Chapter 36. Tumors of the Colon

Tumor is a descriptive term for a growth or mass of cells that are independent of the physiologic function or demand of their surrounding structures. The two characteristic biological growth patterns of tumors include the ability to (1) disrespect tissue boundaries and invade other structures (invasiveness) and (2) gain access to blood and lymph vessels or other structures to spread tumor cells to distant locations, and allow these specially equipped cells to survive and grow new remote tumors (metastases). If a tumor does not have either property, it is benign; if a tumor can invade locally but even at a large size does not have a tendency to metastasize, it is called semimalignant; and if a tumor has the ability to metastasize once a sufficient size is reached, it is a malignant tumor.

Colorectal lesions may be classified as either benign, potentially malignant, or malignant based on their pathological features (Table 36–1); the semimalignant variant with invasion only but no affinity to later form of metastases is not common in the colon. The overwhelming majority of colorectal tumors are of epithelial origin and arise from the mucosal surface, where they become visible descriptively as a polyp. Benign polyps include nonneoplastic polyps (eg, hyperplastic, hamartomatous, or inflammatory polyps); the potentially malignant group consists of adenomatous polyps. Once dysplastic cells in a polyp cross the boundaries of the mucosa (basement membrane and muscularis mucosae) and start to invade the submucosa and the muscularis mucosae, a true cancer (carcinoma) with the potential to metastasize is established. Tumors of nonepithelial or mesenchymal origin are comparably rare and include, among others, lipoma, lymphoma, carcinoid, and sarcoma.1–3

Colonic tumors are important for two reasons: First, they are frequent and account for both a significant mortality rate as well as high cumulative health care costs. Second, the sequence of events leading from a normal mucosa to a manifest cancer occurs through largely preventable precursor stages over the course of several years. This chapter therefore predominantly focuses on the detection, management, and prevention of these conditions.

Colorectal cancer is the most common malignancy in the gastrointestinal tract. In the United States, colorectal cancer ranks fourth in terms of both gender-specific annual cancer incidence and cancer mortality (behind lung, prostate, and breast, respectively).4 With an estimated 146,970 newly diagnosed cases, this disease will be responsible for an estimated 49,920 deaths or 10–15% of cancer-related deaths in the year 2009.4,5 The lifetime risk of approximately 6% in our Western civilization means that 1 in 18 individuals of the general population will be affected by colorectal cancer and many more by polyps, making it an important public health issue.6 Worldwide, colorectal cancer shows large geographical differences, with a crude incidence of 6.5/7.7 cases per 100,000 females/males in less developed areas as opposed to 50.9/60.8 in more developed regions.7

The colorectal cancer incidence has evolved in recent years to an overall predominance of males who represent in 51.4% of the cases.4,8 Rectal cancer is more frequent in males (57.7%), while colon cancer is slightly more frequent in females, resulting in almost identical overall mortality rates for both genders.4 The incidence of colorectal cancer in females is 44.8 per 100,000 and in males 61.2. Regardless of ethnicity, there is an age-dependent increase in incidence with each decade starting at age of 40 years, and the mean age at presentation is around 70–75 years.

In the period between 1975 and 2006, the Surveillance, Epidemiology and End Results (SEER) Registry of the National Cancer Institute (NCI) shows a gradual decline in all cases of colorectal cancer in the United States from 69.7 to 50.6 cases per 100,000.9 However, although these numbers reflect the trend in Caucasians, the incidence of colorectal cancer in the United States for African Americans has remained at the same level of 59.3–61.5 cases per 100,000 individuals. African American males therefore now represent the ethnic subgroup with the highest risk.10,11

The specific cause of colorectal cancer is not known. However, a number of genetic and environmental risk factors have been associated with the disease.12 From a practical and screening standpoint, it has been helpful to group individuals into three risk categories (ie, average risk, increased risk, high risk) based on their presumptive genetic profile as reflected in their individual and family history.13,14 The high-risk and increased-risk groups consist of patients with known hereditary syndromes or bowel diseases or patients with a personal/family history of polyps or cancer, all of which are discussed in a later section of the chapter (Table 36-2).

The majority of cases, however, are sporadic colon cancers that typically arise within a polyp. Geographic and migrational studies have suggested that the Western lifestyle increases the risk for colon cancer, hence suggesting that nutritional and environmental factors may play a key role.15 A large number of epidemiological studies have been undertaken to identify these individual, nutritional, lifestyle, genetic, and environmental factors that would either predispose to or prevent the development of colorectal polyps and cancer (Table 36-3).16–21

Extrinsic Risk Factors

Dietary Fiber, Meat, and Fat

One of the characteristics of a Western diet generally has been the lack of fiber as opposed to the increased amount of meat, total fat, and animal fats.23,24 In view of the known geographic differences, with the highest colorectal cancer incidence in industrialized nations,7 a high-fat and low-fiber diet generally has been considered a risk factor for the development of colorectal cancer.25 This concept gained support from epidemiological studies26 and resulted in common recommendations of high-fiber supplements in order to increase the stool bulk, dilute toxins, and reduce the colonic transit time and thus the exposure time to fecal carcinogens.27–30 More recent prospective trials, however, have questioned the benefit of dietary fiber supplementation in that they were at best inconclusive and did not reduce the incidence of colorectal cancer.31,32 On the other hand, selected fats such as n-3 fatty acids found in fish oils may have a protective effect,33 even though a direct effect to the mucosa could not be observed.34 It therefore could be concluded that the total amount of fats or fibers is of lesser importance than their quality and origin.21,23,35 The protective effect of vegetables and fruits36,37 may come not only from their fiber content but also from the content of antioxidative and antiproliferative agents such as isothiocyanates in cruciferous vegetables (eg, broccoli), which may enhance the expression of carcinogen-metabolizing enzymes and induce apoptosis in neoplastic cells.18,38

Calcium, Vitamins, and Micronutrients

Several prospective studies suggested that increased oral calcium and selenium intake may protect from colorectal polyps and cancers,39–44 whereas other studies could not verify a significant benefit.45 The mechanism by which calcium supplements are thought to reduce the risk of colon cancer is twofold. First, calcium can bind bile and fatty acids in the stool to insoluble complexes that are less likely to attack the colonic mucosa, and second, it can interfere directly with the mucosal cells and decrease their proliferative potential on a cellular level.26

Several vitamins were found to have a cancer-protective effect. Vitamins A, C, and E have been shown to have antioxidant activity. Results from interventional studies, however, have remained somewhat disappointing or controversial.46,47

In a study on postmenopausal women, another correlation was found between dietary heme iron and an increased risk of proximal colon cancer, especially in conjunction with alcohol consumption, whereas intake of dietary zinc reduced the risk of both proximal and distal colon cancer.48

Aspirin and COX-2 Inhibitors

Aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) may interfere with the development of colorectal neoplasms by blocking the cyclooxygenase (COX)-dependent prostaglandin pathway.49 The targets are the constitutive COX-1, as well as the cytokine-inducible COX-2, which has been found at increased expression levels in both polyps and cancers.50 Several trials therefore have studied these agents (eg, aspirin and sulindac) for the chemoprevention of colorectal cancer both in sporadic polyps and cancers51 and in familial adenomatous polyposis (FAP).52–54 In both settings, controlled studies have provided contradictory results.55 Regular prophylactic medication with low-dose aspirin may reduce the risk of sporadic colorectal cancer.51,56 Data from chemoprevention trials in FAP suggest that COX inhibition may delay the onset and number of adenomatous polyps, but it is not yet clear whether it is able to prevent the cancers overall or reduce their respective risk.52–54 COX-2–independent mechanisms may play a role for the beneficial effect of some COX-2 inhibitors.49 A major recent concern, however, has been the documented increased risk of serious cardiovascular events with the use of COX-2 inhibitors.57,58

Because data on the benefits remain conflicting, physicians must decide how to use these pharmacologic tools in the management of their patients. Based on the presumed small risks in general and the supporting data on a possible benefit, most physicians would be inclined to err on the side of a potential benefit in preventing colon polyp formation. Low doses of aspirin and calcium may be helpful in preventing polyps and cancers. However, recent concern about cardiovascular side effects and increased mortality has resulted in a withdrawal of more potent COX-2 inhibitors until further redefinition of the indications and risk groups has been accomplished.57,58

Cholecystectomy and Bile Acids

Evidence that bile acids may act as cocarcinogens or tumor promoters comes from both experimental and epidemiological studies.59,60 Bile acids can induce hyperproliferation of the intestinal mucosa via a number of intracellular mechanisms. Cholecystectomy, which alters the enterohepatic cycle of bile acids, has been associated with a moderately increased risk of proximal colon cancers.61,62 It cannot be ruled out, however, that it is less the effect of the cholecystectomy than the impact of other, not yet identified factors in the lithogenic bile of such patients. A number of cofactors have been identified that may enhance or neutralize the carcinogenic effects of bile acids, for example the amount of dietary fat, fiber,26 or calcium.63 Calcium, in fact, binds bile acids and thus may reduce their negative impact. However, other more intrinsic mucosa-protective mechanisms of calcium supplements probably are more relevant for the demonstrated reduction of recurrent adenomatous colon polyps.

Smoking and Alcohol Consumption

The risk of colorectal cancer is increased, even though only modestly, among long-term smokers compared with nonsmokers.29,48,64,65 The data suggested a dose-response relationship between pack-years of tobacco use and the development of adenomatous polyps.66–69 Equally, excessive alcohol consumption has been associated with an increased risk for colon cancer.29,48,64,65

Other Factors

An ever-increasing number of other factors are accumulating that have been attributed to an increased risk of colon cancer, for example lack of physical activity, diabetes, serum insulin levels, elevated concentrations of insulin-like growth factor 1, and low concentrations of insulin-like growth factor–binding protein 3 (IGFBP-3).70 The complexity of interactions between these factors and the previously mentioned parameters, however, makes it difficult at the present time to draw conclusions that have an impact on the clinical practice.

Intrinsic Risk Factors

Personal and Family History

There is generally little debate on whether the presence of an adenomatous pathology or chronic inflammatory bowel disease (IBD) in itself represents a risk factor for a subsequent colon cancer. In patients with a colon cancer, synchronous colorectal cancers are found in 5–10%, whereas about 10–20% of patients with a history of colorectal cancer will develop metachronous primary cancers in the large intestine. A personal history of adenomatous colonic polyps is an indicator for an increased colonic predisposition to develop subsequent adenomatous or cancerous changes.14,71–75

Compared with the general population, relatives of patients with colon cancer have a two to four times increased risk of developing the disease themselves (Table 36-4).29,76,77 A similar, even though proportionally lesser, risk is observed for family members of individuals with colonic adenomatous polyps.77

Inflammatory Bowel Disease

Inflammatory bowel disease (IBD) is a strong risk factor for colorectal cancer. The risk correlates with the age of onset and extent and duration of active disease (Table 36-5 and Fig. 36-1).79,80 In contrast, however, the disease activity historically was not thought to be correlated with the risk, but recent studies have challenged this view.81 In ulcerative colitis, the risk of colorectal cancer increases from approximately 3% in the first decade to 10–20% in the second decade.79,80 In patients with Crohn's disease with colonic involvement, the disease-associated risk for colorectal cancer is also elevated but generally to a lesser extent.82–84

Other Factors

Less frequent risk factors for colorectal cancers may include a history of a ureterocolostomy85 or previous radiation treatment.86,60 The former requires the combination of fecal bacteria and urine because the microbes degrade urinary metabolites into strong carcinogens.85,87,88 When colonic mucosa is used for bladder augmentation, no increased cancer risk is observed owing to the absence of bacteria. Radiation-induced colorectal cancer is little less clear, but it has been suggested that it may be associated with a mucinous histology and poor prognosis.86

Prevention and Screening

Because symptoms are not reliable for early detection of colorectal cancer, risk-adjusted screening programs for asymptomatic individuals are important. Effective screening has to be based on an understanding of the adenoma-carcinoma sequence, which may take up to 5–10 years from the first molecular change to a clinically manifest cancer, and should reflect an individual's genetic and disease- or age-dependent risk for the development of colorectal cancer.13,14,89–91 Any prevention program has to be sensitive but also practical and cost-effective in order to achieve a broad screening of the population at risk. The term “screening” is applicable only to asymptomatic people; if symptoms are present, it is not screening but diagnostic tests that are initiated. Common tools for screening include fecal occult blood tests (FOBTs), flexible sigmoidoscopies or colonoscopies, and contrast enemas or CT colonography.92

The American Cancer Society, endorsed by the major professional societies, recommends starting colorectal cancer screening in asymptomatic average-risk adults at age 50.13,14,89–91 A slightly earlier screening start at age of 45 has been recommended recently for African American patients based on their statistically increased risk.11 A first baseline colonoscopy is to be performed and, if no pathology is found, repeated every 10 years. In addition, an FOBT should be done at an annual basis, and any positive result should precipitate a full colonic evaluation. Every 5 years, a limited endoscopy (flexible sigmoidoscopy) or barium enema is indicated. If precursor lesions are found, they should be removed and a colonoscopy be performed after 1–3 years to detect missed (20%) or recurrent polyps.93–95

In individuals at increased risk (eg, personal/family history of polyps or cancer or African American ethnicity) or at high risk (eg, cancer syndromes or IBD), the screening has to start earlier (see Table 36-2) and has to be performed at a higher frequency.11 Successful screening programs have been shown to reduce the colorectal cancer incidence by 76–90%.96

Carcinogenesis in the colon is a complex multistep process in which a multitude of alterations must coincide in order to transform a normal cell into a malignant cell. Several categories of genes are involved that normally are regulated in a sophisticated network to keep a tight balance between cell growth and turnover, cell death, DNA replication, and mismatch repair. Disruption of the fine balance between oncogenes, which promote cell proliferation, and tumor suppressor genes, which inhibit excessive growth, results in a growth advantage and allows malignant cells to expand.

