Raw Data for Real-time, In Vivo Skin Cancer Diagnostics by Laser Induced Plasma Spectroscopy Combined with Deep Learning-based Diagnostic Algorithm
Description
In this study, we developed a real-time, minimally invasive, in vivo skin cancer diagnostic method utilizing non-discrete molecular Laser Induced Plasma Spectroscopy (LIPS) combined with deep neural network (DNN)-based diagnostic algorithm. LIPS has the capability of minimally invasively extracting biochemical information of skin lesion using an ultra-short pulsed laser. The objective of this in vivo study was to validate the practical applicability of LIPS and DNN-based diagnostic device to discriminate skin cancers including basal cell carcinoma (n=186), squamous cell carcinoma (n=93) and melanoma (n=14) from various benign lesions. In vivo LIPS spectra were acquired from total 293 skin cancers and 332 benign lesions in multi-site clinical study. The raw data of analysed lesions is listed here.