Chapter 45. Hepatic Colorectal Metastases: Resection, Pumps, and Ablation

Colorectal cancer (CRC) is the third most commonly diagnosed cancer in the United States.1 In spite of an increasing emphasis on screening and prevention of this disease, more than 140,000 new cases will have been diagnosed in 2010, comprising approximately 10% of new cancer diagnoses. Although mortality from this disease has improved over the past two decades, more than 40% of patients with colorectal cancer eventually die of their cancer.1,2 Among those patients with advanced disease, more than half will develop liver metastases, more than any other organ, and many will have disease recognizably confined to this organ. Specifically, approximately 20–40% of patients with metastatic CRC have liver-only metastases at the time of presentation or recurrence, accounting for about 30,000 patients per year in the United States.3,4

While not supported by randomized trials, a preponderance of uncontrolled studies have demonstrated that complete resection in patients with liver metastases is associated with dramatically improved survival compared with patients not undergoing surgical therapy. Advances in imaging technology, surgical techniques, and systemic chemotherapy have brought steady improvements in long-term outcome in patients undergoing resection, with 5-year overall survival exceeding 50%. In addition, other surgically delivered locoregional strategies offer promising directions in improving outcomes in these patients, including ablation, and intra-arterial chemotherapy. Herein, an overview is provided of the important clinical issues relating to the surgical management of patients with colorectal liver metastases.

The extent of evaluation and staging of a patient with hepatic metastases should be determined based on the available treatment options. In patients for whom further treatment is not being considered, either due to comorbid conditions or patient choice, an extensive evaluation for the extent of disease may be unjustified. In patients for whom only noncurative systemic chemotherapy is being considered, an evaluation should establish a baseline to facilitate monitoring of the response to treatment at all sites. In those who are or may become candidates for local therapy directed to the liver, it is important to exclude the presence of extrahepatic disease, particularly in the majority of these patients who are asymptomatic from their liver disease.

When evaluating the patient for extrahepatic disease, computed tomography (CT) is the imaging modality used most frequently. Abdominal CT can detect other intra-abdominal disease, while chest CT is most sensitive for identifying pulmonary metastases, detecting 95% of lesions greater than 1 cm in diameter.5–7 While controversial, chest CT should be strongly considered prior to resection of liver metastases, even in patients with a normal chest x-ray.6 The ability of CT imaging to detect extrahepatic disease within the abdominal cavity or pelvis is lower, with a sensitivity reported between 22 and 41%.8,9 Similarly, magnetic resonance imaging (MRI) can be useful in evaluating evidence of extrahepatic disease.8 In addition, MRI may be useful to characterize indeterminate liver lesions. While most consider a high-quality contrast CT sufficient, some have suggested MRI may be the most sensitive for preoperative staging.10

One of the most informative imaging modalities for the assessment of metastatic sites is whole body positron emission tomography (PET). Unlike standard cross-sectional imaging, PET provides functional information related to metabolic activity. Although a number of positron-emitting radiopharmaceuticals have been developed,18F-fluorodeoxyglucose (FDG) PET is the most widely used in oncology.11 When administered intravenously, 18F-FDG is taken up and accumulated in metabolically active cells. Malignant tissue with relatively increased uptake can be seen as areas of increased signal relative to the surrounding less metabolically active normal tissue.

A wide spectrum of malignancies may be successfully staged by FDG-PET, including colorectal, lung, and breast cancer.12 PET has been reported to have a sensitivity as high as 92–100%, with a specificity of 85–100%.13 In patients with colorectal liver metastases in whom resection is planned, preoperative FDG-PET appears to be especially useful to exclude extrahepatic disease.14,15 Recent studies report identifying such additional disease which alters the patient management in up to 25% of the patients.15,16 Accordingly, FDG-PET is recommended by the National Comprehensive Cancer Network as a routine component of the preoperative evaluation in patient being considered for surgical therapy for liver metastases from CRC.6

When considering any local therapy of the liver, it is particularly important to carefully evaluate the extent of intrahepatic disease and determine the location of metastases relative to major vascular pedicles within the liver. Multiple bilobar metastases as well as involvement of hilar structures or the presence of periportal or celiac nodal disease may preclude resection. Not infrequently, patients who are considered preoperatively to be candidates for resection are found at operation to be unresectable because of previously unrecognized additional hepatic lesions. High-quality contrast-enhanced cross-sectional imaging is imperative for an adequate assessment of the liver prior to surgery. In addition, advanced postacquisition imaging techniques—such as three-dimensional reconstruction, volume-rendering, and digital subtraction angiography—may improve the ability to view the relationship of metastases to vital structures and determine resectability.

