A clinical parameters-based predictive model to predict early efficacy for rectal cancer patients with neoadjuvant chemotherapy alone: Subgroup efficacy analysis of a phase II study.

Abstract

716 Background: To avoid the damage of radiotherapy for a part of patients, we have an ongoing phase II study comparing Neoadjuvant FOLFOX6 chemotherapy with or without radiation in Rectal Cancer. To define the population who could obtain benefit from neoadjuvant chemotherapy alone, we aimed to develop a model in which clinical parameters are used to predict early efficacy of neoadjuvant chemotherapy for rectal cancer, including tumor regression grade (TRG) and ypTNM down-staging. Methods: Until now, 116 patients were enrolled into FOLFOX6 chemotherapy alone group. All the clinical parameters were collected, including gender, age, tumor location and size by MRI imaging, blood routine test, blood biochemistry and serum tumor markers, TRG and ypTNM staging. 96 patients were used to be training set and another 20 patients was the validation set. Binary logistic regression was used to analyze the effect of all clinical factors on TRG and ypTNM stage. Clinical variables were incorporated into the Cox models for multivariate analysis. Results: Among the 116 patients, 9 patients had TRG 0 (complete regression), 26 patients with TRG 1 and 81 patients with TRG 2 to 3 (poor regression). 47 (40.5%) patients had good down-staging with ypTNM stage 0 and stage I. The optimal multivariable predictive model for TRG (0-1 vs. 2-3) was a four-variable model consisting of age, ApoA1, tumor length and tumor encircle size (AUC=0.732, 95% CI, 0.614–0.849). The model for ypTNM (stage 0-I vs. II-III) was also the four parameters (AUC=0.725, 95% CI, 0.615–0.835). Patients with younger age, shorter tumor length, smaller encircle size and higher ApoA1 had better outcome with chemotherapy. Applying the four clinical parameters model to the validation cohort, a good performance (AUC=0.792, 95% CI, 50.54–0.95) was obtained. Conclusions: With this predictive model, the patients who might obtain good efficacy from neoadjuvant FOLFOX6 chemotherapy could avoid radiotherapy damage. And better suggestion would be provided for patients who might be high risk with chemotherapy alone.