Major Impact Of Personalized Dosimetry Using 90y Loaded Glass Microspheres Sirt In Hcc: Final Overall Survival Analysis Of A Multicenter Randomized Phase Ii Study (Dosisphere-01)

516 Background: 90Y loaded microsphere SIRT (radioembolization) is a treatment option in advanced HCC. However, no personalized dosimetric endpoints are currently used. The goal of this study was to compare the efficacy of 90Y loaded glass microsphere SIRT in HCC using a standard versus a personalized dosimetric approach. Methods: DOSISPHERE-01 was a multicenter, randomized phase 2 trial in unresectable HCC patients with at least one tumor ≥7cm. Treatment arm was randomly assigned (1:1) to standard dosimetry arm (SDA), with a goal to deliver 120±20Gy to the treated volume or to personalized dosimetry arm (PDA) with a goal to deliver at least 205Gy to the index lesion. The primary endpoint was the response rate (RR) of the index lesion according to EASL criteria. Secondary endpoints included dose response evaluation, safety and overall survival (OS). Results: Sixty HCC patients were randomized (PDA 31, SDA 29, intent to treat population-ITTP-), and 56 treated (28 in each arm). RR was significantly increased in the PDA versus the SDA, in the ITTP, respectively 64.5% versus 31% (p=0.0095) as in the safety population -SP- (treatment effectively received, personalized 35, standard 21), respectively 74.3% versus 14.3% (p<0.0001). Median OS was significantly increased in the PDA versus the SDA, in the ITTP, respectively 26.7m (CI 95%:11.7-NR) versus 10.6m (CI 95%:6-16.8), p=0.0096, HR=0.421 (95%CI:0.215-0.826), p=0.0119, as in the SP, respectively 26.7m (CI 95%:11.7- NR) versus 9.5m (CI 95%:4.8-14.9), p=0.0015, HR=0.342 (95%CI:0.171-0.683), p=0.0023. Median OS was 26.7m (CI 95%:13.5-NR) versus 6.0m (CI 95%:3.8-14.9) for the patients who received a tumor dose ≥205 Gy or <205 Gy respectively, p=0.0106, HR=0.336 (95%CI:0.154-0.735), p=0.0063. Treatment-related clinically relevant hepatic ≥grade 3 AEs were observed in 5.7% and 14.2% of the patients of the PDA and SDA arms, respectively, (p=ns). Conclusions: MAA SPECT/CT based personalized dosimetry is safe and dramatically increased RR and OS of HCC patients. These results question the interpretation of all phase 3 trials of SIRT designed without personalized dosimetry in HCC. Clinical trial information: 2015-A00894-45.