The goal of the preoperative evaluation is to identify the best candidates for resection and exclude those who will be found at operation to have unresectable disease, thereby obviating unnecessary laparotomy in the latter group. In one report from Johns Hopkins, the non-therapeutic laparotomy rate decreased from 15% in the 1990s to approximately 5% in the more recent years.17

Intraoperative exploration and evaluation of the extent of disease within the abdominal cavity and liver is critical prior to proceeding with surgical resection. The abdominal cavity, including peritoneal surfaces and pelvis should be inspected and palpated if accessible to rule out extrahepatic disease or locoregional recurrence. Special attention should be paid to porta hepatis and periportal nodal region.18,19 Nodal involvement in the periportal area is associated with a significantly poorer long-term outcome.18,20 It probably represents a manifestation of disseminated disease and should preclude curative resection in most cases.

Bimanual and bidigital palpation of the liver is then carried out. In most cases, metastases from CRC are firmer than the surrounding liver and palpable. In some patients, however, a fibrotic or fatty liver, sometimes related to the use of preoperative chemotherapy, may reduce the ability to palpate small metastases. In some cases, an indentation or dimple on the capsule of the liver can provide a clue of the presence of a small metastasis below the surface, particularly in patients following a significant response to chemotherapy.

Intraoperative ultrasonography (IOUS) is the most sensitive modality currently available for detecting otherwise occult liver metastases. Even in the era of high-quality preoperative imaging, IOUS sensitivity is reported to be superior to other imaging modalities.21,22 The sensitivity of IOUS to detect additional lesions depends on multiple factors, including the quality and timing of the preoperative imaging, associated liver disease, and tumor echogenicity.22–24 In one study, van Vledder et al demonstrated in 213 patients undergoing surgical exploration for CRC liver metastases between 1998 and 2009, IOUS alone detected additional tumors in 10% of patients. Detection rate was found to be higher in those patients with multiple metastases (>3) and those in whom the index known lesions were hypoechoic22 (Fig. 45-1). Moreover, they found patients with isoechoic index lesions were associated with significantly higher rate of early intrahepatic recurrence, a surrogate for missed lesions.

By improving the capability of detecting clinically occult metastases, patients with multiple unresectable metastases may be spared unnecessary hepatic resection. IOUS may also contribute to improved survival by helping to detect and excise or ablate otherwise occult residual disease. Moreover, IOUS facilitates the careful examination of intrahepatic vascular structures and their relationship to the hepatic tumors, often facilitating safer resections with more adequate resection margins.25–27

Laparoscopic evaluation of the liver and abdominal cavity just prior to laparotomy has been used in a variety of gastrointestinal malignancies.28–30 This approach is used to potentially identify additional patients who have unresectable disease, reducing the number of patients unnecessarily undergoing full surgical exploration. With recent refinements in laparoscopic ultrasonographic devices, intraoperative hepatic ultrasonography can now compliment visual laparoscopic assessment.29,31 Laparoscopic staging for potentially resectable hepatic colorectal metastases is more controversial and less frequently used. First, non-therapeutic laparotomy rates are lower in this disease than in most other gastrointestinal cancers, resulting in less potential benefit. Yet some continue to recommend it, particularly among those at higher risk.32,33 In one report of 103 consecutive patients evaluated with potentially resectable disease, laparoscopy identified 14 of 26 patients with unresectable disease, 10 of whom were spared an unnecessary laparotomy.33 Despite such enthusiastic initial reports, the role of laparoscopy prior to surgical exploration for patients with isolated hepatic metastases remains to be defined.

Hepatic resection is acknowledged as the most effective therapy for patients with colorectal metastases confined to the liver. A more accurate understanding of liver structure, based on functional segmental anatomy, as well as advances in operative technique and postoperative care, has resulted in the capability to perform even major hepatectomy with very low morbidity and mortality. Thus, the goal of surgery should be complete removal of all metastatic disease. Several issues should be addressed when considering hepatic resection for our patients: (1) How do we select patients for resection? (2) What surgical techniques are recommended? (3) What are the long-term outcomes and prognostic factors? (4) How can we expand the role of liver resection to those with disease that is initially considered to be unresectable? and (5) What is the role of adjuvant or neoadjuvant chemotherapy for patients undergoing hepatic resection?

