Surgical Resection for Colorectal Cancer Liver Metastases: Who, When, How?

A Conversation with Steven A. Curley, MD

Get Permission

The old dogma that we need negative margins to be ≥ 1 cm is not true. We (and others) have shown that a negative margin of only 1 to 2 mm is adequate, and is associated with a very low risk of local recurrence.

—Steven A. Curley, MD

Many patients with colorectal liver metastases can undergo surgical resection with curative intent. Who are these patients and how are they best managed? In an interview with The ASCO Post, Steven A. Curley, MD, Professor of Surgical Oncology at The University of Texas MD Anderson Cancer Center, Houston, offered some answers.

Evolving Approach

How has the thinking about colorectal cancer–related liver metastasis changed in the past decade or so?

The surgical management of colorectal cancer liver metastases has greatly evolved over the course of my career. In 1988, a seminal paper,1 based on the Registry of Hepatic Metastases from 24 institutions, suggested that we not consider resection for any patient with extrahepatic disease (even if deemed resectable), when resection margins would be < 1 cm, with short disease-free intervals, with stage IV disease at the time of diagnosis of the primary tumor, and with more than three metastatic liver lesions.

This created dogma for the time, but we later realized we needed to push the envelope, because many patients have potentially resectable metastases but do not otherwise fit these criteria. For example, I have a patient with eight liver metastases. On imaging, we found that all were in the right lobe of the liver, so technically, these could be resected with a right hepatectomy. If we followed the Registry recommendations, we would not have considered surgery. He is now a 10-year survivor.

So, in 2012, how do you handle the patient with asymptomatic primary colorectal cancer who presents with liver metastasis?

Resection of liver metastasis, usually after neoadjuvant chemotherapy, is the optimal treatment for this patient, and that is always our chief goal. As a surgeon, my first question is whether I can do this operation safely. Can I leave the patient with an adequate volume of functional liver that is well perfused and has good biliary drainage? Today, we do CT volumetry to help answer this question.

This test helps us estimate the volume of the future liver remnant. There is a clear correlation between the volume of liver resection and survival and complication rates. Patients with normal livers can tolerate resections of up to 80% of the organ, but this drops to about 60% to 70% in patients with steatotic livers, and becomes very small for patients with cirrhotic livers, which, thankfully, is not common in patients with colorectal cancer liver metastases.

Impact of Liver Disease

How does underlying liver disease influence outcomes?

Patients with underlying liver disease don’t tolerate resection as well as those without. Clearly, patients with an abnormal liver from, say, steatosis or nonalcoholic steatohepatitis cannot tolerate resection of a large volume. For patients wth a healthy liver, if the future liver remnant is < 20%, the risk for liver-specific complications rises to 50%. If you leave ≥ 20% behind, the complication rate drops to about 20%. Clearly, this is a key factor in surgical planning.

We also are concerned about the blood vessels lying beneath the surface of the organ. With imaging, we want to determine the association of the tumor with these important blood vessels so that we can control for problems during surgery. We now have surgical instruments that can help produce a bloodless surgical field, which has reduced our transfusion rate from over 60% to under 5%.

Radiofrequency Ablation

When do you incorporate radiofrequency ablation into treatment?

Radiofrequency ablation came onto the scene in the 1990s, and it allows us to treat more patients. The MD Anderson database is very informative about the technique, though I emphasize that these are not randomized trial data.

We evaluated 418 patients with liver-only metastases; of these, 348 had a potentially curative procedure, which was resection in 190 patients. For the other 158 patients, a complete resection was not possible, and of this group, 101 had resection plus radiofrequency ablation, whereas 57 had radiofrequency ablation alone. There were also 70 patients deemed unresectable due to extrahepatic disease, and they received no resection or radiofrequency ablation but could receive chemotherapy.

Overall survival at 5 years for the 348 patients treated with curative intent was 44%. When we broke this down, we found that overall survival was best in the patients who received resection only. Overall survival at 3 years was 73% with resection, 43% with resection plus radiofrequency ablation, and 37% with radiofrequency ablation alone. At 5 years, overall survival was 58% with resection but only about 20% in the other arms (P < .0001).2 Recurrence-free survival was also best with resection only.

In the unresectable (for curative intent) group, there also appeared to be some benefit for radiofrequency ablation, alone or with resection, compared to chemotherapy, which resulted in a 3-year overall survival of only 12% and a 5-year overall survival of 7%.

These results have been mirrored in multiple other modern series, all showing about a 58% overall survival rate at 5 years after resection alone for colorectal cancer liver metastases. The data have led surgeons to be much more aggressive about attempting resection. While there may be a role for newer ablation techniques, an advantage over resection has not been show in any randomized trial. For patients with unresectable disease, radiofrequency ablation may have a role in improving overall survival when combined with chemotherapy, but this needs to be confirmed in randomized trials.

Ongoing Challenges

What continues to frustrate you in attempting resection in these patients?

I am frustrated by patients with bilobar tumors that I cannot resect and leave an adequate volume of liver, and patients with one or two tumors that are in locations not amenable to resection with negative margins. We can operate near major blood vessels, but there are patients whose pattern of metastasis is such that we cannot resect them.

Incidentally, the old dogma that we need negative margins to be ≥ 1 cm is not true. We (and others) have shown that a negative margin of only 1 to 2 mm is adequate, and is associated with a very low risk of local recurrence.

Then, does number of metastases no longer matter? Are patients no longer excluded based on number of lesions?

An MD Anderson study of 159 patients with at least four metastatic lesions, who underwent resection or resection plus radiofrequency ablation, found overall survival at 5 years to be 51%,3 suggesting that we should question the dogma that patients with at least five lesions are not surgical candidates. We encourage all patients to be assessed by a surgeon to determine if they are candidates for resection.

Neoadjuvant Chemotherapy

What is the role of neoadjuvant chemotherapy?

Response to neoadjuvant chemotherapy increases the odds of long-term survival 2.5-fold. This can be a response in the form of reduction in tumor volume, increased calcification of lesions on CT scan, or reduction in serum carcinoembryonic antigen levels. Several studies have shown the survival advantage of achieving a pathologic complete response, but 85% of patients with radiographic disappearance of lesions still have viable tumor cells at those metastatic sites. In our own series, major pathologic response (≤ 10% viable cells in the tumor specimen) was an independent predictor of improved disease-free and overall survival, but it is hard to select out this group.

Timing of Resection

What is the time course for treating these patients—that is, how do you time the resection?

The timing of liver resection should be considered by a multidisciplinary team. When a patient has a lesion in a critical location, disease progression while on chemotherapy can render them less resectable. These patients should probably undergo surgery promptly, followed by adjuvant chemotherapy. Other patients may benefit from a neoadjuvant approach, which allows us to evaluate their response and determine whether they are good candidates for resection.

You also have to be cognizant of the effects of systemic chemotherapy on liver quality. Hepatotoxicity has been reported with fluorouracil, irinotecan, and oxaliplatin, which speaks to the need for multidisciplinary management early in the disease course. I often see patients who have had 12 to 18 months of chemotherapy. On imaging, I can see that they once had resectable disease, but now they have irinotecan-related steatohepatitis and the metastases cannot be resected because of treatment-induced liver damage. The chemotherapy-related liver injury has taken them out of the realm of potentially curative surgical therapy.

Role of Biomarkers

Can molecular markers be incorporated into decision-making regarding resection?

Data on tumor biology as it relates to resection outcomes are being collected, but there are no answers yet. We are most interested in identifying the subset of patients who respond to neoadjuvant chemotherapy and undergo resection, yet 6 months later develop new disease. Clearly, in these patients, local therapy did not help. We conduct an ongoing prospective analysis of tissue samples to try to differentiate these patients.

Unanswered Questions

What are the remaining uncertainties about resection of liver metastases?

Unresolved issues include the role of preoperative portal vein embolization to increase the size of the future liver remnant in borderline resectable patients; the benefit of staged resections, which are associated with high morbidity and mortality and should be carefully considered; and timing of the resection in patients with stage IV disease—ie, should we resect the primary or metastasis first? In addition, we need to clarify the assessment of liver damage related to neoadjuvant chemotherapy, the number of cycles and dose of neoadjuvant chemotherapy, the correlation of molecular markers with resectability, and the assessment of liver damage related to comorbid conditions. ■

Disclosure: Dr. Curley reported no potential conflicts of interest.


1. Resection of the liver for colorectal carcinoma metastases: A multi-institutional study of indications for resection. Registry of Hepatic Metastases. Surgery 103:278-288, 1988.

2. Abdalla EK, Vauthey JN, Ellis LM, et al: Recurrence and outcomes following hepatic resection, radiofrequency ablation, and combined resection/ablation for colorectal liver metastases. Ann Surg 239:818-825, 2004.

3. Pawlik TM, Abdalla EK, Ellis LM, et al: Debunking dogma: Surgery for four or more colorectal liver metastases is justified. J Gastrointest Surg 10:240-248, 2006.