Abstract
Deep learning has been extensively applied to medical imaging tasks over the past years, achieving outstanding results. However, the obscure reasoning of the models and the lack of supportive evidence causes both clinicians and patients to distrust the models' predictions, hindering their adoption in clinical practice. In recent years, the research community has focused on developing explanations capable of revealing a model's reasoning. Among various types of explanations, example-based explanations emerged as particularly intuitive for medical practitioners. Despite the intuitiveness and wide development of example-based explanations, no work provides a comprehensive review of existing example-based explainability works in the medical image domain. In this work, we review works that provide example-based explanations for medical imaging tasks, reflecting on their strengths and limitations. We identify the absence of objective evaluation metrics, the lack of clinical validation and privacy concerns as the main issues that hinder the deployment of example-based explanations in clinical practice. Finally, we reflect on future directions contributing towards the deployment of example-based explainability in clinical practice.

Abstract
In the study, a new method for analyzing Electrocardiogram (ECG) signals is suggested, which is vital for detecting and treating heart diseases. The technique focuses on improving ECG signal classification, particularly in identifying different heart conditions like arrhythmias and myocardial infarctions. An enhanced version of the differential evolution (DE) algorithm integrated with neural networks is leveraged to classify these signals effectively. The process starts with preprocessing and extracting key features from ECG signals. These features are then processed by a multi-layer perceptron (MLP), a common neural network for ECG analysis. However, traditional MLP training methods have limitations, such as getting trapped in suboptimal solutions. To overcome this, an advanced DE algorithm is used, incorporating a partition-based strategy, opposition-based learning, and local search mechanisms. This improved DE algorithm optimizes the MLP by fine-tuning its weights and biases, using them as starting points for further refinement by the Gradient Descent with Momentum (GDM) local search algorithm. Extensive experiments demonstrate that this novel training approach yields better results than the traditional method. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

Abstract
Clinical diagnosis of syndromes benefits strongly from objective facial phenotyping. This study introduces a novel approach to enhance clinical diagnosis through the development and exploration of a low-dimensional metric space referred to as the clinical face phenotypic space (CFPS). As a facial matching tool for clinical genetics, such CFPS can enhance clinical diagnosis. It helps to interpret facial dysmorphisms of a subject by placing them within the space of known dysmorphisms. In this paper, a triplet loss-based autoencoder developed by geometric deep learning (GDL) is trained using multi-task learning, which combines supervised and unsupervised learning approaches. Experiments are designed to illustrate the following properties of CFPSs that can aid clinicians in narrowing down their search space: a CFPS can 1) classify syndromes accurately, 2) generalize to novel syndromes, and 3) preserve the relatedness of genetic diseases, meaning that clusters of phenotypically similar disorders reflect functional relationships between genes. The proposed model consists of three main components: an encoder based on GDL optimizing distances between groups of individuals in the CFPS, a decoder enhancing classification by reconstructing faces, and a singular value decomposition layer maintaining orthogonality and optimal variance distribution across dimensions. This allows for the selection of an optimal number of CFPS dimensions as well as improving the classification capacity of the CFPS, which outperforms the linear metric learning baseline in both syndrome classification and generalization to novel syndromes. We further proved the usefulness of each component of the proposed framework, highlighting their individual impact. From a clinical perspective, the unique combination of these properties in a single CFPS results in a powerful tool that can be incorporated into current clinical practices to assess facial dysmorphism.

Abstract
[No abstract available]

Abstract

Abstract
This study presents an explainable content-based image retrieval system for predicting post-surgical aesthetic outcomes in breast cancer patients, comparing state-of-theart vision transformers, convolutional neural networks, and B-cos architectures. Results show that vision transformers, particularly GC ViT and DaViT, outperform convolutional neural networks and B-cos architectures, achieving an adjusted discounted cumulative gain of up to 80.18%. This superior performance is attributed to their ability to model long-range dependencies while effectively capturing local information. Bcos networks underperform (64.28-70.19% adjusted discounted cumulative gain), likely due to oversimplified feature alignment unsuitable for clinical tasks. Explainability analysis using Integrated Gradients reveals that models primarily focus on breast regions but occasionally attend to irrelevant features (e.g., arm positioning, leading to retrieval errors and highlighting a semantic gap between learned visual similarities and clinical relevance. Future work aims to integrate anatomical segmentation and ensemble learning methods to enhance clinical alignment and address attention inaccuracies. Clinical Relevance-The content-based image retrieval system developed in this study aids clinicians by supporting surgical outcome prediction in breast cancer patients and streamlining the traditionally time-intensive task of manually identifying similar reference images for patient consultation. © 2025 IEEE.