Pathologic Response Of Hepatocellular Carcinoma With Yttrium-90 Radioembolization

Yttrium-90 (Y90) radioembolization is often considered for local treatment in the setting of unresectable or chemoresistant hepatocellular carcinoma (HCC). For patients within Milan criteria, the goal of such therapy is to provide disease control until liver transplantation. Previous studies have assessed the dose response of radioembolization using radiographic criteria but, to our knowledge, none have assessed lesion-specific pathologic dose response of HCC with Y90 with dosimetric information. We performed an IRB-approved, single-institution retrospective analysis of patients with HCC who underwent Y90 microsphere radioembolization with post-treatment PET/CT and subsequent liver transplantation from 2016 to 2019. We contoured each treated lesion to determine lesion volume using pre-treatment magnetic resonance and computed tomography imaging and computed dosimetric coverage using post-treatment Y90 PET/CT. Lesion volume, delivered GBq, and mean/D70Gy dosimetric values were obtained and analyzed for complete pathologic response at the time of explant. Optimal cutoff values were generated for factors that demonstrated statistical significance for correlation using paired t-tests for complete pathological response. Sixteen lesions from ten patients met criteria for analysis. Fifteen of sixteen lesions were evaluated for complete pathologic response due to one patient undergoing transplant only 8 days after radioembolization. Lesions had a median volume of 3.3 cm3 (range: 0.9 to 74.5 cm3) and median largest dimension of 1.9 cm (range: 1.0 to 4.6 cm). The median prescription dose was 150 Gy (range: 120 to 150 Gy) which was prescribed to the target lobe volume, expecting much higher doses to individual lesions. Median interval between treatment and transplantation was 143 days (range: 34 to 410 days). Complete pathologic response was seen in ten lesions corresponding to four patients. Of the five lesions with partial response, two showed 80 to 95% necrosis while three showed 50 to 55% necrosis. Of the factors evaluated, only the PET/CT-derived mean dose received per lesion was found to be significantly associated with complete pathologic response (p<0.001). An optimal cut off value of mean dose >600 Gy was statistically significantly associated with higher complete pathologic response rates. Of those with lesions with a complete pathologic response, 10/10 (100%) had a mean dose >600 Gy. Of those lesions with viable tumor, 4/5 (80%) had a mean dose less than 600 Gy (p<0.001). There is limited data on the dosimetric factors associated with pathological complete response in radioembolization treatment for HCC. Here, we correlated increased mean delivered dose to pathological complete response. In this limited set of patients, a mean dose of >600 Gy delivered to the tumor was statistically significantly associated with complete pathologic response.