Management of Metastatic Colorectal Cancer to the Liver

Colorectal cancer is the third most common cancer diagnosed in both males and females in the United States. As of January 2016 there are an estimated 1.4 million men and women living with a diagnosis of colorectal cancer, with an expected 134,490 new diagnoses and 49,190 deaths that same year. Of the projected new cases diagnosed in 2016, approximately 50% to 60% of those patients will develop metastases to the liver. It has been estimated that more than 50% of patients who die of colorectal cancer have liver metastases at autopsy and 35% have isolated hepatic metastases. Metastatic liver disease is the cause of death in most patients. The majority of patients develop metachronous metastatic lesions. However, 20% to 34% of patients present with metastatic disease to the liver at the time of diagnosis and carry a worse prognosis. Hepatic resection for colorectal metastases is the treatment of choice for patients with resected or resectable primary and regional disease if all liver disease can be treated.

In this chapter, we will focus on the management of liver metastases from colorectal cancer. Our primary aim will be to provide practical treatment strategies for various clinical situations encountered by surgeons. Within this overview we will review the approach to liver resection and the supporting data. We will also outline nonresection treatment therapies that can be used in the management of metastases.

Evolution of Treatment for Colorectal Metastasis

The principal spread of metastatic colorectal cancer is through the portal circulation to the liver. Weiss et al. analyzed the metastatic patterns of colorectal cancer in 1541 necropsies and proposed a cascade hypothesis that describes the occurrence of liver metastases through seeding from the portal venous system. These secondary hepatic metastases seed the lungs to form tertiary pulmonary metastases. The tertiary pulmonary metastases then spread through the arterial system to form quaternary metastases in other organs. In 2004 Wang et al. described the liver as a filter that prevents metastatic spread to distant sites. In a study led by Uetsuji et al. metastases were noted in patients without cirrhosis (46 of 210), whereas those patients with cirrhosis had no evidence of metastasis (0 of 40). The authors concluded that decreased portal flow in the cirrhotic liver limited the exposure to metastatic cells.

Untreated colorectal cancer with metastases to the liver has classically shown dismal outcomes. The median survival of this patient population is just 5 to 10 months, with a 3-year survival that is rarely reported. The prognosis is closely linked to both the metastatic burden and the timing of metastasis. In 2015 Adam et al. reported a standardized definition of metastases. Liver metastases detected at or before diagnosis of the primary tumor are defined as synchronous liver metastases. Early metachronous metastases are those detected within 12 months after diagnosis or surgery of the primary, whereas late metachronous metastases are those detected more than 12 months after diagnosis or surgery of the primary. Their review suggests that synchronous metastases have less favorable cancer biology and synchronicity is a sign of poor prognosis.

With a better understanding of hepatic physiology and recent advances in chemotherapy, targeted surgical strategies have been used to successfully treat liver metastases. To date, numerous studies have demonstrated that liver resection for colorectal liver metastases (CRLMs) is safe, with some reporting a 5-year survival of greater than 50%. Based on these studies, liver resection has now become the standard treatment for metastatic colorectal cancer to the liver.

Patient Selection

Many controversies arise when discussing the treatment of liver metastases from colorectal cancer. Some of these include the use of neoadjuvant or adjuvant chemotherapy, the timing of resection for synchronous metastases, how to approach disappearing lesions, the best methods to treat bilobar disease, the treatment of recurrent metastases, and the operative indications for extrahepatic disease (EHD). To select the patients who would benefit most from surgery, a multidisciplinary approach should be used. At our center, patients are presented at a multidisciplinary tumor conference including specialized surgeons, medical oncologists, pathologists, radiologists, and key ancillary staff. Through multidisciplinary discussion, therapeutic options can be discussed and an optimal treatment plan formulated.

Selection of patients for hepatic resection includes the patient's medical fitness for major laparotomy, the stage of the primary tumor, the extent of the hepatic metastases, and the intent of the resection. The morbidity associated with hepatic resection is directly related to patient selection. Although multiple factors are considered in surgical decision making, a study of 747 hepatectomies revealed that the incidence of postoperative complications was significantly influenced by the American Society of Anesthesiologists (ASA) score, the presence of hepatic steatosis, the extent of resection, and an associated extrahepatic procedure. Kamiyama et al. reviewed 793 consecutive hepatectomies and reported that independent relative risk for morbidity was influenced by an operative time greater than 360 minutes, blood loss greater than 400 mL, and serum albumin less than 3.5 g/dL. Thus the patient's clinical performance status and comorbidities of major organ systems should allow them to undergo a major resection with an expected mortality risk of less than 5%.