Colon Cancer: A Genetic Disease

All cells of even such a complex organism as a human being have DNA that is virtually identical to the DNA found in the zygotes. DNA mutations can occur either as a germline mutation or as a somatic mutation. The former may be transmitted from one to the next generation as an inherited defect. More commonly, a spontaneous mutation occurs in a non–germline cell during the growth, development, and maintenance of a tissue or organ (somatic mutation). Even in the cycle of a normally functioning cell, there is a high chance of spontaneous gene mutations, most of which will not result in a growth advantage to the harboring cell. Genesis of a cancer therefore requires several independent accidents to occur in one cell. One can assume that a normal cell will be able to detect damage to its own DNA and maintain an effective repair mechanism. However, if the cell is too severely damaged, it might rather initiate the inherent suicide program called apoptosis. When a cell fails to recognize or correct a DNA damage and continues to replicate, accumulation of faulty gene products within the cell may eventually lead to a proliferative response. If that replication exceeds the growth potential of the neighboring normal cells, the mutation provides a growth advantage that will increase the state of “genetic instability” and hence lead toward a malignant cell.97 Despite this potential, most mutations are silent or lethal to the cell rather than beneficial in terms of providing the cell a biologic advantage. The triggers and the step-by-step cumulative failures that lead to carcinogenesis still are relatively poorly understood.

Two types of genetic instability may occur: at the chromosome level or at the DNA level. A loss of chromosomal material, that is, a chromosomal instability (CIN), results when the chromosomes are not divided symmetrically during mitosis such that one daughter cell receives both copies and the other cell receives none. On an electrophoretic gel, this can be visualized as a loss of one or more bands, which is described as loss of heterozygosity (LOH), and has been associated with a worse prognosis of colorectal cancer.2 The second form of genetic instability, at the DNA level, occurs when replication errors in repetitive short polymorphisms lead to an additional band or bands.98 This phenomenon is described as microsatellite instability (MSI), and it has been a characteristic feature of hereditary nonpolyposis colon cancers (HNPCCs).99

During the process of cell division, DNA is duplicated, with the original DNA serving as a template for the replicated copy. DNA polymerase serves as a “proofreader” that recognizes mismatched genes, halts the DNA synthesis, removes the defective sequence, and then resynthesizes the DNA. Failure of the DNA mismatch repair system predisposes to the development of mutations within daughter cells. Enzymes that monitor newly formed DNA and correct replication errors are called DNA mismatch repair (MMR) systems.

Specific gene functions are lost when both copies (alleles) of a gene are inactivated. Thus, when a germline mutation occurs in a suppressor gene, only the mutation of the remaining normal allele is required for the gene's loss of function. When both copies of the gene are normal, two mutational events are required for the gene's loss of function. This two-hit hypothesis may explain why inherited diseases usually manifest at an earlier age than sporadic disease.6

The Adenoma-Carcinoma Model

After identifying several genetic alterations in colorectal specimens at various stages of their neoplastic transformation and progression, Vogelstein and colleagues in 1988 pioneered a genetic model for colorectal tumorigenesis that since has been known as the adenoma-carcinoma sequence3 (Fig. 36-2). This multistep model described the carcinogenesis as an accumulation of genetic events, uninhibited cell growth, and proliferation and clonal development. Gene mutations and chromosomal/gene losses that were observed in sporadic colon cancer include the APC gene (adenoma–polyposis coli), MMC gene (mutated in colon cancer), K-ras, DCC (deleted in colon cancer), and p53.2,100,101 Mutations of the APC gene, which is involved in the control of cell-to-cell adhesions and intercellular communication, are found in 60% of even small adenomatous polyps, as well as in carcinomas,102 and therefore are believed to occur as a very early event in carcinogenesis. Mutations of K-ras, which under normal function plays a role in intracellular signal transduction and stimulated cell division, occur in larger adenomas and carcinomas and are thought to stimulate cell growth. Deletion of the tumor suppressor gene DCC may be important in the progression from a benign polyp toward a malignant condition.103 Mutations of the p53 gene, which are among the most frequent gene mutations in human cancers, are also common in invasive colon cancers but rare in adenomas, suggesting that p53 mutations occur as a late event in the development of the invasive phenotype.104 The wide range of gene mutations, inactivations, and deletions in the progression to carcinoma seem to hold the secret code for the various tumor behaviors observed in the clinical setting. It is important to note, however, that an increasing number of other genetic events have been observed and reported and that no single event seems to be equally present in all colon cancers. One therefore should caution that the described sequence is only one possible model and that the scenario may not reflect all aspects of colonic carcinogenesis.

Nonhereditary Colon Cancer

Sporadic Colon Cancer

Sporadic colon cancer, that is, colon cancer arising in individuals without a family history or an inherited predisposition, accounts for approximately 60% of all colorectal cancers and affects patients commonly older than 50 years. The risk factors associated with sporadic development of colon cancer have been discussed previously in the epidemiological section of this chapter (see Table 36-3).

Familial Colon Cancer

Familial colon cancer is the second most common (25–30%)6 and at the same time least understood pattern of genetic colon cancer development. In affected families, colon cancer develops too frequently to be considered a sporadic colon cancer, but the pattern is not consistent with the known inherited syndromes.78 An association of familial colon cancer has been found with polymorphisms, which reflect subtle genetic changes in the form of variations in the nucleotide base sequences but which do not affect protein structure.6 Familial colon cancer in the Ashkenazi Jewish population probably is the result of an APC germline mutation on codon 1307 (I1307K). This mutation, which predisposes to sporadic mutations at distant sites of the gene and later results in structural protein abnormalities, is found in 6% of all Ashkenazi Jews and in 28% of those with both a personal and a family history of colon cancer.105

Hereditary Colon Cancer

Familial Adenomatous Polyposis

Familial adenomatous polyposis (FAP) is an autosomal dominant inherited syndrome with near-complete penetrance. The offspring of affected individuals thus have a 50% risk of inheriting FAP. However, up to 20% of patients with FAP are new mutations without a family history. This condition is attributed to a truncating mutation in the germline adenomatous polyposis coli (APC) gene on chromosome 5q21.106 Variants of the polyposis syndrome are classified as Gardener's syndrome (ie, osteomas, desmoid tumors, thyroid neoplasms, and congenital hypertrophy of the retinal pigment epithelium) and Turcot's syndrome (ie, brain tumors).

The inherited syndrome of FAP and its variants accounts for less than 1% of all colon cancers. It is characterized by greater than 100 and often several thousands of adenomatous intestinal polyps that start to develop in the late teens and early twenties and turn into cancer by age 40–45. An attenuated variant of the disease is relatively rare and is characterized by a lower number and a later onset of both the polyps and the resulting cancer (see the following text). Nearly all FAP patients develop duodenal adenomas that are severe in 10% and account for the group's second highest cancer risk, with adenocarcinoma developing in the periampullary region in 3–10% of patients.107,108 Carcinoma arising in the antrum and duodenum after colectomy is the main cause of cancer-related deaths in FAP patients.107,109,110 Nonadenomatous fundic gastric polyps develop in approximately 10–30% of patients with FAP110 but usually do not have a malignant potential. Ten percent of FAP patients develop desmoid tumors either intra-abdominally or on the abdominal wall, extremities, and trunk.111 Histologically, desmoids are fibromatous lesions consisting of large proliferation of myofibroblasts. Even though they do not necessarily carry features of a malignant lesion, the recent literature suggests a low-grade sarcoma-like behavior. Desmoids are lethal in 10% and are the third most frequent cause for mortality of FAP patients, mainly owing to the intra-abdominal variants, which cause small bowel and ureteral obstructions.111,112

Approximately 25% of FAP patients remain without an identified APC mutation (APC-negative)112,113 and, using a detailed analysis, they seem to differ in terms of lower polyp number, later age at diagnosis, and lower occurrence of extracolonic manifestations as compared with classic FAP patients.110,114 This variant of FAP is known as attenuated familial adenomatous polyposis (AFAP).

Hereditary Nonpolyposis Colon Cancers

Hereditary nonpolyposis colon cancer (HNPCC), also known as Lynch I and II syndromes, is an inherited autosomal dominant disease that accounts for 3–5% of all colorectal cancers.115 It is characterized by an early onset of colorectal cancers predominantly but not exclusively on the right side of the colon with synchronous and metachronous cancers. Despite its name, these cancers typically arise from colonic polyps, but a diffuse polyposis is not present. The penetrance of the HNPCC predisposition is high and results in an 80–85% lifetime risk of colorectal cancer and a 40–50% risk of endometrial cancer.17,116,117 Furthermore, HNPCC patients are at increased risk of developing extracolonic malignancies, such as cancer of the small bowel, stomach, hepatobiliary tract, urinary tract, ovary, and brain. The Lynch variants describe patients with predominantly colorectal cancer at a young age (Lynch I) and those with both colorectal and extracolonic cancers (Lynch II).115

An initial observation of expansions and contractions of microsatellite DNA in the genome of colorectal tumor specimens from HNPCC patients established a link between HNPCC and the DNA MMR system.118–120 In contrast to the gatekeeper concept applicable to the APC gene in FAP, the DNA MMR genes belong to the so-called caretakers, which, when inactivated, do not promote tumorigenesis directly but rather lead to a genetic instability that then promotes tumor growth indirectly.121

In order to facilitate the clinical diagnosis of HNPCC, the International Collaborative Group on HNPCC (ICG-HNPCC) proposed the Amsterdam Criteria in 1990.115 Linkage studies in HNPCC families fulfilling Amsterdam Criteria I (Table 36-6) led to the discovery of the first two human MMR genes—hMSH2 and hMLH1. These genes accounted for 45–86% of all classic HNPCC families.122 There also was a higher risk for hMSH2 mutation carriers to develop extracolonic cancers, in particular endometrial cancer, as compared with hMLH1 mutation carriers.117,123 Several other MMR genes have been identified in conjunction with HNPCC and include hPMS1, hPMS2, and hMSH6. A recent study reported that endometrial cancer represents the most common clinical manifestation of HNPCC among female hMSH6 mutation carriers and that colorectal cancer cannot be considered an obligate requisite to define HNPCC.124 The ICG-HNPCC therefore revised the criteria (Amsterdam Criteria II), which now better weighs extracolonic manifestations (eg, endometrial, breast, small bowel, and upper renal tract cancers) as part of the family history (see Table 36-6). In addition, the less restrictive revised Bethesda Criteria (Table 36-7) were adopted to better serve patients who carry hMSH2 or hMLH1 gene mutations but otherwise do not fulfill the Amsterdam Criteria. Testing for MSI has become a valuable diagnostic tool to identify individuals with suspected HNPCC because 85–90% of HNPCC tumors have MSI as opposed to only 15–20% of sporadic colon cancers.99

Hamartomatous Polyposis Syndromes

Approximately 4% of colonic cancers are seen in the context of rare syndromes. Among these are inherited hamartomatous polyposis syndromes that are characterized by the presence of gastrointestinal hamartomatous polyps and an increased risk of gastrointestinal malignancy. Hamartomas result from a disordered differentiation during embryonic development and are characterized morphologically by disrupted representations of normal tissue components.

Peutz-Jeghers Syndrome.

Peutz-Jeghers syndrome is the second most common hamartomatous syndrome, occurring as an autosomal dominant condition with variable penetrance. Genetic alterations in the LKB1/STK (19p13) gene are responsible for approximately 50% of the cases of Peutz-Jeghers syndrome.127 The syndrome is associated with hamartomatous polyps of the gastrointestinal tract and cutaneous melanin deposition. The most common location of Peutz-Jeghers polyps is in the upper gastrointestinal tract, specifically the upper jejunum. One of the most characteristic features is the melanin depositions that is seen most frequently in the perioral region or buccal mucosa but also can occur in the genital region and on the hands and the feet. While a majority of these patients remain relatively asymptomatic, some may present with abdominal pain secondary to obstruction or impending obstruction owing to an intussuscepted polyp and others with gastrointestinal bleeding. Patients with Peutz-Jeghers syndrome have a moderately increased risk in the range of 2–3% to develop gastrointestinal malignancies as well as extraintestinal malignancies.

Juvenile Polyposis Syndrome.

Juvenile polyposis syndrome is the most common hamartomatous syndrome and is inherited as an autosomal dominant trait. The average age of onset is approximately 18 years, and there is an association with congenital birth defects in 15% of patients.128 Although the diagnostic criteria for juvenile polyposis syndrome are somewhat controversial, the most commonly used criteria include three or more juvenile polyps of the colon, polyposis involving the entire gastrointestinal tract, or any number of polyps in a member of a family with a known history of juvenile polyps.129

In infancy, patients may present with acute or chronic gastrointestinal bleeding, intussusception, rectal prolapse, or a protein-losing enteropathy. In adulthood, patients commonly present with either acute or chronic gastrointestinal blood loss. Most of these patients will be found to have polyps, which are located most frequently in the rectosigmoid region.

A germline mutation in the SMAD-4 gene (18q21) accounts for approximately 50% of the reported cases of the syndrome.130 A significant risk of colorectal cancer is associated with juvenile polyposis syndrome, and should not be confused with isolated juvenile polyps because the latter have virtually no malignant potential.

Cowden Syndrome.

Cowden's disease, first described in 1963, is known as multiple hamartoma-neoplasia syndrome. It is an autosomal dominant condition with nearly complete penetrance by age 20 that is caused by germline mutations in the PTEN tumor suppressor gene located at 10q22.131,132 Cowden's disease is unique among the hamartomatous syndromes because polyps arise more commonly from ectodermal rather than endodermal elements. Eighty percent of patients present with tricholemmoma, a benign tumor of the hair shaft. The central nervous system is the second most involved system, with approximately 40% of affected individuals suffering from macrocephaly. Only 35% of patients who meet the diagnostic criteria for Cowden's disease have gastrointestinal polyposis, but no increased risk of invasive gastrointestinal malignancy has been reported to date. The majority of patients with Cowden's disease suffer from benign thyroid or breast disease, on top of which adds a projected lifetime risk of 10% for thyroid cancer and of 30–50% for breast cancer.

