Session Item

Data from phase III prospective trials comparing more conventional RT and SBRT are awaited while the use of SBRT is increasing by extrapolating the results of HDR Brachytherapy. Toxicity and QOL remain the major factors to determine a treatment modality in prostate cancer especially in older patients since until now overall survival was not improved with moderate or extreme hypofractionnation. Our preliminary results and methodology using the binomial law to monitor the safety during such a trial are presented.

150 patients > 70 years were prospectively and consecutively included in a phase II trial from 4/2014-7/2018 (NCT03235557). Low (CAPRA 0-2) and Intermediate (CAPRA 3-5)  risk patients received 5 fractions of 7.25 Gy (36.25 Gy), and High risk (CAPRA 6-10) 5 fractions of 7.5 Gy (37.5 Gy) prescribed on the 80% isodose over 9 days. RT was delivered with a Cyberknife M6 while the tracking was based on 2 pairs of linked fiducial markers in the prostate. 1.5-T MRI was performed at diagnosis for contouring and at 3-year follow up to confirm the resolution of the initial nodules. Toxicities were scored using the CTCAEv4 and quality of life (QOL) with IPSS and IEFF5 scores. Biochemical control follows the Phoenix definition. The preliminary results of the first 73 patients with a median follow-up of 36 months are presented. According to the binomial law with an assumption of <5% risk of severe toxicity, 73 patients is the minimum number of patients with no severe toxicity required to demonstrate the safety of this technique in a larger amount of patients with a 95% confidence interval.

No early nor late grade > 3 toxicity was observed. 19.2% of patients complained of late grade 1-2 toxicity (2 grade 2 (2,7%) proctitis, 1 grade 2 (1,4%) dysuria). Mean urinary QOL at baseline and at the end of follow up were 1.23 ± 1.12 and 0.83 ± 0.94 respectively (p>0.05). Mean IPSS at baseline and at the end of follow up were 4.26 ± 4.35 and 3.53 ± 2.72 (p>0.05). 3-year biochemical control were respectively 100%, 92.5% and 54.5% considering low, intermediate and  high risk prognosis. 5 patients suffered from clinical recurrence with 2 local, 4 nodal and 2 metastatic recurrences. MRI complete response at 3 years was 96.7% (29/30) for the 30 first patients with a complete 3-year-follow-up. As no severe toxicity was observed at 36 months in this population, according to the binomial law, we conclude to the safety of our method at a 95% confidence interval (p<0.05).

Preliminary results are encouraging. SBRT as a short treatment offers a safe and convenient treatment modality, while reducing costs, and could also challenge watchful waiting policies. Using the binomial law in this trial allows us to demonstrate the safety of prostatic SBRT based on preliminary results. We propose this statistical approach to boost clinical implementation of new technologies in radiation treatments while offering the best safety to patients in the ongoing trials.