The extent of hepatic metastatic disease affects patient selection for liver resection. Historically, adequate hepatic reserve in patients without chronic liver disease requires at least two remaining anatomically adjacent segments after resection (with normal vascular inflow and outflow, and normal biliary drainage). If cirrhosis is present, the extent of resection is reduced dramatically, and ablation may assume a larger therapeutic role. Extensive hepatic steatosis without cirrhosis also limits the extent of hepatic resection.

Technical advances in both liver and colonic surgery have enabled simultaneous resection of the primary colorectal cancer and liver metastases in carefully selected patients with synchronous tumors. Undertaking concurrent resection of the primary colorectal cancer and hepatic metastases requires thorough preoperative hepatic imaging, thus allowing for evaluation of the intrahepatic extent of disease and also to exclude EHD. A publication by de Haas et al. suggests that combining colorectal resection with limited hepatectomy is safe in patients with synchronous tumors and is associated with less cumulative morbidity than a delayed liver resection procedure. Extensive liver metastases, including multiple bilobar metastases, require careful planning for the optimal approach. These cases often require a combination of resection and nonresectional therapies to address all liver disease. Involvement of the afferent or efferent vasculature may require advanced vascular resections, whereas involvement of both the afferent and efferent vasculature may be a contraindication to resection.

Hepatic resection for metastases should rarely be undertaken if extensive extrahepatic metastases exist, exclusive of regional lymph node or limited (usually solitary) pulmonary metastases. Although resection of focal EHD that is concurrently resectable with the primary has been reported to show long-term control, it has not been shown to be curative. Current contraindications to resection of hepatic metastases in the setting of extrahepatic metastases include distant metastases, including peritoneal carcinomatosis, osseous or brain metastases, extraabdominal lymph node metastases, and multiple unresectable pulmonary metastases.

Prognostic Determinants

Traditionally it had been suggested that survival rates for patients with metastases from colorectal cancer could be improved through careful selection of operative candidates and the use of neoadjuvant therapy to help with downstaging. However, there continues to be a wide range of oncologic characteristics of the tumors, leading to a variable degree of aggressiveness. Although useful, markers such as tumor number, size, bilobar disease, and high serum carcinoembryonic antigen (CEA) are not as clinically predictive as was once thought. With the evolution of more potent and targeted chemotherapy, emerging determinants include the radiologic and pathologic response of the tumor to chemotherapy, disease-free interval, and synchronicity, as well as the genetic mutational status of the tumor. Poultsides et al. evaluated the pathologic response to preoperative chemotherapy. Even though tumor size may decrease on imaging, this does not always correlate with tumor regression. They suggested that fibrosis is the main determinant of treatment response and that the degree of tumor fibrosis noted in the liver resection specimen correlated with disease-specific survival.

Tumor progression while on neoadjuvant chemotherapy is associated with a poor outcome. Radiologic response has also played a role in evaluating tumor response to chemotherapy. Traditionally the RECIST criteria have been used to evaluate tumor response on imaging ; however, newer morphologic characteristics have been noted in response to chemotherapy that correlate more closely with pathologic response. Shindoh et al. reported on the texture of the tumor noted on computed tomography (CT) and concluded that optimal morphologic response following preoperative chemotherapy correlated with improved overall survival.

With improved techniques in mutational analysis, the identification of specific mutations has also been linked to prognosis. RAS oncogene status is a strong predictor of response to antiepidermal growth factor receptor agents. Some have suggested that KRAS mutational status is associated with aggressive tumor biology. In a recent study, Mise et al. reported that RAS mutational status can be used as a prognostic indicator for those undergoing liver resection for CRLMs. They concluded that major pathologic responses were more common in patients with wild-type RAS in comparison to those with RAS mutations (58.9% vs. 36.8%, respectively; P = .015).

Factors Determining Resectability

Oncologic Evaluation

Traditionally, complete resection of all viable disease is essential to achieve the best long-term outcomes. Preoperative evaluation of the surgical candidate should include the tumor response to chemotherapy, the mutational status of the tumor, and the presence of EHD. Current contraindications to resection of hepatic metastases include diffuse peritoneal carcinomatosis, osseous or brain metastases, extraabdominal lymph node metastases, and multiple, unresectable pulmonary metastases. Focal extrahepatic peritoneal disease that is concurrently resectable was previously thought to be a contraindication to resection but is now considered an option in appropriately selected patients.