Selecting Patients for Surgical Resection

Indications for surgical resection of hepatic colorectal metastases have undergone a major shift in the last decade. Previously, hepatectomy was not recommended in patients who had more than three or four metastases, hilar adenopathy, metastases within 1 cm of major vessels such as the vena cava or main hepatic veins, or extrahepatic disease. More recent studies demonstrate, however, that patients with these clinicopathologic factors can achieve long-term survival following hepatic resection and therefore should not be excluded from surgical consideration.34–37 Specifically, a small number of metastases38–40 and wide margin status are no longer considered necessary for defining resectability.41,42 Similarly, contiguous extension to adjacent anatomical structures and local or regional recurrence at the site of the primary colorectal cancer are not contraindications to resection. An increasing number of studies also indicate that although survival may be reduced in patients with extrahepatic or periportal lymph node metastases, complete resection of these sites in conjunction with resection of hepatic metastases can result in long-term survival.18,43–47

Data such as these have led to a shift in the definition of resectability from criteria based on the characteristics of the metastases (tumor number, size, etc.) to new criteria based on whether a macroscopic and microscopic complete (R0) resection of the liver disease, can be achieved. Currently, hepatic colorectal metastases should be defined as resectable when it is anticipated that (1) all disease can be completely resected, (2) two adjacent liver segments can be spared, (3) adequate vascular inflow and outflow and biliary drainage can be preserved, and (4) the volume of the liver remaining after resection is sufficient.

The health and function of the nontumor bearing liver is clearly one of the most important factors impacting resectability and outcomes following liver resection. In patients undergoing surgery for hepatic metastases, cirrhosis is rarely present. However, increasing use of prolonged preoperative chemotherapy can result in significant steatosis, steatohepatitis, and sinusoidal dilatation. In some situations, these pathologic changes can be associated with increased postoperative morbidity. Assessment of hepatic functional reserve is important when deciding whether resection should be pursued. Staging the extent of hepatic dysfunction using the functional Child-Pugh and model for end-stage liver disease (MELD) classification system may be helpful.48 Thrombocytopenia can be useful to estimate the extent of portal hypertension. Measurement of the indocyanine green retention rate can also provide an estimate of underlying liver function.49 Although this test is used in some centers, it is not employed with great frequency in most cases. Imaging techniques may shed light on the extent of cirrhosis in some cases. Specifically, CT scan or MRI can identify loss of liver volume or hepatic contour changes indicative of more extensive cirrhosis; visualization of portal vein collaterals, splenomegaly, or ascites can indicate more advanced disease.

The size of the remnant volume considered safe varies with the condition of the hepatic parenchyma. In healthy livers, a remnant liver volume greater than 20–25% of the estimated total liver volume is considered adequate.50,51 In contrast, patients with cirrhosis need a greater remnant liver volume (>40%) in order to avoid postoperative liver failure. CT or MRI can now provide an accurate, reproducible method for preoperatively measuring the volume of the future liver remnant.52 In cases where major hepatectomy is planned and there is concern regarding insufficient liver volume, these patients should be considered for ipsilateral portal vein embolization (PVE) to induce hypertrophy of the contralateral liver lobe.51 PVE has been shown to increase the size of the future liver remnant. However, no strong evidence to date has clearly demonstrated improved outcomes following PVE compared to no PVE, particularly in the non-cirrhotic patient. Likely, the selective use of PVE may enable the performance of an extended hepatectomy in a subset of patients who otherwise would not have been candidates for safe resection.

In addition to tumor- and liver-related factors, patient comorbidities also need to be considered. Patients with an increased American Society of Anesthesiology (ASA) score have been shown to have a significantly higher morbidity and mortality as compared to those patients with an ASA of 1.48,53 These patients with higher ASA scores reflect more underlying cardiac disease, renal insufficiency, and other conditions which place this patient population at greater risk for postoperative complications.

Techniques for Liver Resection

Whether for removal of colorectal metastases or other indications, the objective of a liver resection is to remove the involved portion of liver with an acceptable surgical margin and preserve sufficient hepatic reserve. Major hepatic resections can be divided into left and right hepatectomy (or hemihepatectomy) or extended hepatectomy, depending on the number of liver segments removed.54 All or part of the caudate lobe can be included as part of a major liver resection or can be resected separately. When performing major hepatic resection, the vascular structures supplying the liver being removed are typically isolated extrahepatically prior to parenchymal dissection. The relevant portal pedicle can be secured at the hilum, either as a group within the intrahepatic hilum or by individual isolation of the portal vein and hepatic artery.55,56 Selective extrahepatic ligation of inflow allows for the demarcation of the portion of the liver to be removed.

Minor resections can be either nonanatomic wedge resections or segmental resections. Wedge resections typically make no attempt at isolating vascular structures supplying the area being removed and are generally reserved for small peripheral lesions. When such resections are being performed for colorectal metastases, care must be taken to achieve an adequate resection margin as wedge resections are more often associated with positive or close margins.56 Hepatic segmentectomy can be considered an anatomic minor resection. Such resections can be of a single segment or multiple adjacent segments (bisegmentectomy, sectorectomy, or sectionectomy) in one or both hemilivers. Provided adequate tumor margins are achieved, segmental resections have the advantage of preserving hepatic parenchyma compared to major liver resections. Intrahepatic anatomic landmarks visualized with intraoperative ultrasonography can be used to plan segmental resections and vascular pedicles are typically controlled within the liver substance. Temporary total inflow occlusion (Pringle maneuver) can be helpful during these types of resections.

