Predicting Grade 3 Acute Diarrhea During Radiation Therapy for Rectal Cancer Using a Cutoff-Dose Logistic Regression Normal Tissue Complication Probability Model

Purpose Understanding the dose–volume relationship of small bowel irradiation and severe acute diarrhea may help reduce the incidence of this side effect during adjuvant treatment for rectal cancer. Methods and Materials Consecutive patients treated curatively for rectal cancer were reviewed, and the maximum grade of acute diarrhea was determined. The small bowel was outlined on the treatment planning CT scan, and a dose–volume histogram was calculated for the initial pelvic treatment (45 Gy). Logistic regression models were fitted for varying cutoff-dose levels from 5 to 45 Gy in 5-Gy increments. The model with the highest LogLikelihood was used to develop a cutoff-dose normal tissue complication probability (NTCP) model. Results There were a total of 152 patients (48% preoperative, 47% postoperative, 5% other), predominantly treated prone (95%) with a three-field technique (94%) and a protracted venous infusion of 5-fluorouracil (78%). Acute Grade 3 diarrhea occurred in 21%. The largest LogLikelihood was found for the cutoff-dose logistic regression model with 15 Gy as the cutoff-dose, although the models for 20 Gy and 25 Gy had similar significance. According to this model, highly significant correlations ( p