HCC Etiology-Dependent Tumor Control Probability Model After Stereotactic Body Radiation Therapy for Hepatocellular Carcinoma: A Pooled Analysis of an Asian Liver Radiation Therapy Group Study

Abstract

Purpose/Objective(s)Stereotactic body radiation therapy (SBRT) has emerged as an effective local treatment option for hepatocellular carcinoma (HCC). Using a multi-national database from the Asian Liver Radiation Therapy (ALRT) Study Group, we previously reported that higher biologically effective dose (BED) was associated with improved local control (LC). We investigated the tumor control probability (TCP) of HCC using size-adjusted BED (sBED), which has been investigated in prior studies of SBRT for lung cancer.Materials/MethodsALRT database included 519 patients with HCC undergoing SBRT. After excluding 104 patients with vascular invasion, a total of 415 patients without tumor thrombus treated with SBRT for HCC in 7 tertiary-referral hospitals from 2010 to 2016 were retrospectively reviewed. A BED was calculated for each patient using a standard a/b ratio of 10 Gy. Then, sBED was defined as BED minus 10 times the maximal tumor diameter, in centimeters. sBED was incorporated into standard TCP models to predict actuarial 2-year LC rates.ResultsMedian tumor diameter was 2.5 cm (interquartile range, IQR, 1.5-3.9). Median prescribed total and fractional doses for SBRT were 48 (IQR, 40-54) Gy and 10 (IQR, 8-12) Gy, respectively. The median BED was 100 Gy (IQR, 80-116), and the median sBED was 70 Gy (IQR,49-92). The median follow-up duration was 26 months after SBRT, and 73 local failures (18%) were observed. Multivariable analysis showed that sBED was an independent factor associated with LC (hazard ratio, 0.94 per 10 Gy, 95% CI 0.88-0.99, P = 0.028). Initial analyses for non-dosimetric factors demonstrated that hepatitis C virus (HCV) infection was a powerful favorable prognostic factor (HR 1.81, 95% CI 1.01-3.27, P = 0.048) for LC, so TCP modeling was performed separately for HCV-related HCC (n = 125, 2-year LC 88%) and for other tumors (n = 290, 2-year LC 78%). BED values predicted to provide 80%, 85% and 90% 2-year probabilities of local control in various scenarios are provided in Table 1.ConclusionUsing a large, multi-national database, we found a strong association between HCC etiology and local control rates after SBRT. We generated novel TCP models for HCV-related HCC and other HCC etiologies that may be useful for selecting appropriate SBRT schedules. Table. Approximate BEDs required to achieve 80%, 85%, and 90% 2-year tumor control probability for HCV-related and non-HCV-related tumors of selected sizes. Missing values indicate scenarios where results would fall outside the range of commonly used SBRT schedules. Stereotactic body radiation therapy (SBRT) has emerged as an effective local treatment option for hepatocellular carcinoma (HCC). Using a multi-national database from the Asian Liver Radiation Therapy (ALRT) Study Group, we previously reported that higher biologically effective dose (BED) was associated with improved local control (LC). We investigated the tumor control probability (TCP) of HCC using size-adjusted BED (sBED), which has been investigated in prior studies of SBRT for lung cancer. ALRT database included 519 patients with HCC undergoing SBRT. After excluding 104 patients with vascular invasion, a total of 415 patients without tumor thrombus treated with SBRT for HCC in 7 tertiary-referral hospitals from 2010 to 2016 were retrospectively reviewed. A BED was calculated for each patient using a standard a/b ratio of 10 Gy. Then, sBED was defined as BED minus 10 times the maximal tumor diameter, in centimeters. sBED was incorporated into standard TCP models to predict actuarial 2-year LC rates. Median tumor diameter was 2.5 cm (interquartile range, IQR, 1.5-3.9). Median prescribed total and fractional doses for SBRT were 48 (IQR, 40-54) Gy and 10 (IQR, 8-12) Gy, respectively. The median BED was 100 Gy (IQR, 80-116), and the median sBED was 70 Gy (IQR,49-92). The median follow-up duration was 26 months after SBRT, and 73 local failures (18%) were observed. Multivariable analysis showed that sBED was an independent factor associated with LC (hazard ratio, 0.94 per 10 Gy, 95% CI 0.88-0.99, P = 0.028). Initial analyses for non-dosimetric factors demonstrated that hepatitis C virus (HCV) infection was a powerful favorable prognostic factor (HR 1.81, 95% CI 1.01-3.27, P = 0.048) for LC, so TCP modeling was performed separately for HCV-related HCC (n = 125, 2-year LC 88%) and for other tumors (n = 290, 2-year LC 78%). BED values predicted to provide 80%, 85% and 90% 2-year probabilities of local control in various scenarios are provided in Table 1. Using a large, multi-national database, we found a strong association between HCC etiology and local control rates after SBRT. We generated novel TCP models for HCV-related HCC and other HCC etiologies that may be useful for selecting appropriate SBRT schedules. Table. Approximate BEDs required to achieve 80%, 85%, and 90% 2-year tumor control probability for HCV-related and non-HCV-related tumors of selected sizes. Missing values indicate scenarios where results would fall outside the range of commonly used SBRT schedules.