Inverse Problem for the Estimation of Skin Cancerous Region Parameters by Thermal Analysis

Malignant melanoma is the most aggressive skin cancer since it metastasizes rapidly. The existence of a quantitative, non-expensive, and non-invasive diagnostic screening tool for the evaluation of pigmented lesions would be invaluable for the early detection of malignant melanoma. Based on the fact that this cancerous skin lesion has higher metabolism, as well as an increased blood flow, it has been conjectured that it has higher temperature compared to surrounding healthy skin, temperature than can be measured by infrared thermography. Therefore, it is possible to explore the relationship between the characteristics (size and heat produced) of the neoplasm and the resulting temperature distribution on the skin surface to evaluate the development of the lesion. In our investigation the Dual Reciprocity Boundary Element Method (DRBEM) has been coupled to the Simulated Annealing Technique (SAT) in an inverse procedure to estimate simultaneously the depth and heat produced by skin malignant neoplasms at early stages. The SAT was used to estimate neoplasm parameters by minimizing an objective function that involves the skin surface temperature profiles obtained from simulated data to those obtained numerically by the DRBEM. Results are presented for neoplasms of Clark levels II–IV to demonstrate the feasibility of the proposed methodology.