with p53 function may play a role in these families.

Of the known genes in human cancer, p53 is the most commonly mutated. The p53 protein regulates cell-cycle progression as well as apoptotic cell death as part of stress response pathways after exposure to ionizing or ultraviolet (UV) irradiation, chemotherapy, acidosis, growth factor deprivation, or hypoxia. When cells are exposed to stressors, p53 acts as a transcription factor for genes that induce cell-cycle arrest or apoptosis. A majority of p53 mutations are found within a central DNA recognition motif and disrupt DNA binding by p53. Families with germline missense mutations in the DNA-binding domain show a more highly penetrant phenotype than families with other p53 mutations.⁵⁶ Furthermore, proband cancers are linked with significantly younger age at diagnosis in patients with missense mutations in the DNA-binding domain.⁵⁶

BRCA1, BRCA2, and Hereditary Breast-Ovarian Cancer

Syndrome

It is estimated that 5% to 10% of breast cancers are hereditary. Of women with early-onset breast cancer (aged 40 years or younger), nearly 10% have a germline mutation in one of the breast cancer genes BRCA1 or BRCA2.⁵⁷ Mutation carriers are more prevalent among women who have a first- or second-degree relative with premenopausal breast cancer or ovarian cancer at any age. The likelihood of a BRCA mutation is higher in patients who belong to a population in which founder mutations may be prevalent, such as in the Ashkenazi Jewish population. For a female BRCA1 mutation carrier, the cumulative risks of developing breast cancer and ovarian cancer by age 70 have been estimated to be 87% and 44%, respectively.⁵⁸ The cumulative risks of breast cancer and ovarian cancer by age 70 in families with BRCA2 mutation have been estimated to be 84% and 27%, respectively.⁵⁹ Although male breast cancer can occur with either BRCA1 or BRCA2 mutation, the majority of families (76%) with both male and female breast cancer have mutations in BRCA2.⁵⁹ Besides breast and ovarian cancer, BRCA1 and BRCA2 mutations may be associated with increased risks for several other cancers. BRCA1 mutations confer a fourfold increased risk for colon cancer and threefold increased risk for prostate cancer.⁵⁸ BRCA2 mutations confer a fivefold increased risk for prostate cancer, sevenfold in men younger than 65 years.⁶⁰ Furthermore, BRCA2 mutations confer a fivefold increased risk for gallbladder and bile duct cancers, fourfold increased risk for pancreatic cancer, and threefold increased risk for gastric cancer and malignant melanoma.⁶⁰

BRCA1 was the first breast cancer susceptibility gene identified and has been mapped to 17q21. BRCA2, mapped to 13q12.3, was reported shortly afterward. BRCA1 and BRCA2 encode large nuclear proteins, 208 kDa and 384 kDa, respectively, that have been implicated in processes fundamental to all cells, including DNA repair and recombination, checkpoint control of the cell cycle, and transcription.⁶¹ Although early studies suggested that the two proteins function together as a complex, subsequent data demonstrated that they have distinct functions.^62,63 In fact, breast cancers arising from BRCA1 or BRCA2 mutations are different at the molecular level and have been found to have distinct gene expression profiles.⁶⁴ BRCA1-associated tumors are more likely to be estrogen receptor negative, whereas BRCA2-associated tumors are more likely to be estrogen receptor positive. Currently, studies are ongoing to determine whether BRCA1 and BRCA2 status can be used to guide systemic therapy choices for breast cancer.

APC Gene and Familial Adenomatous Polyposis

Patients affected with familial adenomatous polyposis (FAP) characteristically develop hundreds to thousands of polyps in the colon and rectum. The polyps usually appear in adolescence and, if left untreated, progress to colorectal cancer. FAP is associated with benign extracolonic manifestations that may be useful in identifying new cases, including congenital hypertrophy of the retinal pigment epithelium, epidermoid cysts, and osteomas. In addition to colorectal cancer, patients with FAP are at risk for upper intestinal neoplasms (gastric and duodenal polyps, duodenal and periampullary cancer), hepatobiliary tumors (hepatoblastoma, pancreatic cancer, and cholangiocarcinoma), thyroid carcinomas, desmoid tumors, and medulloblastomas.

The product of the adenomatous polyposis coli tumor-suppressor gene (APC) plays an important role in cell-cell interactions, cell adhesion, regulation of β-catenin, and maintenance of cytoskeletal microtubules. Alterations in APC lead to dysregulation of several physiologic processes that govern colonic epithelial cell homeostasis, including cell-cycle progression, migration, differentiation, and apoptosis. Mutations in the APC have been identified in FAP and in 80% of sporadic colorectal cancers.⁶⁵ Furthermore, APC mutations are the earliest known genetic alterations in colorectal cancer progression, which emphasizes its importance in cancer initiation. The germline mutations in APC may arise from point mutations, insertions, or deletions that lead to a premature stop codon and a truncated, functionally inactive protein. The risk of developing specific manifestations of FAP is correlated with the position of the FAP mutations, a phenomenon referred to as genotype-phenotype correlation. For example, desmoids usually are associated with mutations between codons 1403 and 1578.^66,67 Mutations in the extreme 5’ or 3’ ends of APC, or in the alternatively spliced region of exon 9, are associated with an attenuated version of FAP. Better understanding of the genotype-phenotype correlations may assist in patient counseling and therapeutic planning.

Mismatch Repair Genes and Hereditary Nonpolyposis Colorectal Cancer

Hereditary nonpolyposis colorectal cancer (HNPCC), also referred to as Lynch syndrome, is an autosomal dominant hereditary cancer syndrome that predisposes to a wide spectrum of cancers, including colorectal cancer without polyposis. Some have proposed that HNPCC consists of at least two syndromes: Lynch syndrome 1, which entails hereditary predisposition for colorectal cancer with early age of onset (approximately age 44 years) and an excess of synchronous and metachronous colonic cancers; and Lynch syndrome 2, featuring a similar colonic phenotype accompanied by a high risk for carcinoma of the endometrium, transitional cell carcinoma of the ureter and renal pelvis, and carcinomas of the stomach, small bowel, ovary, and pancreas.⁶⁸ The diagnostic criteria for HNPCC are referred to as the Amsterdam criteria, or the 3-2-1-0 rule. The classic Amsterdam criteria were revised to include other HNPCC-related cancers (Table 10-4).⁶⁹ These criteria are met when three or more family members have histologically verified, HNPCC-associated cancers (one of whom is a first-degree relative of the other two), two or more generations are involved, at least one individual was diagnosed before age 50 years, and no individuals have FAP.⁶⁹

During DNA replication, DNA polymerases may introduce single nucleotide mismatches or small insertion or deletion loops. These errors are corrected through a process referred to as mismatch repair. When mismatch repair genes are inactivated, DNA mutations in other genes that are critical to cell growth and proliferation accumulate rapidly. In HNPCC, germline mutations have been identified in several genes that play a key role in DNA nucleotide mismatch repair: hMLH1 (human mutL homologue 1), hMSH2 (human mutS homologue 2), hMSH6, and hPMS1 and hPMS2 (human postmeiotic segregation 1 and 2), of which hMLH1 and hMSH2 are the most common.^{70,71,72,73,74,75} The hallmark of HNPCC is microsatellite instability, which occurs on the basis of unrepaired mismatches and small insertion or deletion loops. Microsatellite instability can be tested by comparing the DNA of a patient’s tumor with DNA from adjacent normal epithelium, amplifying the DNA with polymerase chain reaction (PCR) using a standard set of markers, comparing the amplified genomic DNA sequences, and classifying the degree of microsatellite instability as high, low, or stable. Such microsatellite instability testing may help select patients who are more likely to have germline mutations.

PTEN and Cowden Disease

Somatic deletions or mutations in the tumor-suppressor gene PTEN (phosphatase and tensin homologue deleted on chromosome 10) have been observed in a number of glioma breast, prostate, and renal carcinoma cell lines and several primary tumor specimens.⁷⁶

PTEN encodes a 403-amino-acid protein, tyrosine phosphatase. PTEN negatively controls the PI3K signaling pathway for the regulation of cell growth and survival by dephosphorylating phosphoinositol 3,4,5-triphosphate; thus mutation of PTEN leads to constitutive activation of the PI3K/Akt signaling pathway. The “hot spot” for PTEN mutations has been identified in exon 5. Forty-three percent of CD mutations have been identified in this exon, which contains the tyrosine phosphatase core domain. This suggests that the PTEN catalytic activity is vital for its biologic function. PTEN was identified as the susceptibility gene for the autosomal dominant syndrome Cowden disease (CD) or multiple hamartoma syndrome.⁷⁷ Trichilemmomas, benign tumors of the hair follicle infundibulum, and mucocutaneous papillomatosis are pathognomonic of CD. Other common features include thyroid adenomas and multinodular goiters, breast fibroadenomas, and hamartomatous GI polyps. The diagnosis of CD is made when an individual or family has a combination of pathognomonic major and/or minor criteria proposed by the International Cowden Consortium.⁷⁸ CD is associated with an increased risk of breast and thyroid cancers. Breast cancer develops in 25% to 50% of affected women.⁷⁸

p16 and Hereditary Malignant Melanoma

The gene p16, also known as INK4A, CDKN1, CDKN2A, and MTS1, is a tumor suppressor that acts by binding CDK4 and CDK6 and inhibiting the catalytic activity of the CDK4-CDK6/cyclin D complex that is required for phosphorylation of Rb and subsequent cell-cycle progression. Studies suggest that germline mutations in p16 can be found in 20% of melanoma-prone families.⁷⁹ Mutations in p16 that alter its ability to inhibit the catalytic activity of the CDK4-CDK6/cyclin D complex not only increase the risk of melanoma by 75-fold but also increase the risk of pancreatic cancer by 22-fold.⁸⁰ Interestingly, p16 mutations that do not appear to alter its function increase the risk of melanoma by 38-fold and do not increase the risk of pancreatic cancer.⁸⁰ Genomic characterization of primary tumors has revealed that p16 is inactivated through point mutation, promoter methylation, or deletion in a significant portion of sporadic tumors, including cancers of the pancreas, esophagus, head and neck, stomach, breast, and colon, as well as melanomas.

E-cadherin and Hereditary Diffuse Gastric Cancer

E-cadherin is a cell adhesion molecule that plays an important role in normal architecture and function of epithelial cells. The adhesive function of E-cadherin is dependent on interaction of its cytoplasmic domain with β- and γ-catenins and may be regulated by phosphorylation of β-catenin.

Hereditary diffuse gastric carcinoma is an autosomal dominant cancer syndrome that results from germline mutations in the E-cadherin gene, CDH1. Carriers of CDH1 mutations have a 70% to 80% chance of developing gastric cancer.⁸¹ Furthermore, mutations of CDH1 have been described in sporadic cancers of the ovary, endometrium, breast, and thyroid. However, frequent mutations have been identified in only two particular tumors: diffuse gastric carcinomas and lobular breast carcinomas. Invasive lobular breast carcinomas often show inactivating mutations in combination with a loss of heterozygosity of the wild-type CDH1 allele.⁸² Interestingly, in gastric carcinomas the predominant mutations are exon skipping causing in-frame deletions, whereas most mutations identified in lobular breast cancers are premature stop codons; this suggests a genotype-phenotype correlation.

RET Proto-Oncogene and Multiple Endocrine Neoplasia Type 2

The RET (rearranged during transfection) gene encodes for a transmembrane receptor tyrosine kinase that plays a role in proliferation, migration, and differentiation of cells derived from the neural crest. Gain-of-function mutations in the RET gene are associated with medullary thyroid carcinoma in isolation or multiple endocrine neoplasia type 2 (MEN2) syndromes. MEN2A is associated with medullary thyroid carcinoma and pheochromocytoma (in 50%) or parathyroid adenoma (in 20%), whereas MEN2B is associated with medullary thyroid carcinoma, marfanoid habitus, mucosal neuromas, and ganglioneuromatosis.⁸³RET mutations lead to uncontrolled growth of the thyroid C cells, and in familial medullary cancer, C-cell hyperplasia progresses to bilateral, multicentric medullary thyroid cancer. Mutations in the RET gene have also been identified in half of sporadic medullary thyroid cancers.

Genetic Modifiers of Risk

Individuals carrying identical germline mutations vary in regard to cancer penetrance (whether cancer will develop or not) and cancer phenotype (the tissues involved). It is thought that this variability may be due to environmental influences or, if genetic, to genetic modifiers of risk. Similarly, genetic modifiers of risk also can play a role in determining whether an individual will develop cancer after exposure to carcinogens.

Chemical Carcinogens

The first report indicating that cancer could be caused by environmental factors was by John Hill, who in 1761 noted the association between nasal cancer and excessive use of tobacco snuff.⁸⁴ Currently, approximately 60% to 90% of cancers are thought to be due to environmental factors. Any agent that can contribute to tumor formation is referred to as a carcinogen and can be a chemical, physical, or viral agent. Chemicals are classified into three groups based on how they contribute to tumor formation. The first group of chemical agents, the genotoxins, can initiate carcinogenesis by causing a mutation. The second group, the cocarcinogens, by themselves cannot cause cancer but potentiate carcinogenesis by enhancing the potency of genotoxins. The third group, tumor promoters, enhances tumor formation when given after exposure to genotoxins.