Bannayan-Riley-Ruvalcaba Syndrome.

Formerly known as its subentity, the Ruvalcaba-Myhre-Smith syndrome, this rare autosomal dominant condition includes two other syndromes, both of which, like Cowden's disease, are associated with genetic alterations in the PTEN gene on chromosome 10q23 and may be considered a variant of juvenile polyposis coli.133–135 It is characterized by hamartomatous polyps of the gastrointestinal tract, macrocephaly, mental retardation, delayed psychomotor development, lipid storage myopathy, Hashimoto's thyroiditis, and hyperpigmentation of the skin of the penis. No increased risk of colorectal carcinoma, other gastrointestinal malignancies, or extraintestinal malignancy has been documented in these patients.

Cronkite-Canada Syndrome.

Cronkite-Canada syndrome is characterized by diffuse polyposis and ectodermal abnormalities such as alopecia, onychodystrophy, and skin hyperpigmentation. The syndrome can be distinguished by the diffuse distribution of polyps throughout the entire gastrointestinal tract with exception of the esophagus, which is spared.136 Symptoms include diarrhea, weight loss, nausea, vomiting, and anorexia, as well as paresthesias, seizures, and tetany related to electrolyte abnormalities. Cancer occurs in the stomach, colon, and rectum, but it remains controversial whether polyps in Cronkite-Canada syndrome possess malignant potential. As many as 15% of patients with Cronkite-Canada syndrome have a malignant tumor at the time of diagnosis.

Polyps

Polyp is a descriptive clinical term for any mucosal elevation. Polyps are further categorized along several dimensions, including

1. Size

2. Character of their attachment to the bowel wall (eg, sessile or pedunculated)

3. Cellular architecture (eg, adenomas, hyperplastic, hamartomas, inflammatory) and histologic appearance (eg, tubulous, tubulovillous, villous)

4. Progression from benign to malignant behavior (eg, benign, dysplastic, cancer)

Most polyps are neoplastic but not necessarily malignancies. Neoplastic polyps consist of cells with the potential to acquire over time the ability to invade and to spread, that is, metastasize. Dysplasia is a term used to describe the intervening state between normal tissue and invasive malignancy.

Polyp Size

The most immediate way in which a polyp can be described is by its size. Intuitively, polyps with a larger mass have a greater volume of neoplastic cells, and hence a higher likelihood of harboring cancer. The relationship between adenomatous polyp size and the presence of invasive malignancy was analyzed elegantly by Nusko et al (Table 36-8).137

Polyp Attachment to Bowel Wall

Polyps of any size or architecture may be pedunculated, sessile, or some combination of both. The main clinical relevance of this distinction lies in the ease of endoscopic removal, with pedunculated polyps being clearly more amenable to removal without surgical intervention.138,139

It is important to note that the way in which a polyp is attached to the wall of the colorectum does not accurately predict the presence versus absence of an invasive malignancy.

Malignant polyps of the colon can be either pedunculated or sessile. The type of treatment that should be offered to a patient depends much more on the other characteristics of the polyp.

Polyp Architecture

Based on their histological structure, polyps can be categorized into adenomatous and nonadenomatous polyps, the latter of which consists of hyperplastic, hamartomatous, and inflammatory polyps.

Adenomatous Polyps (Adenomas).

The most common type of polyp in the colon is the adenomatous polyp. Adenomatous polyps are categorized as tubular, tubulovillous, or villous based on the extent to which the dysplastic epithelium is organized with the normal-appearing tubular architecture.140 Tubular adenomas are defined by the presence of tubules within 80% or more of the lesion; adenomas with less than 20% showing a tubular configuration are villous lesions; and the remainder is considered tubulovillous. The majority of polyps are tubular (87%), with a minority either tubulovillous (8%) or villous (5%).141

With few exceptions, the treatment for an adenomatous polyp is endoscopic polypectomy. Colorectal cancer screening programs that include colonoscopy with polypectomy have demonstrated a reduction in the incidence of colorectal cancer and colorectal cancer mortality.142 It is difficult, however, to estimate the likelihood that a small adenoma will progress to a dysplastic adenoma and eventually into cancer. A number of biologic and molecular markers have been analyzed as predictors of a malignant potential, but these are not widely utilized.143 Longitudinal and comparative data suggest that polyps not only progress but also may regress.144 Despite these vagaries, any adenomatous polyp should be considered a premalignant lesion and be treated as such.

Invasive carcinoma is present in 5% of all adenomas, but the incidence correlates with the size and type of the adenoma (Table 36-9).137,145

The Haggitt classification, which defines four levels within the polyp, has evolved as a useful tool to describe the degree of cancer invasion into a pedunculated or sessile adenomatous polyp.147 This classification forms the basis of the management of malignant polyps (Fig. 36-3). In Haggitt's levels 1, 2, and 3, the risk of lymph node metastasis in a surgical specimen is less than 1%, whereas a level 4 invasion of the stalk behaves like a sessile T1 lesion and carries a higher risk of 12–25% of having lymph node metastases. A similar, but less well-known, classification was developed in 1993 by Kudo and associates, who for prognostic purposes suggested to divide the submucosal invasion of sessile malignant lesions into three levels (Sm1, Sm2, Sm3) (Fig. 36-4).148

Flat and/or depressed adenomas are a subtype of colonic adenoma with a propensity for high-grade dysplasia in 10–41% of affected patients regardless of the small size of these lesions.149 The entity was first described in Japan, where they seem to occur at a regular frequency. These lesions, which are flat or slightly raised to less than 2 mm and commonly less than 1 cm in size, may be overlooked easily on colonoscopy and turn into a cancer before having reached a size comparable with classic cancers.149–152 Recent screening studies, which took advantage of chromoendoscopy techniques, have confirmed that flat adenomas represent up to 25–36% of all polyps found in a random cohort and are present in 8–11% of the population.152,153

Hamartomatous Polyps.

A hamartomatous polyp is composed of a spectrum of different cellular elements and is considered a nonneoplastic entity with no significant premalignant potential.154,155 Several clinical syndromes manifest with a polyposis of hamartomatous polyps (Juvenile polyposis, Cowden syndrome, Bannayan-Riley-Ruvalcaba syndrome, Cronkite-Canada syndrome) and these have been discussed earlier in this chapter. These syndromes carry varying risks of intestinal and extraintestinal disease, and several also impose an increased likelihood of developing intestinal cancer due to immature glandular elements in the hamartomatous polyp. Stable estimates of this risk are difficult to calculate because of the relative rarity of these diseases.

Hyperplastic Polyps.

Hyperplastic polyps are small, sessile mucosal outgrowths that display an exaggerated crypt architecture. They are usually small, with only very few (1–4%) larger than 1 cm; however, these larger polyps actually may be serrated adenomas rather than hyperplastic polyps (see the following text).156 Within the colorectum, hyperplastic polyps commonly have a distal distribution pattern, predominantly in the rectum and sigmoid colon, and they have been reported in up to 75% of patients older than 60 years at autopsy.157 It is not unusual to find several of these polyps in a single individual.

Histologically, hyperplastic polyps display well-formed glands and crypts that are lined by nonneoplastic epithelial cells. Because of their small size, hyperplastic polyps are generally clinically silent, but large or multiple hyperplastic polyps occasionally can be responsible for gastrointestinal symptoms.

Historically, hyperplastic polyps have been considered benign and not premalignant.156 This paradigm has been increasingly questioned, beginning in 1990 with work by Longacre and Fenoglio-Preiser.158 The ability of hyperplastic polyps to develop defective mismatch repair genes and foci of microsatellite unstable cancers has been documented, strengthening this concept.159 Additional research has illuminated an epigenetic pathway, whereby a promoter region in the DNA of hyperplastic polyps is methylated, resulting in progression along a sequence of steps that leads to a serrated adenoma and eventually carcinoma.160 The clinical significance of hyperplastic polyps and serrated adenomas is a topic of emerging importance in the field of colorectal cancer prevention.

As with adenomatous polyps, individuals who have a predisposition to developing hyperplastic polyps may be at increased risk for developing colorectal cancer. The endoscopic and radiologic appearance of the mucosal abnormalities in hyperplastic polyposis closely resembles FAP, but the syndrome is not believed to be heritable and does not have any extraintestinal manifestations. The World Health Organization (WHO) has defined criteria for this entity as follows: (1) at least five histologically diagnosed hyperplastic polyps of which two are greater than 20 mm or (2) any number of hyperplastic polyps occurring proximal to the sigmoid colon in someone who has a first-degree relative with hyperplastic polyposis, or (3) more than 30 hyperplastic polyps of any size that are distributed throughout the colon and rectum.161 The risk of colorectal cancer being present or developing subsequently in a patient meeting these criteria are high in case series, but population-based studies have not yet been performed.162 While prophylactic colectomy has been proposed for patients with hyperplastic polyposis, there are no consensus opinions at this time regarding the appropriateness of this approach.163 At a minimum, a program of intensive colonic surveillance is indicated.

Inflammatory Polyps.

Inflammatory polyps are the result of reactive regenerative processes occurring in or next to a damaged epithelium. Because of the extent and chronicity of IBD, inflammatory polyps are most commonly seen in that context. The prominence of inflammatory pseudopolyps often is the result to the presence of adjacent ulcerations. Histologically, a combination of distorted crypt architecture in conjunction with granulation tissue and inflammatory infiltrates is characteristic. Even though the underlying chronic IBD represents a high risk for colorectal cancer, the inflammatory polyps as such do not carry a malignant potential. Biopsies in IBD should therefore also include the more flat-appearing areas rather than the polyps only.

Polyp Transformation

By definition, the neoplastic nature of an adenomatous polyp represents dysplasia. In an effort to quantify the clinical severity/importance of dysplasia, however, the degree of dysplasia is categorized and reported in three grades. This categorization is based on the histopathologic differentiation and architecture of the epithelial cells within the polyp.

Common terms for polyps include low-grade dysplasia, intermediate-grade dysplasia, and high-grade dysplasia (by some also referred to as in situ [Tis] adenocarcinoma). Once there are clear microscopic features of tumor invasion through the muscularis mucosa of the colorectum, an invasive cancer (T1 or greater) is present. This important demarcation is based on the finding that lymphatic vessels are almost never found superficial to the muscularis mucosa. The descriptive terms for invasive cancer include well-differentiated (grade I), moderately differentiated (grade II), or poorly differentiated (grade III) adenocarcinoma.

Management of Colorectal Polyps

The overarching goal of physicians treating patients with colorectal polyps is to minimize the risks associated with invasive malignancy, while simultaneously avoiding complications of diagnosis and treatment. Colorectal cancer prevention programs are widely believed to reduce the risk of colorectal cancer mortality through endoscopic removal of premalignant lesions and the detection of invasive lesions at a point in their progression where they are asymptomatic. The efficacy of colorectal cancer prevention programs has been proven in multiple randomized and nonrandomized studies.142,164–168

The majority of colonic polyps can be removed via colonoscopy, but this may not be the case for one of two reasons. First, a polyp may not be resectable due to size, attachment to bowel wall, or other reasons related to the anatomy of the patient or polyp. In these situations, a careful assessment of the risks of surgical resection versus observational management is warranted, as 12–18% of these polyps harbor an invasive malignancy.169–171 Second, polypectomy may not be reasonable in the presence of innumerable polyps.

When invasive cancer is found in a polyp, the management is based mainly on the level of invasion and the completeness of the polypectomy. Based on Haggitt's observations (see Fig. 36-3), it has been suggested that colonic cancers invasive to Haggitt's levels 1, 2, and 3 can be adequately treated with polypectomy (2-mm margin), whereas polyps with invasion into Haggitt's level 4 should be treated like a sessile lesion.147,172

Management of sessile lesions is more controversial. If a sessile lesion cannot be snared in one intact piece with a microscopically clear margin of at least 2 mm, or if it demonstrates lymphovascular invasion or deep invasion into level Sm3 (lower third of submucosa) (see Fig. 36-4), the patient should undergo a formal oncologic resection of the colon. The approach for an adequately removed lesion with a lesser extent of invasion into the submucosa—Sm1 (invasion only into upper third of submucosa), Sm2 (invasion only into upper two-thirds of submucosa)—should be individualized based on the risk of a surgery versus the risk of lymph node metastases.172,173 It is advisable in any case to tattoo the area of a suspect polyp endoscopically with India Ink for later identification of the site.

Malignant Tumors of the Colon

The vast majority of malignant colon neoplasms are cancers (carcinoma), that is, malignant neoplasms of epithelial origin. Based on the endodermal glandular tissue origin, adenocarcinoma and its histologic variants are by far the predominant histopathology and account for 90–95% of all colorectal malignancies. The majority of this section is therefore devoted to these types of tumors, but it also briefly discusses nonepithelial tumors of the colon.

Adenocarcinoma

Colorectal cancer (adenocarcinoma) is the most frequent malignancy of the gastrointestinal tract, the fourth most frequently diagnosed malignancy, and the fourth most common cause of cancer-related mortality in the world.174 Squamous and adenosquamous carcinomas are exceptionally rare and are located characteristically in the rectoanal junction. The histopathologic classification of colorectal cancer as defined by the WHO is illustrated in Table 36-10.

Macroscopically, most colorectal cancers have either a polypoid or an ulcerative-infiltrating appearance, but combinations are frequent. Very rarely, colorectal cancer may have a dissolute growth pattern and resemble linitis plastica of the stomach, in which case a metastatic lesion from another primary site (eg, lobular breast cancer, stomach cancer) or a nonepithelial neoplasia (eg, lymphoma, carcinoid) would need to be ruled out.

