Authors report no disclosures or funding.
Hepatic resection for colorectal metastases
Article first published online: 7 DEC 2013
© 2013 Wiley Periodicals, Inc.
Journal of Surgical Oncology
Special Issue: Seminars in Surgical Oncology: Role of Surgical Metastasectomy in Oncologic Practice
Volume 109, Issue 1, pages 2–7, January 2014
How to Cite
Frankel, T. L. and D'Angelica, M. I. (2014), Hepatic resection for colorectal metastases. J. Surg. Oncol., 109: 2–7. doi: 10.1002/jso.23371
- Issue published online: 13 DEC 2013
- Article first published online: 7 DEC 2013
- Manuscript Accepted: 19 APR 2013
- Manuscript Received: 15 APR 2013
- colon cancer;
The liver represents a common site for metastasis in colorectal cancer. Improvements in patient selection and surgical techniques has resulted in improved outcomes following hepatic metastasectomy with large series reporting 5- and 10-year overall survival rates of 40% and 20%, respectively. In recent years, criteria for resectability has expanded with the use of forced liver hypertrophy and staged resection. The role of perioperative chemotherapy remains controversial with a slight increase in survival and operative morbidity. J. Surg. Oncol. 2014 109:2–7. © 2013 Wiley Periodicals, Inc.
Colorectal cancer represents the third most common malignancy in men and women in the United States with 100,000 new cases per year [1-3]. Twenty percent of patients present with metastases with many more developing distant spread during the course of their disease. The most common site of metastasis is the liver which is present in nearly 80% of stage IV patients and the sole site of disease in approximately 40% of these . It has been estimated that up to 20% of these patients have resectable disease . As understanding of liver anatomy and anesthetic techniques have evolved over the past three decades, hepatectomy for colorectal metastases has become increasingly adopted with 150–200 resections performed annually in our institution alone. While early series reported operative mortality as high as 30% , recent advancements have brought that number down to 1–3% [6-8]. As perioperative outcomes have improved, so to has the understanding that hepatic resection for limited liver metastases can lead to disease control superior to systemic chemotherapy and perhaps more importantly, long-term cure . Recent data from the United States and Canada have demonstrated 10-year disease-free survivals of 15–25% which is far better than the data on even the best systemic chemotherapy [6, 10-13] (Table I). To achieve the best perioperative and long-term results, a multidisciplinary approach to treatment is required incorporating medical oncology, radiology, anesthesia, and surgery.
|Number of patients||5-year survival (%)||10-year survival (%)|
|Tomlinson et al. ||612||NS||17|
|Choti et al. ||226||40%||26|
|Wei et al. ||423||47%||28|
|Scheele et al. ||350||39%||24|
|Giuliante et al. ||251||39%||24|
Indications and Pre-Operative Considerations
The cornerstone of resection for metastatic disease is patient selection. First popularized by Alexander and Haight , the fundamentals include: resectability of the primary tumor, adequate fitness of the patient to tolerate a major operation, and ability to achieve a margin negative resection at an appropriate disease-free interval. These tenets hold true for hepatic resection for colorectal metastases although as the operation has become safer and systemic chemotherapy more effective, indications have expanded well beyond these rules including resection of synchronous disease. To better select patients most likely to benefit from a potentially morbid operation, many predictive models have been created based on retrospective review of large series data. Fong et al.  reviewed results of 1,001 patients undergoing liver resection for colorectal metastases over a 13-year period. Multivariate analysis identified several factors associated with poor overall survival including: positive margin, presence of extrahepatic disease, number and size of tumors, high pre-operative CEA, lymph node positive primary, and short disease-free interval. Using the five pre-operatively identifiable factors, a score was assigned to patients with low-score patients surviving a median of 74 months and high-score patients 22 months. This model has been validated in independent series [16, 17] although others have not found this score prognostic . A similar scoring system was proposed by Nordlinger et al. [7, 19] using data from 1,600 patients in Europe. Interestingly, many of the same predictors were identified, but their model incorporated age and more detail about the stage of the primary tumor. Similar disparity was found between low and high score patients with 2 years survival of 79% and 43%, respectively. It is important to note that while these scoring systems are able to stratify patients into high and low risk groups, they do not identify patients in whom cure is precluded and should not replace sound clinical judgment.
