2183 - Impact of Gross Tumor Volume (GTV) and Effective Uniform Dose (EUD) on Clinical Outcomes after SBRT for Hepatocellular Carcinoma (HCC)

03:00pm - 04:00pm ET

Poster Hall - Exhibit Hall A

Screen: 29

POSTER

Presenter(s)

Avinash Ramkissoon, MD, MPH - University of Pennsylvania Radiation Oncology, Philadelphia, PA

A. Ramkissoon¹, U. Amit Sr^1,2, M. Perez-Guillermo Cuev Sr¹, K. P. Risolo Jr¹, C. Jiang^1,3, K. Khullar¹, and E. Ben-Josef¹; ¹Department of Radiation Oncology, University of Pennsylvania, Philadelphia, PA, ²Department of Radiation Oncology, Tel Aviv Medical Center, Tel Aviv, Israel, ³Department of Radiation Oncology, Thomas Jefferson University Hospital, Philadelphia, PA

Purpose/Objective(s):

Liver SBRT has emerged as an effective non-surgical treatment option for HCC with excellent local control rates. SBRT inherently leads to dose heterogeneity within the target, and recent studies have suggested that EUD, which models non-uniform dose distributions as a theoretical homogenous dose, is a predictor of outcomes beyond dose-fractionation. This analysis aims to determine the effect of EUD in patients treated with liver SBRT and assess the impact on oncologic outcomes.

Materials/Methods:

70 HCC lesions were analyzed among 60 patients. All lesions were treated with SBRT (50Gy in 5Fx). EUD₅ and EUD₁₅ for each lesion were calculated using the Niemerko model, using a=5 and a=15, respectively. Univariate analyses of demographics (age, gender, ECOG, ethnicity), tumor features (GTV size, etiology of HCC, prior local treatment) and dosimetric parameters (EUD₅ and EUD₁₅) were performed, followed by Cox proportional hazard multivariate modeling for local recurrence (LR) and disease-specific survival (DSS) using clinically relevant variables and those with a statistically significant association on univariate analysis. EUD₁₅ was used in all models, as it emphasizes hotspots and penalizes undercoverage, which is thought to correlate well with local control.

Results:

88% of this cohort were male. 45% were =70 years old at treatment. Hepatitis C was the most common etiology of HCC (48%). The majority of patients were white ethnicity (65%), followed by African-American (18%). 88% were ECOG 0-1.

There was 93% local control at 3-year follow-up, and median survival was 21 months, with 16 HCC-related deaths. For tumors that recurred locally, the median GTV was 55cm³, compared to 18cm³ for those that achieved long-term local control (p=0.09). The median EUD₁₅ delivered to tumors that recurred in-field was 94 Gy compared to 121 Gy in those with long-term local control (p=0.03).

The multivariate models for LR and DSS included patient-related variables (age, sex, ethnicity, ECOG, and HCC etiology), GTV size, EUD₁₅ delivered to the tumor, and an interaction term to account for effect modification between EUD₁₅ and GTV size. In this model, higher EUD₁₅ had a statistically significant association with LR (HR = 0.925 per Gy, 95% CI: 0.856 – 0.999, p = 0.048), but not DSS (HR = 0.986 per Gy, 95% CI: 0.970 – 1.003, p = 0.108), when accounting for effect modification by GTV size. GTV size was significantly associated with both LR (HR = 1.009 per 1cm³, 95% CI: 1.001 – 1.017, p = 0.015) and DSS (HR = 1.009 per 1cm³, 95% CI: 1.002 – 1.015, p = 0.007), when accounting for effect modification by EUD. A larger effect size was observed with EUD₁₅ compared to EUD₅ in all models.

Conclusion:

In this study, we observed that GTV size had a statistically significant relationship with both LR and DSS. Higher EUD can be delivered to smaller tumors, and on multivariate analysis, EUD₁₅ had predictive power for LR, independent of GTV size. This suggests that local control could be improved by optimizing EUD₁₅, particularly when GTV =55cm³.