A variety of techniques have been described for the division of the liver parenchyma.26,57–60 The traditional method utilizes blunt parenchymal dissection either manually, or with a crushing clamp. Vascular and biliary structures are identified within the liver and ligated or clipped. Other dissection techniques utilize devices such as the ultrasonic Cavitron or a saline jet irrigator to divide the liver by disrupting the parenchyma. Newer devices use saline-enhanced radiofrequency energy to achieve coagulative necrosis of the liver substance to achieve hemostasis during or before the dissection.58,59 Other devices thermally precoagulate the resection plane to achieve complete hemostasis prior to transection.60 Surgical stapling devices have also greatly improved the ease in which liver resection can be performed.61 In addition to facilitating division of the extrahepatic vascular pedicles, staplers can be used to divide larger pedicles within the liver parenchyma. The superiority of any one single method has not been established and the choice of technique should be determined by the individual surgeon's expertise.

Results of Liver Resection for Colorectal Metastases

Perioperative mortality of liver resection for colorectal metastases has markedly decreased in recent years, approximating 1% in most recent reported series.62–68 In experienced hands, even major hepatic resections, which are performed in about half of these cases, result in perioperative mortalities of less than 3%.62,64,65 The recommendation is that major liver resection be performed at centers and by surgeons with more than occasional experience with such procedures. One study analyzing the short-term outcome for liver resections in the state of Maryland indicated a clear relationship between hospital procedure volume and perioperative mortality.65 Among 606 patients, the in-hospital mortality was 1.5% in the high-volume hospitals (defined as more than 15 resections per year) compared to a 9.6% mortality in centers performing less than eight resections per year. Although operative mortality should be uncommon, significant complications have been reported in 15–30% of patients.64,67,69,70 The morbidity associated specifically with liver resection includes hemorrhage, perihepatic abscess, bile leak and/or fistula, pleural effusion, and hepatic failure.

The long-term survival reported following hepatic resection with curative intent for metastatic colorectal cancer has improved significantly in the last decade. While older series report 5-year survival rates of 25–40%,69,71,72 more contemporary series report 5-year survival in excess of 50%.62,73–75 Recent large multi-institutional collective series demonstrate trends in improved long-term outcome63,68,76,77 (Table 45-1).

Prognostic Factors Following Resection for Colorectal Metastases

Although surgical resection results in prolonged survival and perhaps cure in some patients, the majority eventually develop recurrent disease. For this reason, many authors have attempted to identify factors that might improve patient selection, thereby improving the long-term outcome of those resected.78,79

Pathologic features of the primary tumor appear to correlate with long-term outcome following liver resection. Both nodal status and histologic grade of the primary tumor are associated with poorer outcome following liver resection in several reported series.62,63,72 Features of the hepatic metastatic disease, including the number, size, and location of the metastases, correlate with prognosis in many series.62,63,73,75 Fong et al report that the number and size of metastases, as well as the preoperative carcinoembryonic antigen (CEA) and disease-free interval each independently correlated with survival by multivariate analysis.33 Although these data suggest that the prognosis appears worse in patients with an increased number of metastases, long-term survival can be achieved, at least in selected patients, with resection of even four or more metastatic lesions. There are insufficient data to allow a definitive threshold to be established at present.78

Technical factors at the time of surgery can also impact on the prognosis. These are of particular significance as they are often under the control of the surgeon. A positive histologic surgical resection margin has been found to be associated with poor long-term survival and higher risk of local recurrence.41,62,80 The optimal width of the negative surgical margin, however, remains controversial. Some investigators have reported an improved survival when clearance margins were 1 cm or greater72 while others have shown no differences, provided the margin is grossly negative.41,81 In recent reports, some have challenged the need even for microscopically negative margins, particularly following preoperative chemotherapy.42 The type of resection performed (wedge resection, segmentectomy, or hemihepatectomy) and the technique of parenchymal dissection do not appear to affect long-term recurrence rates, independent of margin status.62,63,75

