Abstract
Background - All randomised phase 3 studies of selective internal radiation therapy for advanced hepatocellular carcinoma published to date have reported negative results. However, these studies did not use personalised dosimetry. We aimed to compare the efficacy of a personalised versus standard dosimetry approach of selective internal radiation therapy with yttrium-90-loaded glass microspheres in patients with hepatocellular carcinoma. Methods - DOSISPHERE-01 was a randomised, multicentre, open-label phase 2 trial done at four health-care centres in France. Patients were eligible if they were aged 18 years or older and had unresectable locally advanced hepatocellular carcinoma, at least one measurable lesion 7 cm or more in size, a hepatic reserve of at least 30% after selective internal radiation therapy, no extrahepatic spread (other than to the lymph nodes of the hilum, with a lesion <2 cm in size), and no contraindications to selective internal radiation therapy, as assessed by use of a technetium-99m macro-aggregated albumin scan. Patients were randomly assigned (1:1) by use of a permutated block method, with block sizes of four and without stratification, to receive either standard dosimetry (120 ± 20 Gy) targeted to the perfused lobe; standard dosimetry group) or personalised dosimetry (≥205 Gy targeted to the index lesion; personalised dosimetry group). Investigators, patients, and study staff were not masked to treatment. The primary endpoint was the investigator-assessed objective response rate in the index lesion, according to European Association for the Study of the Liver criteria, at 3 months after selective internal radiation therapy in the modified intention-to-treat population. Safety was assessed in all patients who received at least one selective internal radiation therapy injection, and analysed on the basis of the treatment actually received (defined by central dosimetry assessment). The trial is registered with ClinicalTrials.gov, NCT02582034, and has been completed. Findings - Between Dec 5, 2015, and Jan 4, 2018, 93 patients were assessed for eligibility. Of these patients, 60 were randomly assigned: 31 to the personalised dosimetry group and 29 to the standard dosimetry group (intention-to-treat population). 56 (93%) patients (28 in each group) were treated (modified intention-to-treat population). In the modified intention-to-treat population, 20 (71% [95% CI 51-87]) of 28 patients in the personalised dosimetry group and ten (36% [19-56]) of 28 patients in the standard dosimetry group had an objective response (p=0·0074). In the safety analysis population, a least one serious adverse event was reported in seven (20%) of the 35 patients who received personalised dosimetry, and in seven (33%) of the 21 patients who received standard dosimetry. The most frequent (ie, occurring in >5% of patients) grade 3 or higher adverse events were ascites (one [3%] patient who received personalised dosimetry vs two [10%] patients who received standard dosimetry), hepatic failure (two [6%] vs none), lymphopenia (12 [34%] vs nine [43%]), increased aspartate aminotransferase concentrations (three [9%] vs two [10%]), increased alanine aminotransferase concentrations (three [9%] vs none), anaemia (two [6%] vs one [5%]), gastrointestinal haemorrhage (none vs two [10%]), and icterus (none vs two [10%]). One treatment-related death occurred in each group. Interpretation - Compared with standard dosimetry, personalised dosimetry significantly improved the objective response rate in patients with locally advanced hepatocellular carcinoma. The results of this study suggest that personalised dosimetry is likely to improve outcomes in clinical practice and should be used in future trials of selective internal radiation therapy. Funding - Biocompatibles UK, a Boston Scientific Group company.