Nomograms Predicting Survival, Recurrence and Beneficiary Identification of Adjuvant Chemotherapy in Treatment-naïve...

To create and validate nomograms predicting overall survival and recurrence in treatment-naïve rectal cancer (RC) patients who underwent upfront surgery.

Although multidisciplinary treatment is standard for locally advanced RC, understanding surgical efficacy is important for determining indications for perioperative adjuvant therapy.

RC patients who underwent upfront surgery at the Japanese Society for Cancer of the Colon and Rectum institutions were analyzed. A training cohort (n = 1925) of treatment-naïve patients who underwent surgery between 2007 and 2008 was analyzed to construct prognostic models predicting postoperative survival and recurrence. Discrimination and calibration were performed using an external validation cohort (n = 2957; Japanese colorectal cancer registry, procedures in 2005-2006). Effects of adjuvant chemotherapy on survival were evaluated based on nomogram prediction and Surveillance, Epidemiology, and End Results (SEER) data (n = 10,482; upfront surgery for RC in 2010-2015).

In the training cohort, age predicted survival, venous invasion predicted recurrence, and sex, tumor location, histological type, preoperative carcinoembryonic antigen, invasion depth, lymphatic invasion, positive radial margin, and numbers of metastatic nodes and examined nodes predicted both. Internal and external validated Harrell's C-index values were respectively 0.77 and 0.75 for survival and 0.75 and 0.74 for recurrence. RC patients who underwent upfront surgery in the SEER database were stratified into 3 risk levels by nomogram score. Adjuvant chemotherapy did not improve 5-year survival in low-risk patients, but did so for middle-risk (62.4% vs 76.8%) and high-risk (39.4% vs 63.5%) patients.

These nomograms could predict survival and recurrence after upfront curative resection of RC and identify cases expected to benefit more from adjuvant chemotherapy.