Additional Biomarkers beyond RAS That Impact the Efficacy of Cetuximab plus Chemotherapy in mCRC: A Retrospective Biomarker Analysis

Figure 2

The predictive model for objective response. (a) The predictive model is calculated as the sum of treatment and gene mutation predictor values weighted by their regression coefficients. (b) ORR for patients treated with cetuximab plus chemotherapy in training cohort was improved by RAS testing and predictive model. The green bars represent responders (CR + PR); the red bars nonresponders (SD + PD). The size of the bars is in agreement with the corresponding numbers of patients. (c) ORR for patients treated with cetuximab plus chemotherapy in validation cohort was improved by RAS testing and predictive model. (d) ROC curve for the predictive model in training cohort (AUC=0.762, 95%CI 0.658-0.867, P<0.001); (e) ROC curve for the predictive model in validation cohort (AUC=0.751, 95%CI 0.661-0.841, p<0.001). ORR, objective response rate. ROC, receiver operating characteristics.

(a)

(b)

(c)

(d)

(e)