Adenocarcinoma, the exceedingly predominant histopathology of colon cancer, has a less frequent variant of mucinous adenocarcinoma that includes signet ring cell carcinoma and accounts for approximately 10% of all colorectal cancers. Compared to nonmucinous colon cancers, mucinous carcinomas usually present at a more advanced stage and thus have an overall poorer prognosis.8,175,176

A rare variant of colorectal cancer is small cell cancer, which accounts for less than 1% of all cases and, similar to small cell cancer of the lung, and appears to be related to some degree to a neuroendocrine origin. These tumors have a high tendency to develop widespread metastasis early in the course and have an extremely poor prognosis.

The distribution of colorectal cancers among the various segments has seen a continued shift toward right-sided colon cancer.177,178 An estimated 45–55% of colorectal cancers are located in the rectum (10–15%) or sigmoid colon (40%), 25–35% in the cecum or ascending colon, whereas the remaining are equally distributed through the rest of the colon. The local growth pattern for colorectal cancer involves circumferential and transmural invasion of the tumor through the intestinal wall into the peritoneal cavity or surrounding organ structures. Tumor dissemination primarily occurs through access to the lymphatic vessels into the locoregional lymph nodes or through access to the blood stream as hematogenous metastasis to distant organs. The most common site of blood-borne spread is via the portal venous system to the liver; other secondary locations include the lung or, less frequently kidneys, bone, etc. In addition, tumor dissemination can occur by transperitoneal seeding and result in peritoneal carcinomatosis.179 Following gravity, peritoneal seeds may accumulate in the pelvic cul-de-sac or paracolic gutters where they can grow to a considerable size (Blumer's shelf). Growth by perineural infiltration may be seen on microscopic examination and has a negative prognostic impact. About 20% of the patients have evidence of distant metastases (stage IV disease) at the time of presentation.

Staging of Colon Cancer

Modern staging of colorectal cancer defines four clinical stages (I–IV) based on the TNM (tumor-node-metastasis) system, which has just recently been updated by the American Joint Committee on Cancer (AJCC) (Tables 36-11 and 36-12).8,180,181 Independent parameters are (1) the depth of tumor invasion (T) into or through the layers of the intestinal wall with or without invasion of adjacent organs, (2) the number of regional lymph nodes involved (N), and (3) the presence or absence of distant metastases (M). Additional modifiers are used to reflect the method of stage determination (p for pathology, c for clinical, u for ultrasound), and y to indicate a status after neoadjuvant treatment.

Historical classifications such as Dukes and Astler-Coller are still sporadically in use but largely have been and should be abandoned.

Because the extent of tumor resection (complete vs incomplete) strongly correlates with prognosis, the AJCC released additional guidelines to reflect the extent of residual tumor after a surgical resection with the letter R (see Table 36-11).181

Nonepithelial Tumors of the Colon

Benign Nonepithelial Tumors

Lipomas and Lipomatous Polyposis.

Lipomas are submucosal lesions that develop in the fifth or sixth decade of life and are more common in the large than in the small intestine. Histologically, the polyps consist of a submucosal lump of adipose tissue that is covered with a normal colonic mucosa. Whereas solitary lipomas tend to occur more frequently on the right side of the colon in the vicinity of the ileocecal valve or the ascending colon, lipomatous polyposis may diffusely involve the entire small and large intestine.

Lipomas generally are asymptomatic but may be found incidentally on colonoscopy. The characteristic appearance is a smooth mass with normal overlying mucosa. The soft nature of the lipoma can be demonstrated by poking the tumor with an endoscopic instrument (“pillow test”). Asymptomatic, incidentally detected lesions should be left alone.

Occasionally, when lipomas become large enough to protrude into the lumen, they may cause symptoms such as gastrointestinal bleeding, diarrhea, intussusception, or bowel obstruction.182 Endoscopic removal of such a lipoma with a snare often is possible but has a risk of hemorrhage because the fat may prevent the cautery from adequately transmitting the energy to the blood vessels in the stalk. Surgery may be required if such a complication occurs; it should therefore be considered preemptively for very large symptomatic lipomas. Alternatively, the mucosa overlying the lipoma may be opened endoscopically to allow the lipoma to spontaneously enucleate into the lumen.

Potentially Malignant Nonepithelial Tumors of the Colon

Carcinoid or Neuroendocrine Tumors.

Modern nomenclature classifies carcinoids as neuroendocrine tumors, based on their neuroendocrine origin. They are characterized by subepithelial nests of epithelial-appearing cell elements. Carcinoid tumors may occur anywhere in the entire body. A recent study on 11,427 patients from the SEER database found that the gastrointestinal tract is affected in 55%, with the most frequent locations being the small intestine (44.7%), the rectum (19.6%), the appendix (16.7%), and the colon (10.6%),183 a finding that contrasts with traditional reports that the appendix is the most frequent site in the gastrointestinal tract. The annual incidences for the colon and rectum were reported to be 2.0 and 4.2 cases per 100,000 people per year, the risk of metastasis proportional to the size of the carcinoid.183 Unlike most neoplasms, invasiveness of carcinoid tumors is not entirely based on histological criteria (eg, invasion of muscularis propria) but includes clinical aspects. In absence of other definite indicators for malignant behavior, carcinoids smaller than 1 cm are considered benign, lesions larger than 2 cm are likely malignant, and the gray zone in between remains undetermined or potentially malignant.184 Malignant carcinoids may spread locoregionally into the lymph nodes or directly to the liver.

Patients with a gastrointestinal carcinoid tumor may be either completely asymptomatic or present with intestinal obstruction, bleeding, carcinoid syndrome, or carcinoid heart disease, that is, acquired and commonly right-sided valvular heart disease.185,186 Vasoactive substances (eg, serotonin and 5-hydroxyindolacetic acid [5-HIAA]) are released from carcinoid tumors but for the most part are eliminated in a hepatic first-pass effect before reaching the systemic circulation. Carcinoid syndrome is therefore a bad prognostic sign because it does not typically develop until metastatic lesions in the liver directly release their products into the systemic circulation. Hindgut carcinoid tumors (those located in the distal transverse colon and beyond) classically do not cause carcinoid syndrome because they are less endocrinologically active.

Diagnosis of a carcinoid may be suspected clinically but can be difficult to confirm histologically short of a surgical resection because the lesions are submucosal and not commonly in reach of an endoscopic biopsy. A preoperative workup for a carcinoid tumor should include a 24-hour urine collection of 5-HIAA and a plasma chromogranin A. Both parameters can also be used for postoperative surveillance. Cross-sectional imaging and somatostatin receptor scintigraphy are tools to evaluate for systemic disease. Multicentricity and associated high rates of synchronous gastrointestinal and genitourinary malignancies warrant both an upper and lower gastrointestinal endoscopy.187

An oncologic resection should be performed in all carcinoids larger than 2 cm unless contraindicated by clinical circumstances. Tumors of smaller than 1 cm size may be managed locally, whereas the management of lesions measuring 1–2 cm remains controversial.184

Gastrointestinal Stromal Tumors (Gists).

GISTs are the most common mesenchymal tumors of the gastrointestinal tracts and originate from the intestinal pacemaker cells, the interstitial cells of Cajal.188 Sixty percent of GISTs are found in the stomach; 29% in the small intestine; 2% in the colon, rectum, and rectovaginal septum; and 9% in the esophagus.189 Symptoms are nonspecific and include pain, obstruction, bleeding, and a mass. Distinction from other mesenchymal tumors (eg, leiomyosarcoma) is important from a prognostic point of view. Tumor size and light microscopic determination of the mitotic rate (mitotic figures per x number of high-power fields) are the most important conventional prognostic indicators.188 The diagnosis of GISTs is based on morphologic features and immunohistochemical demonstration of c-kit (CD117) expression. This marker is seen in almost all GISTs and is regarded as one of the key diagnostic elements, but a few otherwise characteristic tumors are found to be c-kit negative.190 While the majority of GISTs have activating mutations of the KIT receptor tyrosine kinase, another subset of tumors show mutations in the KIT-related kinase gene platelet-derived growth factor receptor alpha (PDGFRA).191 KIT and PDGFRA mutations appear to be alternative and mutually exclusive oncogenic mechanisms in GISTs.192 Determination of CD117 expression is of practical importance because positivity correlates with a tumor response to treatment with imatinib (Gleevec), which inhibits KIT kinase activity. Surgical resection is the primary treatment for localized GISTs that are resectable without mutilation. Recurrent and locally advanced or metastatic tumors are treated increasingly with imatinib in a palliative, adjuvant, or neoadjuvant setting.

Nodular Lymphoid Hyperplasia.

This condition is characterized by numerous polyps in the small and large intestine, rarely in the stomach, which consist of enlarged submucosal lymphoid follicles. Associated diseases are immune deficiencies of various origins (eg, tumors, hematoproliferative disorders, immunoglobulin A deficiency, and human immunodeficiency virus [HIV] infection), in which case recurrent infections (eg, giardiasis) appear to promote the nodular lymphoid hyperplasia. Immunocompetent patients usually are asymptomatic, and the nodular lymphoid hyperplasia is an incidental finding. Nodular lymphoid hyperplasia has been associated with an increased subsequent incidence of lymphoma (small bowel).193

Malignant Nonepithelial Tumors of the Colon

Lymphoma.

Primary malignant lymphoma of the colon is uncommon and accounts for only 0.2–0.4% of all colonic malignancies and 10–15% of all primary lymphomas of the gastrointestinal tract, which themselves account for about 30% of extranodal lymphomas.194 The most frequent colonic location is the cecum (70%), followed by the rectum and ascending colon. The gross appearance may be a circumferential or polypoid mass, an ulceration, or a diffuse infiltration with stricturing and bowel wall thickening.195 Eighty-six percent of the lesions are solitary, but they can be multiple and diffuse in nature. The intestinal lymphomas may be subclassified into B-cell lymphomas (85%) and T-cell lymphomas (15%). Among the B-cell lymphomas, mantle cell lymphoma has a worse prognosis, whereas mucosa-associated lymphoid tissue (MALT) lymphomas have a better prognosis than other B-cell tumor types.195 While surgical treatment may be indicated for some localized tumors, many authors consider medical management to be the primary treatment. It may include new approaches such as anti-infectious treatment for MALT lymphoma or reconstitution of the patient's immune status, for example by means of antiretroviral treatment in HIV-associated B-cell lymphoma.196

Multiple lymphomatous polyposis of the gastrointestinal tract is a distinct clinicopathologic entity. This rare form of primary gastrointestinal lymphoma occurs most often in elderly patients and accounts for 9% of all gastrointestinal lymphomas.197 The polyps can be widespread throughout multiple segments of the gastrointestinal tract. Histopathologic and immunohistochemical techniques are required to differentiate lymphomatous polyposis from other forms of gastrointestinal polyposis.

Kaposi Sarcoma.

This commonly multifocal angiosarcoma has been associated with herpesvirus-8 (HHV-8) infection in conjunction with immunosuppression (eg, HIV/AIDS, chronic steroid or immunosuppressant medication, etc). The incidence in organ transplant recipients is about 0.5–0.6% but most frequently involves the skin. Extremely rarely, however, the anorectum or intestines are involved and shows characteristic bluish-purple submucosal nodules. Treatment primarily aims at improving the immune status, but chemotherapy and, rarely, radiotherapy may be indicated in patients in whom the immune status cannot be restored.198

Smooth Muscle Tumors.

Smooth muscle tumors of the colon are rare and occur most commonly in the form of a pedunculated leiomyoma of the muscularis mucosa. Leiomyosarcomas, which consist histologically of spindle cells that resemble smooth muscle cells, are even less frequent but are characterized by an extremely aggressive and rapidly fatal growth pattern. Whenever possible, oncologic resection and adjuvant chemotherapy are the treatment of choice.199

Secondary Tumors to the Colon

Endometriosis.

Endometriosis may involve the colon or rectum in approximately 15–20% and may mimic colonic carcinoma. The lesions are rarely larger than 5 cm, involve the subserosa and muscle coats, and may project into the lumen of the bowel. When endometrial tissue extends through to the colonic mucosa, biopsy may be mistaken for adenocarcinoma.

Invasion from Extracolonic Cancers.

Locally advanced tumors from noncolonic primary cancers may directly invade the colon and cause symptoms suggestive of colon cancer (bleeding, obstruction, fistula). These tumors originate from organs in close adjacency to the colon (female organs, bladder, prostate, kidneys, pancreas, duodenum, liver).

Metastatic Cancer.

Carcinomas from other primary sites may metastasize to the colon and occasionally mimic a primary colon cancer. Metastases originate most commonly from lobular breast cancer, stomach cancer, ovarian cancer, malignant melanoma, and leukemia, the latter of which can be diagnosed by the hematopoetic infiltrates.

A fundamental knowledge of the anatomy is unquestionably a key to success-oriented surgical technique aiming at the best oncological outcome and a minimized morbidity. The large intestine starts at the ileocecal junction and extends to the anus. It is about 5–6 ft (125–150 cm) long and can be divided into the cecum with the appendix, the ascending colon, the transverse colon, the descending colon, the sigmoid colon, and the rectum. Definitions of where the sigmoid colon ends and the rectum begins have not always been uniform. The best definition of the rectosigmoid junction from a functional as well as surgical viewpoint is the confluence of the teniae coli.200 However, the inability to visualize this anatomic reference point endoscopically recently led the NCI and other expert committees to define the rectum for the purpose of uniformity in clinical trials as the last 12–15 cm above the anal verge as measured by rigid sigmoidoscopy.201 This endoscopic definition is necessary in order to determine the appropriateness of preoperative (neoadjuvant) chemoradiation for rectal but not sigmoid cancer.201 Obsolete because highly variable and therefore inaccurate definitions relate the rectosigmoid junction to the level of (1) the peritoneal reflection or (2) the sacral promontory.

