NPJ Digital Medicine最新文献

{"title":"Multi-cohort machine learning identifies predictors of cognitive impairment in Parkinson’s disease","authors":"Rebecca Ting Jiin Loo, Lukas Pavelka, Graziella Mangone, Fouad Khoury, Marie Vidailhet, Jean-Christophe Corvol, Enrico Glaab","doi":"10.1038/s41746-025-01862-1","DOIUrl":"https://doi.org/10.1038/s41746-025-01862-1","url":null,"abstract":"<p>Cognitive impairment is a frequent complication of Parkinson’s disease (PD), affecting up to half of newly diagnosed patients. To improve early detection and risk assessment, we developed machine learning models using clinical data from three independent PD cohorts, which are (LuxPARK, PPMI, ICEBERG). Models were trained to predict mild cognitive impairment (<i>PD-MCI</i>) and subjective cognitive decline (<i>SCD</i>) using Explainable Artificial Intelligence (XAI) for classification and time-to-event analysis. Multi-cohort models showed greater performance stability over single-cohort models, while retaining competitive average performance. Age at diagnosis and visuospatial ability were identified as key predictors. Significant sex differences observed highlight the importance of considering sex-specific factors in cognitive assessment. Men were more likely to report <i>SCD</i>. Our findings highlight the potential of multi-cohort machine learning for early identification and personalized management of cognitive decline in PD.</p>","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"214 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2025-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144712173","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AI-driven preclinical disease risk assessment using imaging in UK biobank","authors":"Dmitrii Seletkov, Sophie Starck, Tamara T. Mueller, Yundi Zhang, Lisa Steinhelfer, Daniel Rueckert, Rickmer Braren","doi":"10.1038/s41746-025-01771-3","DOIUrl":"https://doi.org/10.1038/s41746-025-01771-3","url":null,"abstract":"<p>Identifying disease risk and detecting disease before clinical symptoms appear are essential for early intervention and improving patient outcomes. In this context, the integration of medical imaging in a clinical workflow offers a unique advantage by capturing detailed structural and functional information. Unlike non-image data, such as lifestyle, sociodemographic, or prior medical conditions, which often rely on self-reported information susceptible to recall biases and subjective perceptions, imaging offers more objective and reliable insights. Although the use of medical imaging in artificial intelligence (AI)-driven risk assessment is growing, its full potential remains underutilized. In this work, we demonstrate how imaging can be integrated into routine screening workflows, in particular by taking advantage of neck-to-knee whole-body magnetic resonance imaging (MRI) data available in the large prospective study UK Biobank. Our analysis focuses on three-year risk assessment for a broad spectrum of diseases, including cardiovascular, digestive, metabolic, inflammatory, degenerative, and oncologic conditions. We evaluate AI-based pipelines for processing whole-body MRI and demonstrate that using image-derived radiomics features provides the best prediction performance, interpretability, and integration capability with non-image data.</p>","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"29 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2025-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144712316","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The role of face regions in remote photoplethysmography for contactless heart rate monitoring","authors":"Maksym Bondarenko, Carlo Menon, Mohamed Elgendi","doi":"10.1038/s41746-025-01814-9","DOIUrl":"https://doi.org/10.1038/s41746-025-01814-9","url":null,"abstract":"<p>Heart rate (HR) estimation is crucial for early cardiovascular diagnosis, continuous monitoring, and various health applications. While electrocardiography (ECG) remains the gold standard, its discomfort and impracticality for continuous use have spurred the development of non-contact methods like remote photoplethysmography (rPPG). This systematic review (PROSPERO: CRD 42024592157) examines 70 studies to assess the impact of Region of Interest (ROI) selection on HR estimation accuracy. Most methods (36.8%) use the holistic face, while forehead and cheek areas (24.5% and 21.7%) show superior accuracy. Machine learning-based approaches outperform traditional methods under motion artifacts and poor lighting, achieving Mean Absolute Error and Root Mean Square Error below 1.0 for some datasets. Combining multiple patches improves performance, though increasing ROIs beyond 60 patches results in diminishing returns and higher computational complexity. These findings highlight the significance of ROI optimization for robust rPPG-based HR estimation.