Processing of Rare Skin Disease Multispectral Images

Rare disease has become an emerging public health priority in the European Union. The low prevalence leads to lack of knowledge and expertise in diagnosing and treating these conditions. New approaches should be found to identify, assess, and diagnose symptoms associated with rare disease. Neurofibromatosis type I, Fabry disease and Pseudoxanthoma elasticum are rare multi-organ diseases with characteristic skin manifestations. In this study a skin imaging approach previously used for skin cancer imaging was used to assess skin lesions associated with these rare diseases. The device combines diffuse reflectance imaging with 526, 663 and 964 nm illumination and autofluorescence imaging with 405 nm excitation. Image pre-processing was done using MATLAB and ImageJ. Imaging parameters were extracted using Python. The parameters were compared to similar skin lesions which are not associated with disease. A multiparameter analysis was shown to differentiate lesions characteristic of Neurofibromatosis type I from similar lesions, however, this approach was not as effective for differentiating Pseudoxanthoma elasticum plaques and Fabry disease induced angiokeratoma from similar lesions.