A self-supervised learning approach for registration agnostic imaging models with 3D brain CTA

Yingjun Dong; Samiksha Pachade; Xiaomin Liang; Sunil A Sheth; Luca Giancardo

doi:10.1016/j.isci.2024.109004

A self-supervised learning approach for registration agnostic imaging models with 3D brain CTA

iScience. 2024 Feb 1;27(3):109004. doi: 10.1016/j.isci.2024.109004. eCollection 2024 Mar 15.

Authors

Yingjun Dong¹, Samiksha Pachade¹, Xiaomin Liang¹, Sunil A Sheth², Luca Giancardo^{1

3}

Affiliations

¹ McWilliams School of Biomedical Informatics, University of Texas Health Science Center at Houston, 7000 Fannin Street, Houston, TX, USA.
² Department of Neurology, McGovern Medical School, University of Texas Health Science Center at Houston, 6431 Fannin Street, Houston, TX, USA.
³ Institute for Stroke and Cerebrovascular Diseases, University of Texas Health Science Center at Houston, 7000 Fannin Street, Houston, TX, USA.

Abstract

Deep learning-based neuroimaging pipelines for acute stroke typically rely on image registration, which not only increases computation but also introduces a point of failure. In this paper, we propose a general-purpose contrastive self-supervised learning method that converts a convolutional deep neural network designed for registered images to work on a different input domain, i.e., with unregistered images. This is accomplished by using a self-supervised strategy that does not rely on labels, where the original model acts as a teacher and a new network as a student. Large vessel occlusion (LVO) detection experiments using computed tomographic angiography (CTA) data from 402 CTA patients show the student model achieving competitive LVO detection performance (area under the receiver operating characteristic curve [AUC] = 0.88 vs. AUC = 0.81) compared to the teacher model, even with unregistered images. The student model trained directly on unregistered images using standard supervised learning achieves an AUC = 0.63, highlighting the proposed method's efficacy in adapting models to different pipelines and domains.

Keywords: Classification Description Medical imaging; Machine learning; Neuroanatomy.