The International Agency for Research on Cancer (IARC) maintains a registry of human carcinogens that is available through the World Wide Web (http://www.iarc.fr). The compounds are categorized into five groups based on an analysis of epidemiologic studies, animal models, and short-term mutagenesis tests. Group 1 contains what are considered to be proven human carcinogens, based on formal epidemiologic studies among workers who were exposed for long periods (several years) to the chemicals.⁸⁵ Group 2A contains what are considered to be probable human carcinogens. Suggestive epidemiologic evidence exists for compounds in this group, but the data are insufficient to establish causality. There is evidence of carcinogenicity, however, from animal studies carried out under conditions relevant to human exposure. Group 2B contains what are considered to be possible carcinogens, because these substances are associated with a clear statistically and biologically significant increase in the incidence of malignant tumors in more than one animal species or strain. Group 3 agents are not classifiable, and Group 4 agents are probably not carcinogenic to humans. Selected substances that have been classified as proven carcinogens (group 1) by the IARC in an expert panel review in 2009 are listed in Table 10-5.⁸⁶

Physical Carcinogens

Physical carcinogenesis can occur through induction of inflammation and cell proliferation over a period of time or through exposure to physical agents that induce DNA damage. Foreign bodies can cause chronic irritation that can expose cells to carcinogenesis due to other environmental agents. In animal models, for example, subcutaneous implantation of a foreign body can lead to the development of tumors that have been attributed to chronic irritation from the foreign objects. In humans, clinical scenarios associated with chronic irritation and inflammation such as chronic nonhealing wounds, burns, and inflammatory bowel syndrome have all been associated with an increased risk of cancer. H. pylori infection is associated with gastritis and gastric cancer, and thus, its carcinogenicity may be considered physical carcinogenesis. Infection with the liver fluke Opisthorchis viverrini similarly leads to local inflammation and ­cholangiocarcinoma.

The induction of lung and mesothelial cancers by asbestos fibers and nonfibrous particles such as silica are other examples of foreign body-induced physical carcinogenesis.⁸⁷ Animal experiments have demonstrated that the dimensions and durability of the asbestos and other fibrous minerals are the key determinants of their carcinogenicity.⁸⁸ Short fibers can be inactivated by phagocytosis, whereas long fibers (>10 μm) are cleared less effectively and are encompassed by proliferating epithelial cells. The long fibers support cell proliferation and have been shown to preferentially induce tumors. Asbestos-associated biologic effects also may be mediated through reactive oxygen and nitrogen species. Furthermore, an interaction occurs between asbestos and silica and components of cigarette smoke. Polycyclic aromatic hydrocarbons (PAHs) in cigarette smoke are metabolized by epithelial cells and form DNA adducts. If PAH is coated on asbestos, PAH uptake is increased.⁸⁷ Both PAH and asbestos impair lung clearance, potentially increasing uptake further. Therefore, physical carcinogens may be synergistic with chemical carcinogens.

Radiation is the best-known agent of physical carcinogens and is classified as ionizing radiation (X-rays, gamma rays, and alpha and beta particles) or nonionizing radiation (UV). The carcinogenic potential of ionizing radiation was recognized soon after Wilhelm Conrad Roentgen’s discovery of X-rays in 1895. Within the next 20 years, a large number of radiation-related skin cancers were reported. Long-term follow-up of survivors of the atomic bombing of Hiroshima and Nagasaki revealed that virtually all tissues exposed to radiation are at risk for cancer.

Radiation can induce a spectrum of DNA lesions that includes damage to the nucleotide bases and cross-linking, and DNA single- and double-strand breaks (DSBs). Misrepaired DSBs are the principal lesions of importance in the induction of chromosomal abnormalities and gene mutations. DSBs in irradiated cells are repaired primarily by a nonhomologous end-joining process, which is error prone; thus, DSBs facilitate the production of chromosomal rearrangements and other large-scale changes such as chromosomal deletions. It is thought that radiation may initiate cancer by inactivating tumor-suppressor genes. Activation of oncogenes appears to play a lesser role in radiation carcinogenesis.

Although it has been assumed that the initial genetic events induced by radiation constitute direct mutagenesis from radiation, other indirect effects may contribute to carcinogenesis. For example, radiation induces genomic instability in cells that persists for at least 30 generations after irradiation. Therefore, even if cells do not acquire mutations at initial irradiation, they remain at risk for developing new mutations for several generations. Moreover, even cells that have not been directly irradiated appear to be at risk, a phenomenon referred to as the bystander effect.

Nonionizing UV radiation is a potent DNA-damaging agent and is known to induce skin cancer in experimental animals. Most nonmelanoma human skin cancers are thought to be induced by repeated exposure to sunlight, which leads to a series of mutations that allow the cells to escape normal growth control. Patients with inherited xeroderma pigmentosum lack one or more DNA repair pathways, which confers susceptibility to UV-induced cancers, especially on sun-exposed body parts. Patients with ataxia telangiectasia mutated syndrome also have a radiation-sensitive phenotype.

Viral Carcinogens

One of the first observations that cancer may be caused by transmissible agents was by Peyton Rous in 1910 when he demonstrated that cell-free extracts from sarcomas in chickens could transmit sarcomas to other animals injected with these extracts.⁸⁹ This was subsequently discovered to represent viral transmission of cancer by the Rous sarcoma virus. At present, several human viruses are known to have oncogenic properties, and several have been causally linked to human cancers (Table 10-6).⁸⁵ It is estimated that 15% of all human tumors worldwide are caused by viruses.⁹⁰

Viruses may cause or increase the risk of malignancy through several mechanisms, including direct transformation, expression of oncogenes that interfere with cell-cycle checkpoints or DNA repair, expression of cytokines or other growth factors, and alteration of the immune system. Oncogenic viruses may be RNA or DNA viruses. Oncogenic RNA viruses are retroviruses and contain a reverse transcriptase. After the viral infection, the single-stranded RNA viral genome is transcribed into a double-stranded DNA copy, which is then integrated into the chromosomal DNA of the cell. Retroviral infection of the cell is permanent; thus, integrated DNA sequences remain in the host chromosome. Oncogenic transforming retroviruses carry oncogenes derived from cellular genes. These cellular genes, referred to as proto-oncogenes, usually are involved in mitogenic signaling and growth control, and include protein kinases, G proteins, growth factors, and transcription factors (Table 10-7).⁹⁰

Integration of the provirus upstream of a proto-oncogene may produce chimeric virus-cell transcripts and recombination during the next round of replication that could lead to incorporation of the cellular gene into the viral genome.⁹⁰ Then again, many retroviruses do not possess oncogenes but can cause tumors in animals regardless. This occurs by integration of the provirus near a normal cellular proto-oncogene and activation of the expression of these genes by the strong promoter and enhancer sequences in the integrated viral sequence.

Unlike the oncogenes of the RNA viruses, those of the DNA tumor viruses are viral, not cellular, in origin. These genes are required for viral replication using the host cell machinery. In permissive hosts, infection with an oncogenic DNA virus may result in a productive lytic infection, which leads to cell death and the release of newly formed viruses. In nonpermissive cells, the viral DNA can be integrated into the cellular chromosomal DNA, and some of the early viral genes can be synthesized persistently, which leads to transformation of cells to a neoplastic state. The binding of viral oncoproteins to cellular tumor-suppressor proteins p53 and Rb is fundamental to the carcinogenesis induced by most DNA viruses, although some target different cellular proteins.

Like other types of carcinogenesis, viral carcinogenesis is a multistep process. Some retroviruses contain two cellular oncogenes, rather than one, in their genome and are more rapidly tumorigenic than single-gene transforming retroviruses, which emphasizes the cooperation between transforming genes. Furthermore, some viruses encode genes that suppress or delay apoptosis.

Although immunocompromised individuals are at elevated risk, most patients infected with oncogenic viruses do not develop cancer. When cancer does develop, it usually occurs several years after the viral infection. It is estimated, for example, that the risk of hepatocellular carcinoma (HCC) among individuals infected with hepatitis C virus is 1% to 3% after 30 years.⁹¹ There may be synergy between various environmental factors and viruses in carcinogenesis.

Recognition of a viral origin for some tumors has led to the pursuit of vaccination as a preventive strategy. The use of childhood hepatitis B vaccination has already translated into a decrease in liver cancer incidence in the Far East.⁵ Similarly, it is recognized that cervical cancer and its obligate precursors, cervical intraepithelial neoplasia grades 2 and 3, and adenocarcinoma in situ, are caused by oncogenic human papillomavirus (HPV); administration of HPV vaccine to HPV-naive women, substantially reduces the incidence of HPV16/18-related cervical precancers and cervical cancer.⁹² The American Cancer Society now recommends routine HPV vaccination principally for females aged 11 to 12 years, but also for females aged 13 to 18 years to ‘’catch up’’ those who missed the opportunity to be vaccinated or who need to complete the vaccination series.⁹³

Cancer risk assessment is an important part of the initial evaluation of any patient. A patient’s cancer risk not only is an important determinant of cancer screening recommendations but also may alter how aggressively an indeterminant finding will be pursued for diagnosis. A “probably benign” mammographic lesion, for example, defined as one with <2% probability of malignancy (American College of Radiology category III) is usually managed with a 6-month follow-up mammogram in a patient at baseline cancer risk, but obtaining a tissue diagnosis may be preferable in a patient at high risk for breast cancer.⁹⁴

Cancer risk assessment starts with taking a complete history that includes history of environmental exposures to potential carcinogens and a detailed family history. Risk assessment for breast cancer, for example, includes obtaining a family history to determine whether another member of the family is known to carry a breast cancer susceptibility gene; whether there is familial clustering of breast cancer, ovarian cancer, thyroid cancer, sarcoma, adrenocortical carcinoma, endometrial cancer, brain tumors, dermatologic manifestations, leukemia, or lymphoma; and whether the patient is from a population at increased risk, such as individuals of Ashkenazi Jewish descent. Patients who have a family history suggestive of a cancer susceptibility syndrome such as hereditary breast-ovarian syndrome, LFS, or CD would benefit from genetic counseling and possibly genetic testing.

There are several models that can estimate risk based on complex family histories and assist clinicians in estimating breast cancer risk or the likelihood that a BRCA mutation is present, including the Claus model, Tyrer-Cuzick model BRCAPRO model, and the Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA) model.^95,96,97,98 Patients who do have a strong hereditary component of risk can be evaluated on the basis of their age, race, personal history, and exposures. One of the most commonly used models for risk assessment in breast cancer is the Gail model.⁹⁹ Gail and colleagues analyzed the data from 2852 breast cancer cases and 3146 controls from the Breast Cancer Detection and Demonstration Project, a mammography screening project conducted in the 1970s, and developed a model for projecting breast cancer incidence. The model uses risk factors such as an individual’s age, age at menarche, age at first live birth, number of first-degree relatives with breast cancer, number of previous breast biopsy specimens, and whether the biopsy specimen results revealed atypical ductal hyperplasia (Table 10-8).⁹⁹ This model has led to the development of a breast cancer risk assessment tool, which is available on the World Wide Web.¹⁰⁰ This tool incorporates the risk factors used in the Gail model, as well as race and ethnicity, and allows a health professional to ­project a woman’s individualized ­estimated risk for invasive breast cancer over a 5-year period and over her ­lifetime (to age 90 years). Notably, these risk projections assume that the woman is undergoing regular clinical breast examinations and screening mammograms. Also of note is that this ­program ­underestimates the risk for women who have already had a diagnosis of invasive or noninvasive breast cancer and does not take into account specific genetic predispositions such as mutations in BRCA1 or BRCA2. However, risk assessment tools such as this have been validated and are now in widespread clinical use. Similar models are in development or are being validated for other cancers. For example, a lung cancer risk prediction model, which includes age, sex, asbestos exposure history, and smoking history, has been found to predict risk of lung cancer.¹⁰¹ There is now growing interest in using each individuals genotype, such as presence or absence of single nucleotide polymorphisms which each may confer low or intermediate cancer risk. Risk models that include biological as well as environmental factors may accurately predict cancer risk, providing better guidance as to which patients should undergomore intensive screening (e.g., screening with magnetic resonance imaging of the breast, computerized tomography screening of the lung), and should be considered for preventive strategies.

Early detection is the key to success in cancer therapy. Screening for common cancers using relatively noninvasive tests is expected to lead to early diagnosis, allow more conservative surgical therapies with decreased morbidity, and potentially improve surgical cure rates and overall survival rates. Key factors that influence screening guidelines are how prevalent the cancer is in the population, what risk is associated with the screening measure, and whether early diagnosis actually affects outcome. The value of a widespread screening measure is likely to go up with the prevalence of the cancer in a population, which often determines the age cutoffs for screening and explains why screening is done only for common cancers. The risks associated with the screening measure are a significant consideration, especially with more invasive screening measures such as colonoscopy. The consequences of a false-positive screening test result also need to be considered. For example, when 1000 screening mammograms are taken, only 2 to 4 new cases of cancer will be identified; this number is slightly higher (6 to 10 prevalent cancers per 1000 mammograms) for initial screening mammograms.¹⁰² However, as many as 10% of screening mammograms may be potentially suggestive of an abnormality, which requires further imaging (i.e., a 10% recall rate). Of those women with abnormal mammogram findings, only 5% to 10% will be determined to have a breast cancer. Among women for whom biopsy specimen is recommended, 25% to 40% will have a breast cancer. A false-positive screening result is likely to induce significant emotional distress in patients, leads to unnecessary biopsy specimens, and has cost implications for the health care system.