Overall number of extrahepatic metastases resected has a stronger prognostic effect than the location. Carpizo et al. presented their series of 127 patients who underwent hepatectomy with concurrent resection of EHD. They reported a 5-year survival of 26% compared with 49% for those resected without EHD. They concluded that the presence of limited and resectable EHD should not be a contraindication to resection and could be associated with long-term survival. A noticeable feature of their analysis showed that 95% of the patients recurred after complete resection, indicating that concurrent resection of hepatic metastases and EHD is not considered curative. Leung et al. developed an EHD risk scoring system. The score was prognostic of overall and recurrence-free survival. In general, appropriately selected patients with limited EHD could be resected with the expectation of reasonable long-term control with the use of newer adjuvant chemotherapies.

Patients who successfully underwent downstaging with neoadjuvant chemotherapy and then resection have similar survival rates as those patients who were resectable at their initial presentation. This neoadjuvant approach also provides an insight into the response of the tumor, thus revealing its biologic activity or tumor biology. Some have even suggested that in the era of modern chemotherapy, tumor biology is a more important factor in survival than surgical margin clearance. With the use of modern systemic chemotherapy (SC), progression during preoperative treatment is relatively rare. However, progression (with or without the development of new lesions) during preoperative chemotherapy should be regarded as a marker of poor prognosis and an indication for second line chemotherapy before surgery is considered.

Although the neoadjuvant approach can be effective, it is not without risk. As Vauthey et al. have shown, there is a 20% incidence of steatohepatitis in association with irinotecan-based chemotherapy and a 19% incidence of sinusoidal injury after treatment with oxaliplatin. Steatohepatitis was associated with an increased 90-day mortality. Ribero et al. also found that the addition of bevacizumab reduced the incidence and severity of oxaliplatin-related sinusoidal injury while improving clinical response to the chemotherapy. Other series have reported a higher rate of sinusoidal injury and an increase in perioperative blood transfusions and surgical complications in patients who received preoperative chemotherapy than those who did not.

Technical Evaluation for Resectability

The extent of resection depends on the burden of metastases, the intrahepatic site, and the relationship of the tumor to major vasculature and bile ducts. Historically, evaluation for resection was based on resection of all viable disease with (1) preservation of at least two contiguous hepatic segments with (2) adequate vascular inflow and outflow, with biliary drainage and (3) the ability to preserve an adequate future liver remnant (FLR) (>20% in healthy normal liver, >25% for most cases as a minimum FLR).

With advances in resection techniques the development of volumetric and functional measurements has been established to allow a more detailed evaluation of the FLR. As a greater volume of diseased liver is resected, the incidence of postoperative hepatic insufficiency (PHI) increases. The basis of volumetric measurements relies on the use of mutliplanar imaging to assess the volume of the FLR in comparison with the whole liver volume. When considering resection, minimal FLR volume has been set at greater than 20% for a normal liver, greater than 30% in a damaged liver after extensive treatment, and greater than 40% for a cirrhotic liver. A strong correlation between FLR size and PHI, morbidity, and mortality has been shown.

Another method of determining an adequate remnant volume is by assessing the regenerative capacity of the liver. This can be measured with volume of hypertrophy, as well as rate of hypertrophy defined as the kinetic growth rate (KGR); a KGR of 2% per week reduces hepatic complications and liver failure–related death. This typically is in conjunction with portal vein embolization (PVE) or staged hepatectomy. A third measure of overall hepatic regenerative capacity is indocyanine green clearance and hepatic scintigraphy. Both methods effectively estimate the metabolic function of the liver and can be used to accurately assess the FLR.

Risk Scoring Systems

To aid in determining who will benefit most from surgical intervention, several clinical risk scoring (CRS) systems have been developed. The aims of these scoring systems are to optimize patient selection for hepatic resection and to stratify patients for the need of adjuvant therapies. Independent risk factors, such as CEA level, tumor size, and tumor number, have been reported to predict outcome after resection of CRLMs. Different authors have used different variables and cutoff values to discriminate between patient groups with different prognoses.

Nordlinger et al., through a national collective registry of 1568 patients, developed a prognostic scoring system based on seven identified risk factors: (1) age more than 60 years, (2) primary cancer extending into serosa, (3) positive regional lymph nodes, (4) liver metastases confirmed within 24 months of the primary cancer, (5) CEA levels, (6) size of metastasis more than 5 cm, and (7) less than 1-cm resection margin of the metastases. Three risk groups were defined: low risk (zero to two risk factors), intermediate risk (three to four risk factors), and high risk (five to seven risk factors).