The arterial and venous blood supply, as well as the lymphatics of the colon, is summarized in Fig. 36-5. The arterial blood supply to the colon comes from the superior mesenteric artery (SMA) and the inferior mesenteric artery (IMA), which communicate in a watershed area in the splenic flexure (artery of Drummond). The rectum has additional branches from the internal iliac vessels. With a significant degree of anatomic variation, the major vascular stalks to the colonic segments consist of the ileocecal and right colic artery (last branch of the SMA), the middle colic artery (second branch of the SMA), the left colic artery (first branch of the IMA), and the superior hemorrhoidal artery (distal branch of the IMA). The venous blood supply peripherally follows the arterial branches but more centrally divides into the superior mesenteric vein and the inferior mesenteric vein, which connect at separate levels to the portal system. The lymphatic drainage starts with lymphatic follicles in the colonic submucosa, drains through the colonic muscle wall into the epicolic nodes, and continues to the paracolic lymph nodes that follow the blood vessels to the bowel, along the major arteries to the principal lymph nodes at the level of the arterial runoff from the aorta. These lymph node groups consist of the celiac, the superior mesenteric, and the inferior mesenteric groups of lymph nodes.

For a safe surgical technique, the relationship of the colon with adjacent structures, mostly in the retroperitoneum, has to be fully understood. The colon is only a partially intraperitoneal organ. Only the transverse colon and the sigmoid colon are fully peritonealized and have a free mesocolon; the ascending colon and the descending colon, including both flexures, are partially located in the retroperitoneum and therefore reside in proximity to essential anatomic structures. The structures most at risk during a right hemicolectomy include the right ureter and the duodenum; during a transverse colon resection, the SMA/SMV (and its branches) and the gastroepiploic vessels at the gastric curvature; during a takedown of the splenic flexure, the spleen, pancreas, and left kidney; and during a left colon or sigmoid resection, the left ureter, the gonadal vessels, and the hypogastric nerves.

Symptoms and Differential Diagnosis

Colorectal cancer does not have any early signs. In fact, symptoms are often absent until a tumor has grown to a significant size. Unless a patient presents with a tumor complication (eg, bowel obstruction, bleeding, perforation, or fistula formation), symptoms mostly are subtle or uncharacteristic and vague. They may consist of unexplained weight loss, anemia and weakness from chronic blood loss, flatulence, or episodes of colicky abdominal pain. If present, these symptoms therefore always should be suspicious for a locally relatively advanced tumor stage, which is also reflected by the fact that about 20% of colorectal cancer patients at the time of their first presentation already have stage IV disease with distant metastases (Table 36-13). As the stool in the proximal colon is still liquid or at most semisolid, proximal colon tumors may grow to relatively large size before they cause an obstruction. The more distal a lesion is localized (eg, left colon or rectum), the more likely the changes in bowel habits occur. These include rectal bleeding or mucous discharge in or with the stool, sudden onset of constipation, alternating periods of diarrhea and constipation, or a decreasing diameter of the stool. Pelvic or anal pain is an ominous sign because it may occur with increasing size, perforation, or sphincter invasion of a rectal cancer.

Any large bowel obstruction, bleeding per rectum, gas or stool passage other than through the anus, or peritoneal signs should raise the index of suspicion for a colorectal malignancy until proven otherwise. Several other conditions and diseases have to be considered in the differential diagnosis. Obstructive symptoms may result from chronic diverticulitis, benign polyps, Crohn's disease, endometriosis, or a postischemic stricture. A fistula may suggest complicated diverticulitis, Crohn's disease, or tuberculosis. Bleeding per rectum may also be found in hemorrhoids and other benign anorectal conditions, diverticular disease, arteriovenous malformations, endometriosis, and proctitis or colitis. However, even if one of these benign diseases is found on clinical evaluation, the symptoms should not be attributed automatically to them before a malignant disease of the large intestine has been ruled out.

Because symptoms are not reliable for the prevention or early detection of colorectal cancer, risk-adjusted screening programs for otherwise asymptomatic individuals (as discussed in an earlier section of the chapter) are crucial in order to achieve a reduction in cancer mortality.

Not only management planning in a situation with acute cancer complications should include strategies to alleviate symptoms and minimize the morbidity from the complication but also should provide an oncologically adequate treatment for the tumor.

History and Physical Examination

A careful history and physical examination remain the cornerstone in all patients presenting with gastrointestinal symptoms. This review should include questions about changes in bowel habits, time of last stool and gas passage, weight loss, and a personal or family history of cancer, particularly of colorectal cancer or its precursor lesions. Awareness of possible underlying diseases and genetics that predispose to colorectal cancer is of utmost importance not only for the management of the individual patient but also for adequate counseling of potentially affected family members.

A careful physical examination follows to identify any palpable tumor masses and/or signs of tumor complication or dissemination. Apart from vital signs and temperature, the patient's general appearance may reveal evidence of cachexia, dehydration, jaundice, or lymph node enlargements. For example, enlargement of the left supraclavicular nodes may be the first but late sign of a disseminated gastrointestinal malignancy (Troisier's sign). The abdomen is examined for a palpable primary tumor, hepatomegaly (liver metastasis?), distension and/or tympanitic bowel sounds (partial or complete bowel obstruction?). Presence of peritoneal signs such as guarding with local direct and rebound tenderness or percussion tenderness may indicate a tumor perforation. A digital rectal examination and proctoscopy are mandatory to rule out involvement of the rectum or to determine the exact distance of a distal and possibly palpable tumor from the anal verge, its axial and circumferential extent, and the mobility of the tumor against surrounding structures (eg, sacrum, prostate/ vagina, anal sphincter muscle). In addition, the checking finger should assess the rectal vault for the presence of stool, blood, or melena.

A thorough general physical examination is necessary to evaluate the patient's general health status regarding the ability to tolerate a major abdominal procedure under general anesthesia. Particular attention has to be paid to patients who present with acute symptoms in an emergency setting. Prolonged fasting, nausea or vomiting, and translocation of fluids into the third space during a period of bowel obstruction or after a perforation will rapidly result in a state of malnutrition and dehydration. Developing sepsis or acute and recurrent blood loss potentially aggravate these symptoms and may result in a severe volume loss. Alarming signs are a decrease in urine output, tachycardia, hypotension, elevated temperature, short-term weight loss, standing skin folds, dry oral mucosa, and acidosis. Immediate fluid and volume resuscitation has to parallel the further clinical workup and monitoring. Blood tests have to be interpreted with caution; for example dehydration may result in an artificially high hematocrit and mask a significant loss of blood.

Investigations

Patients with symptoms suggestive of colorectal cancer should undergo a series of timely investigations with three goals: (1) to assess the large bowel regarding the primary lesion, concomitant lesions, and a potential underlying colonic disease, (2) to determine whether the tumor has metastasized, and (3) to assess the patient's operability (overall condition and comorbidities).

Complete Evaluation of the Large Intestine

Irrespective of the method used, the primary goal is to document the presence of a malignant pathology and to rule out concomitant lesions in other segments of the large intestine. Both endoscopic and radiological techniques are available for evaluation of the colon and rectum, and each type of examination has inherent strengths and weaknesses.

Rigid Proctoscopy and Flexible Sigmoidoscopy.

These first-line diagnostic tools are used mostly in the outpatient setting to accurately assess lesions in the distal colon and rectum. The two methods are rapid, widely available, and require only minimal bowel preparation (enema). However, they do not provide complete information about the rest of the colon, and therefore a complementary study is indicated before surgery. Furthermore, the flexible sigmoidoscope is notorious for giving inaccurate measurements of the level of the tumor. Determination of the rectal versus colonic location of the tumor should be done with a rigid proctoscope.

Colonoscopy.

Colonoscopy clearly has evolved as the method of choice because of its high sensitivity in detecting tumors and its ability to take biopsies. It provides accurate information about the entire colonic mucosa (ie, polyps, synchronous cancer, colitis, melanosis, and diverticula), and it may be used to remove synchronous neoplastic polyps. Apart from determining the circumferential and longitudinal extent of a colonic lesion, colonoscopy addresses functional aspects such as active bleeding or an imminent obstruction by cauterization, laser ablation, or placement of a self-expanding metallic wall stent, hence allowing for turning an emergency situation into an elective one.

While the overall risk of colonoscopy is very low with a much less than 1% incidence of a bowel perforation, there are some limitations to the technique. There is a 25% risk of smaller lesions to escape detection, and an estimated 10% incidence that the cecum for technical reasons may not be reached. In addition, the precise position of a lesion seen on colonoscopy may not be determined adequately unless one of the two absolute landmarks (dentate line, or the carpet-like villi of the terminal ileum) is in direct proximity. Relative landmarks (eg, assessment of the endoscopic shape of the colon, liver and spleen shadow, ileocecal valve, appendiceal orifice) or the length of instrument insertion from the anal verge vary considerably and should not be used. In practical terms, however, this handicap may be overcome by India Ink tattooing of the area of a lesion for better identification during surgery or repeat endoscopy.

Contrast Enema.

Radiographic contrast enemas alternatively can be used for a colonic evaluation. Contrast enemas are an especially valuable adjunct to colonoscopy in patients with near-obstructing colonic lesions. Furthermore, they have the advantage of more accurately visualizing the anatomic position of a colonic lesion (road map). Ideally, a barium-air double-contrast technique will be used after bowel cleansing; however, in a more acute setting, particularly if there is suspicion of a colonic perforation, administration of barium is contraindicated (risk of barium peritonitis), and instead, a water-soluble contrast material (eg, Gastrografin [diatrizoate meglumine]) should be used in a single-column technique.

The typical aspect of a colon cancer is a fixed filling defect with destruction of the mucosal pattern in an annular configuration (“apple core”), as opposed to an intact mucosal pattern in a filling defect from an extramucosal compression or from chronic diverticulitis. Although preoperative histologic confirmation of a colon cancer is preferable, an unequivocal and characteristic morphology on a barium enema or endoscopy is sufficient evidence to proceed to surgery. Contrast studies have the advantages of a better passage through even severely obstructing lesions and that they commonly reach the cecum. In addition, they are superior in visualizing diverticula or a suspected fistula between the colorectum and other pelvic organs. The major disadvantage of contrast studies is the inability to take biopsies and to detect small lesions.95

Evolving Techniques.

CT colonography (“virtual colonoscopy”)203,204 and the microcapsule study have evolved in the last decade as high-tech alternatives to the two previously described methods. It should be noted that CT colonography still requires patients to undergo a bowel preparation, and that air insufflation is necessary. While there is certainly a lot of promise for both new approaches, which likely will continue to improve over time, the definite role of these techniques awaits further clarification.

Early studies suggested that CT colonography had a considerable rate of false-negative and false-positive results.204,205 In a recent study of 937 patients with risk factors for colorectal cancer, CT colonography had a sensitivity of 85% for lesions 6 mm or larger.92 So far, the technology has not been approved by Medicare for screening purposes, but this may change in the future with additional validation studies. Unfortunately, incidental extracolonic findings may precipitate a large number of unwarranted tests, which add tremendous cost to the health care system. Currently, CT colonography may serve a useful purpose in patients for whom a colonoscopy is undesirable or unsuccessful.

Evaluation of the Local Tumor Extent and of Metastatic Dissemination

Traditionally, the preoperative staging for colon cancer did not mandate further imaging studies because in the majority of cases they do not change the local surgical approach. Increasingly, however, preoperative cross-sectional imaging (CT or MRI) has become the standard of care.206,207 The justification for this shift is twofold. First, patients with a significant burden of liver disease (>50% liver replacement) may carry a prohibitive risk for general anesthesia and should be treated with chemotherapy either in advance of surgery or instead of it. CT scans are the most commonly used cross-sectional imaging technique in the United States and have a 90% and 95% sensitivity and specificity in detecting liver lesions larger than 1 cm.208 Second, the surgeon can be alerted to evidence of advanced locoregional disease that may alter the operative plan and necessitate the involvement of other operative expertise (eg, hepatobiliary surgery, urology, gynecology, etc).

In order to rule out extrahepatic, in particular pulmonary metastases, a chest x-ray in two planes commonly is sufficient, although the yield of this test is relatively low. A CT scan of the chest may be necessary to substantiate a concern from conventional images and is only a minimal incremental burden for a patient who is already undergoing such a study of the abdomen and pelvis.

Positron emission tomography (PET) has an evolving role in the evaluation of metastatic disease. While the routine use of PET scanning in the primary management of colorectal cancer is not recommended at this time, this technology does appear to have greater sensitivity for metastatic disease.209 The extent to which this greater sensitivity can be translated into an algorithmic approach to staging remains to be seen. Its greatest utility at the current time is (1) in patients where systemic disease is suspected (eg, high tumor markers) but not proven, and (2) under special circumstances where the presence of previously unknown tumor manifestations (eg, recurrence vs scar tissue, solitary vs multiple liver metastases, and presence of extrahepatic metastases) would have an impact on the treatment approach (eg, operative vs nonoperative).

Laboratory and Preoperative Tests

Preoperative laboratory tests are aimed at providing evidence for pathophysiologic effects of the tumor and ruling out general health problems that could have an effect on the patient's general operability. A comprehensive workup includes a complete blood count, electrolytes, creatinine/blood urea nitrogen (BUN), glucose, liver function tests (alkaline phosphatase, aspartate aminotransferase [AST], alanine aminotransferase [ALT], bilirubin, total protein, albumin), and coagulation parameters (prothrombin time [PT], partial thromboplastin time [PTT], international normalized ratio [INR]). Arterial blood gas analysis and additional tests will be ordered in an emergency setting or according to the individual patient's risk assessment (eg, cardiac enzymes, etc).