The 2013 American Cancer Society guidelines for the early detection of cancer are listed in Table 10-9.⁹³ These guidelines are updated periodically to incorporate emerging technologies and new data on the efficacy of screening measures. Besides the American Cancer Society, several other professional bodies make recommendations for screening. Although the screening guidelines differ somewhat, most organizations do not emphasize one screening strategy as superior to another, but all emphasize the importance of age-appropriate screening.

Screening guidelines are developed for the general baseline-risk population. These guidelines need to be modified for patients who are at high risk. For example, more intensive colorectal cancer screening is recommended for individuals at increased risk because of a history of adenomatous polyps, a personal history of colorectal cancer, a family history of either colorectal cancer or colorectal adenomas diagnosed in a first-degree relative before age 60 years, a personal history of inflammatory bowel disease of significant duration, or a family history or genetic test result indicating FAP or HNPCC. For some diseases, in higher risk populations, both the screening modality and the screening intensity may be altered. For example, breast magnetic resonance imaging is recommended as an adjunct to mammography for breast cancer screening in BRCA mutation carriers, first-degree relatives of carriers, and women with a lifetime breast cancer risk of 20% to 25% or higher.¹⁰³

More recently, the National Lung Screening Trial demonstrated a 20% reduction in lung cancer deaths in adults aged 55 to 74 years who were at high risk of lung cancer and randomized to low-dose helical computed tomography (LDCT) screening compared with screening with annual CXR.¹⁰⁴ In 2013, the American Cancer Society updated their lung cancer screening recommendations to emphasize that clinicians with access to high-volume, high-quality lung cancer screening and treatment centers should ascertain the smoking history of their patients 55 to 74 years of age, and should discuss lung cancer screening with those who have at least a 30 pack-year smoking history, currently smoke, or have quit within the past 15 years, and who are in relatively good health.¹⁰⁵ It is recommended that this discussion include the benefits, uncertainties, and harms associated with screening for lung cancer with LDCT.

The definitive diagnosis of solid tumors is obtained by performing a biopsy specimen of the lesion. Biopsy findings determine the tumor histology and grade and thus, assist in definitive therapeutic planning. Biopsy specimens of mucosal lesions usually are obtained endoscopically (e.g., via colonoscope, bronchoscope, or cystoscope). Lesions that are easily palpable, such as those of the skin, can either be excised or sampled by punch biopsy ­specimen. Deep-seated lesions can be localized with computed tomographic (CT) scan or ultrasound guidance for biopsy specimen.

A sample of a lesion can be obtained with a needle or with an open incisional or excisional biopsy specimen. Fine-needle aspiration is easy and relatively safe, but has the disadvantage of not giving information on tissue architecture. For example, fine-needle aspiration biopsy specimen of a breast mass can make the diagnosis of malignancy but cannot differentiate between an invasive and noninvasive tumor. Therefore core-needle biopsy specimen is more advantageous when the histologic findings will affect the recommended therapy. Core biopsy specimen, like fine-needle aspiration, is relatively safe and can be performed either by direct palpation (e.g., a breast mass or a soft tissue mass) or can be guided by an imaging study (e.g., stereotactic core biopsy specimen of the breast). Core biopsy specimens, like fine-needle aspirations, have the disadvantage of introducing sampling error. For example, 19% to 44% of patients with a diagnosis of atypical ductal hyperplasia based on core biopsy specimen findings of a mammographic abnormality are found to have carcinoma upon excision of the lesion.¹⁰⁶ It is crucial to ensure that the histologic findings are consistent with the clinical scenario and to know the appropriate interpretation of each histologic finding. A needle biopsy specimen for which the report is inconsistent with the clinical scenario should be either repeated or followed by an open biopsy specimen.

Open biopsy specimens have the advantage of providing more tissue for histologic evaluation and the disadvantage of being an operative procedure. Incisional biopsy specimens are reserved for very large lesions in which a definitive diagnosis cannot be made by needle biopsy specimen. Excisional biopsy specimens are performed for lesions for which either core biopsy specimen is not possible or the results are nondiagnostic. Excisional biopsy specimens should be performed with curative intent, that is, by obtaining adequate tissue around the lesion to ensure negative surgical margins. Marking of the orientation of the margins by sutures or clips by the surgeon and inking of the specimen margins by the pathologist will allow for determination of the surgical margins and will guide surgical re-excision if one or more of the margins are positive for microscopic tumor or are close. The biopsy specimen incision should be oriented to allow for excision of the biopsy specimen scar if repeat operation is necessary. Furthermore, the biopsy specimen incision should directly overlie the area to be removed rather than tunneling from another site, which runs the risk of contaminating a larger field. Finally, meticulous hemostasis during a biopsy specimen is essential, because a hematoma can lead to contamination of the tissue planes and can make subsequent follow-up with physical examinations much more challenging.

Cancer staging is a system used to describe the anatomic extent of a malignant process in an individual patient. Staging systems may incorporate relevant clinical prognostic factors such as tumor size, location, extent, grade, and dissemination to regional lymph nodes or distant sites. Accurate staging is essential in designing an appropriate treatment regimen for an individual patient. Staging of the lymph node basin is considered a standard part of primary surgical therapy for most surgical procedures and is discussed later in this chapter. Cancer patients who are considered to be at high risk for distant metastasis ­usually undergo a preoperative staging work-up. This involves a set of imaging studies of sites of preferential metastasis for a given cancer type. For a patient with breast cancer, for example, a staging work-up would include a chest radiograph, bone scan, and liver ultrasound, or CT scan of the abdomen to evaluate for lung, bone, and liver metastases, respectively. A distant staging work-up usually is performed only for patients likely to have metastasis based on the characteristics of the primary tumor; for example, a staging work-up for a patient with ductal carcinoma in situ of the breast or a small invasive breast tumor is likely to be low yield and not cost effective.

Recently there also is interest in using molecular imaging with positron emission tomography (PET) scanning, or PET/CT, for cancer staging. Most commonly PET scanning is performed with fluorine 18 incorporated into fluorodeoxyglucose (FDG). FDG PET assesses the rate of glycolysis. FDG uptake is increased in most malignant tissues but also in benign pathologic conditions such as inflammatory disorders, trauma, infection, and granulomatous disease. It may be especially useful in the staging and management of lymphoma, lung cancer, and colorectal cancer. The role of PET in evaluating many other cancers is evolving, and additional molecular tracers, such as 3′-deoxy-3′-¹⁸ F-fluorothymidine, used to assess proliferation, are being actively pursued.

Standardization of staging systems is essential to allow comparison of results from different studies from different institutions and worldwide. The staging systems proposed by the American Joint Committee on Cancer (AJCC) and the Union Internationale Contre le Cancer (International Union Against Cancer, or UICC) are among the most widely accepted staging systems. Both the AJCC and the UICC have adopted a shared tumor, node, and metastasis (TNM) staging system that defines the cancer in terms of the anatomic extent of disease and is based on assessment of three components: the size of the primary tumor (T), the presence (or absence) and extent of nodal metastases (N), and the presence (or absence) and extent of distant metastases (M).

The TNM staging applies only to tumors that have been microscopically confirmed to be malignant. Standard TNM staging (clinical and pathologic) is completed at initial diagnosis. Clinical staging (cTNM or TNM) is based on information gained up until the initial definitive treatment. Pathologic staging (pTNM) includes clinical information and information obtained from pathologic examination of the resected primary tumor and regional lymph nodes. Other classifications, such as retreatment staging (rTNM) or autopsy staging (aTNM), should be clearly identified as such.

The clinical measurement of tumor size (T) is the one judged to be the most accurate for each individual case based on physical examination and imaging studies. For example, in breast cancer the size of the tumor could be obtained from a physical examination, mammogram, or ultrasound, and the tumor size is based only on the invasive component.

If even one lymph node is involved by tumor, the N component is at least N1. For many solid tumor types, simply the absence or presence of lymph node involvement is recorded, and the tumor is categorized either as N0 or N1. For other tumor types, the number of lymph nodes involved, the size of the lymph nodes or the lymph node metastasis, or the regional lymph node basin involved also has been shown to have prognostic value. In these cancers, the designations N1, N2, and N3 suggest an increasing abnormality of lymph nodes based on size, characteristics, and location. NX indicates that the lymph nodes cannot be fully assessed.

Cases in which there is no distant metastasis are designated M0, cases in which one or more distant metastases are detected are designated M1, and cases in which the presence of distant metastasis cannot be assessed are designated MX. In clinical practice, negative findings on clinical history and examination are sufficient to designate a case as M0. However, in clinical trials, routine follow-up often are performed to standardize the detection of distant metastases.

The practice of dividing cancer cases into groups according to stage is based on the observation that the survival rates are higher for localized (lower-stage) tumors than for tumors that have extended beyond the organ of origin. Therefore, staging assists in selection of therapy, estimation of prognosis, evaluation of treatments, and exchange of information among treatment centers. Notably, the AJCC regularly updates its staging system to incorporate advances in prognostic technology to improve the predictive accuracy of the TNM system. Therefore it is important to know which revision of a staging system is being used when evaluating studies.

Prognostic and Predictive Tissue Markers

Tumor markers are substances that can be detected in higher than normal amounts in the serum, urine, or tissues of patients with certain types of cancer. Tumor markers are produced either by the cancer cells themselves or by the body in a response to the cancer.

Over the past decade, there has been an especially high interest in identifying tissue tumor markers that can be used as prognostic or predictive markers. Although the terms prognostic marker and predictive marker are sometimes used interchangeably, the term prognostic marker generally is used to describe molecular markers that predict disease-free survival, disease-specific survival, and overall survival, whereas the term predictive marker often is used in the context of predicting response to certain therapies.

The goal is to identify prognostic markers that can give information on prognosis independent of other clinical characteristics and therefore can provide information to supplement the projections based on clinical presentation. This would allow practitioners to further classify patients as being at higher or lower risk within clinical subgroups and to identify patients who may benefit most from adjuvant therapy. For example, ideal prognostic tumor markers would be able to help determine which patients with node-negative breast cancer are at higher risk of relapse so that adjuvant systemic therapy could be given only to that group. However, although a large number of studies have identified potential novel prognostic markers, most have not been tested with enough vigor to be shown to be of clinical utility. In the 2007 American Society of Clinical Oncology (ASCO) guidelines, it was decided that level of uPA/PAI-1 measured by enzyme-linked immunosorbent assay could be used to determine prognosis in cases of newly diagnosed ­node-negative breast cancer.¹⁰⁷ In contrast, the data for many other markers, including DNA content, proportion of tumor cells in S phase, Ki-67, cyclin E, p27, p21, thymidine kinase, topoisomerase II, HER2, p53, and cathepsin D, were felt to be insufficient to support their use in the management of breast cancer patients.¹⁰⁷ Similarly, in the 2006 ASCO GI tumor guidelines, the data were felt to be insufficient to recommend the routine use of p53, ras, thymidine synthase, dihydropyrimidine dehydrogenase, thymidine phosphorylase, microsatellite instability, 18q loss of heterozygosity, or deleted-in-colon-cancer protein in the management of patients with colorectal cancer.¹⁰⁸

Predictive markers are markers that can prospectively identify patients who will benefit from a certain therapy. For example in breast cancer, estrogen receptor (ER) and HER2 assessment can identify patients who can benefit from antiestrogen therapies (e.g., tamoxifen) and anti-HER2 targeted therapies (e.g., trastuzumab), respectively, and the 2007 ASCO guidelines recommend that these markers be routinely assessed.¹⁰⁷ High-throughput techniques such as transcriptional profiling allow for assessment of the relative mRNA levels of thousands of genes simultaneously in a given tumor using microarray technology. With the advent of such molecular profiling technologies, researchers have focused on identifying expression profiles that are prognostic for different cancer types. For breast cancer, although many such multiparameter tests are under development, few have reached the large-scale validation stage.¹⁰⁹ In 2007, ASCO guidelines suggested that one of these, the Oncotype DX assay, can be used to predict recurrence in women with node-negative, ER-positive breast cancer who are treated with tamoxifen.¹⁰⁷ Oncotype DX is a quantitative reverse-transcriptase polymerase chain reaction (RT-PCR) test that used paraffin-fixed tissue. A 21-gene recurrence score (RS) is generated based on the expression of 16 cancer genes and 5 reference genes. The levels of expression are used to derive an RS that ranges from 0 to 100, using a prospectively defined mathematical algorithm. This novel quantitative approach to the evaluation of the best-known molecular pathways in breast cancer has produced impressive results. Use of this multigene assay to predict recurrence was validated in the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-14 trial, in which ER-positive, node-negative patients had received tamoxifen.¹¹⁰ By multivariate Cox proportional analysis, RS was found to be independently associated with recurrence risk, with a hazard ratio of 3.21 (95% confidence interval of 2.23 to 4.65, P<.001). The RS was indeed able to stratify patients by freedom from distant recurrence (Fig. 10-13).¹¹⁰ The Trial Assessing Individualized Options for Treatment for breast cancer (TAILORx) is evaluating the utility of Oncotype DX for predicting prognosis in patients with ER-positive, node-negative tumors and will focus on women with intermediate RS scores in whom the role of chemotherapy is unclear. Several other multigene predictors for breast cancer are available including MammaPrint, a gene expression profiling platform assessing a 70-gene transcriptional signature.¹¹¹ This assay was approved by the Food and Drug Administration (FDA) in February 2007. The usefulness of this assay in making therapy-related decisions is being tested prospectively in a large-scale study, the Microarray in Node-Negative Disease May Avoid Chemotherapy (MINDACT) trial.

