American Statistical Association
Adaptive treatment strategy (or dynamic treatment regimen) is an active research area that leverages machine learning, causal inference, and other statistical methodologies to identify the optimal medical treatment or sequence of treatments that are tailored to individual patient's characteristics. Cancer treatment is a highly complicated process due to the complex mechanism of the underlying disease and the great heterogeneity of cancer patients. Multiple lines of treatments are usually prescribed based on the patient's genetic profile, clinical characteristics, and response to previous treatments. Thus, it is natural to apply adaptive treatment strategy to cancer research to provide oncologists with evidence-based treatment decision rules. In this talk, I will provide an overview of current research in the adaptive treatment regimen and discuss its potential application to cancer research at MSKCC. I will introduce Q-learning, G-estimation, marginal structural model, and outcome-weighted learning in both single-stage and multi-stage settings. The primary focus is on observational studies, but I will briefly touch upon the Sequential Multiple Assignment Randomized Trial (SMART).
|Date:||Wednesday, June 7, 2017|
|Time:||4:00 - 5:00 P.M.|
Memorial Sloan Kettering Cancer Center
Department of Epidemiology and Biostatistics
485 Lexington Avenue
(Between 46th & 47th Streets)
2nd Floor, Conference Room B
New York, New York
**Outside visitors please email email@example.com for building access.
You must be on the security list to enter the floor.