Even though tumor markers such as carcinoembryonic antigen (CEA) are determined routinely, their role is limited because of the low sensitivity and specificity for colonic carcinoma and because the measured value virtually never changes the management. CEA can also be elevated in proximal gastrointestinal cancers, benign inflammatory conditions of the bowel, lung and breast cancer, and smoking. Nonetheless, CEA level determination may prove helpful in some settings, for example when the return of an elevated preoperative CEA level to normal indicates a complete tumor resection or when a postoperatively elevated level may indicate residual or recurrent disease.210

Preoperative standard evaluation includes a chest x-ray in two planes for cardiopulmonary assessment and for detection of pulmonary metastases (see previous sections). Electrocardiogram (ECG) and pulmonary function tests (forced vital capacity [FVC], forced expiratory volume in 1 second [FEV1], residual volume [RV], and diffusion capacity) are indicated in patients either older than 40 years or with a respective personal history. Specialized tests such as cardiac stress tests, echocardiogram, perfusion scintigraphy, or interventional cardiologic studies depend on the individual patient's history and risk assessment.

Principles of Surgical Management

As a basic principle, any colorectal cancer is an indication for surgery unless widespread tumor dissemination or general contraindications from the patient's overall health status are present. Furthermore, any precursor pathology with statistical risk for cancer (eg, large sessile polyp in an otherwise healthy individual or dysplasia in a patient with ulcerative colitis) that cannot be managed nonoperatively is an indication for surgery.

The general goal for surgical management is either to achieve cure from the tumor and extension of survival or at least disease-free survival or, in the case of a precursor pathology with or without an underlying disease (eg, ulcerative colitis or FAP), to prevent the cancer and ideally to remove the risk-bearing disease. In a palliative setting, the goal is to prolong the period of symptom-free survival.

Local tumor control generally is the primary treatment objective to prevent local tumor complications, that is, obstruction, perforation, fistula formation, bleeding, and pain. Even in the presence of distant metastases in the liver or lung, resection of the primary tumor remains a reasonable priority. Because solitary or a limited number of metastases in the liver or lung often may be treated surgically by partial organ resection or metastasectomy with a cure rate of up to 35%, their presence should not necessarily alter the surgical approach to do a curative resection at the primary site. However, if there are extensive metastases or peritoneal carcinomatosis and cancer cure is not a reasonable goal, alleviation of symptoms and prevention of impending local complications, for example by restoring the intestinal continuity, is the best palliation.

The specific surgical and oncologic strategy planning is based on a number of factors. It has to take into account the exact localization of the tumor, the tumor stage, the presence of synchronous colonic lesions or an underlying colonic disease, the risk for metachronous lesions, the patient's age and general condition, the extent of the local procedure, and the timing. Only after the extent of the operation has been defined can the method and approach to be used be discussed as to whether the procedure is only suitable for an open laparotomy approach or laparoscopy may be reasonable and beneficial.

In contrast to rectal cancer, neoadjuvant treatment (ie, preoperative chemoradiation) is not indicated in the overwhelming majority of colonic cases. In patients with resectable metastases, preoperative chemotherapy followed by a combined colon and liver resection may be an attractive alternative to a staged resection and may help in assessing the tumor response to a particular chemotherapy regimen. Only rarely is a locally very advanced lesion treated with chemotherapy in anticipation of an otherwise unresectable mass. Adjuvant (ie, postoperative) treatment is discussed in a later section.

Preparation for Surgery

When a patient is considered an operative candidate, several preparatory steps need to be addressed.

Transfusion.

Most colonic operations can be performed without a blood transfusion. Blood-spearing surgical techniques have reduced the need while the threshold to transfuse has substantially increased. The indication will depend on the starting hemoglobin, the patient's age and physiologic status, a history of ischemic events (coronary, stroke, etc), and the extent of expected and real intraoperative blood loss. As a routine, it is recommended to have the patient's blood typed and screened, but to reserve crossmatching units of blood for these higher-risk situations.

While the risk of blood-borne infections is very low, there is some controversy as to the immunologic effect of blood transfusions on the overall prognosis of colorectal cancer. Because the initial report that transfusion may be associated with an increased likelihood of recurrence,211 many subsequent reports have reached conflicting conclusions. Meta-analysis studies have strongly questioned whether there is a true causal effect present.212 Other factors such as extent of resection required, tumor location, and experience of the surgeon actually may be the more relevant cause for recurrence, but transfusion may be an indirect reflection of extensive disease and surgery. Furthermore, a randomized trial comparing the use of autologous versus allogenic blood in patients undergoing colorectal resections did not show any statistical difference in prognosis.213

Bowel Cleansing.

Traditionally, bowel cleansing was considered an essential preparation to any elective colon surgery. The rational is based on the colon being a large reservoir for numerous anaerobic and aerobic bacteria. However, recent prospective, randomized, controlled studies and meta-analyses comparing mechanical preparation versus no preparation for elective colorectal surgery have failed to demonstrate any appreciable decrease in infection rates, anastomotic leaks, or mortality rates in patients undergoing mechanical bowel preparation.214–220 Contrasting with the evidence, however, the majority of colorectal surgeons still perform bowel cleansing in their patients. The indisputable advantages of a bowel preparation remain (1) the intraoperative ability to perform a colonoscopy if that were needed, and (2) the absence of a preanastomotic stool load if a primary anastomosis or the tissue quality were unexpectedly less than optimal and asked for a fecal diversion.

There are a wide variety of laxatives, washouts, and enemas available on the market for mechanical cleansing, but the products used generally are based on either polyethylene glycol (eg, GoLYTELY) or sodium phosphate (Fleet Phospho Soda), the latter of which is contraindicated in patients with renal failure and has come under more broad scrutiny in the United States. In the absence of a consensus regarding the best regimens (ie, orthograde cleansing alone or combined with retrograde enemas), the choice often is a matter of personal preference. Depending on an individual patient's constitution and the degree of obstruction, the bowel cleansing should be started 1 or even 2 days before surgery. The cathartic may result in significant fluid and electrolyte imbalances. Elderly patients, who are more prone to this adverse effect, therefore should preemptively be given intravenous fluids and electrolytes.

Antibiotic Prophylaxis.

Perioperative administration of prophylactic antibiotics aims at reducing colonic and dermal bacterial concentrations and is considered a crucial component of colorectal procedures. The benchmark is the rate of surgical site infections in relation to the level of wound contamination. Prophylaxis has to be distinguished from therapeutic antibiotic treatment in patients who already have an established infection. Prophylaxis (ie, in patients who do not primarily suffer from an infection) should be targeted, adequately dosed, and short (ie, start within 1 hour of the incision and be limited to less than 24 hours) in order to minimize antibiotic side effects and propagation of resistances. Coverage should include both aerobic bacteria (eg, Staphylococcus, Escherichia coli, Klebsiella, Proteus, etc) and anaerobic bacteria (eg, Bacteroides fragilis, Clostridium).

Intravenous administration of broad-spectrum antibiotics is the most common form of prophylaxis and includes several acceptable antibiotic selections: (1) single antibiotics (ertapenem, piperacillin-tazobactam); (2) combination of two antibiotics (second- or third-generation cephalosporin + metronidazole, fluoroquinolone + metronidazole, clindamycin + aminoglycoside, clindamycin + quinolone, clindamycin + aztreonam); or (3) triple combinations such as amoxicillin-clavulanic acid + metronidazole + aminoglycoside. Oral antibiotics (eg, metronidazole combined with nonabsorbable neomycin) in conjunction with a mechanical bowel preparation may yield similar results but may increase the risk of nosocomial superinfections, in particular with Clostridium difficile.

Special considerations according to national guidelines have to be followed for prophylaxis in patients at risk for endocarditis (eg, patients with mechanical heart valve).

Thromboembolic Prophylaxis.

Thromboembolic prophylaxis is recommended in all patients undergoing major surgical procedures to reduce the incidence of postoperative deep venous thrombosis and pulmonary embolism. Both pharmacologic and physical prophylaxis (eg, pneumatic calf compression) have been proven to be effective,221 but the use of pharmacologic prophylaxis has recently been endorsed by a task force recommendation.222 Both low-dose unfractionated heparin and low-molecular-weight heparins (LMWHs) have been shown to be equally effective in reducing the incidence of postoperative thromboembolic events without resulting in significant complications.223 A recent randomized study, however, showed that LMWHs have a slightly higher rate of minor bleeding events.224 Based on economic analysis, the data favor the use of subcutaneous heparin as being more cost-effective than LMWHs.225 It is recommended that these drugs be commenced at least 2 hours before surgery and continued postoperatively until the patient has obtained full ambulation. Intermittent pneumatic calf-compression boots are an alternative to heparin that has been demonstrated to be equally successful in preventing deep venous thrombosis and possessing the advantage of no risk of increased bleeding.226 It remains to be determined whether a combination of chemical agents and pneumatic calf-compression boots for patients undergoing colonic resection will be an advantage.

Anticoagulated patients who need to take warfarin (eg, owing to a mechanical heart valve) should be switched perioperatively to intravenous heparin to allow for stopping the warfarin medication and antagonizing its effect with vitamin K. Four hours before incision, the heparin may be discontinued and resumed within 24 hours postoperatively with a stepwise increase in the dose.

Urinary Catheters/Stents.

After induction of general anesthesia, bladder catheterization should be performed in all major cases to adequately monitor the urine output peri- and postoperatively. In selected patients with a previous history of colorectal or pelvic dissections, placement of ureteral stents allows better intraoperative identification and protection of these crucial structures. Laparoscopic colon procedures do not routinely need ureteral stents; however, selective use of lighted ureteral stents during challenging laparoscopic procedures may facilitate identification of these structures.

Nasogastric Tube.

Placement of a nasogastric tube is not necessary on a routine basis for patients undergoing resection of the colon or rectum and should be avoided unless they present with a complete or partial bowel obstruction.227

Preoperative Marking of Ostomy Site.

In patients who may need permanent or temporary placement of an ostomy during the surgical procedure, preoperative marking of the ideal stoma site by a stoma nurse helps to facilitate postoperative ostomy handling by the patient.

Preemptive Pain Management.

Effective pain management is an important factor not just for patient comfort but to reduce the incidence of postoperative pulmonary complications. Preoperative placement of epidural analgesia is a very valuable strategy, which, in addition to its pain-relieving effect, promotes the earlier resumption of postoperative bowel function as a result of its suppression of sympathetic nerves. The relevant segments that need to be blocked for an abdominal incision are located at a thoracic level (T6–T12).

Surgery

General Technical Principles

The objective of surgery for colonic cancer is to perform a curative resection by removing the cancerous segment of colon, the mesentery with the primary feeding vessel and the lymphatics, and any organ with direct tumor involvement. Because the lymphatics run with the arterial supply of the colon, the primary artery supplying the segment of the colon to be resected is divided at its origin. Ligation at the origin of the vessel ensures inclusion of apical nodes, which may convey prognostic significance for the patient.228 While careful dissection in the right place is the mainstay of a successful surgery, the historical Turnbull no-touch technique with early vascular ligation and occlusion of the bowel with tapes to prevent embolization of tumor and improve survival has not shown any advantage.229

The length of bowel and mesentery resected is dictated by tumor location and distribution of the primary artery (Table 36-14), but a radical resection of a colonic tumor should achieve at least a 5-cm clearance at the proximal and distal margin. Extended resections for confined tumors outside of high-risk patients have not been shown to confer additional survival benefit230; however, tumors located in “border zones” should be resected with both neighboring lymphatics to encompass possible bidirectional spread. If a tumor is adherent to or invading an adjacent organ such as the kidney or small bowel, an en bloc resection should be performed where technically feasible. Because adhesions between the tumor and adjacent organ may not necessarily be inflammatory, but, because of carcinoma, mere division or “pinching” of a tumor from an adjacent organ is not an acceptable surgical technique because it may reduce the chance of cure.

When synchronous cancers are present in the colon, an extended resection or even total colectomy, with ideally only one anastomosis, should be performed. Occasionally, two separate resections (eg, right hemicolectomy and low anterior resection) with two anastomoses are preferable to preserve colon length and to avoid postcolectomy diarrhea. Cancer on the basis of an underlying pancolonic disease (eg, ulcerative colitis or FAP) requires a total proctocolectomy with either an ileoanal pull-through procedure or an ileostomy80; young patients (<50 years, with/without proven HNPCC gene constellation) presenting with tumors proximal to the sigmoid colon should be offered a total abdominal colectomy to reduce the risk of metachronous cancers and to facilitate surveillance.231

A limited wedge resection may be considered for an unfit patient or for palliative resection in those with widespread tumor. This will relieve the patient's symptoms and prevent future obstruction and bleeding from the primary tumor.

Intraoperative Surgical Technique

Positioning.

For all left-sided colonic resections, it is advisable to place the patient in a modified lithotomy position, which gives access to the anus (eg, for a stapled anastomosis) and allows an assistant or the surgeon to stand between the legs for retraction or an excellent view to mobilize the splenic flexure, respectively. The same positioning obviously also can be used for all other colon resections, but a supine position usually is sufficient and faster. Laparoscopic procedures typically require the operating table to be tilted and moved to steep Trendelenburg's position; appropriate fixation and securing of the patient is therefore mandatory.

Incision.

For an open procedure, the peritoneal cavity is most commonly entered through a midline laparotomy incision. For a proctocolectomy, we usually recommend the use of an infraumbilical incision in order to provide good exposure for the pelvic dissection. For a more proximal segmental colon resection, however, an equally short but higher midline incision may be more convenient. In addition, a transverse incision or even a subcostal incision may give excellent exposure for a right hemicolectomy.

For a laparoscopic procedure, a first camera trocar is placed in either Veress needle or in open Hasson technique. The site should be chosen such that additional working ports can be placed along a circle with the target in the center.

Exploration.

After the peritoneal cavity is entered (open or laparoscopically), the abdomen is explored systematically to determine the resectability of the tumor. Special attention is addressed to the presence of distant metastases in the liver, peritoneal carcinomatosis, or additional synchronous lesions throughout the large intestine. Other accessible organ systems are assessed equally, for example the gallbladder and the female reproductive organs.

Colon Resection.

The surgical technique has been standardized for three segments: right colon, left colon, and rectosigmoid. Depending on the extent of the resection eventually needed in an individual patient, the technique for those segments may be combined (see Table 36-14). With a detailed description of the maximal resection, that is, an open total colectomy/proctocolectomy, all information about the individual steps necessary to perform any colorectal resection of lesser extent will therefore be provided.

