Differential diagnosis of cutaneous carcinomas by infrared spectral micro-imaging combined with pattern recognition.

Non-melanoma skin cancer includes basal cell carcinoma (BCC), squamous cell carcinoma (SCC) and Bowen's disease. The differential diagnosis of these lesions is sometimes difficult and relies on the histopathological examination of surgical specimens. However, a precise differential diagnosis is crucial for an accurate therapy and thus better patient care. FTIR spectral micro-imaging was applied directly on formalin-fixed paraffin-embedded samples of non-melanoma skin cancers. Chemometric and multivariate statistical analyses were developed to generate an automated IR-based histology without any chemical dewaxing. Different prediction models were developed using linear discriminant analysis combined with data reduction by Principal Component Analysis (PCA) or by wavenumber selection using statistical tests or genetic algorithms. Pseudo-colour maps were reconstructed and compared to conventional histology procedures. High correlation was obtained between the prediction maps and the histology which proves the great potential of FTIR spectroscopy for the differential diagnosis of skin carcinomas.