Dual-imaging system for burn depth diagnosis

Currently, determination of burn depth and healing outcomes has been limited to subjective assessment or a single modality, e.g., laser Doppler imaging. Such measures have proven less than ideal. Recent developments in other non-contact technologies such as optical coherence tomography (OCT) and pulse speckle imaging (PSI) offer the promise that an intelligent fusion of information across these modalities can improve visualization of burn regions thereby increasing the sensitivity of the diagnosis. In this work, we combined OCT and PSI images to classify the degree of burn (superficial, partial-thickness and full-thickness burns). Algorithms were developed to integrate and visualize skin structure (with and without burns) from the two modalities. We have completed the proposed initiatives by employing a porcine burn model and compiled results that attest to the utility of our proposed dual-modal fusion approach. Computer-derived data indicating the varying burn depths were validated through immunohistochemical analysis performed on burned skin tissue. The combined performance of OCT and PSI modalities provided an overall ROC-AUC = 0.87 (significant at p < 0.001) in classifying different burn types measured after 1-h of creating the burn wounds. Porcine model studies to assess feasibility of this dual-imaging system for wound tracking are underway. (C) 2013 Elsevier Ltd and ISBI. All rights reserved.