Polarimetry based partial least square classification of ex vivo healthy and basal cell carcinoma human skin tissues

- Ex vivo basal cell carcinoma (BCC) tissue samples were analyzed by optical polarimetry.
- Depolarization and retardance for BCC samples were significantly lower (p < 0.001) compared to healthy tissue group.
- Histopathology was used to interpret polarimetry results.
- Partial least square analyses classified the samples with 100% accuracy, sensitivity and specificity.

Optical polarimetry was employed for assessment of ex vivo healthy and basal cell carcinoma (BCC) tissue samples from human skin. Polarimetric analyses revealed that depolarization and retardance for healthy tissue group were significantly higher (p < 0.001) compared to BCC tissue group. Histopathology indicated that these differences partially arise from BCC-related characteristic changes in tissue morphology. Wilks lambda statistics demonstrated the potential of all investigated polarimetric properties for computer assisted classification of the two tissue groups. Based on differences in polarimetric properties, partial least square (PLS) regression classified the samples with 100% accuracy, sensitivity and specificity. These findings indicate that optical polarimetry together with PLS statistics hold promise for automated pathology classification.