Deep learning enables accurate soft tissue deformation estimationin vivo

Abstract

ABSTRACTImage-based deformation estimation is an important tool used in a variety of engineering problems, including crack propagation, fracture, and fatigue failure. These tools have been instrumental in biomechanics research where measuringin vitroandin vivotissue deformations help evaluate tissue health and disease progression. However, accurately measuring tissue deformationin vivois particularly challenging due to limited image signal-to-noise ratio. Therefore, we created a novel deep-learning approach for measuring deformation from a sequence ofin vivoimages calledStrainNet. Utilizing a training dataset that incorporates image artifacts,StrainNetwas designed to maximize performance in challengingin vivosettings. Artificially generated image sequences of human flexor tendons undergoing known deformations were used to compareStrainNetagainst two conventional image-based strain measurement techniques.StrainNetoutperformed the traditional techniques by nearly 90%. High-frequency ultrasound imaging was then used to acquire images of the flexor tendons engaged during contraction. OnlyStrainNetwas able to track tissue deformations under thein vivotest conditions. Findings revealed strong correlations between tendon deformation and contraction effort, highlighting the potential forStrainNetto be a valuable tool for assessing preventative care, rehabilitation strategies, or disease progression. Additionally, by using real-world data to train our model,StrainNetwas able to generalize and reveal important relationships between the effort exerted by the participant and tendon mechanics. Overall,StrainNetdemonstrated the effectiveness of using deep learning for image-based strain analysisin vivo.