Liver Resection in the Presence of Extrahepatic Colorectal Metastases

Extrahepatic disease, even when limited and resectable is generally acknowledged to be associated with worse prognosis following liver resection of colorectal metastases, with many in the past considering this a contraindication to surgery.63,71 However, reduced surgical morbidity as well as the development of more effective systemic chemotherapy regimens have prompted many to recommend resection of all disease when possible, including extrahepatic sites.45,47 Certainly, local extension to adjacent structures and locoregional recurrence should not be considered true extrahepatic disease and are not contraindication to surgery if complete resection is possible. In contrast, offering hepatic resection in the presence of extrahepatic metastases is somewhat more controversial. The lungs are the second most frequent site of metastatic disease, accounting for approximately 20% of those with metastases from colorectal cancer, and about 5–10% of patients who present with metastatic disease will have both liver and lung metastases. In the past, multiple studies have reported favorable long-term survival rates after resection of localized lung-only disease.82,83 More recently, however, several centers have reported their results for patients undergoing combined lung and liver resection.84,85 These studies have reported 5-year survival rates in excess of 30%. Prognosis following resection of pulmonary and liver metastasis does not appear to be affected by synchronous versus metachronous presentation. Several studies have reported no significant difference in survival between patients who present with simultaneous versus synchronous disease.86 Patients with multiple pulmonary metastases (>3) have a significantly higher risk for disease-specific death compared with other patients.

Perihepatic lymph nodes are felt to be “metastases from metastases” and are generally associated with a poor outcome. Until recently, many considered such nodal involvement an absolute contraindication to hepatic resection.87 However, more recent studies have reported long-term survival in some patients with such nodal metastases and have concluded that this patient population may still benefit from hepatic resection.18,47,88,89

Metastatic spread to the peritoneal surfaces can occur in 15–20% of patients with advanced colorectal cancer and is not uncommonly seen in those with liver metastases. Although usually a particularly ominous indicator of poor prognosis, some centers have reported success with cytoreductive surgery and intraperitoneal chemotherapy.90 Surgical therapy of combined liver and peritoneal metastases has been advocated by some.91 However, concomitant peritoneal disease should be considered a contraindication for resection of liver metastases.

Elias et al92 reported that the 5-year survival rate following hepatectomy for colorectal liver metastasis and simultaneous resection of extrahepatic disease with curative intent was 29%, including selected cases with pulmonary, nodal, and peritoneal metastasis. Patients with peritoneal disease or extrahepatic disease at multiple sites were found to have worse survival rate than patients with single-site extrahepatic disease.92 In contrast, multiple studies have shown that patients with only pulmonary metastases have the most favorable outcome following liver and lung resection.82,83

While preoperative factors may be generally instructive, these factors should not be used to exclude patients from surgical consideration. Patients with one or multiple negative prognostic factors can still derive a significant survival advantage from hepatic resection of their colorectal metastases. However, the presence of disseminated extrahepatic disease, probably including periportal nodal disease or lack of control of the locoregional primary disease, is generally considered a contraindication to resection with curative intent. In patients with metastatic disease confined to the liver in whom surgical resection can completely and safely remove all evident disease with negative margins, surgical resection should be advocated.

Current combination chemotherapy regimens incorporating oxaliplatin or irinotecan in addition to 5-fluorouracil (5-FU) and leucovorin (LV) or capecitabine have produced response rates in excess of 40%.93–95 The addition of biologic agents such as those targeting epithelial and vascular endothelial growth factor pathways has been shown to further improve response rates and increase survival in the metastatic setting.96,97 The impressive results of these regimens have led to much enthusiasm for their use in combination with hepatic resection.

Optimal integration of chemotherapy and surgical therapy of liver metastases is an important and controversial topic. While extensive randomized controlled trials have shown the benefit of postoperative adjuvant chemotherapy in locoregional colon cancer,98–100 the role of chemotherapy combined with potentially curative resection of liver metastases is less well studied. Specifically, integration of chemotherapy may include (1) postoperative adjuvant therapy in the resected patient, (2) neoadjuvant or perioperative chemotherapy in the patient with initially resectable disease, and (3) preoperative chemotherapy in the initially unresectable patient in order to achieve a conversion to a resectable state.

Postoperative Adjuvant Chemotherapy

It has been known for decades that adjuvant chemotherapy improves disease-free and overall survival in earlier-stage colon cancer.99,100 In stage III patients, the risk of relapse with surgery alone is reduced by 40% with 5-FU-based adjuvant therapy, and the risk of death is reduced by 30%.99 The addition of oxaliplatin to infusion 5-FU/LV (FOLFOX regimen) significantly improves disease-free survival.98 However, the role of adjuvant chemotherapy is not as clearly defined after potentially curative resection of liver metastases (stage IV with no evidence of disease). Given that these patients have overall higher risk of recurrence than stage III patients, the rationale for use of adjuvant therapy in this setting is clear, yet few studies have evaluated whether chemotherapy improves outcomes following liver resection. Two European randomized trials closed early because of slow accrual. A pooled analysis of these included 278 patients who had undergone complete resection of the primary tumor and up to four liver or lung metastases and randomized patients to receive either six cycles of bolus 5-FU/LV or surgery alone.101 There was a trend toward improved 5-year progression-free survival with chemotherapy compared with surgery alone, but the results did not reach statistical significance. Portier et al reported a French trial of 173 patients who were randomized to 6 months of postoperative bolus 5-FU/LV versus observation following R0 resection.102 The 5-year disease-free survival was 34% for patients in the chemotherapy group and 27% for patients in the surgery-alone group (OR 0.66, p = .028). There was a trend toward increased 5-year overall survival with chemotherapy (51 vs 41% with surgery alone), but this difference did not reach statistical significance. These trials were conducted with bolus 5-FU/LV, the standard regimen at the time; the benefit of more effective postoperative oxaliplatin-based regimens such as FOLFOX as well as with biologic therapies has yet to be determined.

