Detecting Brain Anomalies in Clinical Routine with the $β$-VAE: Feasibility Study on Age-Related White Matter Hyperintensities

Sophie Loizillon, Yannick Jacob, Maire Aurélien, Didier Dormont, Olivier Colliot, Ninon Burgos, Apprimage Study Group
Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning, PMLR 250:903-917, 2024.

Abstract

This experimental study assesses the ability of variational autoencoders (VAEs) to perform anomaly detection in clinical routine, in particular the detection of age-related white matter lesions in brain MRIs acquired at different hospitals and gathered in a clinical data warehouse (CDW). We pre-trained a state-of-the-art $\beta$-VAE on a healthy cohort of over 10,000 FLAIR MR images from the UK Biobank to learn the distribution of healthy brains. The model was then fine-tuned on a cohort of nearly 700 healthy FLAIR images coming from a CDW. We first ensured the good performance of our pre-trained model compared with the state-of-the-art using a widely used public dataset (MSSEG). We then validated it on our target task, age-related WMH detection, on ADNI3 and on a curated clinical dataset from a single-site neuroradiology department, for which we had manually delineated lesion masks. Next, we applied the fine-tuned $\beta$-VAE for anomaly detection in a CDW characterised by an exceptional heterogeneity in terms of hospitals, scanners and image quality. We found a correlation between the Fazekas scores extracted from the radiology reports and the volumes of the lesions detected by our model, providing a first insight into the performance of VAEs in a clinical setting. We also observed that our model was robust to image quality, which strongly varies in the CDW. However, despite these encouraging results, such approach is not ready for an application in clinical routine yet due to occasional failures in detecting certain lesions, primarily attributed to the poor quality of the images reconstructed by the VAE.

Cite this Paper

BibTeX
@InProceedings{pmlr-v250-loizillon24a, title = {Detecting Brain Anomalies in Clinical Routine with the $β$-VAE: Feasibility Study on Age-Related White Matter Hyperintensities}, author = {Loizillon, Sophie and Jacob, Yannick and Aur\'elien, Maire and Dormont, Didier and Colliot, Olivier and Burgos, Ninon and Group, Apprimage Study}, booktitle = {Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning}, pages = {903--917}, year = {2024}, editor = {Burgos, Ninon and Petitjean, Caroline and Vakalopoulou, Maria and Christodoulidis, Stergios and Coupe, Pierrick and Delingette, Hervé and Lartizien, Carole and Mateus, Diana}, volume = {250}, series = {Proceedings of Machine Learning Research}, month = {03--05 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v250/main/assets/loizillon24a/loizillon24a.pdf}, url = {https://proceedings.mlr.press/v250/loizillon24a.html}, abstract = {This experimental study assesses the ability of variational autoencoders (VAEs) to perform anomaly detection in clinical routine, in particular the detection of age-related white matter lesions in brain MRIs acquired at different hospitals and gathered in a clinical data warehouse (CDW). We pre-trained a state-of-the-art $\beta$-VAE on a healthy cohort of over 10,000 FLAIR MR images from the UK Biobank to learn the distribution of healthy brains. The model was then fine-tuned on a cohort of nearly 700 healthy FLAIR images coming from a CDW. We first ensured the good performance of our pre-trained model compared with the state-of-the-art using a widely used public dataset (MSSEG). We then validated it on our target task, age-related WMH detection, on ADNI3 and on a curated clinical dataset from a single-site neuroradiology department, for which we had manually delineated lesion masks. Next, we applied the fine-tuned $\beta$-VAE for anomaly detection in a CDW characterised by an exceptional heterogeneity in terms of hospitals, scanners and image quality. We found a correlation between the Fazekas scores extracted from the radiology reports and the volumes of the lesions detected by our model, providing a first insight into the performance of VAEs in a clinical setting. We also observed that our model was robust to image quality, which strongly varies in the CDW. However, despite these encouraging results, such approach is not ready for an application in clinical routine yet due to occasional failures in detecting certain lesions, primarily attributed to the poor quality of the images reconstructed by the VAE.} }
Endnote
%0 Conference Paper %T Detecting Brain Anomalies in Clinical Routine with the $β$-VAE: Feasibility Study on Age-Related White Matter Hyperintensities %A Sophie Loizillon %A Yannick Jacob %A Maire Aurélien %A Didier Dormont %A Olivier Colliot %A Ninon Burgos %A Apprimage Study Group %B Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2024 %E Ninon Burgos %E Caroline Petitjean %E Maria Vakalopoulou %E Stergios Christodoulidis %E Pierrick Coupe %E Hervé Delingette %E Carole Lartizien %E Diana Mateus %F pmlr-v250-loizillon24a %I PMLR %P 903--917 %U https://proceedings.mlr.press/v250/loizillon24a.html %V 250 %X This experimental study assesses the ability of variational autoencoders (VAEs) to perform anomaly detection in clinical routine, in particular the detection of age-related white matter lesions in brain MRIs acquired at different hospitals and gathered in a clinical data warehouse (CDW). We pre-trained a state-of-the-art $\beta$-VAE on a healthy cohort of over 10,000 FLAIR MR images from the UK Biobank to learn the distribution of healthy brains. The model was then fine-tuned on a cohort of nearly 700 healthy FLAIR images coming from a CDW. We first ensured the good performance of our pre-trained model compared with the state-of-the-art using a widely used public dataset (MSSEG). We then validated it on our target task, age-related WMH detection, on ADNI3 and on a curated clinical dataset from a single-site neuroradiology department, for which we had manually delineated lesion masks. Next, we applied the fine-tuned $\beta$-VAE for anomaly detection in a CDW characterised by an exceptional heterogeneity in terms of hospitals, scanners and image quality. We found a correlation between the Fazekas scores extracted from the radiology reports and the volumes of the lesions detected by our model, providing a first insight into the performance of VAEs in a clinical setting. We also observed that our model was robust to image quality, which strongly varies in the CDW. However, despite these encouraging results, such approach is not ready for an application in clinical routine yet due to occasional failures in detecting certain lesions, primarily attributed to the poor quality of the images reconstructed by the VAE.
APA
Loizillon, S., Jacob, Y., Aurélien, M., Dormont, D., Colliot, O., Burgos, N. & Group, A.S.. (2024). Detecting Brain Anomalies in Clinical Routine with the $β$-VAE: Feasibility Study on Age-Related White Matter Hyperintensities. Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 250:903-917 Available from https://proceedings.mlr.press/v250/loizillon24a.html.

Related Material