An automated machine learning framework to optimize radiomics model construction validated on twelve clinical applications
Abstract
Predicting clinical outcomes from medical images using quantitative features (``radiomics'') requires many method design choices, Currently, in new clinical applications, finding the optimal radiomics method out of the wide range of methods relies on a manual, heuristic trial-and-error process. We introduce a novel automated framework that optimizes radiomics workflow construction per application by standardizing the radiomics workflow in modular components, including a large collection of algorithms for each component, and formulating a combined algorithm selection and hyperparameter optimization problem. To solve it, we employ automated machine learning through two strategies (random search and Bayesian optimization) and three ensembling approaches. Results show that a medium-sized random search and straight-forward ensembling perform similar to more advanced methods while being more efficient. Validated across twelve clinical applications, our approach outperforms both a radiomics baseline and human experts. Concluding, our framework improves and streamlines radiomics research by fully automatically optimizing radiomics workflow construction. To facilitate reproducibility, we publicly release six datasets, software of the method, and code to reproduce this study.
Cite
@article{arxiv.2108.08618,
title = {An automated machine learning framework to optimize radiomics model construction validated on twelve clinical applications},
author = {Martijn P. A. Starmans and Sebastian R. van der Voort and Thomas Phil and Milea J. M. Timbergen and Melissa Vos and Guillaume A. Padmos and Wouter Kessels and David Hanff and Dirk J. Grunhagen and Cornelis Verhoef and Stefan Sleijfer and Martin J. van den Bent and Marion Smits and Roy S. Dwarkasing and Christopher J. Els and Federico Fiduzi and Geert J. L. H. van Leenders and Anela Blazevic and Johannes Hofland and Tessa Brabander and Renza A. H. van Gils and Gaston J. H. Franssen and Richard A. Feelders and Wouter W. de Herder and Florian E. Buisman and Francois E. J. A. Willemssen and Bas Groot Koerkamp and Lindsay Angus and Astrid A. M. van der Veldt and Ana Rajicic and Arlette E. Odink and Mitchell Deen and Jose M. Castillo T. and Jifke Veenland and Ivo Schoots and Michel Renckens and Michail Doukas and Rob A. de Man and Jan N. M. IJzermans and Razvan L. Miclea and Peter B. Vermeulen and Esther E. Bron and Maarten G. Thomeer and Jacob J. Visser and Wiro J. Niessen and Stefan Klein},
journal= {arXiv preprint arXiv:2108.08618},
year = {2025}
}
Comments
22 pages, 3 figures, 2 tables, 1 algorithm, 3 supplementary figures, 4 supplementary tables, 1 supplementary algorithm