Personalizing Maintenance Rituximab in Follicular Lymphoma: A Machine Learning Framework for Risk-Benefit
Junyi Gao1,2,3, Jiaxin Liu1, Xinze Li4
1Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing), Department of Lymphoma, Peking University Cancer Hospital & Institute, Beijing, China.
Health Data Science
|August 13, 2026
Summary
Maintenance rituximab therapy (MT) for follicular lymphoma shows varied patient benefit. A new machine learning model identifies patients unlikely to benefit, potentially reducing overtreatment and improving outcomes in oncology.
Area of Science:
- Oncology
- Hematology
- Machine Learning in Medicine
Background:
- Maintenance rituximab therapy (MT) improves progression-free survival in follicular lymphoma.
- Individual patient benefit from MT is variable, necessitating personalized treatment approaches.
- Overtreatment with MT carries burdens that can be avoided through tailored strategies.
Purpose of the Study:
- To develop a machine learning framework for estimating individualized treatment effects of MT.
- To identify patients with follicular lymphoma who are unlikely to benefit from MT.
- To reduce overtreatment by personalizing MT based on predicted benefit.
Main Methods:
- A double machine learning framework was developed.
- A retrospective cohort of 404 patients with follicular lymphoma was analyzed.
- Individualized treatment effects of MT on 24-month progression risk were estimated.
Main Results:
- The model identified significant heterogeneity in MT benefit among patients.
- A "low-risk, low-benefit" subgroup was identified, potentially suitable for forgoing MT.
- Alignment with model recommendations correlated with a lower progression rate (14.8% vs. 38.0%).
Conclusions:
- A data-driven framework personalizes MT by distinguishing prognostic risk from predictive benefit.
- The framework supports a clinical decision tool to guide personalized MT strategies.
- This approach can guide future trials and reduce unnecessary cancer treatment.
