1214 - Proton vs. Photon EBRT for Hepatocellular Carcinoma: A Multi-Institutional Individual Patient Data Analysis Using Propensity Score Weighting

03:05pm - 03:15pm ET

Room 254

Presenter(s)

Ted Yanagihara, MD, PhD - University of North Carolina, Chapel Hill, NC

T. K. Yanagihara¹, A. M. Moon², L. A. Dawson³, J. I. Yu⁴, T. S. Lawrence⁵, T. H. Kim⁶, M. Yan³, H. Iwata⁷, N. Nabavizadeh⁸, S. Apisarnthanarax⁹, E. M. Dunne¹⁰, M. I. Lock¹¹, M. D. Chuong¹², C. L. Chiang¹³, M. Scorsetti¹⁴, N. Katoh¹⁵, S. Sioshansi¹⁶, K. Numata¹⁷, H. Liu¹⁸, H. Iwamoto¹⁹, M. Wakatsuki²⁰, Y. Chen²¹, E. L. Pollom²², E. Gkika²³, S. K. Jabbour²⁴, R. Kumar²⁵, P. Muñoz-Schuffenegger²⁶, C. Hajj²⁷, D. Dutta²⁸, M. Ueno²⁹, C. L. Hallemeier³⁰, A. M. Feldman³¹, A. Mendez Romero³², X. Tan³³, M. Molla³⁴, J. E. Tepper³⁵, F. Torres³⁶, and M. Reig³⁷; ¹University of North Carolina, Department of Radiation Oncology, Chapel Hill, NC, ²University of North Carolina, Division of Gastroenterology/Hepatology, Chapel Hill, NC, ³Department of Radiation Oncology, Princess Margaret Cancer Centre, University of Toronto, Toronto, ON, Canada, ⁴Department of Radiation Oncology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea, Republic of (South), ⁵Department of Radiation Oncology, University of Michigan, Ann Arbor, MI, ⁶National Cancer Center, Goyang-Si, Korea, Republic of (South), ⁷Department of Radiation Oncology, Nagoya Proton Therapy Center, Nagoya City University West Medical Center, Nagoya, Aichi, Japan, ⁸Department of Radiation Medicine, Oregon Health & Science University, Portland, OR, ⁹Department of Radiation Oncology, University of Washington, Seattle, WA, ¹⁰BC Cancer Vancouver, Vancouver, BC, Canada, ¹¹Department of Radiation Oncology, London Health Sciences Centre, London, ON, Canada, ¹²Department of Radiation Oncology, Miami Cancer Institute, Baptist Health South Florida, Miami, FL, ¹³Department of Clinical Oncology, School of Clinical Medicine, LKS Faculty of Medicine, The University of Hong Kong, Hong Kong, China, Hong Kong, Hong Kong, ¹⁴IRCCS Humanitas Research Hospital , via Manzoni 56, 20089, Rozzano, Milan, Italy, ¹⁵Department of Radiation Oncology, Hokkaido University Faculty of Medicine, Sapporo, Japan, ¹⁶The Warren Alpert Medical School of Brown University, Providence, RI, ¹⁷Department of gastroenterology, Yokohama City University Medical Center, Yokohama, Japan, ¹⁸Princess Alexandra Hospital, Brisbane, Australia, ¹⁹Kurume University School of Medicine, Kurume City, Japan, ²⁰QST Hospital, National Institutes for Quantum Science and Technology, Chiba, Japan, ²¹Zhongshan Hospital, Fudan University, Shanghai, Shanghai, China, ²²Department of Radiation Oncology, Stanford University School of Medicine, Stanford, CA, ²³Department of Radiation Oncology, University Hospital Bonn, Bonn, Germany, ²⁴Department of Radiation Oncology, Rutgers Cancer Institute, New Brunswick, NJ, ²⁵Department of Radiation Oncology, Rutgers Cancer Institute of New Jersey, New Brunswick, NJ, ²⁶Radiation Oncology Unit, Department of Hematology - Oncology, Pontificia Universidad Catolica de Chile, Santiago, Chile, ²⁷Cleveland Clinic Abu Dhabi, Abu Dhabi, United Arab Emirates, ²⁸Amrita Institute of Medical Science, Kochi, India, ²⁹Kurashiki Central Hospital, Kurashiki, Japan, ³⁰Department of Radiation Oncology, Mayo Clinic, Rochester, MN, ³¹Department of Radiation Oncology, Henry Ford Health, Detroit, MI, ³²Department of Radiotherapy, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, Netherlands, ³³Gillings School of Global public health, University of North Carolina, Chapel Hill, NC, ³⁴Hospital Clinic Barcelona, Barcelona 08036, Spain, ³⁵Department of Radiation Oncology, University of North Carolina School of Medicine, Chapel Hill, NC, ³⁶Universitat Autònoma de Barcelona, Barcelona, Spain, ³⁷BCLC group, Liver Unit, ICMDM, IDIBAPS, Hospital Clinic of Barcelona, University of Barcelona, Barcelona, Spain

