SLCCC-Net: Hybrid steganography and AI system for secure cancer classification from histopathological images in internet of medical things applications.

Lung and colon cancer contribute to cancer-related deaths globally. Early detection through histopathological image analysis is pivotal in improving patient outcomes. However, challenges exist in ensuring the confidentiality and security of sensitive medical data and achieving accurate classification. There are dual issues of maintaining data privacy during transmission and achieving high accuracy in automated cancer classification in existing methods.•This work introduces a novel Secure Lung-Colon Cancer Classification Network (SLCCC-Net) approach, combining secure communication and advanced machine learning techniques to classify lung and colon cancer from histopathological images within a secured IoMT environment.•Here, the Two-Level Encryption Adopted Image Steganography (TLE-IS) method integrates Quantum Key Distribution (QKD) and Fully Homomorphic Encryption (FHE) for encrypting patient messages and medical images, ensuring both confidentiality and integrity of the data.•To enhance security, Hybrid-Inverse Wavelet Transform (HIWT)- based steganography embeds the encrypted data into histopathological images for secure transmission. On the receiver side, a Sand Cat Swarm Optimization with Genetic Weight Updating (SCSO-GWU) algorithm is employed for feature extraction from the stego image.•These extracted features are classified using an Interpretable Convolutional Neural Network (ICNN). Additionally, the inverse TLE-IS operation is performed to retrieve the original message.•The proposed SLCCC-Net demonstrates exceptional performance with an accuracy of 99.672 %, precision of 99.258 %, and recall rates of 99.900 % and 99.701 %, highlighting its effectiveness in lung and colon cancer classification.