Intrahepatic infusional chemotherapy also has been proposed as adjuvant therapy following liver resection, based in part on the high risk of intrahepatic recurrence and the ability of regional chemotherapy to deliver higher drug concentrations within the liver. Several studies have examined the role of adjuvant hepatic artery infusion (HAI) using 5-FU or floxuridine (FUDR) with or without systemic chemotherapy.103 This therapy was compared with either observation or systemic chemotherapy with bolus 5-FU/LV. Two small trials demonstrated some benefit for HAI over this systemic chemotherapy alone.103 Although some improved efficacy was observed in these studies as well as phase II studies, comparisons with newer, more active systemic regimens in randomized trials are lacking.104 In addition, the toxicity and technical aspects of implanting and maintaining a hepatic arterial pump have limited its applicability in practice. One recent cooperative trial in the United States closed early after failing to accrue many patients, and the current role for adjuvant therapy using regional infusion chemotherapy in this disease is limited outside of a clinical trial.

Perioperative Chemotherapy in the Initially Resectable Patient

The use of preoperative chemotherapy for patients with initially resectable hepatic metastases has generated significant controversy. Several arguments have been made in favor of preoperative chemotherapy (Table 45-2). First, it provides some time for biologically aggressive disease to declare itself, potentially avoiding unnecessary surgery. Second, determination of response in the macroscopic disease provides prognostic information105–108 as well as potentially allowing for modification of the postoperative therapy based on response. Finally, neoadjuvant chemotherapy may allow a smaller resection or increase the possibility of an R0 resection. Achieving negative margins is important when resecting for curative intent and the rate of positive surgical margins may be lower with preoperative systemic chemotherapy than with immediate resection.42

While theoretically appealing, the use of preoperative chemotherapy in an initially resectable patient may have some disadvantages (Table 45-2). Some consider it important to offer the surgical therapy first in order to avoid a lost opportunity. If neoadjuvant therapy fails to stabilize the disease, it may progress to an unresectable status. This situation is rare, however, with most series reporting less than 10% occurrence. In addition, patients in this group may have less to gain from aggressive surgical therapy given the poor prognosis that early progression may indicate.106 Another disadvantage of preoperative chemotherapy is the risk of chemotherapy-associated liver injury, which could increase the morbidity of surgery.109–112 Seen in two forms, vascular sinusoidal dilatation and fatty liver changes (steatosis or steatohepatitis), hepatotoxicity can be limited with shorter duration of chemotherapy.

Another potential disadvantage of neoadjuvant chemotherapy is related to the complete radiologic response that can be observed in some lesions. In one study, van Vledder et al found 23% of patients receiving preoperative chemotherapy had at least one disappearing liver metastases.113 This finding was more common in small lesions, in patients with multiple metastases, and in those with longer chemotherapy duration (Fig. 45-2). Only approximately one-half were found at the time of surgery and, when not identified, the majority of these were not treated. In most cases local recurrence developed in these sites. Similarly, Benoist et al, reported the presence of residual disease at the site of a radiologic complete response in 82%.114 Thus, if the hepatic lesions are rendered so small that the surgeon can no longer find them, perhaps the patient may have been done a dis-service by preoperative chemotherapy.

The impact of perioperative chemotherapy in conjunction with liver resection was reported in a randomized, multi-institution trial sponsored by the European Organization for Research and Treatment of Cancer (EORTC 40983).115 In this study, 364 chemonaive patients with up to four colorectal liver metastases (median of 1) were randomized to either perioperative chemotherapy with six cycles of FOLFOX before surgery and six cycles after surgery, or to surgery alone. Difference in 3-year progression-free survival between groups was 7.3% in randomized patients (p = .058), 8.1% in eligible patients (p = .041), and 9.2% in patients successfully undergoing resection (p = .025). There were no differences in perioperative mortality between the treatment groups, but postoperative complications occurred significantly more often in those who received neoadjuvant therapy (25 vs 16%), including increased risk of biliary fistula and intra-abdominal infection. These findings support a role for chemotherapy in patients undergoing liver resection, although the benefit was perhaps less than may be expected based on the benefit seen with chemotherapy in stage III disease. Overall survival was not significant, so it is unclear whether perioperative FOLFOX actually contributes to cure or simply delays recurrence. Furthermore, the trial was not designed to test the value of preoperative versus postoperative chemotherapy and did not definitively answer the important question about the optimal sequencing of chemotherapy around curative liver resection.