Once the decision is made to operate, one must ensure medical fitness of the patient and that an adequate future liver remnant (FLR) will be present after resection. Given that a partial hepatectomy is a major operation and patients with colorectal cancer tend to be of advanced age, thorough medical clearance should be obtained to ensure optimal cardiac and pulmonary function. Patients should undergo an assessment of performance status and any co-morbid medical conditions. Liberal use of medical sub-specialty physicians should be used if there are suspicions of specific underlying medical issues.
The FLR can best be determined using pre-operative cross-sectional imaging with or without some measurement of hepatic synthetic function. Using standard radiographic software, the proposed line of resection is drawn and the remaining liver volume is calculated as a percentage of the whole liver volume (Fig. 1). Tumors, cysts, and ablation cavities are excluded. While no firm number exists regarding absolute minimal FLR, most agree that an FLR of approximately 40% and 30% for chemo treated with significant parenchymal damage and naïve patients, respectively ensures a low (but not absent) risk of post-operative liver failure [19, 20]. When synthetic function is compromised from underlying liver disease or heavy pre-treatment, one should consider quantitative measures of liver function including indocyanine green clearance  and lidocaine conversion tests  although these are not commonly used in Western centers.
Patient selection is often more important than technical considerations when operating for metastatic disease. The biology of the tumor must always be in the forefront including number of tumors, rate of growth, presence of extra-hepatic disease and likelihood of obtaining a margin negative resection. While retrospective reviews of series have helped identify subgroups of patients likely to benefit from intervention, institutional and surgeon experience as well as multidisciplinary input are essential to ensuring success.
Pre-operative knowledge of tumor size, number, and location are essential to identify patients suitable for resection and determining FLR. While surgeons historically relied on intraoperative exploration and ultrasound at the time of resection, improvements in cross-sectional imaging have provided increasingly in-depth information pre-operatively. There is no gold standard technique for identifying liver metastases or anatomic structures and much of it depends on institutional bias. Multi-detector computed tomography (MDCT) is fast, relatively inexpensive, and with the addition of iodinated intravenous contrast and thin slices through the liver, provides an adequate view of metastases which appear hypovascular to the surrounding liver parenchyma  (Fig. 2a). Because of the speed at which images are acquired, the entire chest, body, and pelvis can be obtained in less than a minute allowing for evaluation of extra-hepatic disease and requiring very little immobility from patients. Drawbacks of MDCT include radiation exposure, toxicity of contrast and inability to characterize lesions less than 1 cm in size.
Magnetic resonance imaging (MRI) uses the differing properties of water and fat protons to generate a series of images based on the proportion of both in physiologic and pathologic tissues. While the images can be of lower resolution than MDCT, they are better able to distinguish small liver lesions using the amount of water present. Colorectal metastases appear hypointense to surrounding liver parenchyma on T1 images (Fig. 2b) and hyperintense on T2 and diffusion weighted sequences. Unlike MDCT, MRI requires a great degree of patient compliance, is relatively expensive and typically only images the region of interest. A significant advantage of MRI over other imaging modalities is its ability to identify even small lesions in steatotic livers.
Positron emission testing (PET) relies on the propensity of hypermetabolic tissue to take up radioactive tagged glucose molecules. It is most often performed in combination with MDCT to allow better localization of hypermetabolic foci. Because of the high physiologic metabolic activity of liver parenchyma, PET is not particularly useful in imaging intrahepatic tumors, but can help identify and characterize radiologically occult extra-hepatic disease .
