Diagnosis of BCC by multiphoton laser tomography

Abstract

Background/purpose: Multiphoton Laser Tomography (MPT) is a non-linear optical technique that gives access to morphology and structure of both cells and extracellular matrix of the skin through the combination of autofluorescence imaging and second harmonic generation (SHG). The aim of this study was to identify MPT descriptors on ex vivo specimens of basal cell carcinoma (BCC) to assess the sensitivity and specificity of these criteria for the diagnosis of BCC and its differentiation from other skin tumours, inflammatory diseases and healthy skin.

Methods: In the preliminary study, MPT images referring to 24 BCCs and 24 healthy skin samples were simultaneously evaluated by three observers for the identification of features characteristic of BCC. In the main study, the presence/absence of the descriptors identified in the preliminary study was blindly evaluated on a test set, comprising 66 BCCs, 66 healthy skin samples and 66 skin lesions, including 23 nevi, 8 melanomas, 17 skin tumours and other skin lesions by 3 independent observers.

Results: In the preliminary study, three epidermal descriptors and six descriptors for BCC were identified. The latter included aligned elongated cells, double alignment of cells, cell nests with palisading and phantom islands. From the test set, 56 BCCs were correctly diagnosed, whereas in 10 cases the diagnosis was 'other lesions'. However, it was always possible to exclude the diagnosis of BCC in healthy skin and other lesion samples. Thus, overall sensitivity of the method was 84.85, whereas a specificity of 100% was observed with respect to both healthy skin and 'other lesions'.

Conclusions: This study describes new morphological descriptors of BCC enabling its characterization and its distinction from healthy skin and other skin lesions in ex vivo samples, and demonstrates for the first time that MPT represents a sensitive and specific technique for the diagnosis of BCC.