Multigene profiles to predict prognosis are in development or in validation phases for many other solid tumor types, including lung cancer, ovarian cancer, pancreatic cancer, colorectal cancer, and melanoma. Gene signatures and genomic alterations also are being studied for their ability to predict response to specific chemotherapy regimens or targeted therapies. Many of these multigene marker sets will likely be incorporated into clinical practice in the years to come.

Serum Markers

Serum markers are under active investigation because they may allow early diagnosis of a new cancer or may be used to follow cancer response to therapy or monitor for recurrence. Unfortunately, identification of serum markers of clinical value has been challenging. Many of the tumor markers proposed so far have had low sensitivities and specificities.¹⁰⁹ Tumor marker levels may not be elevated in all patients with cancer, especially in the early stages, when a serum marker would be most useful for diagnosis. Therefore when a tumor marker is used to monitor recurrence, it is important to be certain that the level of the tumor marker was elevated before primary therapy. Moreover, tumor marker levels can be elevated in benign conditions. Many tumor markers are not specific for a certain type of cancer and can be elevated with more than one type of tumor. Since there may be significant laboratory variability, it is important to obtain serial results from the same laboratory. In spite of these many clinical limitations, several serum markers are in clinical use. A few of the commonly measured serum tumor markers are discussed in the following sections.

Prostate-Specific Antigen

Prostate-specific antigen (PSA) is an androgen-regulated serine protease produced by the prostate epithelium. PSA is normally present in low concentrations in the blood of all adult males. PSA levels may be elevated in the blood of men with benign prostate conditions such as prostatitis and benign prostatic hyperplasia, as well as in men with prostate cancer. PSA levels have been shown to be useful in evaluating the effectiveness of prostate cancer treatment and monitoring for recurrence after therapy. In monitoring for recurrence, a trend of increasing levels is considered more significant than a single absolute elevated value.

Although PSA has been widely used for prostate cancer screening, the utility of PSA screening remains controversial. There is concern that the number of men who avoid dying from prostate cancer due to screening is small, while the harms relatedto the treatment of screen-detected cancers, including incontinence and erectile dysfunction are at least moderate. In 2012, the US Preventive Services Task Force concluded with moderate certainty that the harms of PSA testing outweigh the benefits and on that basis recommended against PSA-based screening for all men.¹¹² In 2010, the American Cancer Society updated its guidelines for the early detection of prostate cancer to state that men who have at least a 10-year life expectancy should have an opportunity to make an informed decision with their health care provider about whether to be screened for prostate cancer with digital rectal exam and serum PSA, after receiving information about the benefits, risks, and uncertainties associated with prostate cancer screening;¹¹³ this recommendation was reinforced in their 2013 guidelines.⁹³

Carcinoembryonic Antigen

Carcinoembryonic antigen (CEA) is a glycoprotein found in the embryonic endodermal epithelium. Elevated CEA levels have been detected in patients with primary colorectal cancer as well as in patients with breast, lung, ovarian, prostate, liver, and pancreatic cancer. Levels of CEA also may be elevated in benign conditions, including diverticulitis, peptic ulcer disease, bronchitis, liver abscess, and alcoholic cirrhosis, especially in smokers and in elderly persons.

CEA measurement is most commonly used in the management of colorectal cancer. However, the appropriate use of CEA testing in patients with colorectal cancer has been debated. Use of CEA level as a screening test for colorectal cancer is not recommended. CEA levels may be useful if obtained preoperatively and postoperatively in patients with a diagnosis of colorectal cancer. Preoperative elevation of CEA level is an indicator of poor prognosis. However, the 2007 ASCO clinical practice guidelines state that the data are insufficient to support the use of CEA to determine whether to give a patient adjuvant therapy; the data are stronger for the use of CEA for monitoring for postoperative recurrence.¹⁰⁷ CEA measurement is the most cost-effective approach for detecting metastasis, with 64% of recurrences being detected first by an elevation in CEA level. Therefore, in cases in which the patient would be a candidate for resection of recurrent colorectal cancer or systemic therapy, the 2006 ASCO guidelines recommend that postoperative CEA testing be performed every 3 months in patients with stage II or III disease for at least 3 years.¹⁰⁸ CEA is the marker of choice for monitoring metastatic colorectal cancer during systemic therapy.¹⁰⁸

There is also interest in using CEA levels for monitoring patients with breast cancer. However, the 2007 ASCO guidelines state that the routine use of CEA for screening, diagnosis, staging, or surveillance of breast cancer is not recommended because available data are insufficient.¹⁰⁷ For monitoring patients during active therapy, CEA can be used in conjunction with diagnostic imaging and history and physical examination.¹⁰⁷ In the absence of measurable disease, an increase in CEA level may be taken to indicate treatment failure. However, caution is advised when interpreting rising levels in the first 4 to 6 weeks of therapy.¹⁰⁷

Alpha-Fetoprotein. Alpha-fetoprotein (AFP) is a glycoprotein normally produced by a developing fetus. AFP levels decrease soon after birth in healthy adults. An elevated level of AFP suggests the presence of either primary liver cancer or a germ cell tumor of the ovary or testicle. Rarely, other types of cancer such as gastric are associated with an elevated AFP level. Benign conditions that can cause elevations of AFP include cirrhosis, hepatic necrosis, acute hepatitis, chronic active hepatitis, ataxia-telangiectasia, Wiskott-Aldrich syndrome, and pregnancy.¹¹⁴

The sensitivity of an elevated AFP level for detecting HCC is approximately 60%. AFP is considered to be sensitive and specific enough to be used for screening for HCC in high-risk populations. Current consensus recommendations are to screen healthy hepatitis B virus carriers with annual or semiannual measurement of AFP level and to screen carriers with cirrhosis or chronic hepatitis and patients with cirrhosis of any etiology with twice-yearly measurement of AFP level and liver ultrasonography.¹¹⁵ Although AFP testing has been used widely for a long time, its efficacy in early diagnosis of HCC is limited. With improvements in imaging technology, a larger proportion of patients diagnosed with HCC are now AFP seronegative.

Cancer Antigen 19-9. Cancer antigen 19-9 (CA 19-9) is a tumor-related antigen that was originally defined by a monoclonal antibody produced by a hybridoma prepared from murine spleen cells immunized with a human colorectal cancer cell line.¹⁰⁸ The data are insufficient to recommend use of CA 19-9 for screening, diagnosis, surveillance, or monitoring of therapy for colon cancer.¹⁰⁸ Based on the 2006 ASCO guidelines, there are also insufficient data to recommend use of CA 19-9 for screening, diagnosis, or determination of the operability of pancreatic cancer.¹⁰⁸ However, for patients with locally advanced or metastatic cancer receiving active therapy, CA 19-9 can be measured at the start of therapy and every 1 to 3 months while therapy is given; elevations in serial CA 19-9 levels may indicate progressive disease and should be confirmed by additional studies.¹⁰⁸

Cancer Antigen 15-3. Cancer antigen 15-3 (CA 15-3) is an epitope of a large membrane glycoprotein encoded by the MUC1 gene that tumor cells shed into the bloodstream. The CA 15-3 epitope is recognized by two monoclonal antibodies in a sandwich radioimmunoassay. CA 15-3 levels are most useful in following the course of treatment in women diagnosed with advanced breast cancer. CA 15-3 levels are infrequently elevated in early-stage breast cancer. CA 15-3 levels can be increased in benign conditions such as chronic hepatitis, tuberculosis, sarcoidosis, pelvic inflammatory disease, endometriosis, systemic lupus erythematosus, pregnancy, and lactation, and in other types of cancer such as lung, ovarian, endometrial, and GI cancers.

The sensitivity of CA 15-3 is higher for metastatic disease, and in these cases studies have shown sensitivity to be between 54% and 87%, with specificity as high as 96%. This has led to interest in using CA 15-3 for monitoring patients with advanced breast cancer for recurrence. Elevated CA 15-3 levels have been reported before relapse in 54% of patients, with a lead time of 4.2 months. Therefore, detection of elevated CA 15-3 levels during follow-up should prompt evaluation for recurrent disease. However, 6% to 8% of patients without recurrence will have elevated CA 15-3 levels that require evaluation. Furthermore, monitoring with the use of CA 15-3 levels has shown no demonstrated impact on survival. Therefore, the 2007 ASCO guidelines state that the routine use of CA 15-3 for screening, diagnosis, staging, or surveillance of breast cancer is not recommended because available data are insufficient.¹⁰⁷ For monitoring patients during active therapy, CA 15-3 can be used in conjunction with diagnostic imaging and history and physical examination.¹⁰⁷ In the absence of measurable disease, an increase may be interpreted to indicate treatment failure. However, caution is advised when interpreting rising levels in the first 4 to 6 weeks of therapy.¹⁰⁷

Cancer Antigen 27-29. The MUC-1 gene product in the serum may be quantitated by using radioimmunoassay with a monoclonal antibody against the cancer antigen 27-29 (CA 27-29). CA 27-29 levels can be elevated in breast cancer as well as in cancers of the colon, stomach, kidney, lung, ovary, pancreas, uterus, and liver. First-trimester pregnancy, endometriosis, benign breast disease, kidney disease, and liver disease also may be associated with elevated CA 27-29 levels.

CA 27-29 has been reported to have a sensitivity of 57%, a specificity of 98%, a positive predictive value of 83%, and a negative predictive value of 93% in detecting breast cancer recurrences.¹¹⁶ Although CA 27-29 has been found to predict recurrence an average of 5.3 months before other symptoms or tests, testing of CA 27-29 levels has not been demonstrated to affect disease-free and overall survival rates.^116,117 Therefore, the 2007 ASCO guidelines state that, as with CA 15-3, the routine use of CA 27-29 for screening, diagnosis, staging, or surveillance of breast cancer is not recommended because available data are insufficient.¹⁰⁷ CA 27-29 levels can be used together with diagnostic imaging and history and physical examination to monitor patients during active therapy.¹⁰⁷ When no measurable disease is present, an increase in level may be considered to indicate treatment failure. However, rising levels in the first 4 to 6 weeks of therapy should be interpreted with caution.¹⁰⁷

Circulating Tumor Cells

Circulating tumor cells (CTCs) are cells present in the blood that possess antigenic or genetic characteristics of a specific tumor type.¹⁰⁷ One CTC detection methodology is capture and quantitation of CTCs with immunomagnetic beads coated with antibody specific for cell-surface, epithelial, or cancer antigens. Another methodology used to detect cancer cells in the peripheral blood is RT-PCR. It has been suggested that measurement of CTCs can be an effective tool for selecting patients who have a high risk of relapse and for monitoring efficacy of cancer ­therapy.

CTCs have probably been most extensively studied in breast cancer.¹⁰⁷ The most promising data come from the use of CTC measures in metastatic breast cancer. In a prospective multicenter trial, the number of CTCs (≥5 CTCs vs. <5 CTCs per 7.5 mL of whole blood) before treatment of metastatic breast cancer was an independent predictor of progression-free and overall survival rates.¹¹⁸ The presence of >5 CTCs after the first course of therapy predicted lack of response to treatment. This technology, known as CellSearch, has been approved by the FDA for clinical use. Further, in a recent single institutional study, detection of one or more CTCs in Stage I-III breast cancer patients was associated with both decreased progression-free survival and overall survival.¹¹⁹

However, there is limited data to prove that the use of CTC testing leads to improved survival or improved quality of life; thus the ASCO 2007 guidelines update did not recommend the use of CTC measurement in any clinical setting.¹⁰⁷ The clinical utility of measuring CTC response to initial therapy is now being tested prospectively in a multicenter clinical trial. The use of CTC levels as a tool in treating many other types of tumor is also under active investigation.

The prognostic implications of detection of CTCs by RT-PCR have been intensively studied for melanoma. In the recent multicenter Sunbelt Melanoma Trial, serial RT-PCR was performed on peripheral blood samples using four markers—tyrosinase, melanoma antigen reacting to T cell (MART-1), melanoma antigen 3 (MAGE3), and gp 100—to detect occult melanoma cells in the bloodstream.¹²⁰ Although there were no differences in survival between patients in whom at least one marker was detected and those in whom no markers were detected, the disease-free survival and distant disease-free survival were worse for patients in whom more than one marker was detected at any time during follow-up.¹²⁰ The detection of occult cancer cells with RT-PCR remains investigational, however, and is not used to direct therapy for melanoma and other cancer types at this time.

Bone Marrow Micrometastases

Micrometastatic disease in the bone marrow, also referred to as minimal residual disease, also is being investigated as a potential prognostic marker. Bone marrow micrometastatic disease usually is detected by staining bone marrow aspirates with monoclonal antibodies to cytokeratin, but other methodologies such as flow cytometry and RT-PCR are being explored. Breast cancer patients with bone marrow micrometastasis have larger tumors, tumors with a higher histologic grade, more lymph node metastases, and more hormone receptor-negative tumors than patients without bone marrow micrometastasis. In 4700 patients with stage I, II, or III breast cancer, micrometastasis was a significant prognostic factor associated with poor overall survival, breast cancer-specific survival, disease-free survival, and distant disease-free survival during a 10-year observation period.¹²¹ Recently, in the American College of Surgeons Oncology Group Z0010 trial enrolled women with clinical T1 to T2N0M0 invasive breast carcinoma in a prospective observational study to determine to determine the association between survival and metastases detected by immunochemical staining of bone marrow specimens from patients with early-stage breast cancer.¹²² Of 3413 bone marrow specimens examined by immunocytochemistry, only 104 (3.0%) were positive for tumor. Bone marrow involvement was associated with a decreased overall survival but this association was not significant on multivariable analysis. The prognostic implication of bone marrow involvement is also being studied by the National Surgical Adjuvant Breast and Bowel Project Protocol BP-59.

