Use of Early Tumor Shrinkage to Predict Long-Term Outcome in Metastatic Colorectal Cancer Treated With Cetuximab

Abstract

Early tumor shrinkage (ETS) is associated with long-term outcome in patients with chemorefractory metastatic colorectal cancer (mCRC) receiving cetuximab. This association was investigated in the first-line setting in the randomized CRYSTAL and OPUS mCRC trials, after controlling for KRAS tumor mutation status. Radiologic assessments at week 8 were used to calculate the relative change in the sum of the longest diameters of the target lesions. Time-dependent receiver operating characteristics provided Cτ-indices (time-dependent c-index). Cox regression models and subpopulation treatment effect pattern plot analysis investigated associations between ETS (radiologic tumor size decrease at week 8) and survival and progression-free survival (PFS). In both trials, in patients with KRAS wild-type mCRC, Cτ values for PFS and survival were higher (P < .001) in those receiving chemotherapy plus cetuximab versus chemotherapy alone, indicating a stronger predictive value of ETS for long-term outcome in these patients. In the CRYSTAL and OPUS trials, respectively, the cutoff value of ETS ≥ 20% (v < 20%) identified patients with KRAS wild-type mCRC receiving chemotherapy plus cetuximab with longer PFS (medians 14.1 v 7.3 months, hazard ratio [HR] = 0.32; P < .001, and medians 11.9 v 5.7 months, HR = 0.22; P < .001) and survival (medians 30.0 v 18.6 months, HR = 0.53; P < .001 and medians 26.0 v 15.7 months, HR = 0.43; P = .006). ETS was significantly associated with long-term outcome in patients with KRAS wild-type mCRC treated first-line with chemotherapy plus cetuximab. Validation in prospective trials is required to assess the value of this on-treatment marker in the clinical decision-making process.