AI-Powered Imaging Tracks Wound Healing Under Skin in Real Time: OCT + Deep Learning in 2026

AI-Powered Imaging Tracks Wound Healing Under Skin in Real Time: OCT + Deep Learning in 2026

AI-powered imaging now tracks wound healing under the skin in real time, marking a transformative leap in regenerative medicine. Researchers from Duke University’s Pratt School of Engineering, in collaboration with Nokia Bell Labs, have developed a novel system that merges optical coherence tomography (OCT) with artificial intelligence to visualize and quantify subsurface tissue repair without invasive procedures. This breakthrough addresses a longstanding clinical challenge: the inability to accurately assess what occurs beneath the skin during healing.

How OCT Enables Subsurface Imaging

Custom-built OCT hardware captures high-resolution, cross-sectional images of tissue layers as wounds heal—penetrating up to 2mm beneath the skin’s surface. Unlike ultrasound or MRI, OCT provides micron-level resolution without ionizing radiation, making it ideal for continuous monitoring. The system’s speed allows for real-time imaging during hydrogel therapy applications, enabling immediate feedback on tissue response.

Deep Learning Models for Tissue Quantification

AI models trained on over 50,000 annotated biological datasets identify subtle patterns in collagen alignment, vascular density, and inflammatory cell migration. These deep learning OCT algorithms generate quantitative healing metrics—such as tissue thickness changes and neovascularization rates—with 94% accuracy compared to histology. Human analysts would require weeks to process what the AI analyzes in seconds.

Clinical Applications in Hydrogel Therapy

Duke’s Sharon Gerecht, chair of Biomedical Engineering, used the system to validate that stiffer hydrogels enhance healing outcomes by 40% compared to softer variants—a discovery previously obscured by subjective assessments. Clinicians now use the platform to tailor hydrogel compositions to individual patient biomarkers, optimizing regeneration timelines for diabetic ulcers and post-surgical wounds.

Non-Invasive Monitoring Reduces Biopsy Dependence

Traditional wound evaluation relies on painful biopsies and surface observations, which miss critical subsurface events. The AI-OCT system eliminates this gap, offering non-invasive monitoring that reduces patient discomfort and infection risk. Early trials show a 60% reduction in unnecessary interventions for chronic wounds.

Real-Time Diagnostics for Personalized Medicine

By generating digital biomarkers of healing, the technology enables dynamic treatment adjustments. For example, if neovascularization lags, clinicians can trigger growth factor boosts or adjust hydrogel stiffness via connected smart dressings. FDA regulatory pathways are underway, with pilot deployments planned in 12 trauma centers by Q4 2026.

As AI-powered imaging tracks wound healing under the skin in real time, it transforms subjective clinical judgment into data-driven precision. This innovation doesn’t just accelerate research—it redefines standards of care for millions living with non-healing wounds, offering hope through subsurface imaging, quantitative healing metrics, and intelligent therapy personalization.