At this time the routine use of bone marrow testing is not recommended.¹⁰⁷ Ongoing clinical trials are evaluating the role of routine assessment of bone marrow status in the care of patients with early and advanced breast cancer. The utility of assessment of bone marrow micrometastasis is also being evaluated in other tumor types, including gastric, esophageal, colorectal, lung, cervical, and ovarian cancer.¹²³

Multidisciplinary Approach to Cancer

Although surgery is an effective therapy for most solid tumors, patients who die from cancer usually die of metastatic disease. Therefore, to improve patient survival rates, a multimodality approach, including systemic therapy and radiation therapy is key for most tumors. It is important that surgeons involved in cancer care not only know the techniques for performing a cancer operation but also know the alternatives to surgery and be well versed in reconstructive options. It is also crucial that the surgeon be familiar with the indications for and complications of preoperative and postoperative chemotherapy and radiation therapy. Although the surgeon may not be delivering these other therapies, as the first physician to see a patient with a cancer diagnosis, he or she is ultimately responsible for initiating the appropriate consultations. For this reason, the surgeon often is responsible for determining the most appropriate adjuvant ­therapy for a given patient as well as the best sequence for therapy. In most instances, a multidisciplinary approach beginning at the patient’s initial presentation is likely to yield the best result.

Surgical Management of Primary Tumors

The goal of surgical therapy for cancer is to achieve oncologic cure. A curative operation presupposes that the tumor is confined to the organ of origin or to the organ and the regional lymph node basin. Patients in whom the primary tumor is not resectable with negative surgical margins are considered to have inoperable disease. The operability of primary tumors is best determined before surgery with appropriate imaging studies that can define the extent of local-regional disease. For example, a preoperative thin-section CT scan is obtained to determine resectability of pancreatic cancer, which is based on the absence of extrapancreatic disease, the absence of tumor extension to the superior mesenteric artery and celiac axis, and a patent superior mesenteric vein-portal vein confluence.¹²⁴ Disease involving multiple distant metastases is deemed inoperable because it is usually not curable with surgery of the primary tumor. Therefore patients who are at high risk of having distant metastasis should undergo a staging work-up before surgery for the primary tumor. On occasion, primary tumors are resected in these patients for palliative reasons, such as improving the quality of life by alleviating pain, infection, or bleeding. An example of this is toilet mastectomies for large ulcerated breast tumors. Patients with limited metastases from a primary tumor on occasion are considered surgical candidates if the natural history of isolated distant metastases for that cancer type is favorable or the potential complications associated with leaving the primary tumor intact are significant.

In the past it was presumed that the more radical the surgery, the better the oncologic outcome would be. Over the past three decades, this has been recognized as not necessarily being true, which has led to more conservative operations, with wide local excisions replacing compartmental resections of sarcomas, and partial mastectomies, skin-sparing mastectomies, and breast-conserving therapies replacing radical mastectomies for breast cancer. The uniform goal for all successful oncologic operations seems to be achieving widely negative margins with no evidence of macroscopic or microscopic tumor at the surgical margins. The importance of negative surgical margins for local tumor control and/or survival has been documented for many tumor types, including sarcoma, breast cancer, pancreatic cancer, and rectal cancer. Thus it is clear that every effort should be made to achieve microscopically negative surgical margins. Inking of the margins, orientation of the specimen by the surgeon, and immediate gross evaluation of the margins by a pathologist using frozen-section analysis when necessary may assist in achieving negative margins at the first operation. In the end, although radiation therapy and systemic therapy can assist in decreasing local recurrence rates in the setting of positive margins, adjuvant therapy cannot substitute for adequate surgery.

Although it is clear that the surgical gold standard is negative surgical margins, the appropriate surgical margins for optimal local control are controversial for many cancer types. In contrast, in melanoma the optimal margin width for any tumor depth has been better defined, owing to the systematic study of this question in randomized clinical trials.^125,126 Although such randomized studies may not be possible for all tumor types, it is important to determine optimum surgical margins for each cancer type so that adjuvant radiation and systemic therapy can be offered to patients deemed to be at increased risk for local treatment failure. There are also ongoing studies on approaches to assess margins intraoperatively, to allow immediate intraoperative reexcisions as needed, and maximizing local control.

Surgical Management of the Regional Lymph Node Basin

Most neoplasms have the ability to metastasize via the lymphatics. Therefore, most oncologic operations have been designed to remove the primary tumor and draining lymphatics en bloc. This type of operative approach usually is undertaken when the lymph nodes draining the primary tumor site lie adjacent to the tumor bed, as is the case for colorectal cancers and gastric cancers. For tumors in which the regional lymph node basin is not immediately adjacent to the tumor (e.g., melanomas), lymph node surgery can be performed through a separate incision. Unlike most carcinomas, soft tissue sarcomas rarely metastasize to the lymph nodes (<5%); therefore lymph node surgery usually is not necessary.

It is generally accepted that a formal lymphadenectomy is likely to minimize the risk of regional recurrence of most cancers. For example, the introduction of total mesorectal excision of rectal cancer has been associated with a large decline in local-regional recurrence, and this procedure has become the new standard of operative management.¹²⁷ On the other hand, there have been two opposing views regarding the role of lymphadenectomy in survival of cancer patients. The traditional Halsted view states that lymphadenectomy is important for staging and survival. The opposing view counters that cancer is systemic at inception and that lymphadenectomy, although useful for staging, does not affect survival. For most cancers, involvement of the lymph nodes is one of the most significant prognostic factors. Interestingly, in some studies removal of a larger number of lymph nodes has been found to be associated with an improved overall survival rate for many tumors, including breast cancer, colon cancer, and lung cancer. Although this seems to support the Halsted theory that more extensive lymphadenectomy yielding of nodes reduces the risk of regional recurrence, there may be alternative explanations for the same finding. For example, the surgeon who performs a more extensive lymphadenectomy may obtain wider margins around the tumor or even provide better overall care, such as ensuring that patients receive the appropriate adjuvant therapy or undergo a more thorough staging work-up. Alternatively, the pathologist may perform a more thorough examination, identifying more nodes and more accurately staging the nodes. The effect of appropriate staging on survival is twofold. Patients with nodal metastases may be offered adjuvant therapy, which improves their survival chances. Further, the improved staging can improve perceived survival rates through a “Will Rogers effect”; that is, identification of metastases that had formerly been silent and unidentified leads to stage migration and thus to a perceived improvement in chances of survival. Clearly the impact of lymphadenectomy on survival will not be easily resolved.

Surgical management of the clinically negative regional lymph node basin has evolved with the introduction of lymphatic mapping technology (Fig. 10-14).¹²⁸ Lymphatic mapping and sentinel lymph node biopsy specimen were first reported in 1977 by Cabanas for penile cancer.¹²⁹ Now, sentinel node biopsy specimen is the standard of care for the management of melanoma and breast cancer. Moreover, the utility of sentinel node biopsy specimen in other cancer types is being explored.

The first node to receive drainage from the tumor site is termed the sentinel node. This node is the node most likely to contain metastases, if metastases to that regional lymph node basin are present. The goal of lymphatic mapping and sentinel lymph node biopsy specimen is to identify and remove the lymph node most likely to contain metastases in the least invasive fashion. The practice of sentinel lymph node biopsy specimen followed by regional lymph node dissection for selected patients with a positive sentinel lymph node avoids the morbidity of lymph node dissections in patients with negative nodes. An additional advantage of the sentinel lymph node technique is that it directs attention to a single node, which allows more careful analysis of the lymph node most likely to have a positive yield and increases the accuracy of nodal staging. Two criteria are used to assess the efficacy of a sentinel lymph node biopsy specimen: the sentinel lymph node identification rate and the false-negative rate. The sentinel lymph node identification rate is the proportion of patients in whom a sentinel lymph node was identified and removed among all patients undergoing an attempted sentinel lymph node biopsy specimen. The false-negative rate is the proportion of patients with regional lymph node metastases in whom the sentinel lymph node was found to be negative. False-negative biopsy specimen results may be due to identifying the wrong node or to missing the sentinel node (i.e., surgical error) or they may be due to the cancer cells’ establishing metastases not in the first node encountered but in a second-echelon node (i.e., biologic variation). Alternatively, false-negative biopsy specimen results may be due to inadequate histologic evaluation of the lymph node. The false-negative rates for sentinel lymph node biopsy specimen in study series range between 0% and 11%. Both increases in the identification rate and decreases in the false-negative rate have been observed as surgeons gain experience with the technique.

Lymphatic mapping is performed by using isosulfan blue dye, technetium-labeled sulfur colloid or albumin, or a combination of both techniques to detect sentinel nodes. The combination of blue dye and technetium has been reported to improve the capability of detecting sentinel lymph nodes. The nodal drainage pattern usually is determined with a preoperative lymphoscintigram, and the “hot” and/or blue nodes are identified with the assistance of a gamma probe and careful nodal basin exploration. Careful manual palpation is a crucial part of the procedure to minimize the false-negative rate.

The nodes are evaluated with serial sectioning, hematoxylin and eosin staining, and immunohistochemical analysis with S-100 protein and homatropine methylbromide staining for melanoma and cytokeratin staining for breast cancer. The utility of molecular techniques such as RT-PCR to assess the sentinel nodes is still being explored.

Another area of active investigation is the prognostic value of minimal nodal involvement. For example, in breast cancer, nodes with isolated tumor cell deposits of <0.2 mm (also called nanometastasis) are considered to be N0 by the sixth edition of the AJCC staging manual. However, some retrospective studies have suggested that even this amount of nodal disease burden has negative prognostic implications.¹³⁰ Molecular ultrastaging with RT-PCR for patients with node-negative disease was assessed in a prospective multicenter trial and was found not to be prognostic in malignant melanoma.¹²⁰ However, a recent meta-analysis of 22 studies enrolling 4019 patients found that PCR positivity was associated with worse overall and disease-free survival.¹³¹ Further study of the utility of ultrastaging of nodes in breast cancer, melanoma, and several other tumor types is ongoing.

Until recently, in breast cancer management, when sentinel node mapping revealed a positive sentinel node, this was followed by a completion axillary lymph node dissection. Recently results of the American College of Surgeons Oncology Group Z0011 trial, challenged this practice. ACOSOG Z11 was a phase 3 multicenter noninferiority trial conducted to determine the effects of complete axillary lymph node dissectionon survival of patients with sentinel lymph node metastasis of breast cancer.¹²² Patients were women with clinical T1-T2 invasive breast cancer, no palpable adenopathy, and 1 to 2 SLNs containing metastases identified by frozen section, touch preparation, or hematoxylin-eosin staining on permanent section. All patients underwent breast-conserving surgery and tangential whole-breast irradiation. Those with sentinel node metastases identified by sentinel node biopsy specimen were randomized to undergo axillary lymph node dissection or no further axillary treatment. At a median follow-up of 6.3 years, 5-year overall survival was 91.8% (95% confidence interval [CI], 89.1%–94.5%) with axillary lymph node dissection and 92.5% (95% CI, 90.0%–95.1%) with sentinel node alone. The 5-year disease-free survival was 82.2% (95% CI, 78.3%–86.3%) with axillary lymph node ­dissection, and 83.9% (95% CI, 80.2%–87.9%) with sentinel node alone. Thus ACOSOGZ11 demonstrated that among breast cancer patients with limited sentinel node metastasis treated with breast conservation and systemic therapy, the use of sentinel node alone compared with axillary lymph node dissection did not result in inferior survival. This study challenges the traditional surgical dictum of regional management, and has led to a selective utilization of completion axillary lymph node dissection in breast cancer patients undergoing breast conservation. The role of completion lymph node dissections in melanoma is under investigation.

Surgical Management of Distant Metastases

The treatment of a patient with distant metastases depends on the number and sites of metastases, the cancer type, the rate of tumor growth, the previous treatments delivered and the responses to these treatments, and the patient’s age, physical condition, and desires. Although once a tumor has metastasized it usually is not curable with surgical therapy, such therapy has resulted in cure in selected cases with isolated metastases to the liver, lung, or brain.

Patient selection is the key to the success of surgical therapy for distant metastases. The cancer type is a major determinant in surgical decision making. A liver metastasis from a colon cancer is more likely to be an isolated and thus resectable lesion than a liver metastasis from a pancreatic carcinoma. The growth rate of the tumor also plays an important role and can be determined in part by the disease-free interval and the time between treatment of the primary tumor and detection of the distant recurrence. Patients with longer disease-free intervals have a higher survival rate after surgical metastasectomy than those with a short disease-free interval. Similarly, patients who have synchronous metastases (metastases diagnosed at the initial cancer diagnosis) do worse after metastasectomy than patients who develop metachronous metastases (metastasis diagnosed after a disease-free interval). The natural history of metastatic disease is so poor for some tumors (e.g., pancreatic cancer) that there is no role at this time for surgical metastasectomy. In cancers with a more favorable outlook, observation for several weeks or months, potentially with initial treatment with systemic therapy, can allow the surgeon to monitor for metastases at other sites.

