Conversion strategies with chemotherapy plus targeted agents for colorectal cancer liver-only metastases: A systematic review

Background: There is no consensus on the optimal systemic conversion therapy in patients with unresectable colorectal cancer liver-only metastases (CRLM) to achieve a complete resection. Interpretation of trials is complicated by heterogeneity of patients caused by emerging prognostic and predictive characteristics, such as RAS/BRAF mutation status, lack of consensus on unresectability criteria and lack of data on clinical outcome of secondary resections. A systematic review was performed of characteristics of study populations and methodology of trials regarding patients with initially unresectable colorectal cancer liver-only metastases. Methods: Phase II/III randomised trials, published after 2008, regarding first-line systemic conversion therapy in patients or subgroups of patients with CRLM were included. Data on secondary resection outcomes were collected. Results: Overall, 20 trials were included for analysis: seven prospective trials in patients with unresectable CRLM and 13 trials in the overall population of unresectable metastatic colorectal cancer (mCRC) with retrospective subgroup analysis of CRLM patients. Fourteen trials did not provide unresectability criteria at baseline, and criteria differed among the remaining studies. Trials and study populations were heterogeneous in prognostic/predictive factors, use of primary end-points, and reporting on long-term clinical outcomes. R0-resection rates in CRLM patients varied between CRLM studies and mCRC studies, with rates of 22–57% and 11–38%, respectively. Conclusions: Cross-study comparison of (subgroups of) studies regarding first-line systemic treatment in patients with unresectable CRLM is hampered by heterogeneity in study populations, trial designs, use of (K)RAS/BRAF mutational tumour status, and differences/absence of unresectability criteria. No optimal conversion systemic regimen can be selected from available data. Prospective studies with well-defined criteria of these issues are warranted.