A Decade of Prospective SBRT Protocols For Prostate: Dose Response Models of Late Effects For The Emulating HDR Study

Abstract

As long-term survival post-treatment for localized prostate cancer is excellent, analysis of late effects from radiation is essential. We present late effect dose response models from the first prospective four fraction prostate SBRT study (IJROBP 2008 Apr 1;70(5):1588-97). All treated patients received 38 Gy in 4 fractions (BED_1.5 = 279 Gy, linear quadratic model, alpha/beta = 1.5 Gy), with “HDR-like” intraprostatic dose escalation (Dmax > = 150% of prescribed dose), and with commensurate HDR-like limitations of urethra, rectal wall, rectal mucosa and bladder wall doses. The first 100 patients had a median follow-up of 7 years (range: 0.75 – 10.19 years), and are the subject of this analysis. Logistic dose response models were examined with maximum likelihood parameter fitting, for bladder, urethra, rectum, rectal mucosa, small bowel, sigmoid colon. Relationship of late effect to treatment was scored from "definitely not related” (1), "probably not related” (2), "possibly related” (3), "probably related” (4), "definitely related” (5), and "unknown” (6). All late effects that occurred 6 months or later following treatment, and rated (3) "probably related” and higher were included in the model. Toxicities were graded per CTCAE version 3.0. Two grade 3 GU late toxicity effects were identified and included in the analysis, one with bladder Dmax 1.7 Gy below the protocol limit of 45.6 Gy, and one with bladder Dmax 1 Gy per day above the limit. The dose response model did not reach statistical significance due to the paucity of complications, but did estimate the risk at the 45.6 Gy protocol limit to be 2%. All 100 patients met the bladder D10% = 41.8 Gy protocol limit, and the two cases with late effects had D10% of 31.7 and 33.2 Gy. Regarding the urethra, one of the grade 3 late effects had a Dmax 0.7 Gy below the protocol limit of 45.6 Gy, the other 1.4 Gy over the limit. Grade 3 GU dose response was steep at the upper end of the range, with estimated 8% risk at urethra Dmax = 45.6 Gy. None of the late effects occurred when urethral Dmax was below 44 Gy, estimated risk = 1.3% at < = 44Gy urethra Dmax. Therefore, it may be possible to reduce the GU complication rate further by limiting the urethral Dmax constraint to 44 Gy. Volume effects including equivalent uniform dose are also presented. There was no grade 3 or higher late effects that were probably related to treatment for: rectum, rectal mucosa, small bowel, sigmoid colon. As this prospective prostate SBRT protocol matures, with 7-year median follow-up in this cohort, late effect rates remain acceptably low. There is no specific DVH bladder correlate identified for the 2% of patients that experienced late grade 3 GU toxicity. The model does predict significantly increased risk at and above the protocol urethra Dmax limit of > = 45.6Gy, with Dmax < = 44 Gy suggested as a safer urethra Dmax limit. There are zero grade 3 GI toxicity events, thus no dosimetry model for that domain.