Multiple studies have compared the three modalities with regards to liver disease with most finding greatest sensitivity and specificity with MRI over CT or PET . At our institution, CT is the imaging modality of choice due to its ease of acquisition and tolerability by patients. CT is performed as a triple phase study with 2.5 mm cuts through the liver. If lesions are identified that are difficult to characterize or if significant hepatic steatosis is present, an MRI is obtained to differentiate metastases from benign entities such as cysts, adenomas, and hemangiomas. PET is similarly used selectively to characterize suspicious extrahepatic lesions identified on MDCT.
Surgical Technique and Post-Operative Considerations
Once the decision to operate has been made and an appropriate pre-operative work-up is complete, an operative plan is constructed with the goal of complete resection with negative margins while preserving as much functional liver parenchyma as possible. This is highly dependent on the distribution of lesions and their proximity to inflow pedicles and outflow veins. Recently, construction of these plans has shifted from the operating room to pre-operative multidisciplinary meetings where films are presented and radiologists help identify the location of tumors and anatomic structures. Medical oncologists and interventional radiologists are often present to discuss potential strategies of downsizing tumors pre-operatively and occasionally planning for treatment of small volume intra or extra-hepatic disease with adjunct therapies such as ablation or radiation.
When considering extent of resection, one must consider if limited segmental resections can preserve liver tissue while ensuring a microscopically negative margin. In the past we relied more heavily on hemi-hepatectomy and trisectionectomy led in large part by data suggesting a higher positive margin rate with wedge resections . The pendulum, however, has shifted back to more limited resections as attempts are made at treating a greater number and broader distribution of tumors and newer series have demonstrated the oncologic safety of more focused resections [27-29].
It has become increasingly recognized that much of the progress made in decreasing the morbidity, mortality, and blood loss associated with hepatic resection is due to maintenance of low central venous pressure (CVP) during hepatic transection. As such, one of the most important pre-operative plans is the anesthetic management which should limit the pre-transection fluid administration to a minimum and have medications ready if a high CVP is encountered during the operation. Previously, we relied on central venous catheters to measure CVP. We routinely use visual inspection of the inferior vena cava (IVC) intraoperatively and identify respiratory variability and distention to estimate appropriate CVP. A sentinel study by Melendez et al.  from Memorial Sloan-Kettering Cancer Center identified a dramatic decrease in operative blood loss, morbidity, and mortality when liver resection is performed under low CVP conditions. Importantly, this study did not reveal any increase in post-operative renal insufficiency. Because of the importance of low CVP, good communication between the anesthesiologist and the surgeon is essential to minimize blood loss during resection.
Open exploration of colorectal metastases should begin with a midline laparotomy or subcostal incision large enough to allow for access to the entire liver irrespective of the planned resection. This ensures adequate manual palpation and ultrasound to locate occult lesions. Incision choice must also take into account any planned combined procedure such as colectomy for synchronous disease. The issue of laparoscopic exploration prior to formal laparotomy has been addressed in previous trials and the yield found to be quite low [31, 32]. Once the abdominal cavity is entered, portal lymph nodes and peritoneal surfaces including the diaphragm and deep pelvis should be inspected as tumors here are often missed on pre-operative cross sectional imaging. Once extra-hepatic disease has been excluded, focus is turned to the liver and surrounding structures. After division of the ligamentum teres and falciform ligament, hepatic exploration begins with bi-manual palpation as well as intraoperative ultrasound (IOUS) to identify deeper lesions, anatomic structures, and variants. It is imperative that all liver surgeons be well versed in intraoperative hepatic ultrasonography. Once a complete assessment of disease burden and location is completed, preparation for hepatic transection can occur. Control of the hepatic inflow at the hilum should be obtained in all liver resections so that a Pringle maneuver can be performed if necessary. If a hemi-hepatectomy or tri-sectionectomy is planned, control of the hepatic venous outflow should be considered, particularly if tumors are in close proximity to their origin. For major resections inflow and outflow are typically controlled prior to hepatic transection.
