Instructor in Investigation Radiation Oncology, Mass General Research Institute

Instructor in Radiation Oncology Harvard Medical School

markov model; optimal stopping of radiation therapy; radiation dosage; radiation therapy; radiotherapy

Treatment personalization according to individual patient's biological characteristics remain the ultimate goal in cancer care. Three pillars of treatment personalization are (i) Discovery of predictive biomarkers of treatment response, (ii) Developing accurate predictive models for biomarker-based response prediction, and (iii) Devising dynamic and robust optimization methods for treatment adaptation. My research interest lies in the intersection of these areas.

I am interested in using advanced (big) data analytics on medical imaging, genomic, and proteomics data for discovery of novel predictive biomarkers of response to radiation therapy. I use state-of-the-art machine learning tools, with an additional focus on "interpretable machine learning",  to derive predictive models of RT response by synthesizing patient-specific information from clinical, pathological, and biomarker data. Furthermore, I heavily rely on advanced optimization methods, with a focus on robustness and adaptability, to adapt the treatment plans according to biomarker information and model's predictions.