Enhanced diagnosis of skin lesions through torsional wave propagation and probabilistic inverse problem algorithms: An experimental study

Abstract

This study introduces an innovative method for detecting skin lesions using Probabilistic Inverse Problem (PIP) methodologies combined with Torsional Wave Elastography (TWE) for rheological modeling. Its primary aim was to identify the most effective way to measure mechanical properties of the skin. The Spring Pot (SP) model and the Kelvin-Voigt Fractional Derivative (KVFD) were compared to find the best rheological model for skin tissue simulation. A numerical model based on these rheological models was developed and experimental data was collected from 15 patients aged 46 to 72 years using a TWE sensor. This data provided crucial information for reconstructing and diagnosing skin tissue parameters using PIP methodologies. The results, comparing experimental to numerical data and pathological tissue to either symmetrical healthy (control) tissue or adjacent healthy tissue, show that this approach effectively reconstructs skin tissue parameters and diagnoses skin lesions. The SP model revealed values from 7.5 to 12 Pa.s for control tissue and 14 to 67.5 Pa.s for pathological tissue, while the KVFD model showed values of 3.5 to 9.5 Pa.s for control tissue and 10 to 46 Pa.s for pathological tissue. For the KVFD model, control tissue had values of 0.93 to 8.7 kPa, and pathological tissue had values of 7.04 to 197.56 kPa. The SP model, which requires only two parameters compared to KVFD’s three, proved more efficient and straightforward. Consequently, this work contributes significantly to dermatology, offering important insights for both academic research and practice.