Colorectal hepatic metastases: adjuvant chemotherapy and survival.

“Pathologic Response to Preoperative Chemotherapy: A New Outcome End Point After Resection of Hepatic Colorectal Metastases,” with great interest. This study confirmed that pathologic response to preoperative chemotherapy predicts survival for patients with colorectal hepatic metastases, establishing pathologic response as a prognostic indicator and as a potential end point in clinical trials evaluating neoadjuvant chemotherapy. However, we are concerned by the omission of adjuvant chemotherapy in the survival analysis. Patients with colorectal hepatic metastases are frequently treated with both pre- and postoperative chemotherapy, an approach that is supported by the European Organisation for Research and Treatment of Cancer 40983 trial. In this study, 364 patients with resectable colorectal hepatic metastases were randomized to six cycles of neoadjuvant infusional fluorouracil/leucovorin (FOLFOX4) plus six cycles of adjuvant FOLFOX4 versus surgery alone, resulting in a 9.2% improvement in 3-year diseasefree survival for the group that received perioperative chemotherapy and were resected (P .03). 2 Unfortunately, we do not know if both neoadjuvant and adjuvant approaches are necessary, and to date no clinical trial has addressed this issue. In the study by Blazer et al, 1 197 (64.6%) of 305 patients received adjuvant chemotherapy after hepatic resection, with a median time to treatment of 7 weeks and a median duration of adjuvant chemotherapy of 12 weeks. However, adjuvant chemotherapy is not included as a variable in the univariate or multivariate models reported in the article. One of the inherent flaws of constructing a prognostic model from retrospective data is that it is difficult to include all known clinical predictive markers, but the omission of adjuvant chemotherapy from the survival analysis is important. Its inclusion could alter the results. The benefit of adjuvant chemotherapy after resection of colorectal hepatic metastases has been reported. In a multivariate analysis, adjuvant chemotherapy and size of metastases were significantly associated with overall survival (OS; hazard ratio [HR] for surgery alone 1.39; 95% CI, 1.00 to 1.93; P .046; HR for two or more metastases 1.49; 95% CI, 1.06 to 2.11; P .023). 3 A retrospective review of 792 patients added further support for the importance of adjuvant chemotherapy. 4 Even for poor-risk patients (four or more bilobar metastases) adjuvant chemotherapy improved OS compared with surgery alone (51.5 months v 23 months; P .01), 5 and for patients with synchronous disease, treatment with postoperative fluorouracil or hepatic arterial infusion floxuridine-based therapy demonstrated improved OS compared with surgery alone (HR 0.62; 95% CI, 0.50 to 0.78; HR 0.51; 95% CI, 0.28 to 0.97, respectively). 6 Three randomized studies showed an increase in disease-free survival with the use of hepatic arterial infusion and systemic therapy after liver resection. 7-9 In our opinion, this data mandates the inclusion of adjuvant chemotherapy in any multivariate model investigating survival. We recently reported that radiologic response to neoadjvuant chemotherapy did not correlate with OS in patients with operable synchronous colorectal hepatic metastases. 10 In our univariate analysis, response to neoadjuvant chemotherapy was not associated with OS, and in the multivariate Cox model, only margins, stage of primary and postoperative carcinoembryonic antigen 5 ng/dL were significant (P .04, P .03, and P .01, respectively). For patients who progressed on neoadjuvant chemotherapy, there was a trend towards improved 5-year OS for those who received adjuvant hepatic arterial infusion chemotherapy. An association between pathological response to neoadjuvant chemotherapy and survival is not a new concept. It is well established, for example, in the management of muscle-invasive urothelial carcinoma, and it is likely that the association reported in this study is true. 11 However, we believe that a survival analysis must consider the impact of adjuvant chemotherapy and are interested to learn if its inclusion in the univariate and multivariate models in the Blazer et al 1 study changes the results. It would also be of clinical interest to test if duration of chemotherapy was related to pathologic response, and if predictors of complete response such as burden of disease can be identified. We would also welcome a comment from the authors on why different cutoff points were used for carcinoembryonic antigen (200 ng/mL v 5 ng/mL), and tumor size (5 cm v 3 cm) and number (multiple/solitary v 3/ 4) in the two analyses. There was a high death rate (8%) in the postoperative period. Was there an association between type of neoadjuvant chemotherapy used and postoperative death? Lastly, as we saw in our study, some patients who progress on neoadjuvant chemotherapy appear to be responsive with adjuvant salvage treatment. It would be interesting to know how many of the 149 patients (55%) with minor response or progression achieved long-term survival, and if the use of adjuvant chemotherapy contributed to this.