After preparations have been made, the line of transection is marked and liver stitches used to provide traction and open the transection plane. There have been many techniques described to divide the hepatic parenchyma including crush/clamp, ultrasonic vibration, water jets, and radiofrequency ablation. The goal of all techniques is to quickly and safely divide the parenchyma while minimizing damage to inflow and outflow structures from the parenchyma to be left behind. Transection of the liver should be performed as a careful dissection with division of relevant vascular structures rather than blind coagulation. Frequent reapplication of the ultrasound probe to confirm the trajectory of resection and ensure adequate margin and safe distance from remnant pedicles and outflow is prudent. Following transection, the specimen should be inspected for adequate margin and the cut surface treated with the argon beam coagulator and/or hemostatic agents. Following completion of transection and assurance of hemostasis and biliostasis, patients are rehydrated to a physiologic CVP. Adequate hydration should take place prior to closure of the abdomen as the rise in CVP may occasionally identify bleeding from the hepatic veins that were hidden do to the low pressure.
There is one scenario where disease is intentionally left behind at the end of operation which is the planned two-stage resection. At a time when bilobar disease was largely felt to be unresectable, this two-stage approach was popularized by Adam et al. in the 1990s . Treatment strategies were developed where one hemiliver was cleared of disease at an initial operation with parenchymal sparing wedge resections allowing preservation of the major inflow and outflow structures to the FLR. Portal vein embolization (PVE) or ligation is then typically used to induce hypertrophy of the FLR. Patients are restaged and taken back to the operating room at a minimum of 6–8 weeks later and the contralateral lobe resected. In their initial study, authors identified survival rates similar to patients with unilobar disease ushering in the era of two-stage hepatectomy. Since that time, multiple series have reported on the safety and efficacy of this strategy [34-36].
PVE or portal vein ligation are helpful strategies when a proposed one or two stage resection would leave a remnant liver too small for regeneration and adequate function. It has long been recognized that occlusion of portal flow to a hepatic lobe leads to compensatory hypertrophy of the contralateral lobe [37, 38]. It was first used preoperatively in humans in the mid 1980s but has become commonplace as more large volume liver resections in diseased livers are attempted . Because it can be done safely and leads to reliable growth of the FLR, pre-operative PVE has become an important tool in liver surgery . While multiple techniques exist, PVE is most often performed via percutaneous transhepatic access . Most series report a 30–50% relative increase at 6 weeks although this varies based on the overall health of the patient and the liver [42, 43]. Although liver hypertrophy may continue for up to 6 months, most plateau at around 6–8 weeks which is typically when the planned resection is performed. A recent study by Vauthey et al. has suggested that the degree and rate of hypertrophy can be used to predict post-operative liver failure .
In an effort to accentuate FLR hypertrophy, recent strategies have aimed at obstructing flow to the lobar portal vein and in addition, intrahepatic collaterals using associating liver partition and portal vein ligation for staged hepatectomy (ALPPS) [45, 46]. At an initial operation, portal vein ligation and hepatic transection is performed leaving the arterial inflow, biliary drainage, and hepatic venous outflow intact. After 7–10 days a reoperation is performed and the inflow and outflow divided and specimen removed. Initial reports suggest a 74% hypertrophy of the FLR in 1 week, although very little data exist and mortality rates in initial reports are concerning [47, 48].
In the immediate post-operative period, patients are monitored for evidence of bleeding and adequate pain control. Post-operative hepatic function is assessed by measurement of synthetic ability and clearance using INR and serum bilirubin, respectively. A trend of increased INR and bilirubin over the first few post-operative days should trigger suspicion of post-operative liver failure and supportive measures taken such as phosphorous repletion and avoidance of hypotension [49, 50]. Imaging with duplex ultrasound or contrast-enhanced CT should be considered to rule out biliary and inflow or outflow vascular issues that may be correctable.