</p>","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"21 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2025-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144710632","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Personalized federated learning for predicting disability progression in multiple sclerosis using real-world routine clinical data.","authors":"Ashkan Pirmani,Edward De Brouwer,Ádám Arany,Martijn Oldenhof,Antoine Passemiers,Axel Faes,Tomas Kalincik,Serkan Ozakbas,Riadh Gouider,Barbara Willekens,Dana Horakova,Eva Kubala Havrdova,Francesco Patti,Alexandre Prat,Alessandra Lugaresi,Valentina Tomassini,Pierre Grammond,Elisabetta Cartechini,Izanne Roos,Cavit Boz,Raed Alroughani,Maria Pia Amato,Katherine Buzzard,Jeannette Lechner-Scott,Joana Guimarães,Claudio Solaro,Oliver Gerlach,Aysun Soysal,Jens Kuhle,Jose Luis Sanchez-Menoyo,Daniele Spitaleri,Tunde Csepany,Bart Van Wijmeersch,Radek Ampapa,Julie Prevost,Samia J Khoury,Vincent Van Pesch,Nevin John,Davide Maimone,Bianca Weinstock-Guttman,Guy Laureys,Pamela McCombe,Yolanda Blanco,Ayse Altintas,Abdullah Al-Asmi,Justin Garber,Anneke Van der Walt,Helmut Butzkueven,Koen de Gans,Csilla Rozsa,Bruce Taylor,Talal Al-Harbi,Attila Sas,Cecilia Rajda,Orla Gray,Danny Decoo,William M Carroll,Allan G Kermode,Marzena Fabis-Pedrini,Deborah Mason,Angel Perez-Sempere,Mihaela Simu,Neil Shuey,Bhim Singhal,Marija Cauchi,Todd A Hardy,Sudarshini Ramanathan,Patrice Lalive,Carmen-Adella Sirbu,Stella Hughes,Tamara Castillo Trivino,Liesbet M Peeters,Yves Moreau","doi":"10.1038/s41746-025-01788-8","DOIUrl":"https://doi.org/10.1038/s41746-025-01788-8","url":null,"abstract":"Early prediction of disability progression in multiple sclerosis (MS) remains challenging despite its critical importance for therapeutic decision-making. We present the first systematic evaluation of personalized federated learning (PFL) for 2-year MS disability progression prediction, leveraging multi-center real-world data from over 26,000 patients. While conventional federated learning (FL) enables privacy-aware collaborative modeling, it remains vulnerable to institutional data heterogeneity. PFL overcomes this challenge by adapting shared models to local data distributions without compromising privacy. We evaluated two personalization strategies: a novel AdaptiveDualBranchNet architecture with selective parameter sharing, and personalized fine-tuning of global models, benchmarked against centralized and client-specific approaches. Baseline FL underperformed relative to personalized methods, whereas personalization significantly improved performance, with personalized FedProx and FedAVG achieving ROC-AUC scores of 0.8398 ± 0.0019 and 0.8384 ± 0.0014, respectively. These findings establish personalization as critical for scalable, privacy-aware clinical prediction models and highlight its potential to inform earlier intervention strategies in MS and beyond.","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"114 1","pages":"478"},"PeriodicalIF":15.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144701079","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Equity-enhanced glaucoma progression prediction from OCT with knowledge distillation","authors":"Sulaiman O. Afolabi, Leila Gheisi, Jing Shan, Lucy Q. Shen, Mengyu Wang, Min Shi","doi":"10.1038/s41746-025-01884-9","DOIUrl":"https://doi.org/10.1038/s41746-025-01884-9","url":null,"abstract":"<p>Glaucoma is a progressive disease that can lead to permanent vision loss, making progression prediction vital for guiding effective treatment. Deep learning aids progression prediction but may yield unequal outcomes across demographic groups. We proposed a model called FairDist, which utilized baseline optical coherence tomography scans to predict glaucoma progression. An equity-aware EfficientNet was trained for glaucoma detection, which was then adapted for progression prediction with knowledge distillation. Model accuracy was measured by the AUC, Sensitivity, Specificity, and equity was assessed using equity-scaled AUC, which adjusts AUC by accounting for subgroup disparities. The mean deviation, fast progression, and total deviation pointwise progression were explored in this work. For both progression types, FairDist achieved the highest AUC and equity-scaled AUC for gender and racial groups, compared to methods with and without unfairness mitigation strategies. FairDist can be generalized to other disease progression prediction tasks to potentially achieve improved performance and fairness.