Converting Unresectable to Resectable Disease with Preoperative Chemotherapy

The prospect of converting initially unresectable metastases to resectable disease has become more attainable with improved response rates from modern systemic regimens. Combination chemotherapy, including FOLFOX and FOLFIRI, achieves radiologic response rates of approximately 50%.97,116 More aggressive regimens (eg, FOLFOXIRI) as well as the addition of targeted therapies increase response rates even further.117–119

The goal of tumor downsizing with a chemotherapeutic response is to be able to remove the gross residual disease of all original sites of disease. Liver metastases located near major vascular pedicles which need to be salvaged are ideal candidates (Fig. 45-3). Currently, it is estimated that approximately one-fourth of patients with liver metastases are initially resectable and conversion from unresectable to resectable disease through tumor downsizing can be achieved in approximately 20% of those initially considered unresectable.120,121

Outcomes following liver resection in initially unresectable patients following chemotherapy response appear similar to initially resectable patients. In one report, 5- and 10-year overall survival rates were 33% and 23%, respectively in initially unresectable patients who subsequently underwent resection.120 In a more recent study, unresectable patients with liver-only disease were randomized to either FOLFOX or FOLFIRI in combination with cetuximab.122 Complete resection was achieved in 34% overall and was comparable in both groups.

Strategies for tumor downsizing and resection need to be further refined. We still do not know what circumstances indicate optimally converted patients. For example, it is unclear whether all areas where the tumor initially existed should be resected. Furthermore, it is still unclear what role ablation might play in such a setting. The optimal duration of preoperative chemotherapy in this setting also remains uncertain. For example, should patients be treated until disease is resectable or to a maximal response? It is likely that residual visible disease is of benefit in identifying all initial sites that need to be resected to prevent recurrent disease. Similarly, it is unclear how to manage disease in the case of complete radiologic response. Should patients undergo surgery at that time or wait until some disease becomes radiologically evident?

Although surgical resection may afford the best potential for cure in patients with hepatic metastases, many patients may not be candidates for surgical resection for a variety of reasons. Novel methods for local ablation have been developed with a goal of increasing the number of patients eligible for local, potentially curative therapy. The early experiences with metastasis ablation have been primarily with hepatic cryosurgery. This technique relies on the destruction of a defined area within the liver by freeze or thawing, using probes cooled by liquid nitrogen to subzero temperatures. Placement and monitoring of the freezing process is monitored with ultrasound. More recently, radiofrequency ablation (RFA) has been applied for the treatment of liver tumors in much the same manner as cryotherapy. With this technique, a needle-probe is inserted within the selected tumor under image guidance and electric current is employed to generate heat, resulting in interstitial thermal destruction. RFA can be performed laparoscopically or percutaneously, as well as via open laparotomy. Advantages of RFA over cryotherapy include the smaller probe and fewer related complications. For these reasons, RFA has replaced cryotherapy as the preferred method of interstitial ablation in many centers. While most studies report outcomes of RFA for treatment of colorectal metastases, newer ablative approaches are being used with increasing frequency. These include microwave ablation and nonthermal irreversible electroporation. While potentially promising, these newer ablative modalities await larger controlled reports to determine their role in therapy of hepatic colorectal metastases.

When performing ablation of colorectal metastases, RFA in particular, careful planning of the zone of destruction is necessary to achieve complete necrosis of the target lesion.123 In some cases when using an expandable multielectrode RFA needle of sufficient size, complete ablation can be achieved with a single application, deploying the electrode from the center of the tumor. In cases of larger lesions, multiple overlapping applications may be required.124 Tumor size as well as location can preclude effective ablation with curative intent. Tumor sizes larger than 3 cm are associated with an increased incidence of local recurrence.125 Similarly, tumors near major vascular structures such as the vena cava are more difficult to achieve long-term local control with ablation.

Once the target tumor is identified with the IOUS transducer during a surgical ablation, the RFA electrode needle is inserted under ultrasound guidance (Fig. 45-4). Optimally, the electrode is advanced in a track parallel and within the plane of the transducer, so the entire path of the needle can be visualized. Monitoring during thermal ablation can be performed using a variety of methods. Some RFA devices have the capacity to measure tissue temperatures with thermistors located at the tips of the electrodes. Alternatively, tissue impedance and current can be monitored during treatment. The ablation zone is visualized by ultrasound during treatment. Typically, local miniscule gas bubble formation results in hyperechogenicity within the treated tissue.

