TY - JOUR
T1 - Towards a safe and efficient clinical implementation of machine learning in radiation oncology by exploring model interpretability, explainability and data-model dependency
AU - Barragan-Montero, Ana
AU - Bibal, Adrien
AU - Dastarac, Margerie Huet
AU - Draguet, Camille
AU - Valdes, Gilmer
AU - Nguyen, Dan
AU - Willems, Siri
AU - Vandewinckele, Liesbeth
AU - Holmstrom, Mats
AU - Lofman, Fredrik
AU - Souris, Kevin
AU - Sterpin, Edmond
AU - Lee, John A.
N1 - Funding Information:
Ana Barragán and Margerie Huet are funded by the Walloon region in Belgium (PROTHERWAL/CHARP, grant 7289). Gilmer Valdés was supported by the National Institute of Biomedical Imaging and Bioengineering of the National Institutes of Health under Award Number K08EB026500. Dan Nguyen is supported by the National Institutes of Health (NIH) R01CA237269 and the Cancer Prevention & Research Institute of Texas (CPRIT) IIRA RP150485. Liesbeth Vandewinckele is supported by a PhD fellowship of the research foundation-Flanders (FWO), mandate 1SA6121N. Kevin Souris is funded by the Walloon region (MECATECH/BIOWIN, grant 8090). John A. Lee is a Senior Research Associate with the F.R.S.-FNRS.
Publisher Copyright:
© 2022 The Author(s). Published on behalf of Institute of Physics and Engineering in Medicine by IOP Publishing Ltd.
PY - 2022/6/7
Y1 - 2022/6/7
N2 - The interest in machine learning (ML) has grown tremendously in recent years, partly due to the performance leap that occurred with new techniques of deep learning, convolutional neural networks for images, increased computational power, and wider availability of large datasets. Most fields of medicine follow that popular trend and, notably, radiation oncology is one of those that are at the forefront, with already a long tradition in using digital images and fully computerized workflows. ML models are driven by data, and in contrast with many statistical or physical models, they can be very large and complex, with countless generic parameters. This inevitably raises two questions, namely, the tight dependence between the models and the datasets that feed them, and the interpretability of the models, which scales with its complexity. Any problems in the data used to train the model will be later reflected in their performance. This, together with the low interpretability of ML models, makes their implementation into the clinical workflow particularly difficult. Building tools for risk assessment and quality assurance of ML models must involve then two main points: interpretability and data-model dependency. After a joint introduction of both radiation oncology and ML, this paper reviews the main risks and current solutions when applying the latter to workflows in the former. Risks associated with data and models, as well as their interaction, are detailed. Next, the core concepts of interpretability, explainability, and data-model dependency are formally defined and illustrated with examples. Afterwards, a broad discussion goes through key applications of ML in workflows of radiation oncology as well as vendors' perspectives for the clinical implementation of ML.
AB - The interest in machine learning (ML) has grown tremendously in recent years, partly due to the performance leap that occurred with new techniques of deep learning, convolutional neural networks for images, increased computational power, and wider availability of large datasets. Most fields of medicine follow that popular trend and, notably, radiation oncology is one of those that are at the forefront, with already a long tradition in using digital images and fully computerized workflows. ML models are driven by data, and in contrast with many statistical or physical models, they can be very large and complex, with countless generic parameters. This inevitably raises two questions, namely, the tight dependence between the models and the datasets that feed them, and the interpretability of the models, which scales with its complexity. Any problems in the data used to train the model will be later reflected in their performance. This, together with the low interpretability of ML models, makes their implementation into the clinical workflow particularly difficult. Building tools for risk assessment and quality assurance of ML models must involve then two main points: interpretability and data-model dependency. After a joint introduction of both radiation oncology and ML, this paper reviews the main risks and current solutions when applying the latter to workflows in the former. Risks associated with data and models, as well as their interaction, are detailed. Next, the core concepts of interpretability, explainability, and data-model dependency are formally defined and illustrated with examples. Afterwards, a broad discussion goes through key applications of ML in workflows of radiation oncology as well as vendors' perspectives for the clinical implementation of ML.
KW - clinical implementation
KW - interpretability and explainability
KW - machine learning
KW - radiation oncology
KW - uncertainty quantification
UR - http://www.scopus.com/inward/record.url?scp=85131371379&partnerID=8YFLogxK
U2 - 10.1088/1361-6560/ac678a
DO - 10.1088/1361-6560/ac678a
M3 - Article
C2 - 35421855
AN - SCOPUS:85131371379
SN - 0031-9155
VL - 67
JO - Physics in Medicine and Biology
JF - Physics in Medicine and Biology
IS - 11
M1 - 11TR01
ER -