Deep Learning‐Based OCT Segmentation for Stiffness Quantification in Evaluating Low‐Level Laser Therapy for Wound Healing

Background This study evaluated the short‐term biomechanical response of wound tissue following low‐level laser therapy (LLLT) by examining changes in skin stiffness, a surrogate biomechanical indicator of short‐term tissue response, across different limb regions. Methods A 660 nm LLLT protocol was applied to wound sites. Skin stiffness was quantified using optical coherence tomography (OCT) combined with an air‐jet indentation system, enabling non‐contact measurement of tissue deformation. For accurate layer‐specific assessment, a U‐Net–based model was employed to automate OCT image segmentation. Results The automated segmentation by the U‐Net model achieved a segmentation accuracy of 92%, facilitated precise segmentation of skin layers. LLLT significantly reduced skin stiffness after treatment, indicating an acute modulation of tissue compliance. Conclusion Short‐duration LLLT reduces skin stiffness immediately post‐treatment, indicating its potential as a non‐invasive intervention to modulate the biomechanical environment of wounds. Trial Registration: ClinicalTrials.gov identifier: NCT07177274