Personalized digital pipeline for basal cell carcinoma delineation using hyperspectral imaging.

Abstract

Basal cell carcinoma (BCC) is the most common form of skin cancer with the current diagnostic procedures relying on surgical excisions and histopathological analyses. Incomplete removal of the tumor necessitates additional surgery, which is both costly and time-consuming. To address this, non-invasive diagnostic techniques capable of determining the dimensions of skin cancer preoperatively are urgently needed. This work presents the development of an automated diagnostic pipeline for delineating BCC, using hyperspectral imaging (HSI), combined with machine learning (ML). HSI is a non-invasive, high-resolution imaging technique capable of differentiating tissue structures based on their molecular composition. ML is a branch of AI where algorithms learn patterns from data to make predictions without being explicitly programmed. The proposed pipeline is designed to adapt for individual patients, as universal skin cancer characteristic signals are challenging to define. The pipeline predicted tumor sizes with high accuracy, as validated by histopathological measurements in the 21 samples evaluated. The median absolute errors were low (0.25-0.30 mm), meaning that the predicted tumor dimensions were within the clinically accepted safety margins for radical excisions (≤0.4 mm). This work is the first to use unsupervised ML for tumor delineation in BCC using HSI, which enables personalized preoperative assessment and facilitates surgical precision.