In curative surgery for distant metastases, as with surgery for primary tumors, the goal is to resect the metastases with negative margins. In patients with hepatic metastases that are unresectable because their location near intrahepatic blood vessels precludes a margin-negative resection, or because they are multifocal or hepatic function is inadequate, tumor ablation with cryotherapy or radiofrequency ablation is an alternative.^132,133 Curative resections or ablative procedures should be attempted only if the lesions are accessible and the procedure can be ­performed safely.

Clinical Use of Chemotherapy

In patients with documented distant metastatic disease, chemotherapy is usually the primary modality of therapy. The goal of therapy in this setting is to decrease the tumor burden, thus prolonging survival. It is rare to achieve cure with chemotherapy for metastatic disease for most solid tumors. Chemotherapy administered to a patient who is at high risk for distant recurrence but has no evidence of distant disease is referred to as adjuvant chemotherapy. The goal of adjuvant chemotherapy is eradication of micrometastatic disease, with the intent of decreasing relapse rates and improving survival rates.

Adjuvant therapy can be administered after surgery (postoperative chemotherapy) or before surgery (preoperative chemotherapy, neoadjuvant chemotherapy, or induction therapy). A portion or all of the planned adjuvant chemotherapy can be administered before the surgical removal of the primary tumor. Preoperative chemotherapy has three potential advantages. The first is that preoperative regression of tumor can facilitate resection of tumors that were initially inoperable or allow more conservative surgery for patients whose cancer was operable to begin with. In the NSABP B-18 project, for example, women were randomly assigned to receive adjuvant doxorubicin and cyclophosphamide preoperatively or postoperatively. More patients treated before surgery than after surgery underwent breast-conserving surgery (68% vs. 60%).¹³⁴ The second advantage of preoperative chemotherapy is the treatment of micrometastases without the delay of postoperative recovery. The third advantage is the ability to assess a cancer’s response to treatment clinically, after a number of courses of chemotherapy, and pathologically, after surgical resection. This is especially important if alternative treatment regimens are available to be offered to patients whose disease responded inadequately. Molecular characterization of the residual disease may also give insight into mechanisms of chemoresistance and possible therapeutic targets.

There are some potential disadvantages to preoperative chemotherapy, however. Although disease progression while the patient is receiving preoperative chemotherapy is rare in chemotherapy-sensitive tumors such as breast cancer, it is more frequent in relatively chemotherapy-resistant tumors such as sarcomas.¹³⁵ Thus, patient selection is critical to ensure that the opportunity to treat disease surgically is not lost by giving preoperative chemotherapy. Often, rates of postoperative wound infection, flap necrosis, and delays in postoperative adjuvant therapy do not differ between patients who are treated with preoperative chemotherapy and patients who are treated with surgery first. However, preoperative chemotherapy can introduce special challenges to tumor localization, margin analysis, lymphatic mapping, and pathologic staging.

Response to chemotherapy is monitored clinically with imaging studies as well as physical examinations. Response usually is defined as complete response, partial response, stable disease, or progression. Response generally is assessed using the Response Evaluation Criteria in Solid Tumors (RECIST) criteria.¹³⁶ Objective tumor response assessment is critical, because tumor response is used as a prospective endpoint in clinical trials and tumor response is a guide to clinicians regarding continuation of current therapy.

Principles of Chemotherapy

Chemotherapy destroys cells by first-order kinetics, which means that with the administration of a drug a constant percentage of cells is killed, not a constant number of cells. If a patient with 10¹² tumor cells is treated with a dose that results in 99.9% cell kill (3-log cell kill), the tumor burden will be reduced from 10¹² to 10⁹ cells (or 1 kg to 1 g). If the patient is re-treated with the same drug, which theoretically could result in another 3-log cell kill, the cells would decrease in number from 10⁹ to 10⁶ (1 g to 1 mg) rather than being eliminated totally.

Chemotherapeutic agents can be classified according to the phase of the cell cycle during which they are effective. ­Cell-cycle phase-nonspecific agents (e.g., alkylating agents) have a linear dose-response curve, such that the fraction of cells killed increases with the dose of the drug.¹³⁷ In contrast, the cell-cycle phase-specific drugs have a plateau with respect to cell killing ability, and cell kill will not increase with further increases in drug dose.

Anticancer Agents

Alkylating Agents

Alkylating agents are cell-cycle–nonspecific agents, that is, they are able to kill cells in any phase of the cell cycle. They act by cross-linking the two strands of the DNA helix or by causing other direct damage to the DNA. The damage to the DNA prevents cell division and, if severe enough, leads to apoptosis. The alkylating agents are composed of three main subgroups: classic alkylators, nitrosoureas, and miscellaneous DNA-binding agents (Table 10-10).

Antitumor Antibiotics

Antitumor antibiotics are the products of fermentation of microbial organisms. Like the alkylating agents, these agents are cell-cycle nonspecific. Antitumor antibiotics damage the cell by interfering with DNA or RNA synthesis, although the exact mechanism of action may differ by agent.

Antimetabolites

Antimetabolites are generally cell-cycle–specific agents that have their major activity during the S phase of the cell cycle and have little effect on cells in G₀. These drugs are most effective, therefore, in tumors that have a high growth fraction. Antimetabolites are structural analogues of naturally occurring metabolites involved in DNA and RNA synthesis. Therefore, they interfere with normal synthesis of nucleic acids by substituting for purines or pyrimidines in the metabolic pathway to inhibit critical enzymes in nucleic acid synthesis. The antimetabolites include folate antagonists, purine antagonists, and pyrimidine antagonists.

Plant Alkaloids

Plant alkaloids are derived from plants such as the periwinkle plant, Vinca rosea (e.g., vincristine, a vinca alkaloid), or the root of American mandrake, Podophyllum peltatum (e.g., etoposide, a podophyllotoxin).¹³⁷ Vinca alkaloids affect the cell by binding to tubulin in the S phase. This blocks microtubule polymerization, which results in impaired mitotic spindle formation in the M phase. Taxanes such as paclitaxel, on the other hand, cause excess polymerization and stability of microtubules, which blocks the cell cycle in mitosis. The epipodophyllotoxins act to inhibit a DNA enzyme called topoisomerase II by stabilizing the DNA-topoisomerase II complex. This results in an inability to synthesize DNA, and thus the cell cycle is stopped in the G₁ phase.¹³⁷

Combination Chemotherapy

Combination chemotherapy may provide greater efficacy than single-agent therapy by three mechanisms: (a) it provides maximum cell kill within the range of toxicity for each drug that can be tolerated by the host, (b) it offers a broader range of coverage of resistant cell lines in a heterogeneous population, and (c) it prevents or delays the emergence of drug-resistant cell lines.¹³⁷ When combination regimens are devised, drugs known to be active as single agents usually are selected. Drugs with different mechanisms of action are combined to allow for additive or synergistic effects. Combining cell-cycle–specific and cell-cycle–nonspecific agents may be especially advantageous. Drugs with differing dose-limiting toxic effects are combined to allow for each drug to be given at therapeutic doses. Drugs with different patterns of resistance are combined whenever possible to minimize cross-resistance. The treatment-free interval between cycles is kept to the shortest possible time that will allow for recovery of the most sensitive normal tissue.

Drug Toxicity

Tumors are more susceptible than normal tissue to chemotherapeutic agents, in part because they have a higher proportion of dividing cells. Normal tissues with a high growth fraction, such as the bone marrow, oral and intestinal mucosa, and hair follicles, are also sensitive to chemotherapeutic effects. Therefore, treatment with chemotherapeutic agents can produce toxic effects such as bone marrow suppression, stomatitis, ulceration of the GI tract, and alopecia. Toxic effects usually are graded from 0 to 4 on the basis of World Health Organization standard criteria.¹³⁸ Significant drug toxicity may necessitate a dosage reduction. A toxic effect requiring a dose modification or change in dose intensity is referred to as a dose-limiting toxic effect. Because maintaining dose intensity is important to preserve as high a tumor cell kill as possible, several supportive strategies have been developed, such as administration of colony-stimulating factors and erythropoietin to treat poor bone marrow reserve and administration of cytoprotectants such as mesna and amifostine to prevent renal dysfunction.

Administration of Chemotherapy

Chemotherapy usually is administered systemically (IV, IM, SC, or PO). Systemic administration treats micrometastases at widespread sites and prevents systemic recurrence. However, it increases the drug’s toxicity to a wide range of organs throughout the body. One method to minimize systemic toxicity while enhancing target organ delivery of chemotherapy is regional administration of chemotherapy. Many of these approaches require surgical access, such as intrahepatic delivery of chemotherapy for hepatic carcinomas or metastatic colorectal cancer using a hepatic artery infusion pump, limb perfusion for extremity melanoma and sarcoma, and intraperitoneal hyperthermic perfusion for pseudomyxoma peritonei. Alternately, percutaneous access may be utilized, such as limb infusion with percutaneously placed catheters.

Some tumors, most notably breast and prostate cancers, originate from tissues whose growth is under hormonal control. The first attempts at hormonal therapy were through surgical ablation of the organ producing the hormones involved, such as oophorectomy for breast cancer. Currently, hormonal anticancer agents include androgens, antiandrogens, antiestrogens, estrogens, glucocorticoids, gonadotropin inhibitors, progestins, aromatase inhibitors, and somatostatin analogues. Hormones or hormone-like agents can be administered to inhibit tumor growth by blocking or antagonizing the naturally occurring substance, such as with the estrogen antagonist tamoxifen. Other substances that block the synthesis of the natural hormone can be administered as alternatives. Aromatase inhibitors, for example, block the peripheral conversion of endogenous androgens to estrogens in postmenopausal women. Hormonal therapy provides a highly tumor-specific form of therapy in sensitive tissues. In breast cancer, estrogen and progesterone receptor status is used to predict the success of hormonal therapy. Androgen receptor is also being pursued as a therapeutic target for breast cancer treatment.

Over the past decade, increased understanding of cancer biology has fostered the emerging field of molecular therapeutics. The basic principle of molecular therapeutics is to exploit the molecular differences between normal cells and cancer cells to develop targeted therapies. Thus targeted therapies usually are directed at the processes involved in tumor growth rather than directly targeting the tumor cells. The ideal molecular target would be exclusively expressed in the cancer cells, be the driving force of the proliferation of the cancer cells, and be critical to their survival. A large number of molecular targets are currently being explored, both preclinically and in clinical trials. The major groups of targeted therapy agents are inhibitors of growth factor receptors, inhibitors of intracellular signal transduction, cell-cycle inhibitors, apoptosis-based therapies, and antiangiogenic compounds.

Protein kinases have come to the forefront as attractive therapeutic targets with the success of imatinib mesylate (Gleevec) in treating chronic myelogenous leukemia and GI stromal tumors, and trastuzumab (Herceptin) in treating breast cancer, and vemurafanib in treating melanoma. These drugs work by targeting bcr-abland c-kit (imatinib) and HER2 and Braf, respectively. For example, recently a phase III randomized trial demonstrated that, compared with dacarbazine, standard of care chemotherapy option for patients with metastatic melanoma with a V600E BRAF mutation, the BRAF inhibitor vemurafenib led to significantly higher response rates (48% vs. 5%).¹³⁹ At 6 months, overall survival was 84% (95% CI, 78 to 89) in the vemurafenib group and 64% (95% CI, 56 to 73) in the dacarbazine group. The hazard ratio for tumor progression in the vemurafenib group was 0.26 (95% CI, 0.20 to 0.33; P<0.001). The estimated median progression-free survival was 5.3 months in the vemurafenib group and 1.6 months in the dacarbazine group. This trial highlights the fact that in at least some tumor types targeted therapies that inhibit a genomic alteration that is a driver is likely to be more effective than an unselected therapeutic option.

Sequencing of the human genome has revealed approximately 500 protein kinases. Several tyrosine kinases have been shown to have oncogenic properties and many other protein kinases have been shown to be aberrantly activated in cancer cells.⁹⁰ Therefore, protein kinases involved in these aberrantly activated pathways are being aggressively pursued in molecular therapeutics. Potential targets like HER2 can be targeted via different strategies, such as transcriptional downregulation, targeting of mRNA, RNA inhibition, antisense strategies, direct inhibition of protein activity, and induction of immunity against the protein. Most of the compounds in development are monoclonal antibodies like trastuzumab or small-molecule kinase inhibitors like imatinib or vemurafanib. Some other agents, such as sunitinib, are multi-targeted kinase inhibitors. Selected FDA-approved targeted therapies are listed in Table 10-11. Many of the promising pathways, such as the PI3K/Akt/mTOR pathway are being pursued as therapeutic targets with several drugs in development, targeting different aspects of the pathway (Fig. 10-15).¹⁴⁰

Development of molecularly targeted agents for clinical use presents several unique challenges. Once an appropriate compound is identified and confirmed to have activity in preclinical testing, predictive markers for activity in the preclinical setting must be defined. Expression of a target may not be sufficient to predict response, because the pathway of interest may not be activated or critical to the cancer’s survival. Although in traditional phase I trials the goal is to identify the maximum tolerated dosage, the maximum dosage of biologic agents may not be necessary to achieve the desired biologic effect. Thus assays to verify modulation of the target need to be developed to determine at what dosage the desired effect is achieved. When phase II and III clinical trials are initiated, biomarker modulation studies should be integrated into the trial to determine whether clinical response correlates with target modulation and thus to identify additional parameters that impact response. Rational dose selection and limitation of study populations to patients most likely to respond to the molecular therapy as determined by predictive markers are most likely to lead to successful clinical translation of a product. Finally, most biologic agents are cytostatic, not cytotoxic. Thus rational combination therapy mixing new biologic agents with either established chemotherapeutic agents that have synergy or with other biologic agents is more likely to lead to cancer cures.