One potential actionable cause of post-operative liver failure is prevention and early recognition of infection. The systemic inflammatory response triggered by infection impairs the ability of the liver to regenerate and can potentiate failure. It is imperative to identify the infectious source early and treat as necessary. Cross-sectional imaging should be considered and all suspicious intraabdominal fluid collections should be promptly drained. When an elevated bilirubin is identified (particularly in the absence of elevated INR) biliary leak should be suspected and appropriate imaging and drainage performed. Occasionally there is no definable cause of liver failure and although rare this can be a lethal complication associated with major hepatic resection.
Other complications which may occur following hepatic resection include pleural effusion, wound breakdown, or the development of ascites . The latter can range from asymptomatic to large volume ascites which may result in early satiety, wound leakage, and protein wasting. Prompt identification and treatment with diuretics and aldosterone antagonists can help minimize the sequelae of this condition.
Considerable controversy exists regarding the incidence and prophylaxis of venous thromboembolism (VTE) following hepatic resection. It has long been thought that the coagulopathy following liver resection was protective against VTE and that risk of hemorrhage from prophylaxis makes its use relatively contraindicated. Recently reviews using single institution and National Surgical Quality Improvement Program (NSQIP) databases have shown the risk of VTE is similar to that of other major abdominal operations and that prophylaxis should be considered [52, 53].
Hepatic Artery Infusion Chemotherapy
Because the liver is a common site of recurrence after metastasectomy, we have advocated use of regionally delivered chemotherapy as adjuvant therapy via a hepatic artery infusion (HAI) pump. Placed at the time of operation into the gastroduodenal artery, chemotherapy (typically Fluorodeoxyuridine; FUdR) is delivered continuously to the functional liver remnant beginning approximately 4 weeks after surgery. High first pass metabolism ensures low systemic exposure and therefore minimal toxicity allowing concurrent administration of systemic chemotherapy . Complications such as biliary toxicity are uncommon and typically managed with dose reduction and steroids . A randomized controlled trial of adjuvant HAI therapy and systemic therapy versus systemic therapy alone demonstrated improved hepatic relapse free survival and overall survival in the combined therapy patients [56-58]. Regional therapy has also shown efficacy in unresectable patients with liver only or dominant disease  with high rates of down staging and conversion to resectability . Despite these promising results, HAI therapy has yet to gain wide acceptance, likely due to the complexity in overall management.
Adjuvant Systemic Chemotherapy
Considerable controversy exists regarding the role of adjuvant systemic chemotherapy following complete resection of metastatic colon cancer. The first attempt to shed light on this topic was the European multicenter randomized trial FFCD ACHBTH AURC 9002 which evaluated surgery and adjuvant 5-FU versus surgery alone . Despite closing early due to poor accrual, authors identified a trend towards improved overall and recurrence free survival in the adjuvant therapy arm. To improve the sample size and power, Mitry el al.  pooled results of this trial with a similar study (ENG (EORTC/NCIC CTG/GIVIO)). With a total of 278 patients in the combined analysis, authors showed a 9- and 15-month improvement in progression free and overall survival, respectively. Despite the larger sample sizes, they were still unable to reach statistical significance leaving considerable doubt about the efficacy of adjuvant chemotherapy. In 2008, Nordlinger et al.  published the results of European Organization for Research and Treatment of Cancer (EORTC) 40,983 in which 364 patients were randomized to receive surgery + perioperative (pre and post-operative) 5-FU, Leucovorin and Oxaliplatin or surgery alone. The primary endpoint was improved progression free survival which at 3 years was 7.1% higher in the chemotherapy group although not statistically different (P = 0.058). There was however a significant increase in post-operative complications in those receiving chemotherapy including intra-abdominal infection (7% vs. 2%) and biliary fistula (8% vs. 4%). The secondary endpoint of the trial was overall survival and was reported 4 years later . At a median follow-up of 8.5 years, there was no significant difference in overall survival between the groups.
