This cluster reflects an emerging trend toward low-cost, non-invasive, image-based diagnostics for peripheral neuropathies, exemplified by the New Energy Vision camera system for Carpal Tunnel Syndrome (CTS) detection. Rather than relying solely on established but resource-intensive gold standards—Nerve Conduction Tests, Semmes-Weinstein Monofilament Testing, and the Boston Carpal Tunnel Questionnaire—researchers are validating a multispectral RGB imaging approach that captures skin color and texture changes plausibly linked to median nerve compression. This signals a broader trajectory in neuropathy diagnostics: shifting from electrophysiological and symptom-based assessment toward computer-vision and machine-learning-assisted screening tools that could be deployed in primary care or remote settings.
Mechanistically, the approach hinges on the premise that median nerve dysfunction produces detectable microvascular and dermal texture alterations in the innervated palm territory, distinguishable from the ulnar-innervated (nerve-normal) region within the same hand. Quantitative descriptors—Haralick texture features and red proportion (a color-based proxy for perfusion/skin change)—are shown to differ significantly between median nerve-damaged and ulnar nerve-normal palm areas, providing a within-subject biological rationale that strengthens the imaging biomarker's validity beyond simple case-control contrasts.
The two-part study design demonstrates a translational pipeline: Part 1 establishes feasibility by training a Support Vector Machine classifier on images from 103 participants (50 controls, 53 CTS patients), achieving 93.33% accuracy and 81.79% cross-validation accuracy, with a balanced confusion matrix ([[14,1],[1,14]]) indicating symmetric sensitivity and specificity. Part 2 extends the mechanistic evidence by focusing intra-patient comparisons in 32 CTS patients, reinforcing that texture and color signatures are localized to the nerve-affected region rather than generalized hand differences. This staged validation—paired with correlation against conventional diagnostic standards—positions multispectral imaging and machine learning classification as a promising adjunct or triage tool, part of a wider movement toward accessible, camera-based biomarker discovery for compressive neuropathies and potentially other peripheral nerve disorders.