A case-matched study of toxicity outcomes after proton therapy and intensity-modulated radiation therapy for prostate cancer.

BACKGROUND: The authors assessed whether proton beam therapy (PBT) for prostate cancer (PCa) was associated with differing toxicity compared with intensity-modulated radiation therapy (IMRT) using case-matched analysis.

METHODS: From 2010 to 2012, 394 patients who had localized PCa received 79.2 Gray (Gy) relative biologic effectiveness (RBE) delivered with either PBT (181 patients) or IMRT (213 patients). Patients were case-matched on risk group, age, and prior gastrointestinal (GI) and genitourinary (GU) disorders, resulting in 94 matched pairs. Both exact matching (risk group) and nearest-neighbor matching (age, prior GI/GU disorders) were used. Residual confounding was adjusted for by using multivariable regression. Maximum acute and late GI/GU Common Terminology Criteria for Adverse Events-graded toxicities were compared using univariate and multivariable logistic and Cox regression models, respectively.

RESULTS: Bladder and rectum dosimetry variables were significantly lower for PBT versus IMRT (P ≤ .01). The median follow-up was 47 months (range, 5-65 months) for patients who received IMRT and 29 months (range, 5-50 months) for those who received PBT. On multivariable analysis, which exploited case matching and included direct adjustment for confounders and independent predictors, there were no statistically significant differences between IMRT and PBT in the risk of grade ≥ 2 acute GI toxicity (odds ratio, 0.27; 95% confidence interval [CI], 0.06-1.24; P = .09), grade ≥ 2 acute GU toxicity (odds ratio, 0.69; 95% CI, 0.32-1.51; P = .36), grade ≥ 2 late GU toxicity (hazard ratio, 0.56; 95% CI, 0.22-1.41; P = .22), and grade ≥ 2 late GI toxicity (hazard ratio, 1.24; 95% CI, 0.53-2.94; P = .62).

CONCLUSIONS: In this matched comparison of prospectively collected toxicity data on patients with PCa who received treatment with contemporary IMRT and PBT techniques and similar dose-fractionation schedules, the risks of acute and late GI/GU toxicities did not differ significantly after adjustment for confounders and predictive factors.