Streamlined Quantitative Imaging Biomarker Development

Generalization of radiomics through automated machine learning

Abstract

Due to the paradigm shift in health care towards personalized medicine, there is an increased demand for biomarkers. Radiomics leverages quantitative medical imaging features and machine learning to create biomarkers based on medical imaging. While many radiomics methods have been described in the literature, these are generally designed for a single application. The overall aim of this thesis is to streamline radiomics research, facilitate its reproducibility, and simplify its application. In this thesis, we exploit recent advances in automated machine learning to develop an adaptive radiomics framework and demonstrate its use to develop radiomics biomarkers in eight different, independent clinical applications. See also this blog I wrote for one of my thesis sponsors, Quantib.

Martijn P. A. Starmans
Martijn P. A. Starmans
Assistant Professor & PI Artificial Intelligence for Integrated Diagnostics (AIID) focused on Medical Imaging in Oncology

My research interests include radiomics, pathomics, meta-learning, and AutoML.