In a similar method, Fong et al. through a single institution study of 1001 patients, devised a CRS system based on: (1) nodal status of primary, (2) disease-free interval from the primary to discovery of liver metastases less than 12 months, (3) number of tumors greater than 1, (4) size of the largest tumor greater than 5 cm, and (5) preoperative CEA level greater than 200 ng/mL. Each positive criterion is assigned a point, with the total score out of 5 being predictive of outcome. The Fong prognostic scoring system has been validated by numerous independent databases and shows a significant survival difference between groups scoring 0 to 2 when compared with groups that scored 3 to 5. Merkel et al. compared the three score systems of Nordlinger, Fong, and the extended TNM classification. They identified the CRS system developed by Fong to be the most important tool for estimating prognosis and selecting patients for surgical resection. This CRS system has also been helpful in selecting patients that could benefit from neoadjuvant therapy or be stratified into clinical trials. Fong et al. demonstrated that use of this CRS was shown to be useful in guiding preoperative evaluation of the patient with metastatic colorectal cancer, including when to use positron emission tomography (PET)/CT and diagnostic laparoscopy to prevent unneeded laparotomy.

Pretherapeutic Imaging Evaluation

Computed Tomography

Contrast-enhanced CT (ce-CT) has become the primary tool used for evaluation of CRLMs. Chest, abdomen, and pelvis can all be evaluated in a matter of minutes. This modality identifies CRLMs as hypovascular lesions when compared with surrounding normal liver parenchyma. Although CRLMs can often be identified on ce-CT imaging, the timing of the contrast is vital to identify these hypovascular lesions. A ce-CT that includes arterial, portal venous, and delayed venous phases will help to define the burden of disease and relevant anatomy and aid in preoperative planning. Although ce-CT can be quick and relatively inexpensive, it does expose the patient to radiation and potential contrast allergies and has a limited ability to evaluate lesions less than 10 mm in size.

Magnetic Resonance Imaging

Historically, magnetic resonance imaging (MRI) has been used as a problem-solving modality in lesions that are equivocal on ce-CT or ultrasound imaging. CRMs often appear hypointense on T1-weighted images and hyperintense on T2- and diffusion-weighted sequences. In 2014 Zech et al. evaluated the efficacy of gadoxetic acid–enhanced MRI (Eovist) compared with conventional MRI with extracellular contrast medium (ECCM) and ce-CT in detecting CRLMs. They concluded that Eovist MRI is superior as a first line imaging modality in detecting CRLMs. They noted that the use of Eovist imaging in preoperative planning was associated with the lowest proportion of patients for whom the surgical plan had to be changed at the time of surgery. They also noted that when used to help to better define ce-CT imaging characteristics, it prevented unnecessary surgery and clarified metastatic characteristics, allowing patients to undergo successful hepatic resection. Although Eovist imaging is more sensitive and specific for detecting CRLMs, it is also associated with a higher cost and lower patient compliance and is not readily available in all institutions.

Positron Emission Tomography

PET relies on the activity of hypermetabolic malignant tissue to take up radiolabeled glucose molecules. This modality is often combined with CT to better define areas of concern. In evaluating liver tumors, it has been shown to be less sensitive than MRI or CT; however, it can be useful for defining radiologically occult EHD. PET/CT is not currently recommended as a staging or baseline modality except in those circumstances in which previous imaging has revealed surgically resectable metastatic disease and the patient needs to be evaluated for unrecognized metastatic disease that would preclude surgical management.

General Principles for Resection

Resection or ablation of metastases should never put the liver at risk for irreversible dysfunction. The extent of resection will depend on the size of the metastases, intrahepatic site, and on the relationship of the tumor to major afferent and efferent vasculature and bile ducts. In patients with deeply seated metastases, formal anatomic resections are indicated. Moreover, metastatic disease manifesting indistinct margins mandates formal resection. A negative margin is vital to reduce the risk of intrahepatic recurrence at the margins of resection. However, margins of resection should never risk damage to major hepatic vasculature. The afferent and efferent vasculature of the liver remnant must be protected. The liver parenchyma can be transected by a variety of methods: compression (finger fracture or clamp fracture), contact (Cavitron Ultrasonic Aspirator [Cavitron, Long Island, New York]), thermal (electrocautery, laser, radiofrequency ablation [RFA], LigaSure, Harmonic scalpel), or stapled techniques. Each approach has advantages and disadvantages and are often selected based on surgeon preference. Most methods disrupt parenchyma to expose vessels and bile ducts for ligation, whereas newer methods including TissueLink (TissueLink Medical, Dover, New Hampshire), Harmonic scalpel (Ethicon Endo-Surgery, Cincinnati, Ohio), and LigaSure (Covidien Ltd., Dublin, Ireland) fuse small vessels and ducts while transecting parenchyma. Although the extent of parenchymal necrosis adjacent to the transection plane varies among techniques, such devitalized parenchyma is not clinically significant. Vessels or ducts of diameter larger than 2 mm generally require ligation with suture or clips. Major hepatic or portal veins are best occluded securely with the use of vascular staples or alternatively a running monofilament permanent suture.