Improvements in anesthetic and surgical techniques have made liver surgery for colorectal metastases increasingly safe. While no randomized controlled trials have been performed, countless series have demonstrated survival and cure rates unmatched by modern chemotherapy. While outcomes have improved, the risks of liver surgery such as bleeding, infection, and liver failure remain and these operations should be done in experienced centers with appropriate ancillary support. Although commonly used, perioperative chemotherapy continues to be controversial and while reasonable to consider, its efficacy is largely unproven.
- 23Pre-operative imaging for hepatic resection of colorectal cancer metastasis. J Gastrointest Oncol 2012; 3:11–18. PubMed PMID: 22811865. Pubmed Central PMCID: 3397633. Epub 2012/07/20.eng., ,
- 25Diagnostic imaging of colorectal liver metastases with CT, MR imaging, FDG PET, and/or FDG PET/CT: A meta-analysis of prospective studies including patients who have not previously undergone treatment. Radiology 2010; 257:674–684. PubMed PMID: 20829538. Epub 2010/09/11.eng., ,
- 28Increased use of parenchymal-sparing surgery for bilateral liver metastases from colorectal cancer is associated with improved mortality without change in oncologic outcome: Trends in treatment over time in 440 patients. Ann Surg 2008; 247:109–117. PubMed PMID: 18156930. Epub 2007/12/25.eng., , , et al.:
- 30Perioperative outcomes of major hepatic resections under low central venous pressure anesthesia: Blood loss, blood transfusion, and the risk of post-operative renal dysfunction. J Am Coll Surg 1998; 187:620–625. PubMed PMID: 9849736. Epub 1998/12/16.eng., , , et al.:
- 34A two-stage hepatectomy procedure combined with portal vein embolization to achieve curative resection for initially unresectable multiple and bilobar colorectal liver metastases. Ann Surg 2004; 240:1037–1049; discussion 49–51. PubMed PMID: 15570209. Pubmed Central PMCID: 1356519. Epub 2004/12/01.eng., , , et al.:
- 44Kinetic growth rate after portal vein embolization predicts posthepatectomy outcomes: Toward zero liver-related mortality in patients with colorectal liver metastases and small future liver remnant. J Am Coll Surg 2013; 216:201–209. PubMed PMID: 23219349. Epub 2012/12/12.eng., , , et al.:
- 45Rapid improvement in liver volume induced by portal vein ligation and staged hepatectomy: The ALPPS procedure. HPB (Oxford) 2012; 14:874. PubMed PMID: 23134191. Pubmed Central PMCID: 352 1918. Epub 2012/11/09.eng.
- 48Right portal vein ligation combined with in situ splitting induces rapid left lateral liver lobe hypertrophy enabling 2-staged extended right hepatic resection in small-for-size settings. Ann Surg 2012; 255:405–414. PubMed PMID: 22330038. Epub 2012/02/15.eng., , , et al.:
- 52Risk of venous thromboembolism outweighs post-hepatectomy bleeding complications: Analysis of 5651 National Surgical Quality Improvement Program patients. HPB (Oxford) 2012; 14:506–513. PubMed PMID: 22762398. Pubmed Central PMCID: 3406347., , , et al.:
- 56Phase I/II study of hepatic arterial therapy with floxuridine and dexamethasone in combination with intravenous irinotecan as adjuvant treatment after resection of hepatic metastases from colorectal cancer. J Clin Oncol 2003; 21:3303–3309. PubMed PMID: 12947066. Epub 2003/08/30.eng., , , et al.:
- 63Perioperative chemotherapy with FOLFOX4 and surgery versus surgery alone for resectable liver metastases from colorectal cancer (EORTC Intergroup trial 40983): A randomised controlled trial. Lancet 2008; 371:1007–1016. PubMed PMID: 18358928. Pubmed Central PMCID: 2277487., , , et al.: