In patients with colorectal liver metastases, can we still rely on number to define treatment and outcome?

We congratulate Dr. Weiser and colleagues for an excellent and concise review of the approaches available for colorectal cancer patients with a limited number of metastases, defined as “oligometastatic disease.”[1] This commentary provides us with the opportunity to revisit the concept of “oligometastatic disease” that was developed in the era predating the use of effective chemotherapy, when surgery predominantly defined patient outcomes. At that time, a small number of metastases was considered a sign of favorable biology. Today up to 25 % of patients with metastatic colorectal cancer are candidates for resection of colorectal liver metastases (CLM). Patients with CLM encompass a broad spectrum of clinical presentations, and new concepts of prognostication are emerging. To achieve optimal outcomes, patients with CLM should undergo a planned sequence of therapeutic interventions that include perioperative chemotherapy and surgery. The surgeon and the medical oncologist are pivotal in keeping the patient on a therapeutic track that will optimize the benefits derived from chemotherapy and surgery. In this commentary, we provide approaches and perspectives that represent alternatives to the important concepts discussed by Weiser et al. We have organized our remarks around five key questions raised by the authors. 1. Should the number of lesions still be included in the criteria used to select patients for surgical resection? In 2003, Altendorf-Hofmann et al demonstrated no survival differences between patients who underwent R0 resection for 1 to 3, 4, 5 to 7, or 8 to 11 metastases (Figure 1).[2] In light of these data, we think that response to modern chemotherapy may help determine the biology of the disease better than the number of lesions. In recent studies, pathologic response to chemotherapy predicted outcome and outperformed the traditional predictors of outcome, including the number of metastases.[3]These findings may explain why the resection of multiple CLM was associated with more favorable long-term outcomes in previous studies. Because the number of lesions does not necessarily correlate with the biology of the disease, we use high-quality CT scans to assess the radiologic response to chemotherapy—which correlates with pathologic response.[4] The specific radiologic criteria we use go beyond the Response Evaluation Criteria In Solid Tumors and include morphologic changes in the CLM (eg, homogeneous loss of enhancement with well-defined margins) as a size-independent predictor of outcome.[5] In addition, we have shown that RAS mutations (KRAS and NRAS) predict poorer overall and disease-free survival, as well as a pattern of early lung recurrence in patients undergoing resection for CLM.[6,7] On multivariate analysis, RAS mutational status outperformed traditional predictors of outcome, such as the number and size of CLM. Because RAS mutational status is highly concordant (> 90%) between the primary tumor and the CLM, this prognostic information can be obtained early in the evaluation of the disease via pretreatment biopsy or from the resected primary tumor. 2. Does a small number of lesions define the extent of resection and the approach? One or two lesions may call for limited liver surgery in some patients, but if unfavorably located (eg, involvement of two hepatic veins), these few lesions may necessitate extensive resection. This could have a direct impact on decisions regarding the type and sequence of surgical procedures or the use of perioperative therapy. In patients with unfavorably located lesions, preoperative portal vein embolization may be useful prior to surgery and should be integrated into the multidisciplinary