Personalized medicine typically takes one of two forms: (i) identifying subgroups of patients who benefit from a particular treatment, or (ii) determining the optimal treatment for an individual patient. My primary interest lies in the latter, the intersection of personalized medicine and reinforcement learning, where I focus on developing novel statistical methods to estimate optimal, adaptive interventions tailored to the uniquely evolving health status of each patient over time.

• Oh, E. J., Qian, M., and Cheung, Y. K. (2022). Generalization error bounds of dynamic treatment regimes in penalized regression-based learning. Annals of Statistics, 50(4), 2047-2071. [pdf] [supp]
• Oh, E. J., Qian, M., Cheung, K., and Mohr, D. C. (2020). Building health application recommender system using partially penalized regression. Statistical Modeling in Biomedical Research, Springer, 105-123.

Another key area of my research is risk prediction in oncology using machine learning (ML) algorithms, which plays a crucial role in improving patient outcomes and guiding clinical decisions. We aim to address the following key questions: How can we more effectively account for heterogeneous effects in high-dimensional data, particularly in the context of rare diseases? How can we ensure proper model calibration to enhance the reliability of the model’s estimates when data for the target samples is limited? How can we develop accurate, yet clinically intuitive, ML models to inform risk-based follow-up care for time-to-event outcomes (e.g., cause-specific survival, relapse-free survival)?

• Oh, E. J., Ahn, S., Tham, T., and Qian, M. (2025). Leveraging two-phase data for improved prediction of survival outcomes with application to nasopharyngeal cancer. Biometrics, 81(2), ujaf080. [link] [supp]
• Oh, E. J., Alfano, C. M., Esteva, F. J., Baron, P. L., Xiong, W., Brooke, T. E., Chen, E. I., and Chiuzan, C. (2025). Risk stratification using tree-based models for recurrence-free survival in breast cancer. JCO Oncology Advances, 2, e2400011. [pdf]
• Oh, E. J., Parikh, R. B., Chivers, C., and Chen, J. (2021). Two-stage approaches to accounting for patient heterogeneity in machine learning risk prediction models in oncology. JCO Clinical Cancer Informatics, 5, 1015-1023. [pdf]

Beyond oncology, I have collaborated with researchers in various fields, including otolaryngology, immunology, and surgery. Collaborative projects that I would like to highlight include, but not limited to, the following publications:

• Ahn, S., Oh, E. J., Saleem, M., and Tham, T. (2024). Machine learning methods in classification of prolonged radiation therapy in oropharyngeal cancer: national cancer database. Otolaryngology-Head and Neck Surgery, 171(6), 1764-1772.
• Overdevest, J., Irace, A. L., Mazzanti, V., Oh, E. J., Joseph, P. V., Devanand, D. P., Bitan, Z. C., Hod, E. A., Gudis, D. A., and Chiuzan, C. (2022). Chemosensory deficits are best predictor of serologic response among individuals infected with SARS-CoV-2. PLoS ONE, 17(12), e0274611.
• Gartrell, R. D., Enzler, T., Kim, P. S., Fullerton, B. T., Fazlollahi, L., Chen, A. X., Minns, H. E., Perni, S., Weisberg, S. P., Rizk, E. M., Wang, S., Oh, E. J., Guo, X. V., …, and Saenger, Y. M. (2022). Neoadjuvant chemoradiation alters the immune microenvironment in pancreatic ductal adenocarcinoma. OncoImmunology, 11(1), 2066767.
• Zhu, D., Wong, A., Oh, E. J., Ahn, S., Wotman, M., Sahai, T., Bottalico, D., Frank, D., and Tham, T. (2022). Impact of treatment parameters on racial survival differences in oropharyngeal cancer: national cancer database study. Otolaryngology–Head & Neck Surgery, 166(6), 1134-1143.
• Toyoda, Y., Oh, E. J., Premaratne, I. D., Chiuzan, C., and Rohde, C. H. (2020). Affordable care act state-specific medicaid expansion: impact on health insurance coverage and breast cancer screening rate. Journal of the American College of Surgeons, 230(5), 775-783.