The aim of immunotherapy is to induce or potentiate inherent antitumor immunity that can destroy cancer cells. Central to the process of antitumor immunity is the ability of the immune system to recognize tumor-associated antigens present on human cancers and to direct cytotoxic responses through humoral or T-cell–mediated immunity. Overall, T-cell–mediated immunity appears to have the greater potential of the two for eradicating tumor cells. T cells recognize antigens on the surfaces of target cells as small peptides presented by class I and class II MHC molecules.

Several antitumor strategies are under investigation. One approach to antitumor immunity is nonspecific immunotherapy, which stimulates the immune system as a whole through administration of bacterial agents or their products, such as bacille Calmette-Guérin. This approach is thought to activate the effectors of antitumor response such as natural killer cells and macrophages, as well as polyclonal lymphocytes.¹⁴¹ Another approach to nonspecific immunotherapy is systemic administration of cytokines such as interleukin-2, interferon-α, and interferon-γ. Interleukin-2 stimulates proliferation of cytotoxic T lymphocytes and maturation of effectors such as natural killer cells into lymphokine-activated killer cells. Interferons, on the other hand, exert antitumor effects directly by inhibiting tumor cell proliferation and indirectly by activating host immune cells, including macrophages, dendritic cells, and natural killer cells, and by enhancing human leukocyte antigen (HLA) class I expression on tumor cells.¹⁴¹

Antigen-specific immunotherapy can be active, as is achieved through antitumor vaccines, or passive. In passive immunotherapy, antibodies to specific tumor-associated antigens can be produced by hybridoma technique and then administered to patients whose cancers express these antigens, inducing antibody-dependent cellular cytotoxicity.

The early attempts at vaccination against cancers used allogeneic cultured cancer cells, including irradiated cells, cell lysates, and shed antigens isolated from tissue culture supernatants. An alternate strategy is the use of autologous tumor vaccines. These have the potential advantage of being more likely to contain antigens relevant for the individual patient but have the disadvantage of requiring a large amount of tumor tissue for preparation, which restricts eligibility of patients for this modality. Strategies to enhance immunogenicity of tumor cells include the introduction of genes encoding cytokines or chemokines, and fusion of the tumor cells to allogeneic MHC class II-bearing cells.¹⁴² Alternatively, heat shock proteins derived from a patient’s tumor can be used, because heat shock protein peptide complexes are readily taken up by dendritic cells for presentation to T cells.¹⁴²

Identification of tumor antigens has made it possible to perform antigen-specific vaccination. For example in the case of melanoma, several antigens have been identified that can be recognized by both CD8⁺ cytotoxic T cells and CD4⁺ helper T cells, including MART-1, gp 100, MAGE1, tyrosinase, TRP-1, TRP-2, and NY-ESO-1.¹⁴³ Antigens tested usually are ­overexpressed or mutated in cancer cells. Tissue specificity and immunogenicity are important determinants in choosing an appropriate target. Vaccines directed at defined tumor antigens aim to combine selected tumor antigens and appropriate routes for delivering these antigens to the immune system to optimize antitumor immunity.¹⁴⁴ Several different vaccination approaches are under study, including tumor cell-based vaccines, peptide-based vaccines, recombinant virus-based vaccines, DNA-based vaccines, and dendritic cell vaccines.

In adoptive transfer, antigen-specific effector cells (i.e., cytotoxic T lymphocytes) or antigen-nonspecific effector cells (i.e., natural killer cells) can be transferred to a patient. These effector cells can be obtained from the tumor (tumor-infiltrating lymphocytes) or the peripheral blood.

Clinical experience in patients with metastatic disease has shown objective tumor responses to a variety of immunotherapeutic modalities. It is thought, however, that the immune system is overwhelmed with the tumor burden in this setting, and thus adjuvant therapy may be preferable, with immunotherapy reserved for decreasing tumor recurrences. Trials to date suggest that immunotherapy is a potentially useful approach in the adjuvant setting. How to best select patients for this approach and how to integrate immunotherapy with other therapies are not well understood for most cancer types.

Tolerance to self-antigens expressed in tumors is a limitation in generating antitumor responses.¹⁴⁵ Recently, several pathways that modulate tolerance and approaches to manipulating these pathways have been identified: pathways that activate professional antigen-presenting cells such as Toll-like receptors, growth factors, and the CD40 pathway; cytokines to enhance immunoactivation; and pathways that inhibit T-cell inhibitory signals or Tregs.¹⁴⁵

A new strategy being actively explored involves the use of cytotoxic T-lymphocyte antigen 4 (CTLA-4). CTLA-4 exists on the surfaces of T cells and has a homeostatic immunosuppressive function, downregulating the response of T cells to stimuli.¹⁴⁶

In a recent phase 3 study, ipilimumab—which blocks CTLA-4, was administered with or without glycoprotein 100 (gp100) peptide vaccine and was compared with gp100 alone in HLA-A*0201-positive patients with previously treated metastatic melanoma. The median overall survival was significantly longer for patients receiving ipilimumab with or without gp100, compared with patients who receiving gp100.¹⁴⁷ In another Phase III trial, ipilimumab in combination with dacarbazine, compared with dacarbazine plus placebo, improved overall survival in patients with previously untreated metastatic melanoma.¹⁴⁸ Anti-CTLA-4 antibodies are under study for use in melanoma as well as several other cancer types as single agents, in combination with targeted therapies, interleukin-2, chemotherapy, or peptide vaccines.¹⁴⁶

Programmed death ligand 1 (PD-L1) is a 40kDa type 1 transmembrane protein that is thought to play an important role in suppressing the immune system. PD-L1 binds to its receptor, PD-1, which is found on activated T cells, B cells, and myeloid cells. The PD1/PDL1 pathway is increasingly recognized as a key contributor to tumor-mediated immune suppression. Thus both anti-PD1 and anti-PD-L1 strategies are actively being pursued for cancer therapy.

Gene therapy is being pursued as a possible approach to modifying the genetic program of cancer cells as well as treating metabolic diseases. The field of cancer gene therapy uses a variety of strategies, ranging from replacement of mutated or deleted tumor-suppressor genes to enhancement of immune responses to cancer cells.¹⁴⁹ Indeed, in preclinical models, approaches such as replacement of tumor-suppressor genes leads to growth arrest or apoptosis. However, the translation of these findings into clinically useful tools presents special challenges.

One of the main difficulties in getting gene therapy technology from the laboratory to the clinic is the lack of a perfect delivery system. An ideal vector would be administered through a noninvasive route and would transduce all of the cancer cells and none of the normal cells. Furthermore, the ideal vector would have a high degree of activity, that is, it would produce an adequate amount of the desired gene product to achieve target cell kill. Unlike genetic diseases in which delivery of the gene of interest into only a portion of the cells may be sufficient to achieve clinical effect, cancer requires either that the therapeutic gene be delivered to all of the cancer cells or that a therapeutic effect be achieved on nontransfected cells as well as transfected cells through a bystander effect. But then, treatment of a metabolic disease requires prolonged gene expression, whereas transient expression may be sufficient for cancer therapy.

Several vector systems are under study for gene therapy; however, none is considered ideal. One of the promising approaches to increase the number of tumor cells transduced is the use of a replication-competent virus like a parvovirus, human reovirus, or vesicular stomatitis virus that selectively replicates within malignant cells and lyses them more efficiently than it does normal cells. Another strategy for killing tumor cells with suicide genes exploits tumor-specific expression elements, such as the MUC-1, PSA, CEA, or VEGF promoters, that can be used to achieve tissue-specific or tumor-specific expression of the desired gene.

Because the goal in cancer therapy is to eradicate systemic disease, optimization of delivery systems is the key to success for gene therapy strategies. Gene therapy is likely to be most successful when combined with standard therapies, but it will provide the advantage of customization of therapy based on the molecular status of an individual’s tumor.

Several tumor factors influence tumor cell kill. Tumors are heterogeneous, and, according to the Goldie-Coldman hypothesis, tumor cells are genetically unstable and tend to mutate to form different cell clones. This has been used as an argument for giving chemotherapy as soon as possible in treatment to reduce the likelihood that resistant clones will emerge. Tumor size is another important variable. Large tumor, may have greater heterogeneity, although heterogeneity may also differ based on biological subtype. Moreover, according to the gompertzian model, cancer cells initially grow rapidly (exponential growth phase), then the growth slows down owing to hypoxia and decreased nutrient supply. Because of the larger proportion of cells dividing, smaller tumors may be more chemosensitive.

Multiple mechanisms of systemic therapy resistance have been identified (Table 10-12).¹⁵⁰ Cells may exhibit reduced sensitivity to drugs by virtue of their cell-cycle distribution. For example, cells in the G₀ phase are resistant to drugs active in the S phase. This phenomenon of “kinetic resistance” usually is temporary, and if the drug level can be maintained, all cells will eventually pass through the vulnerable phase of the cell cycle.¹³⁷ Alternatively, tumor cells may exhibit “pharmacologic resistance,” in which the failure to kill cells is due to insufficient drug concentration. This may occur when tumor cells are located in sites where effective drug concentrations are difficult to achieve (such as the central nervous system) or can be due to enhanced metabolism of the drug after its administration, decreased conversion of the drug to its active form, or decrease in the intracellular drug level caused by increased removal of the drug from the cell associated with enhanced expression of P-glycoprotein, the protein product of multidrug resistance gene 1. Other mechanisms of resistance include decreased affinity of the target enzyme for the drug, altered amount of the target enzyme, or enhanced repair of the drug-induced defect. For drug-sensitive cancers, another factor limiting optimal killing is improper dosing. A dose reduction of 20% because of drug toxicity can lead to a decline in the cure rate by as much as 50%.¹³⁷ Furthermore, a twofold increase in dose can be associated with a tenfold (1 log) increase in tumor cell kill.

Cancer cells demonstrate adaptive responses to targeted therapy, like activating alternate pathways of survival; thus these alterations may blunt therapeutic efficacy. Cancer cells also acquire resistance upon prolonged treatment with targeted therapy through a variety of mechanisms. One mechanism is through the loss of the target. For example, this was observed in a study of patients with HER2-positive breast cancer patients who were treated with neoadjuvant trastuzumab-based chemotherapy.¹⁵¹ Post-neoadjuvant treatment, a third of the samples from patients who did not have a complete pathologic response displayed loss of the HER2 amplification that had been present in their pretreatment-biopsy specimens¹⁵¹. Another means bywhich cancers develop resistance is the acquisition of additional genomic aberrations. In lung cancer, a second mutation in EGFR (T790M) and MET amplification have been described as two main mechanisms of drug resistance to EGFR inhibitors erlotinib and gefinitib.^152,153,154 Other mechanisms like novel genetic changes, including HER2 and EGFR amplification, PIK3CA mutations, and markers of epithelial-to-mesenchymal transition have also been reported in EGFR inhibitor resistant lung.^155,156 Analysis of metastases from patients with colorectal cancer who developed resistance to cetuximab or panitumumab showed the emergence of KRAS amplification in one sample and acquisition of secondary KRAS mutations in 60% of the cases.¹⁵⁷ These studies emphasize the utility of repeat tumor biopsy specimens at the time of relapse or progression to identify mechanisms of resistance and best combinatorial therapies.

Physical Basis of Radiation Therapy

Ionizing radiation is energy strong enough to remove an orbital electron from an atom. This radiation can be electromagnetic, like a high-energy photon, or particulate, such as an electron, proton, neutron, or alpha particle. Radiation therapy is delivered primarily as high-energy photons (gamma rays and X-rays) and charged particles (electrons). Gamma rays are photons that are released from the nucleus of a radioactive atom. X-rays are photons that are created electronically, such as with a clinical linear accelerator. Currently, high-energy radiation is delivered to tumors primarily with linear accelerators. X-rays traverse the tissue, depositing the maximum dose beneath the surface, and thus spare the skin. Electrons are used to treat superficial skin lesions, superficial tumors, or surgical beds to a depth of 5 cm. Gamma rays typically are produced by radioactive sources used in brachytherapy.

The dose of radiation absorbed correlates with the energy of the beam. The basic unit is the amount of energy absorbed per unit of mass (joules per kilogram) and is known as a gray(Gy). One gray is equivalent to 100 rads, the unit of radiation measurement used in the past.

Biologic Basis of Radiation Therapy

Radiation deposition results in DNA damage manifested by single- and double-strand breaks in the sugar phosphate backbone of the DNA molecule.¹⁵⁸ Cross-linking between the DNA strands and chromosomal proteins also occurs. The mechanism of DNA damage differs by the type of radiation delivered. Electromagnetic radiation is indirectly ionizing through short-lived hydroxyl radicals produced primarily by the ionization of cellular hydrogen peroxide (H₂O₂).¹⁵⁸ Protons and other heavy particles are directly ionizing and directly damage DNA.