Anatomy

Safe hepatic resection depends on a clear understanding of the hepatic anatomy. Although hepatic regenerative capacity and metabolic reserve permit many types of resections, resection based on preservation of residual anatomic integrity best reduces the operative risk and optimizes function. Couinaud's description of hepatic anatomy highlights the anatomic features of the liver relevant to resection and in adults provides anatomic terminology that is clinically useful. Fig. 172.1 details the functional divisions of the liver, according to Couinaud's nomenclature. Although the regenerative capacities and metabolic reserve of the liver are important, hepatic resection based on anatomic considerations reduces operative risk and optimizes postoperative liver function. In general, anatomic resections are the oncologically approved method to ensure cancer-free margins and to lower the risk of potential sites for intrahepatic spread. The major anatomic features of the liver relevant to resection have been detailed elsewhere.

FIGURE 172.1, The functional division of the liver and of the liver segments according to Couinaud's nomenclature. (A) As seen in the patient. (B) In the ex vivo position.

The hilar plate is the extension of a vasobiliary sheath that is particularly relevant to hepatic resection ( Fig. 172.2 ). The vasobiliary sheath represents a fusion of the endoabdominal fascia around the bile ducts, portal vein, and hepatic artery at the porta hepatis. These fibrous sheaths invest the components of the pedicles from the portal vein bifurcation to the sinusoids. By contrast, the hepatic veins lack endoabdominal fascial investment and, consequently, are more fragile than their portal counterparts. The density of the vasculobiliary sheaths decreases as the pedicles extend intrahepatically. At the hepatic hilus, these sheaths fuse to form plates that surround the portal pedicles, both anteriorly and posteriorly. Three primary hepatic plates are recognized: the cystic, the hilar, and the umbilical plates ( Fig. 172.3 ). Recognition of the vasculobiliary sheaths and the hepatic plates facilitates precise access to the hilar structures. Division of these plates is needed to expose and mobilize the portal pedicle during resection.

FIGURE 172.2, (A) Relationship between the posterior aspect of the quadrate lobe and the biliary confluence. The hilar plate is formed by the fusion of the connective tissue enclosing the biliary and vascular elements with Glisson capsule. (B) Biliary confluence and left hepatic duct exposed by lifting the quadrate lobe upward after incision of Glisson capsule at its base. This technique (lowering of the hilar plate) generally is used to display a dilated bile duct above an iatrogenic stricture or hilar cholangiocarcinoma. (C) Line of incision (left) to allow extensive mobilization of the quadrate lobe. This maneuver is of particular value for high bile duct stricture and in the presence of liver atrophy or hypertrophy. The procedure consists of lifting the quadrate lobe upward (see [A] and [B]), then not only opening the umbilical fissure, but also incising the deepest portion of the gallbladder fossa. Right, Incision of Glisson capsule to gain access to the biliary system (arrow).

FIGURE 172.3, Sketch of the anatomy of the plate system. Note the cystic plate (A) above the gallbladder, the hilar plate (B) above the biliary confluence and at the base of the quadrate lobe, and the umbilical plate (C) above the umbilical portion of the portal vein. Arrows indicate the plane of dissection of the cystic plate during cholecystectomy and of the hilar plate during approaches to the left hepatic duct.

Number of Metastases

The number of metastases is no longer a contraindication to resection provided an adequate FLR remains and all metastases are ultimately completely resected or ablated. Kokudo et al. concluded that resection in patients with four or more CRM can achieve long-term survival and that the number of hepatic metastases alone should not be used as a sole contraindication to resection. Similarly, although long-term prognosis in patients with metastases to lymph nodes is unfavorable, hepatic resection combined with lymphadenectomy may be beneficial in occasional patients whose disease has been downstaged or completely eliminated clinically by chemotherapy and can be resected completely. Timing of resection in patients presenting with synchronous disease should allow for liver resection before lung resection to permit complete pulmonary function prior to lung resection. In addition, hepatic resection before lung resection allows for surgical inspection of the abdomen, ruling out advanced disease. Extensive peritoneal metastases are considered by most advanced centers to be a contraindication to surgical resection. Although there are limited data reporting a small survival advantage of combined resection and hyperthermic intraperitoneal chemotherapy (HIPEC), this is not standard of care and is usually practiced in the setting of clinical trials.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here