Prediction of R0/R+ surgery by different classifications for locally recurrent rectal cancer.

A widely adopted classification system for locally recurrent rectal cancer (LRRC) is currently missing, and the indication for surgery is not standardized. To evaluate all the published classification systems in a large monocentric cohort of LRRC patients, assessing their capability to predict a radical (R0) resection. A total of 152 consecutive LRRC patients treated at the National Cancer Institute of Milan (NCIM) from 2009 to 2017 were classified according to Pilipshen, Mayo Clinic, Memorial Sloan-Kettering Cancer Center (MSKCC), Wanebo, Yamada, Boyle, Dutch TME Trial, Royal Marsden and National Cancer Institute of Milan (NCIM) classification systems. Central location of LRRC was significantly predictive of R0 resection across all classification systems. R + resection was predicted by the "anterior" category of MSKCC (OR 2.66, p = 0.007), the "S2b" (OR 3.50, p = 0.04) and the "S3" (OR 2.70, p = 0.01) categories of NCIM, "pelvic disease through anastomosis" of Pilipshen (OR 2.89, p = 0.002), "fixed at 2 sites" of Mayo Clinic (OR 2.68, p = 0.019), and "TR4" of Wanebo (OR 3.39, p = 0.002). The NCIM was the most predictive classification for R0 surgery. The NCIM classification seems to be superior among the others in predicting R0 surgery. Generally, lateral invasive and high sacral invasive relapses are associated with reduced probability of R0 surgery and unfavorable outcomes.