The same steps should be achieved with laparoscopic resections; however, depending on the surgeon's preference and skills, a medial-to-lateral mobilization of the colon (ie, starting at the feeding vascular stalks before moving to the retroperitoneal attachments) supports the autoretraction of the colon throughout the critical steps.

On careful exploration of the abdomen, mobilization of the colon starts on the right side. Use of a mobile (eg, Richardson retractor) instead of a fixed (eg, Balfour or Bookwalter retractor) abdominal wall retractor in this first phase will allow a more flexible and unidirectional exposure according to rapidly changing needs. The small bowel is eviscerated from the abdomen and moved to the left. The abdominal wall is retracted to the right side while exerting countertraction on the cecum and ascending colon. A small incision is made at the exposed white line of Toldt to enter the retroperitoneum. Elevating the ascending colon from the retroperitoneal structures, the peritoneum is divided along the lateral gutter from the terminal ileum to the hepatic flexure. On the right side, the ureter is at fairly low risk and routinely falls away; however, special care is needed to avoid damage to the third part of the duodenum. The mobilization is facilitated by firm traction placed on the colon and the surgeon's left hand inserted into the retroperitoneum as a guide to divide along the peritoneal reflection. Because of the limited view around the hepatic flexure and the presence of small vessels at this level, transsection of the peritoneum with cautery is often advisable.

As the right edge of the gastrocolic ligament is reached, it may be easier to complete the dissection of the hepatic flexure in retrograde direction. The abdominal wall retractor is moved quickly into the upper end of the incision in order to pull in a cephalad direction. The lesser sac is entered far to the left in an avascular portion of the omentum, and the greater omentum is divided inferior to the gastroepiploic vessels between clamps and ligatures. While the omentum may be preserved in benign diseases, its resection with the respective colon segment is part of an oncologic resection. Dissection of the gastrocolic ligament is carried out from the left to the right. Connective tissue attachments between the antrum, duodenum, and transverse mesocolon and the hepatic flexure are divided stepwise by a combination of blunt digital tunneling and sharp dissection using both hands. Care should be taken at this point to avoid dissecting too deeply into the retroperitoneum, where large blood vessels can be encountered. Once the mobilization has been completed around the hepatic flexure, the right colon and transverse colon are attached only to their vascular supply and are ready for resection. This would be used for any standard right hemicolectomy or the first part of an extended transverse colectomy. For total colectomy, mobilization of the whole colon commonly is continued before dividing the major vessels.

At this point, the abdominal wall retractor is moved to the left side of the abdomen, and traction is placed to expose the left portion of the colon. The dissection is initiated at the level of the sigmoid, where the white line of Toldt again is incised and the retroperitoneum entered. Once the areolar tissues are identified, a small sponge is taken, and with firm pressure against the sigmoid mesentery, the retroperitoneal tissues are bluntly reflected, and the left ureter is exposed. Only after the ureter has been clearly identified and moved out of the way is incision of the peritoneum continued into the pelvis for a short distance and up to the splenic flexure along the left gutter. The colon is reflected bluntly from the retroperitoneal tissues, and with firm traction the peritoneal incision is continued. Gentle traction on the transverse and descending colon will help to lower the splenic flexure until it can be visualized fully. A hand placed in the retroperitoneum will help to mobilize the splenic flexure, and under direct vision the peritoneum over the splenic flexure can be incised. Care must be taken at this point to protect the spleen from direct or traction injury. The final attachments of the splenocolic ligament that hold the splenic flexure are clamped and divided in appropriate tissue portions. Clamping and ligating this tissue are recommended because even small vessels retracting into the left upper quadrant can be a nuisance.

After completion of the first two parts, the colon is mobilized completely from its retroperitoneal attachments from the terminal ileum to the upper rectum. Elevation of the colon allows identification of all primary feeding vessels. In order to ligate the inferior mesenteric vessels, the surgeon is on the patient's left and the colon is reflected to the left. The attachments that run over the sacral promontory and up along the left gutter are incised, and a hand is used to dissect the tissues bluntly from behind the inferior mesenteric vessels. By identifying the inferior mesenteric vessels and making the window just under those, the hypogastric nerves going down into the pelvis are protected routinely. Sometimes, for example if there is concern about cancer in the rectum or if the patient is very obese, these structures need to be freed up more to elevate the nerves initially and later to dissect them out under direct vision. The avascular window around the origin of feeding vessels then is opened. In the case of the inferior mesenteric vessels, the left hand is placed behind the inferior mesenteric stalk, and the thumb and opposing index finger can clear a window of avascular tissue above it. Dissection of redundant adipose tissue around the vessels is carried out under direct vision, before the vessels are clamped. Before transsection and ligature of the vessels, the remote location of the ureter is confirmed once more. If the ureter is not identified properly before dividing the vascular pedicle, accidental dissection of the ureter can occur and requires a repair. If unrecognized intraoperatively, the ureter injury may result in a urinoma. In difficult cases (eg, repeat operation or recurrence), it is therefore advisable to place preoperative ureteral stents to allow better identification. The whole vascular stalk may be ligated with a double ligature or a suture ligature. Individual ligature of the artery and vein is optional and has not been shown to provide an advantage. For the reason mentioned earlier, it is recommended to ligate the vessels as proximally as possible, but from an oncologic standpoint a high ligation of the IMA does not provide any advantage in comparison with a low ligature distal to the origin of the left colic artery.232,233

The vascular dissection is then continued around the colonic mesentery. The avascular tissue can be divided sharply while clamping is applied to vessels when they are encountered. The vascular anatomy of the colon is quite variable. However, if one is truly in the retroperitoneum and ligating named vessels at their origin, the colon can be taken out with as few as three to four clamps. In particular, the inferior mesenteric, middle colic, and ileocolic vessels need to be ligated. The presence of additional right and left colic vessels sometimes requires the use of five or six clamps. By taking the vessels closer to their origin, that is, before they branch off into multiple subsegments, fewer clamps are necessary and the dissection proceeds more rapidly.

Once the vessels have been ligated, the bowel may be divided by means of cutting linear stapling devices at the previously determined levels. In patients with an underlying disease (eg, ulcerative colitis or FAP), the dissection at this point would be continued as a total mesorectal excision down into the pelvis to the pelvic floor (see respective chapters). It is strongly recommended to have the specimen assessed macroscopically to verify the pathology. Tumor in the resection margin means an inadequate cancer operation requiring a re-resection. Intraoperative frozen sections of the resection margins should be requested whenever there is any doubt about the completeness of the resection.

Reconstruction/Diversion.

After the resection has been completed, either the bowel ends can be reanastomosed or the proximal end may be brought out as an ostomy. Prerequisites for a successful anastomosis are meticulous technique, well-vascularized and healthy appearing tissues, apposition of bowel ends without any tension, and good nutritional status of the patient with an albumin level greater than 3.0 mg/dL. Constructing an anastomosis under tension and/or with poor blood supply increases the risk of an anastomotic leak that may cause an infection and sepsis. A protective diverting ostomy does not prevent the leak as such but should diminish the life-threatening complications of an anastomotic leak. While a stapled functional end-to-end anastomosis between the ileum and the colon (ie, an enterocolonic anastomosis) is reasonable, this type of anastomosis may potentially be less desirable between two colon segments (ie, a colocolonic anastomosis) because it can result in an iatrogenic giant diverticulum that may interfere with the propulsion of formed stool or impede the performance of a surveillance colonoscopy. Performing an end-to-end anastomosis, either hand-sewn or by means of a circular stapler, will avoid these problems. An ileocolonic anastomosis in most instances can be performed in an unprepared bowel, whereas a colocolonic anastomosis on the left side traditionally requires pre- or intraoperative reduction in the stool load unless a colostomy was performed. As mentioned previously, this view has come under scrutiny.

Drains.

Placement of drains is more often a matter of personal preference than of scientific objectiveness.234–236 Most bowel anastomoses, even colocolonic anastomoses, do not need to be drained. The use of drains generally may be recommended when a pelvic dissection and anastomosis have been performed and accumulation of fluid and blood in the dependent areas around the anastomosis should be avoided. Whether prospective, but underpowered, studies are sufficient evidence to effectuate a change in this practice needs to be determined.237,238

Technical Considerations

Laparotomy versus Laparoscopy.

Laparoscopic colon surgery has a clearly established place in the management of both benign and malignant colon diseases. In many specialized centers, it is even regarded the first-line approach unless patient-specific factors suggest otherwise. The path to a nearly unanimous endorsement of the technique at least for right-sided, left-sided, and sigmoid resections for colon cancer started in the early 1990s,239 moved from palliative resections to institutional case series in curative intent, and culminated in several prospective randomized trials throughout the world,240–245 the first large-scale trial being a multicenter study by the NCI.241 This study, which enrolled 872 patients with stages I–III colon cancer, confirmed that there was a moderate quality-of-life benefit for the laparoscopic approach246 but otherwise no difference in oncological outcome and survival between the laparoscopic and open-resection groups.241 Subsequently, two large-scale European prospective multicenter trials (ie, the COLOR [COlon cancer Laparoscopic or Open Resection] trial with 1248 and the CLASICC [Conventional versus Laparoscopic-Assisted Surgery In Colorectal Cancer] trial with 794 patients) have confirmed similar results.247,248 This equality of the study results offered the unique opportunity for both opponents and proponents of the laparoscopic approach to justify their personal preference for either the open or laparoscopic technique depending on their background and skills. In contrast to one early report of a high incidence of port-site recurrences, it has become clear subsequently that with appropriate surgical technique, the incidence is in the range of 0.8–1.3% and, on a stage-by-stage comparison, not higher than wound implants after open surgery.

For the laparoscopic procedure, about three to five trocars are inserted. Lacking the tactile sensation of open procedures, tattooing of the target lesion should generally be performed prior to the surgery. The colon should be mobilized to the same extent as during open surgery, but it may be advantageous to start with the vascular pedicle rather than with the retroperitoneal attachments. The technical equipment to perform an intracorporeal resection and anastomosis is available, but it is questionable whether there is any advantage to this because at some point an incision must be made anyway to retrieve the specimen. In the laparoscopically assisted technique, the segment, once it has been mobilized to the required extent, therefore is exteriorized through a small sleeve-protected abdominal incision, and an extra-abdominal resection and anastomoses are performed. The bowels are returned into the abdomen, the fascia is closed, and the pneumoperitoneum may be reinstalled to inspect the peritoneal cavity again. To facilitate complex resections, some surgeons use hand-assisted laparoscopic surgery (HALS) to combine tactile sensation with a minimally invasive approach.

Sentinel Lymph Node Mapping.

Although the interest in lymphatic mapping and sentinel lymph nodes has been derived from favorable experiences in breast cancer and melanoma, most recent data do not support the value of this technique for colon cancer. In particular, analysis of the recent intergroup study 0114 demonstrated a lack of correlation in an alarming 54% of the patients.249 Sentinel lymph node mapping not only may be misleading and therefore not useful in the management of colorectal cancer, but there also is simply no need for this technique in colon resections because the lymphadenectomy—in contrast to breast and melanoma surgery—is not associated with any morbidity.

Special Circumstances in Emergency Surgery

Approximately 20% of patients with colon cancer present as an emergency requiring an urgent operation for a tumor-related complication (eg, bowel obstruction, perforation, or massive bleeding).250 Morbidity and mortality are significantly higher than under elective conditions. Contributing factors are the lack of a mechanical bowel preparation and the patient's impaired overall status, which typically is characterized by dehydration, third spacing of fluids, anemia, a deranged metabolism with electrolyte imbalances, and possible sepsis. The risks for wound and intra-abdominal infections and anastomotic leakages are three to six times higher.251

Tumor Obstruction.

Sixteen percent of patients with colon cancer present with a bowel obstruction and complain of colicky abdominal pain, abdominal distension, vomiting, constipation, and, occasionally, paradoxical diarrhea. Imaging studies (abdominal x-ray or CT scan) characteristically demonstrate the features of a large or small bowel obstruction depending on how proximal in the colon the obstruction is located and whether the ileocecal valve is competent. Attention should be paid to the diameter of the cecum, which presents a risk of cecal perforation if the diameter reaches 12 cm or more. Urgent intervention is required in such circumstances to prevent cecal perforation. The most important differential diagnosis is pseudo-obstruction (Ogilvie's syndrome), which is seen as a result of various medical conditions and may mimic the features of bowel obstruction. Every patient therefore should have a rigid proctoscopy, followed by a water-soluble contrast enema, which should visualize only the colon up to the site of obstruction but not beyond the stenosis because the hyperosmolar nature of the contrast material can result in an increase in the intraluminal volume and trigger a perforation.

If the level of obstruction in the colon is proximal enough, a resection with primary enterocolonic anastomosis, for example right hemicolectomy, extended right hemicolectomy, or subtotal colectomy, may be carried out. If the tumor is located on the left side of the colon, adjustments to the surgical approach are necessary because the stool load proximal to the obstruction is of concern for a colocolonic anastomosis and because that segment of the colon could not be cleared before the operation. Synchronous lesions, which in the setting of an obstructing lesion may occur in up to 15%, may be missed and necessitate further intervention in the future. Strategies then include either (1) a subtotal colectomy, (2) an on-table lavage with segmental colon resection, intraoperative colonoscopy, and primary anastomosis, or (3) performance of a two- or even three-stage procedure instead of the elective one-stage approach. Historically, obstructed left-sided tumors were treated with a three-stage approach starting with a defunctioning loop colostomy, followed by resection and anastomosis and last by closure of the defunctioning stoma. The Hartmann procedure, the classic example among several two-stage procedures, consists of a discontinuous rectosigmoid resection with creation of a terminal colostomy and a blind rectal stump in the first stage, followed by a colostomy takedown and reanastomosis in a second operation.

