CT-based online adaptive radiotherapy improves target coverage and organ at risk (OAR) avoidance in stereotactic body radiation therapy (SBRT) for prostate cancer

INTRODUCTION: Stereotactic body radiation therapy (SBRT) is an emerging treatment modality for clinically localized prostate cancer (PCa). Online daily adaptive radiotherapy (ART) could potentially improve the therapeutic ratio of prostate SBRT by accounting for inter-fraction variation in target and OAR volumes. To our knowledge, no group has evaluated the clinical utility of a novel AI-augmented CT-based ART system for prostate SBRT. In this study we hypothesized that adaptive prostate SBRT plans would result in improved target coverage and lower dose to OARs in comparison to unadapted treatment plans. METHODS: Seven patients with favorable intermediate to oligometastatic PCa treated with 5-fx prostate adaptive SBRT were retrospectively reviewed. Patients were treated with 3625 cGy to the prostate and seminal vesicles. 6 patients additionally received 2500 cGy to the pelvic nodes, 5 patients underwent a boost to 4000 cGy to the prostate. For each fraction, a CBCT was acquired and OARs (rectum, bladder, bowel, sigmoid, femurs) were segmented/deformed using AI. CTVs were rigidly registered. Volumes were adjusted manually and PTV expansions added. Adaptive treatment plans were developed based on the contoured targets and OARs and dose to these volumes for the adapted vs. initial plans were compared for each fraction. V100 and the D0.03 cc between scheduled and adapted treatment plans were compared using a Student’s t-test, with significance threshold of P < 0.05. RESULTS: Seven patients completed 35 Fx’s of adaptive RT. Daily adaptation resulted in a statistically significant mean improvement in PTV V100 for all targets: [21.4 % ± 4.3 % for PTV 4000 (p < 0.0001); 8.7 % ± 1.1 % for PTV 3625 (p < 0.0001); and 11.5 % ± 3.1 % for PTV 2500 (p = 0.0013)]. Mean rectal D0.03 was significantly reduced by 38.8 cGy ± 5.95 cGy (p < 0.0001) per fraction (194 cGy/5 fractions) compared to the initial plans. There was a modest increase in bladder dose of 10.9 cGy ± 4.93 cGy per fraction (p = 0.0424) for the adaptive plans. The adaptive plans met bladder constraints for every fraction. There were no statistically significant differences between sigmoid or bowel dose for adapted vs. initial plans. No patients experienced acute CTCAE grade ≥ 3 GI/GU adverse events (median F/U 9.5 months). All statistically significant differences were maintained in the presence and absence of rectal hydrogel spacer (p < 0.05). CONCLUSIONS: CT-based online adaptive SBRT resulted in statistically significant and clinically meaningful improvements in PTV coverage and D0.03 cc dose to the rectum. A trial evaluating CT adaptive whole-pelvis prostate SBRT is underway.