Radiation damage is manifested primarily by the loss of cellular reproductive integrity. Most cell types do not show signs of radiation damage until they attempt to divide, so slowly proliferating tumors may persist for months and appear viable. Some cell types, however, undergo apoptosis.

The extent of DNA damage after radiation exposure is dependent on several factors. The most important of these is cellular oxygen. Hypoxic cells are significantly less radiosensitive than aerated cells. Because the presence of oxygen is thought to prolong the half-life of free radicals produced by the interaction of X-rays and cellular H₂O₂, indirectly ionizing radiation is less efficacious in tumors with areas of hypoxia.¹⁵⁸ In contrast, radiation damage from directly ionizing radiation is independent of cellular oxygen levels.

The extent of DNA damage from indirectly ionizing radiation is dependent on the phase of the cell cycle. The most radiation-sensitive phases are G₂ and M, whereas G₁ and late S phases are less sensitive. Thus irradiation of a population of tumor cells results in killing of a greater proportion of cells in G₂ and M phases. However, delivery of radiation in divided doses, a concept referred to as fractionation, allows the surviving G₁ and S phase cells to progress to more sensitive phases, a process referred to as reassortment. In contrast to DNA damage after indirectly ionizing radiation, that after exposure to directly ionizing radiation is less dependent on the cell-cycle phase.¹⁵⁹

Several chemicals can modify the effects of ionizing radiation. These include hypoxic cell sensitizers such as metronidazole and misonidazole, which mimic oxygen and increase cell kill of hypoxic cells.¹⁵⁸ A second category of radiation sensitizers are the thymidine analogues iododeoxyuridine and bromodeoxyuridine. These molecules are incorporated into the DNA in place of thymidine and render the cells more susceptible to radiation damage; however, they are associated with considerable acute toxicity. Several other chemotherapeutic agents sensitize cells to radiation through various mechanisms, including 5-fluorouracil, actinomycin D, gemcitabine, paclitaxel, topotecan, doxorubicin, and vinorelbine.¹⁵⁸

Radiation Therapy Planning

Radiation therapy is delivered in a homogeneous dose to a well-defined region that includes tumor and/or surrounding tissue at risk for subclinical disease. The first step in planning is to define the target to be irradiated as well as the dose-limiting organs in the vicinity.¹⁶⁰ Treatment planning includes evaluation of alternative treatment techniques, which is done through a process referred to as simulation. Once the beam distribution that will best achieve homogeneous delivery to the target volume and minimize the dose to the normal tissue is determined, immobilization devices and markings or tattoos on the patient’s skin are used to ensure that each daily treatment is given in the same way. Conventional fractionation is 1.8 to 2 Gy/d, administered 5 days each week for 3 to 7 weeks.

Radiation therapy may be used as the primary modality for palliation in certain patients with metastatic disease, primarily patients with bony metastases. In these cases, radiation is recommended for symptomatic metastases only. However, lytic metastases in weight-bearing bones such as the femur, tibia, or humerus also are considered for irradiation. Another circumstance in which radiation therapy might be appropriate is spinal cord compression due to metastases to the vertebral body that extend posteriorly to the spinal canal.

The goal of adjuvant radiation therapy is to decrease local-regional recurrence rates. Adjuvant radiation therapy can be given before surgery, after surgery, or, in selected cases, during surgery. Preoperative radiation therapy has several advantages. It may minimize seeding of the tumor during surgery and it allows for smaller treatment fields because the operative bed has not been contaminated with tumor cells. Also, radiation therapy for inoperable tumors may achieve adequate reduction to make them operable. The disadvantages of preoperative therapy are an increased risk of postoperative wound healing problems and the difficulty in planning subsequent radiation therapy in patients who have positive surgical margins. If radiation therapy is given postoperatively, it is usually given 3 to 4 weeks after surgery to allow for wound healing. The advantage of postoperative radiation therapy is that the surgical specimen can be evaluated histologically and radiation therapy can be reserved for patients who are most likely to benefit from it. Further, the radiation therapy can be modified on the basis of margin status. The disadvantages of postoperative radiation therapy are that the volume of normal tissue requiring irradiation may be larger owing to surgical contamination of the tissue planes and that the tumor may be less sensitive to radiation owing to poor oxygenation. Postlaparotomy adhesions may decrease the mobility of the small bowel loops, increasing the risk for radiation injury in abdominal or pelvic irradiation. Given the potential advantages and disadvantages of both approaches, the roles of preoperative and postoperative radiation therapy are being actively evaluated and compared for many cancer types.

Another mode of postoperative radiation therapy is brachytherapy. In brachytherapy, unlike in external beam therapy, the radiation source is in contact with the tissue being irradiated. The radiation source may be cesium, gold, iridium, or radium. Brachytherapy is administered via temporary or permanent delivery implants such as needles, seeds, or catheters. Temporary brachytherapy catheters are placed either during open surgery or percutaneously soon after surgery. The implants are loaded interstitially, and treatment usually is given postoperatively for a short duration, such as 1 to 3 days. Although brachytherapy has the disadvantages of leaving scars at the catheter insertion site and requiring special facilities for inpatient brachytherapy the advantage of patient convenience owing to the shorter treatment duration, has made intracavitary treatment approaches very popular for the treatment of breast cancer.

Another short delivery approach is intraoperative radiotherapy (IORT), often used in combination with external beam therapy. The oncologic consequences of the limited treatment volume and duration associated with brachytherapy and IORT are not well understood. Accelerated partial breast irradiation with interstitial brachytherapy, intracavitary brachytherapy (MammoSite), IORT, and three-dimensional conformal external beam radiotherapy is being compared with whole breast irradiation in an intergroup phase III trial (NSABP B-39/Radiation Therapy Oncology Group 0413). Several additional studies of adjuvant IORT also are ongoing internationally. There has also been increasing interest in utilizing intensity-modulated radiation therapy (IMRT). IMRT is a complex technique for the delivery of radiation therapy preferentially to target structures while minimizing doses to adjacent normal critical structures.¹⁶¹ It is widely utilized for the treatment of a variety of tumor types, including the central nervous system, head and neck, breast, prostate, gastrointestinal tract, and gynecologic organs, as well as in patients where previous radiation therapy has been delivered.

It is thought that chemotherapy given concurrently with radiation improves survival rates. Chemotherapy before radiation has the advantage of reducing the tumor burden, which facilitates radiation therapy. On the other hand, some chemotherapy regimens, when given concurrently with radiation, may sensitize the cells to radiation therapy. Chemoradiation is being pursued in many tumor types, including rectal cancer, pancreatic cancer, and esophageal cancer.^162,163,164 In a recent Cochrane review of six randomized controlled trials, it was demonstrated that in patients T3/4 rectal cancer, chemoradiation was associated with a significantly lower local recurrence rate compared with radiation therapy alone (OR 0.56, 95% CI 0.42-0.75, P<0.0001), but was not associated with improved survival.¹⁶²

Side Effects

Both tumor and normal tissue have radiation dose-response relationships that can be plotted as a sigmoidal curve (Fig. 10-16).¹⁶⁰ A minimum dose of radiation must be given before any response is seen. The response to radiation then increases slowly with an increase in dose. At a certain dose level the curves become exponential, with increases in tumor response and normal tissue toxicity with each incremental dose increase. The side effects of radiation therapy can be acute, occurring during or 2 to 3 weeks after therapy, or chronic, occurring weeks to years after ­therapy. The side effects depend on the tissue included in the target ­volume. Some of the major acute and chronic sequelae of radiation are summarized in Table 10-13.^160,165 In addition to these effects, a small increase in the risk for secondary malignancies is attributable to radiation therapy.

The truth of the old axiom, “An ounce of prevention is worth a pound of cure” is being increasingly recognized in oncology. Cancer prevention can be divided into three categories: (a) primary prevention (i.e., prevention of initial cancers in healthy individuals), (b) secondary prevention (i.e., prevention of cancer in individuals with premalignant conditions), and (c) tertiary prevention (i.e., prevention of second primary cancers in patients cured of their initial disease).

The systemic or local administration of therapeutic agents to prevent the development of cancer, called chemoprevention, is being actively explored for several cancer types. In breast cancer, the NSABP Breast Cancer Prevention Trial demonstrated that tamoxifen administration reduces the risk of breast cancer by one half and reduces the risk of estrogen receptor-positive tumors by 69% in high-risk patients.¹⁶⁶ Therefore, tamoxifen has been approved by the FDA for breast cancer chemoprevention. The subsequent NSABP P-2 trial demonstrated that raloxifene is as effective as tamoxifen in reducing the risk of invasive breast cancer and is associated with a lower risk of thromboembolic events and cataracts but a non-statistically significant higher risk of noninvasive breast cancer; these findings led the FDA to approve raloxifene for prevention as well. Several other agents are also under investigation.¹⁶⁷ Celecoxib has been shown to reduce polyp number and polyp burden in patients with FAP, which led to its approval by the FDA for these patients. In head and neck cancer, 13-cis-retinoic acid has been shown both to reverse oral leukoplakia and to reduce second primary tumor development.^168,169 Thus, the chemoprevention trials completed so far have demonstrated success in primary, secondary, and tertiary prevention. Although the successes of these chemoprevention studies are impressive, much remains to be done over the next few years to improve patient selection and decrease therapy-related toxic effects. It is important for surgeons to be aware of these preventive options, because they are likely to be involved in the diagnosis of premalignant and malignant conditions and will be the ones to counsel patients about their chemopreventive options.

In selected circumstances, the risk of cancer is high enough to justify surgical prevention. These high-risk settings include hereditary cancer syndromes such as hereditary breast-ovarian cancer syndrome, hereditary diffuse gastric cancer, multiple endocrine neoplasia type 2, FAP, and hereditary nonpolyposis colorectal cancer, as well as some nonhereditary conditions such as chronic ulcerative colitis. Most prophylactic surgeries are large ablative surgeries (e.g., bilateral risk-reducing mastectomy or total proctocolectomy). Therefore, it is important that the patient be completely informed about potential surgical complications as well as long-term lifestyle consequences. Further, the conservative options of close surveillance and chemoprevention need to be discussed. The patient’s cancer risk needs to be assessed accurately and implications for survival discussed. Ultimately, the decision to proceed with surgical prevention should be individualized and made with caution.

Cancer Screening and Diagnosis

It is clear that the practice of oncology will change dramatically over the next few decades, because our understanding of the molecular basis of cancer and available technologies are evolving rapidly. One of the critical changes expected is earlier detection of cancers. With improvements in available imaging modalities and development of newer functional imaging techniques, it is likely that many tumors will be detected at earlier, more curable stages in the near future.

Another area of rapid development is the identification of serum markers. High-throughput technologies such as matrix-assisted laser desorption ionization time-of-flight mass spectroscopy and liquid chromatography ion-spray tandem mass spectroscopy have revolutionized the field of proteomics and are now being used to compare the serum protein profiles of patients with cancer with those of individuals without cancer. Identification of unique proteins as well as unique proteomic profiles for most cancer types is being pursued actively by many researchers and, if successful, could dramatically enhance our ability to detect cancers early.¹⁷⁰ In addition, there is greater interest placed in leveraging circulating free DNA as a potential approach for cancer screening.

Surgical Therapy

The current trend in surgery is toward more conservative resections. With earlier identification of tumors, more conservative operations may be possible. The goal, however, is always to remove the tumor en bloc with wide negative margins. Another interesting area being explored is the destruction of tumors by techniques such as radiofrequency ablation, cryoablation, and heat-producing technologies like lasers, microwaves, or focused ultrasound. Pilot studies have demonstrated that radiofrequency ablation is effective for destruction of small primary breast cancers. Although this approach remains experimental and potentially of limited applicability because of the need for expertise in breast imaging, the development of imaging technologies that can accurately map the extent of cancer cells, these types of noninvasive interventions are likely to come to the forefront. However, use of these techniques will be limited to treatment of cancers not involving hollow viscera.

The debate over how to manage the regional lymph node basins for certain cancer types continues. With an increasing understanding of the metastatic process, surgeons may be able to stratify patients on the basis of the likelihood that their disease will spread metastatically, based on the gene expression profile of their primary tumors, and offer regional therapy accordingly. There is also a growing interest in minimally invasive surgical treatments for a variety of cancer types.

Systemic Therapy

The current trend in systemic therapy is toward individualized therapy. It is now presumed that all cancers of a certain cell origin are the same. Thus all patients are offered the same systemic therapy. Not all patients respond to these therapies; however, this emphasizes the biologic variability within the tumor groups. Therefore, the intent is to determine the underlying biology of each tumor to tailor therapy accordingly. Genomic, transcriptional, and proteomic profiling approaches are being used to identify molecular signatures that correlate with response to certain agents. It is likely that in the near future all tumors can be tested and treatments individualized. Patients who will respond to conventional therapies can be treated with these regimens, whereas patients who will not respond will not, which spares them the toxicity. Instead, the latter patients can be offered novel therapies. Furthermore, with emerging biologic therapies, it is likely that patients may be given a combination of biologic therapies that specifically target the alterations in their own tumors. Patients can be genotyped for critical alleles that may affect drug metabolism and thus, may influence the efficacy as well as the side effect of the drugs given. Finally, stratification of patients by gene expression profile for prognosis may assist in determining which patients are at higher risk of relapse, so that patients whose tumors have less aggressive biologic characteristics can be spared further therapy.