The relationship between tumor radiological (RX) response and long-term outcome is unclear for current drugs. Early response is a proposed efficacy end point for drug screening and on-treatment evaluation and warrants investigation as a predictor of long-term outcomes. We previously showed early tumor shrinkage (ETS) in chemotherapy + cetuximab regimens identified KRAS wild-type (wt) mCRC patients with improved outcome. We used the CRYSTAL and OPUS studies to quantify the predictive power of tumor size changes at week 8 for long-term outcome, investigating heterogeneity across treatment arms.Methods
Tumor KRAS status and RX data at week 8 were available in patients from the CRYSTAL (n = 820) and OPUS (n = 297) studies. Using investigators' assessments, a relative change in the sum of the longest diameters of target lesions versus baseline was calculated. Time-dependent receiver-operating characteristics (TDROC) were used to obtain Cτindices. Cox regression models and Subpopulation Treatment Effect Pattern Plots (STEPP) identified interaction between ETS and treatment. Outcome measures were progression-free survival (PFS) and overall survival (OS).Results
In both studies, TDROC indicated a strong relationship between change in tumor size at week 8 and PFS in KRAS wild-type mCRC patients receiving chemotherapy + cetuximab (CRYSTAL Cτ = 0.688, OPUS Cτ = 0.703). This was weaker in the chemotherapy arm (CRYSTAL Cτ = 0.604, OPUS Cτ = 0.547). Data were similar for OS in the chemotherapy + cetuximab (CRYSTAL Cτ = 0.606, OPUS, Cτ = 0.655) and chemotherapy arms (CRYSTAL Cτ = 0.553, OPUS Cτ = 0.566) but with lower Cτ indices. Treatment interaction was confirmed by significant interaction terms for PFS in the Cox regression models (CRYSTAL P = 0.020, OPUS P = 0.004) and for both PFS (CRYSTAL P = 0.048, OPUS P = 0.053) and OS (CRYSTAL P = 0.001, OPUS P = 0.036), with STEPP analysis.Conclusions
These findings show a strong relationship between ETS and long-term outcome in KRAS wild-type mCRC patients receiving chemotherapy + cetuximab versus chemotherapy alone. This suggests a possible use for ETS as a predictor of outcome in this setting. Further studies are needed to see how this impacts on clinical decision-making.