The kidney is an abdominal organ in the human body that supports filtering excess water and waste from the blood. Kidney diseases generally occur due to changes in certain supplements, medical conditions, obesity, and diet, which causes kidney function and ultimately leads to complications such as chronic kidney disease, kidney failure, and other renal disorders. Combining patient metadata with computed tomography (CT) images is essential to accurately and timely diagnosing such complications. Deep Neural Networks (DNNs) have transformed medical fields by providing high accuracy in complex tasks. However, the high computational cost of these models is a significant challenge, particularly in real-time applications. This paper proposed SpinalZFNet, a hybrid deep learning approach that integrates the architectural strengths of Spinal Network (SpinalNet) with the feature extraction capabilities of Zeiler and Fergus Network (ZFNet) to classify kidney disease accurately using CT images. This unique combination enhanced feature analysis, significantly improving classification accuracy while reducing the computational overhead. At first, the acquired CT images are pre-processed using a median filter, and the pre-processed image is segmented using Efficient Neural Network (ENet). Later, the images are augmented, and different features are extracted from the augmented CT images. The extracted features finally classify the kidney disease into normal, tumor, cyst, and stone using the proposed SpinalZFNet model. The SpinalZFNet outperformed other models, with 99.9% sensitivity, 99.5% specificity, precision 99.6%, 99.8% accuracy, and 99.7% F1-Score in classifying kidney disease. Graphical Abstract

Deep learning models have been instrumental in extracting critical indicators for breast cancer diagnosis - the prevalent malignancy among women worldwide - from baseline magnetic resonance imaging. However, many existing models do not fully leverage the rich spatial information available in the 3D structure of medical imaging data, potentially overlooking important contextual details. This develops an explainable deep learning framework for classifying breast cancer that leverages the complete 3D and provides classification results alongside visual explanations of the decision-making process. The preprocessing pipeline is fed with 3D sequences containing ‘tumour’ and ‘non-tumour’ regions. It includes a 3D Adaptive Unsharp Mask (AUM) filter to reduce noise and augment image class, followed by normalisation and data augmentation. Classification is then achieved by training an augmented ResNet150 model. Three explainable artificial intelligence (XAI) techniques, including Shapley Additive Explanations, 3D Gradient-Weighted Class Activation Mapping, and Contextual Importance and Utility, are employed to provide improved interpretability. The model demonstrates state-of-the-art performance over the QIN-BREAST dataset, achieving testing accuracies of 98.861% for ‘tumours’ and 99.447% for ‘non-tumours’, as well as over the Duke Breast Cancer Dataset, where it achieves 99.104% for ‘tumours’ and 99.753% for ‘non-tumours’, while offering enhanced interpretability through XAI methods.

The early detection of breast cancer plays a critical role in improving survival rates and facilitating precise medical interventions. Therefore, the automated identification of breast abnormalities becomes paramount, significantly enhancing the prospects of successful treatment outcomes. To address this imperative, our research leverages multiple modalities such as MRI, CT, and mammography to detect and screen for breast cancer. Our proposed methodology integrates innovative modules such as the Distributed Load-based Dilated Convolution (DLM) and Multi-Head Dense Dilated Convolution (MHD) modules. These modules are designed to capture multi-scale information crucial for accurate tumor detection while simultaneously minimizing the parameter count, thus enhancing computational efficiency. Additionally, an Optimal Multi-Level Attention (OMLA) and Key Aspect Attention (KA) modules are adept at aggregating task-relevant information effectively, contributing to improved model performance and accuracy. Our experimental evaluation utilizing multimodal clinical data underscores the effectiveness of the proposed method, achieving an impressive Dice coefficient of 86.756%, 96.931%, and 97.813% for segmentation and classification accuracy of 89.982%, 97.028%, and 98.760% for Mammogram, CT, and MRI datasets, respectively. These results validate the efficacy and potential clinical utility of our approach in enhancing breast cancer detection and diagnosis. We introduce a novel framework for the classification and segmentation of multimodal medical images acquired by a variety of modalities for breast cancer detection and identification. In addition to improving segmentation accuracy, our proposed Computer-Aided Diagnosis (CAD) method utilizes the segmentation mask as an additional channel feature to enhance the classification outcomes of breast tumors. Additionally, we have meticulously curated a comprehensive CT breast cancer dataset, comprising various clinical data types related to breast tumors, including detailed segmentation and classification annotations. This dataset surpasses previous efforts in its depth and breadth, providing a robust foundation for future research endeavors.

Accurate breast tumor diagnosis using MRI is essential for effective treatment planning. While traditional methods rely on radiologists’ expertise, the adoption of AI is often limited by concerns over decision-making transparency. The role of Explainable AI (XAI) is critical in addressing these concerns, as it helps make AI-based decisions more interpretable for clinicians. Advances in deep learning, particularly through convolutional neural networks (CNNs), have led to highly effective models for digital image analysis. In this work, an advanced preprocessing module—employing Fourier transform and edge-detection techniques— optimizes the input data for the EfficientNetB7 architecture. EfficientNetB7 first extracts localized textures and patterns through its MBConv and Squeeze-and-Excitation blocks. The transformer component then employs multi-head self-attention to integrate distant image patches, effectively capturing global context. A custom dense layer generates three distinct output probabilities—representing normal, benign, or malignant tissues— thereby enhancing clinical interpretability. Finally, Grad-CAM and Eigen-CAM methodologies are applied to highlight the regions most influential to the model’s decisions, improving both transparency and practical reliability. The model’s performance was assessed using the publicly accessible RIDER and QIN-BREAST datasets, achieving 99.39% and 99.51% accuracy, respectively.