The optimal method for evaluating the efficacy of RFA treatment using imaging modalities is not well defined. The presence of residual viable tumor tissue after RFA can be detected using contrast-enhanced CT or MRI.126,127 However, as with other ablative approaches, interpretation of these images can be difficult at times, as a hypoattenuating lesion may persist for months to years despite complete tumor destruction. In most cases, a local recurrence is characterized by an increase in the lesion size on serial scans, or evidence of new areas of contrast enhancement. FDG-PET may be useful in assessing recurrent disease provided it is acquired after the postablation inflammation has subsided, typically after 3 months.127

Unlike hepatic resection where long-term efficacy is relatively established, evidence of the benefit of ablation for treatment of hepatic colorectal metastases has been limited and inconsistent.123 There are no published randomized controlled trials examining its use in this disease so the data are largely based on single-arm, retrospective, and prospective studies.

Local recurrence rates published in the literature range from less than 10% to as high as 40–50%.73,123,128,129 Data on the survival benefit of RFA have been similarly contradictory. Some studies have reported 5-year survival of less than 20% following RFA whereas other studies have reported 5-year survival rates in the range of 40% or more.123 In addition, several studies have attempted to compare outcome of resection versus ablation, most demonstrating RFA to be associated with a worse disease-free and overall survival compared with resection. One must realize, however, that important prognostic and treatment-related variables differ between the two cohorts when compared retrospectively. In a review of RFA for colorectal cancer by the American Society of Clinical Oncology (ASCO), the authors recognized a scarcity of quality evidence and emphasized a compelling need for more clinical trials to determine the efficacy and utility of RFA in these patients.123

Chemotherapy delivered systemically to treat hepatic metastases may be limited by the incapability to deliver high concentrations of drug to tumor cells without systemic toxicity. This provides the rationale for regional chemotherapy. The normal liver derives its blood supply from both the hepatic arterial and portal venous vessels. Macroscopic hepatic tumors, however, derive most of their afferent blood supply from the hepatic artery.130 Directing a high-dose infusion of chemotherapy into the hepatic artery therefore increases the concentration of drug to which the tumor is exposed, proportional to the hepatic parenchyma, as well as to the body as a whole, and should thereby improve the therapeutic index.131 Drugs such as fluorodeoxyuridine (FUDR), an active metabolite of 5-FU, which are rapidly metabolized within the liver on the first pass, are the most commonly used regional chemotherapeutic agents.132

While evidence regarding the role of HAI therapy was summarized earlier, several older randomized trials have compared the results of continuous HAI chemotherapy alone to systemic chemotherapy in patients with unresectable liver metastases.133–135 Most reported increased response rates compared to systemic therapy but survival benefit was inconsistent, in part because of inadequate sample sizes, crossover study designs, or inadequately administered systemic chemotherapy.

In 2003, Kerr et al reported a European trial randomizing 290 patients to systemic versus HAI 5-FU.136 While the study was criticized for the technique and study design, the median overall survival was no different between treatment groups. In another randomized multi-institutional trial, Kemeny et al reported a higher response rate with HAI FUDR with dexamethasone (47%) compared to the systemic group receiving 5-FU/LV (24%), along with an improved overall survival in the HAI group (24.4 vs 20 months, p = .003).103 As with HAI therapy in the adjuvant setting, limited level 1 evidence exists comparing regional infusion chemotherapy to current combination systemic regimens. While response rates are high with this approach, even following tumor progression on systemic therapies, the biliary toxicity and technical aspects of implanting and maintaining an hepatic arterial pump have limited its applicability in current practice beyond few centers with experience in this approach.

Surgical therapies for hepatic colorectal metastases have been shown to be increasingly safe and effective resulting in more frequent and aggressive application of this local approach. Preoperative and intraoperative assessment and planning are important to achieve safe and complete resection of all evident disease. Current methods for increasing the ability to offer liver resection include preoperative chemotherapy, staged resection, preoperative portal vein embolization, and ablative strategies. Perioperative chemotherapy may play a role in the optimal treatment of initially resectable disease, but the sequencing of chemotherapy and surgery remains unclear.

In the near future, we are likely to see expanding use of local therapies of hepatic metastases, particularly as systemic chemotherapy improves. Minimally invasive approaches for resection, including laparoscopic resection, will likely be increasingly utilized, as well as other nonextirpative techniques. However, until the role of cytoreduction or incomplete local therapies is defined, complete, curative-intent therapy must be advocated.