Purpose/Objective(s): For patients with hepatocellular carcinoma (HCC), comparisons between proton and photon external beam radiotherapy (EBRT) for have been confounded by systematic differences in patient populations. To address this, we performed a multi-institutional individual patient data (IPD) analysis using propensity score weighting to estimate the average treatment effect of proton versus photon EBRT for overall survival (OS) and local progression-free survival (LPFS).

Materials/Methods: International IPD were pooled from patients with HCC treated with EBRT across multiple centers. To mitigate measured confounding, we estimated propensity scores for receipt of proton versus photon EBRT using baseline demographics, albumin-bilirubin score, Child-Pugh, alpha fetoprotein, size and number of intrahepatic lesions, performance status, prior liver-directed therapies, and treatment center. Missing baseline covariates were handled using multiple imputation (30 datasets), and propensity score weighting was performed within each imputed dataset to target the average treatment effect. Treatment effects were estimated using weighted Cox proportional hazards models, and results were pooled across imputations using Rubin’s rules. Covariate balance was assessed using standardized mean differences (SMDs). In BCLC stage–specific analyses, propensity score models and weighting were performed within each stage.

Results: A total of 4,777 patients were included (photon n=2,988; proton n=1,789). Following multiple imputation and propensity score weighting, covariate balance was achieved (most mean absolute SMDs <0.10). In the overall cohort, proton EBRT showed no significant improvement in OS or LPFS. Stage-specific analyses suggested heterogeneity of treatment effect. Proton EBRT was associated with improved OS in BCLC-D, was at borderline significance for BCLC-A, and was non-significant for BCLC-0, -B, or -C (Table) using stage-specific propensity score weighting. For LPFS, proton EBRT was associated with improvement in BCLC-D, but not BCLC-0, -A, -B, or -C (Table).

Conclusion: In this large international IPD analysis using propensity score weighting, proton therapy was associated with improved OS and LPFS in advanced (BCLC-D) HCC. There may be non-proportional hazards in some stage-specific strata and future analyses will complement HRs with restricted mean survival time and time-specific survival estimates. Residual confounding and limited toxicity detail constrain mechanistic inference and prospective randomized comparisons, including NRG-GI003, are needed to confirm these observations.

	Variable	HR	95% CI	p-value
OS	Overall cohort	0.84	0.64-1.11	N.S.
	BCLC-0	0.94	0.75-1.18	N.S.
	BCLC-A	0.72	0.53-0.99	0.046
	BCLC-B	0.96	0.76-1.21	N.S.
	BCLC-C	0.84	0.6-1.18	N.S.
	BCLC-D	0.6	0.43-0.85	<0.01
LPFS	Overall cohort	0.91	0.73-1.13	N.S.
	BCLC-0	1.07	0.96-1.2	N.S.
	BCLC-A	0.92	0.85-1.01	N.S.
	BCLC-B	0.91	0.61-1.35	N.S.
	BCLC-C	0.83	0.59-1.17	N.S.
	BCLC-D	0.55	0.42-0.7	<0.01