X-CaD: Explainable AI for Skin Cancer Diagnosis in Healthcare 4.0 Telesurgery

The advent of healthcare 4.0 has catalyzed a paradigm shift in medical practices, ushering in innovative approaches such as telesurgery, a groundbreaking method for remote patient surgery and monitoring. This transformative technique extends beyond traditional surgeries, finding application in dermatological procedures. The success of telesurgery in skin-related surgeries hinges on accurate and efficient skin cancer detection using dermoscopic images. Recognizing the inherent complexities in interpreting deep learning models, especially in the context of healthcare, Explainable Artificial Intelligence (X-AI) becomes imperative. In this context, we propose a novel CNN-powered X-AI mechanism i.e., X-CaD, tailored for skin cancer detection in telesurgery environments, leveraging ResNet and MobileNet for feature extraction. To enhance interpretability and bridge the gap between model predictions and clinical decision-making, we employ X-AI techniques such as Local Interpretable Model-agnostic Explanations (LIME) and Integrated Gradient (IG). LIME provides granular insights into model predictions, elucidating decision-making processes, while IG offers a comprehensive view of feature attributions. X-CaD relies on the synergistic integration of advanced CNN architectures, i.e. ResNet and MobileNet along with X-AI techniques to identify skin cancer accurately and provide clinicians with clear insights. The effective impact of X-CaD is evaluated through the observed loss and accuracy values for the DL models, and heat map outputs for X-AI. This represents a significant advancement in the integration of state-of-the-art technology and healthcare, offering a dependable telesurgery solution for the diagnosis of skin cancer in surgical procedures.