Masoud Zarepisheh
Associate Professor/Attending
Department of Medical Physics
Memorial Sloan Kettering Cancer Center
321 East 61st Street, New York, NY 10065
I develop mathematical optimization methods for radiotherapy treatment planning and translate them into clinical practice. Our automated planning system, ECHO, is part of the daily clinical routine at Memorial Sloan Kettering, where it has been used to treat more than 12,000 patients, it was a finalist for the 2021 INFORMS Franz Edelman Award, and its technology has been licensed to RaySearch Laboratories. I also lead PortPy, an open-source Python platform that gives researchers benchmark data and algorithms for treatment planning optimization and is downloaded more than 1,000 times a month.
Research
Radiotherapy treats cancer with precisely directed radiation beams, and every patient’s plan (beam angles, beam shapes, and intensities) must be tailored to that patient’s anatomy and the physician’s prescription for the tumor and the surrounding healthy tissue. Mathematically, this is a large-scale, multi-criteria, and often non-convex optimization problem, and my work is about solving it well enough, and fast enough, to be used in the clinic.
Automated radiotherapy treatment planning: ECHO
Treatment planning has traditionally required hours of manual parameter tuning, with plan quality depending on the planner’s skill and experience. ECHO (Expedited Constrained Hierarchical Optimization) automates this with hierarchical constrained optimization. ECHO performs plan optimization independently of the treatment planning system and uses Eclipse, the FDA-approved commercial system in our clinic, only for the final dose calculation. It runs in our daily clinical routine and has been used to treat more than 12,000 patients. ECHO was a finalist for the 2021 INFORMS Franz Edelman Award (Video · Slides · Paper · Podcast · Press), and its technology has been licensed to RaySearch Laboratories.
Open-source planning research: PortPy
PortPy is an open-source Python package that provides research-ready data and code to speed up the development and clinical translation of treatment planning optimization algorithms, and is downloaded more than 1,000 times a month. It ships benchmark data of 329 patients (200 lung and 129 prostate) pre-calculated with Eclipse and benchmark implementations of automated planning, beam-angle, aperture, and robust optimization.
Selected publications
- Sparse plus low-rank matrix embedding with applications in cancer radiotherapy optimization Sparse + low-rank embedding · curse of dimensionality
- Compressed radiotherapy treatment planning (CompressRTP): A new paradigm for rapid and high-quality treatment planning optimization Matrix compression · curse of dimensionality
- Randomized Sparse Matrix Compression for Large-Scale Constrained Optimization in Cancer Radiotherapy Randomized sketching · large-scale constrained optimization
- Automated VMAT treatment planning using sequential convex programming: algorithm development and clinical implementation Sequential convex programming · VMAT · in clinical use
- Distributed and scalable optimization for robust proton treatment planning ADMM · robust proton planning
- Domain knowledge driven 3D dose prediction using moment-based loss function Deep learning dose prediction · moment-based loss
- Automated and Clinically Optimal Treatment Planning for Cancer Radiotherapy ECHO · Franz Edelman Award finalist · in clinical use
- Automating Proton Treatment Planning with Beam Angle Selection Using Bayesian Optimization Bayesian optimization · beam angle selection
- Automated Proton Treatment Planning with Robust Optimization Using Constrained Hierarchical Optimization Robust optimization · setup and range uncertainty
- Integrating Soft and Hard Dose-Volume Constraints into Hierarchical Constrained IMRT Optimization Dose-volume constraints · mixed-integer programming · in clinical use
-
Automated Intensity Modulated Treatment Planning: The Expedited Constrained Hierarchical Optimization (ECHO) System
ECHO · hierarchical constrained optimization · in clinical use
Medical Physics, 2019 · Paper
Full list on the CV page and on Google Scholar.