Artificial Intelligence in Medicine最新文献

Conventional performance metrics in clinical decision support systems, such as accuracy or sensitivity, fail to reflect the reliability of individual predictions-an essential concern for clinicians operating in high-stakes environments. We introduce a calibration-informed framework featuring two novel metrics: the Local Predictive Value (LPV) and the Credible Predictive Value (CPV). LPV estimates the empirical reliability of a prediction by assessing the observed correctness frequency in the neighborhood of its confidence score. CPV refines this estimate using a Bayesian approach, integrating global predictive values as priors to produce a posterior distribution over correctness probabilities. LPV offers a descriptive, data-driven view of local reliability, while CPV provides a belief-adjusted estimate that mitigates overfitting to sparse local data. Applied to benchmark medical imaging datasets, these metrics yielded locally adaptive, interpretable reliability estimates. Divergences between LPV and CPV identified cases where local evidence was insufficient or misleading, highlighting how Bayesian smoothing improves stability against sparse or misleading local evidence. By combining local calibration with Bayesian inference, LPV and CPV advance the development of medical AI systems that are not only accurate but also interpretable and trustworthy at the individual case level.

Accurate detection of the QRS complex, a crucial reference for heartbeat localization in electrocardiogram (ECG) signals, remains inadequate in wearable ECG devices due to complex noise interference. In this study, we propose a novel QRS complex detection method based on dynamic Bayesian network (DBN), integrating the probability distribution of RR intervals. Unlike methods focusing solely on ECG waveforms, our approach explicitly integrates ECG waveform and heart rhythm information into a unified probability model, enhancing noise robustness. Additionally, an unsupervised parameter optimization using expectation maximization (EM) adapts to individual differences of patients. Furthermore, several simplification strategies improve reasoning efficiency, and an online detection mode enables real-time applications. Our method outperforms other state-of-the-art QRS detection methods, including deep learning (DL) methods, on noisy datasets. In conclusion, the proposed DBN-based QRS detection algorithm demonstrates outstanding accuracy, noise robustness, generalization ability, real-time capability, and strong scalability, indicating its potential application in wearable ECG devices.

Fundus images are widely used in early retinopathy examination to prevent visual impairment caused by retinopathy. The retinopathy examination process based on fundus images can be mainly summarized in three steps: (1) ophthalmologists obtain comprehensive fundus information by jointly analyzing multi-view fundus images; (2) ophthalmologists obtain complementary lesion information by contrastingly analyzing multi-modal fundus images; (3) ophthalmologists diagnose retinopathy categories and write specialized fundus reports. To simulate the clinical fundus image examination process, we introduce an efficient multi-view and multi-modal fundus image joint ancillary diagnosis framework that can simultaneously accept fundus images of different views and modalities for pathology classification and symptom report generation tasks. In our framework, we propose jointly employing self-attention in intra-view local and inter-view sparse global windows to extract comprehensive fundus information among different views. We propose a multi-modal fusion transformer via shunted multi-scale cross-attention to model lesions of various scales by splitting attention granularity at query and queried modalities to fuse complementary lesion information among different modalities. The experimental results of retinopathy classification and report generation tasks indicate that our proposed method is superior to other benchmarking methods, achieving a classification accuracy of 83.96% and a report generation CIDEr of 0.934.

As artificial intelligence (AI) is increasingly integrated into medical diagnostics, it is essential that predictive models provide not only accurate outputs but also reliable estimates of uncertainty. In clinical applications, where decisions have significant consequences, understanding the confidence behind each prediction is as critical as the prediction itself. Uncertainty modelling plays a key role in improving trust, guiding decision-making, and identifying unreliable outputs, particularly under dataset shift or in out-of-distribution settings. The primary aim of uncertainty metrics is to align model confidence closely with actual predictive performance, ensuring confidence estimates dynamically adjust to reflect increasing errors or decreasing reliability of predictions. This study investigates how different ensemble learning strategies affect both performance and uncertainty estimation in a clinically relevant task: classifying Normal, Mild Cognitive Impairment, and Dementia from electroencephalography (EEG) data. We evaluated the performance and uncertainty of ensemble methods and Monte Carlo dropout on a large EEG dataset. The models were assessed in three settings: (1) in-distribution performance on a held-out test set, (2) generalisation to three out-of-distribution datasets, and (3) performance under gradual, EEG-specific dataset shifts simulating noise, drift, and frequency perturbation. Ensembles consisting of multiple independently trained models, such as deep ensembles, consistently achieved higher performance in both the in-distribution test set and the out-of-distribution datasets. These models also produced more informative and reliable uncertainty estimates under various types of EEG dataset shifts. These results highlight the benefits of ensemble diversity and independent training to build robust and uncertainty-aware EEG classification models. The findings are particularly relevant for clinical applications, where reliability under distribution shift and transparent uncertainty are essential for safe deployment.

Epilepsy is a chronic brain disorder characterized by recurrent seizures resulting from abnormal brain cell activity. The unpredictability of these seizures underscores the criticality of anticipating and promptly addressing them to enhance the patient's overall quality of life. Electroencephalography (EEG) is a frequently employed technique for seizure prediction, leveraging its economic viability and high temporal resolution. However, the complexity of EEG signals has driven interest in machine learning and deep learning for automated seizure prediction systems. Nevertheless, conventional approaches that employ predefined methodologies for analyzing seizures may not adequately account for the variability in spectral and spatial characteristics among patients. To address these limitations and present a more effective and interpretable approach, we introduce the patient-tailored seizure prediction network (PSP-Net) for adaptive spectral-spatial-temporal EEG feature representation learning. PSP-Net combines patient-tailored bandpass filters, a patient-tailored spatial coupling matrix, and an attentive temporal convolution network-based feature extractor in a unified framework to automatically extract patient-specific spectral-spatial-temporal features from EEG data. The proposed method achieves state-of-the-art performance on multiple publicly available seizure datasets, which highlights its potential as a reliable tool for personalized clinical applications.