More recently, there has been a trend toward attempting to relieve the acute obstruction at the tumor-bearing segment by colonoscopic insertion of a self-expanding metallic stent. Successful decompression of the prestenotic colon converts the emergency situation into an elective setting, allowing for stabilization of the patient and performance of bowel preparation. The risk of a colonic perforation during stent placement is relatively low but acceptable because an emergency operation would be necessary anyway if the stent could not be placed successfully. Several nonrandomized, noncontrolled case series have demonstrated that colonic stenting for acute obstruction is safe and highly successful.252–255 A proximal diversion hence may be avoided with this procedure.

Tumor-Related Perforation.

Colonic perforation secondary to a tumor occurs in two different settings. Either a transmural tumor perforates itself, or the proximal colon becomes overdistended, particularly in the case of a competent ileocecal valve. Both conditions may result in diffuse fecal peritonitis with significant morbidity and mortality. In addition, the tumor perforation results in spillage of tumor cells and thus has to be considered a stage IV tumor. Surgical management is indicated in every case and requires not only addressing the site of colonic perforation but also removing the tumor in an oncologically correct fashion.250 The same tactical principles described in the preceding section apply.

Massive Colonic Bleeding.

Massive bleeding from a colonic tumor is a relatively rare complication. The general algorithms for the workup and management of lower gastrointestinal bleedings apply, but, most commonly, the bleeding site can be easily identified. If the bleeding is minor or self-limited, the standard workup can be performed. If the patient is or remains unstable and requires repeated transfusions, surgical management is indicated.

Management of Advanced Disease

Locally Advanced Disease.

It has been estimated that approximately 15% of colonic tumors will be adherent to adjacent organs.256 With locally advanced colon tumors, it is still possible to achieve cure if the surgeon is prepared to resect involved adjacent organs. Unfortunately, it is often impossible to distinguish between malignant and inflammatory adhesions, but at least 40% of these adhesions are expected to harbor malignant cells. The surgeon therefore has to consider them malignant until proven otherwise and perform an en bloc resection to achieve a tumor-free margin.257

Operable Metastases.

At the time of presentation, 20% of patients with colorectal cancer have stage IV disease. Distant metastasis, particularly liver and lung, is a major cause of death in patients with colorectal carcinoma. However, patients with asymptomatic liver metastases may have a statistically natural life expectancy of several months up to almost 2 years without any treatment. Chemotherapy and surgical metastasectomy in selected patients may improve disease-free and overall survival substantially, resulting in a cure rate of 30%.258 In the case of potentially resectable metastases, resection of the colonic primary tumor therefore should be performed in an oncologic fashion.

Inoperable Disseminated Disease.

In patients with unresectable metastatic disease, the surgical treatment goal is to provide palliation and to prevent predictable complications. In contrast to the oncologically defined standard resections, a limited segmental wedge resection of the colon is acceptable in this setting. In particular, tumors located in the sigmoid colon or in the cecum and ascending colon are suitable for a laparoscopic or laparoscopically assisted resection because these segments can be mobilized easily to a sufficient extent to ensure a safe anastomosis. If a tumor in a patient with metastatic disease is too advanced locally to be resected safely (eg, infiltration of other organs), palliation may be achieved by creating an internal bypass or a proximal diversion.

Postoperative Management

Postoperative fast-track management after a colorectal resection has become very straightforward and routine. The immediate postoperative monitoring of vital signs, fluids, and electrolytes, as well as adequate pain control, is not different from any other major surgery. However, there has been an increased emphasis on epidural pain management, early mobilization and regular spirometry exercises, avoidance of tubes and drains (eg, nasogastric tubes), and early resumption of oral intake no later than on the first or second postoperative day with advancement to a regular diet as tolerated. Daily assessment of the abdomen and bowel activity is crucial, including careful auscultation and palpation of the abdomen to assess bowel sounds or peritoneal signs. Unless soaking, a wound dressing may be left in place until the second postoperative day or even for 5–10 days if an occlusive transparent dressing is used. The incision has to be checked daily for the presence of induration, hematoma, redness, dehiscence, or discharge of fluids (eg, pus, hematoma, or serosanguineous fluid). Large amounts of serous fluids draining from the wound should not be mistaken for a seroma but indicate a fascial dehiscence until proven otherwise. The average length of stay after colorectal resections depends on the patient's constitution but generally is in the range between 5 and 7 days for an open standard procedure, and 2 and 5 days for a laparoscopic approach. Before discharge, further tumor treatment should have been addressed with the patient. Adjuvant chemotherapy (and rarely radiation therapy) typically are not initiated before 3–4 weeks after surgery and may be delayed if infectious complications or anastomotic leaks occur.

Complications of Surgery

The overall perioperative mortality within 30 days of colorectal resections is between 3.5 and 6%,259 with less than 2% after elective but up to 20% after emergency operations. Complications of surgery may be of a general or surgery-specific nature and can be classified based on the time of their occurrence as either early (within the first 30 days) or late (after 30 days). Intraoperative complications like injury to relevant anatomic structures such as ureters, spleen, bowel, and duodenum are related to the surgical technique, to blurred anatomic landmarks and layers owing to the disease (eg, peritonitis or massive adhesions), or to the patient's habitus (eg, obesity). Early surgery-specific complications include bleeding, most frequently within the first few days of the resection, nonspecific infections, or infections related to an anastomotic dehiscence. Other more general complications in the early postoperative period (postoperative days 1–3) commonly are related to the cardiopulmonary system and include pulmonary problems (eg, atelectasis, pneumonia, aspiration, and pulmonary embolism) and cardiac events (eg, arrhythmia, myocardial ischemia, and dysfunction). Insufficient pain control has been recognized as an important factor promoting these conditions because it results in a poor respiratory effort by the patient and the inability to cough up sputum, leading to superficial respiration and suboptimal saturation. High fever in the 3 days therefore may be related to the development of an atelectasis rather than to an early infection.

Infectious complications usually occur after the third postoperative day and may be located either intra-abdominally, in the wound, in the urinary tract, or in the lungs. The primary workup therefore includes bacteriologic cultures and stains, blood and urine analysis, and a chest x-ray.

Abdominal complications consist of delayed return of upper and lower gastrointestinal function (also referred to as postoperative ileus), fascial dehiscence, and anastomotic breakdown. Clinical leaks occur in 1–2% of all colonic resections, but subclinical leaks are more frequent and may be seen incidentally on contrast studies in otherwise asymptomatic patients. A leak may present with insidious symptoms such as fever, tachycardia, abdominal distension, ileus, feces draining through a drain or the wound, or local and generalized peritonitis. Occasionally, a leak may present with sudden deterioration, generalized peritonitis, and septic shock as the result of a significant and rapid contamination of the peritoneal cavity. Owing to the heterogeneous symptoms, a leak should be suspected in any patient who is not progressing to the expected degree. Blood parameters such as white blood cell counts and C-reactive protein may be elevated but are nonspecific and difficult to distinguish from a normal postoperative reaction. After an abdominal operation, normal free air should be resorbed within 7–10 days.260 The presence of substantial free subdiaphragmatic air later in the course should therefore raise the index of suspicion for an anastomotic leak.

Imaging studies to define the presence of an anastomotic leak include a water-soluble contrast enema to visualize extravasation of the contrast material and/or a CT scan with oral, intravenous, and possibly rectal contrast material. Apart from antibiotic treatment, the management of an anastomotic leak depends on its presumed extent and the clinical presentation. A patient with generalized peritonitis requires a relaparotomy after appropriate resuscitation. Depending on its location, the anastomosis either should be taken down and the ends should be exteriorized or, in more favorable conditions, resected, and a new anastomosis performed with healthy-looking bowel ends, either with or without proximal diversion. A local repair alone carries a high risk of failure but may succeed in combination with drain placement and a proximal diverting ostomy. By the time of the reexploration, the prolonged peritonitis in some cases already may have transformed the bowel loops into rigid pipes that would not allow any mobilization for an ostomy or for a new anastomosis. In such a case, creation of a confined leak by means of a catheter enterostomy may be a desperate attempt for local control. A fecal fistula can be managed in a conservative manner if there is no evidence of generalized peritonitis or uncontrolled sepsis. Under favorable conditions, including good nutritional support and absence of a distal obstruction or disease of the involved bowel segment, the fistula may close spontaneously. The surrounding skin will need special care, and a stoma therapist will be helpful in this regard.

Adjuvant Chemotherapy and Radiotherapy

The rationale for adjuvant chemotherapy is based on the fact that we are clearly not as successful with surgical treatment as we would like to be. 5-Fluorouracil (5-FU) was the first and most extensively evaluated drug for the treatment of colorectal cancer. Multiple studies had been completed without proof of value until Krook's study.261 Subsequently, a review of 29 randomized trails concluded that adjuvant chemotherapy for colon cancer resulted in a 5% improvement in survival.262 When studies using 5-FU–based regimens are analyzed, there is a 2.3–5.7% absolute improvement in 5-year overall survival. However, when just those at high risk of recurrence are treated, the improvement in survival in this group is closer to 30%. Patients with stage III colon cancer are recognized to be at high risk for recurrence, and administration of 5-FU/leucovorin (LV) for 6 months after surgery has proven to decrease recurrence and improve long-term survival.263 The combination treatment of 5-FU/LV for 6 months was proven to be equivalent in efficacy to 12 months, and the addition of levamisole to 5-FU/LV did not seem to add any benefit.264 Low-dose LV also was demonstrated to be equally efficacious as high-dose LV when used in combination with 5-FU. Thus the first-line standard of treatment from 1998 to 2000 was a combination of 5-FU and low-dose LV (folinic acid) given for 6 months on either a weekly schedule or 5 consecutive days every 4 weeks. At present, there is not enough evidence to recommend the routine use of adjuvant chemotherapy in stage II disease. Lenz and colleagues have demonstrated that molecular or genetic markers may better identify subgroups of patients who are likely to benefit from adjuvant chemotherapy.265,266,267

Several new agents, for example irinotecan (CPT-11)268,269 and oxaliplatin,270–272 have demonstrated significantly superior activity in combination with 5-FU/LV in the metastatic setting. Irinotecan/5-FU/LV (IFL)268 and oxaliplatin/5-FU/LV (FOLFOX) have been entered into randomized clinical trials against 5-FU/LV in resected stage III colon cancer.273 Both these studies prove that the new agents in association with 5-FU/LV were superior to 5-FU/LV alone. Because of these successes, IFL was approved as first-line chemotherapy in 2000. In 2005, 5-FU/LV with oxaliplatin (FOLFOX) was approved for adjuvant therapy and has evolved in most centers as the treatment of choice. The FOLFOX regime has been compared in a large randomized, controlled trial with IFL and irinotecan/oxaliplatin (IROX) in patients with previously untreated metastatic colorectal cancer.273 This study showed significantly superior results of the FOLFOX regime for all end points. The median time to progression observed for FOLFOX was 8.7 months, response rate was 45%, and the median survival time was 19.5 months. The FOLFOX regimen had significantly lower rates of severe nausea, vomiting, diarrhea, febrile neutropenia, and dehydration. Sensory neuropathy and neutropenia were common with the regimens containing oxaliplatin.

Capecitabine (Xeloda), an oral agent designed to generate 5-FU preferentially in tumor tissue, is an exciting new development with improved convenience. A randomized phase III study comparing oral capecitabine versus intravenous 5-FU/LV concluded that capecitabine demonstrated a statistically significantly greater response rate compared with 5-FU/LV (26 vs 17%; p < .002) and an equivalent time to progression and overall survival.274 This study demonstrated that capecitabine is a suitable alternative to IV 5-FU and perhaps a replacement in the future. There are currently phase II trails being conducted on capecitabine/oxaliplatin (CAPEOX) and capecitabine/irinotecan (CAPEIRI).275–279

Two of the most fascinating targets in the treatment of colorectal cancer are the epithelial growth factor receptor (EGFR) and vascular endothelial growth factor (VEGF) blockers.280,281 Agents that inhibit the EGFR or bind to VEGF have demonstrated clinical activity as single agents and in combination with chemotherapy in phases II and III clinical trials. The most promising of these agents are the monoclonal antibodies cetuximab, which blocks the binding of epithelial growth factor, and bevacizumab, which binds free VEGF.280,281 However, the benefit of cetuximab is limited to patients with a tumor bearing wild-type K-ras while tumors bearing mutated K-ras do not show any response.282,283 Both agents have proven benefit and seem to work best as first-line therapy for metastatic colorectal cancer. Introduction into the primary adjuvant treatment after curative resection of stages II and III tumors will remain a subject of future trials. We await future developments of these and other newer drugs and their impact in the fight against colorectal cancer.

Generally, radiotherapy does not play a primary role in the adjuvant treatment of colon cancer. However, it may be considered as a locoregional field radiation in selected locally advanced T4N0–N1 tumors.284–286

Outcome and Prognosis

Recent years have produced a trend toward better outcome and survival in patients diagnosed with colorectal cancer. This may be related to safer and more successful surgical treatment in combination with better nonoperative and adjuvant treatments. The perioperative mortality within 30 days of elective colorectal resections is less than 2% even though it still may be relatively high after an emergency operation, thus resulting in an overall mortality of 3.5–5.5%.259 SEER data demonstrate an overall decline in colorectal cancer mortality. While the overall 5-year survival of patients with colon cancer was at 41% between 1950 and 1952, it has since increased steadily to 63.8% between 1995 and 2000. Analyzed for each stage as defined by the AJCC sixth edition system (Table 36-15) separately, 5-year survival was 93.2% for stage I, 84.7% for stage IIa, 72.2% for stage IIb, 83.4% for stage IIIa, 64.1% for stage IIIb, 44.3% for stage IIIc, and 8.1% for stage IV.8 The prognosis of patients with synchronous primary colon tumors is not different from that of patients with solitary tumors if they are compared on the basis of the most advanced stage (see Table 36-15).202