</p>","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"47 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144701903","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting outcomes following endovascular aortoiliac revascularization using machine learning","authors":"Ben Li, Badr Aljabri, Derek Beaton, Leen Al-Omran, Mohamad A. Hussain, Douglas S. Lee, Duminda N. Wijeysundera, Ori D. Rotstein, Charles de Mestral, Muhammad Mamdani, Mohammed Al-Omran","doi":"10.1038/s41746-025-01865-y","DOIUrl":"https://doi.org/10.1038/s41746-025-01865-y","url":null,"abstract":"<p>Endovascular aortoiliac revascularization is a common treatment option for peripheral artery disease that carries non-negligible risks. Outcome prediction tools may support clinical decision-making but remain limited. We developed machine learning algorithms that predict 30-day post-procedural outcomes. The National Surgical Quality Improvement Program targeted vascular database was used to identify patients who underwent endovascular aortoiliac revascularization between 2011–2021. Input features included 37 pre-operative demographic/clinical variables. The primary outcome was 30-day post-procedural major adverse limb event (MALE) or death. Data were split into training (70%) and test (30%) sets. Using 10-fold cross-validation, 6 machine learning models were trained using pre-operative features. Overall, 6601 patients were included, and 30-day MALE/death occurred in 470 (7.1%) individuals. The best-performing model was XGBoost, achieving an AUROC (95% CI) of 0.94 (0.93–0.95). In comparison, logistic regression had an AUROC (95% CI) of 0.74 (0.73–0.76). The XGBoost model accurately predicted 30-day post-procedural outcomes, performing better than logistic regression.</p>","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"14 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144693972","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificially intelligent nasal perception for rapid sepsis diagnostics.","authors":"Joonchul Shin,Gwang Su Kim,Seongmin Ha,Taehee Yoon,Junwoo Lee,Taehoon Lee,Woong Heo,Kyungyeon Lee,Seong Jun Park,Sunyoung Park,Jaewoo Song,Sunghoon Hur,Hyun-Cheol Song,Ji-Soo Jang,Jin-Sang Kim,Hyo-Il Jung,Chong-Yun Kang","doi":"10.1038/s41746-025-01851-4","DOIUrl":"https://doi.org/10.1038/s41746-025-01851-4","url":null,"abstract":"Sepsis, a life-threatening disease caused by infection, presents a major global health challenge due to its high morbidity and mortality rates. A rapid and precise diagnosis of sepsis is essential for better patient outcomes. However, conventional diagnostic methods, such as bacterial cultures, are time-consuming and can delay sepsis diagnosis. Considering these, researchers investigated alternative techniques that detect volatile organic compounds (VOCs) produced by bacteria. In this study, we designed colorimetric gas sensor arrays, which change color upon interaction with biomarkers, offer a direct visual signal, and demonstrate high sensitivity and specificity in detecting sepsis-related VOCs. Furthermore, an artificial intelligence (AI) based algorithm, Rapid Sepsis Boosting (RSBoost), was employed as an analytical technique to enhance diagnostic accuracy (96.2%) in blood sample. This approach significantly improves the speed and accuracy of sepsis diagnostics within 24 h, holding great potential for transforming clinical diagnostics, saving lives, and reducing healthcare costs.","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"19 1","pages":"476"},"PeriodicalIF":15.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144701293","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Personalized and real time hemodynamic management in critical care using Dynamic Cohort Ensemble Learning (DynaCEL)","authors":"Lingzhong Meng, Jiangqiong Li, Xiang Liu, Yanhua Sun, Zuotian Li, Jinjin Cai, Ameya D. Parab, George Lu, Aishwarya Budhkar, Saravanan Kanakasabai, David C. Adams, Ziyue Liu, Xuhong Zhang, Jing Su","doi":"10.1038/s41746-025-01863-0","DOIUrl":"https://doi.org/10.1038/s41746-025-01863-0","url":null,"abstract":"<p>Effective hemodynamic management in the intensive care unit requires individualized targets that adapt to dynamic clinical conditions. We developed Dynamic Cohort Ensemble Learning (DynaCEL), a real-time framework that recommends personalized heart rate and systolic blood pressure targets by modeling each time point post-intensive care unit admission as a distinct temporal cohort. Trained on eICU data and validated on MIMIC-IV and Indiana University Health datasets, DynaCEL demonstrated robust predictive performance (AUCs 0.83–0.91). In the MIMIC-IV cohort, proximity to DynaCEL-predicted targets was associated with lower 24-hour mortality compared to fixed targets, after adjustment using propensity score matching. Dose-response and comparative analyses revealed that greater deviations from personalized targets were associated with higher mortality. Case studies illustrated temporal and inter-individual variation in optimal targets. DynaCEL offers interpretable and scalable support for exploring precision hemodynamic management, although its clinical utility remains to be established in prospective trials.</p>","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"19 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144693965","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Synthetic data trained open-source language models are feasible alternatives to proprietary models for radiology reporting","authors":"Aakriti Pandita, Angela Keniston, Nikhil Madhuripan","doi":"10.1038/s41746-025-01658-3","DOIUrl":"https://doi.org/10.1038/s41746-025-01658-3","url":null,"abstract":"<p>The study assessed the feasibility of using synthetic data to fine-tune various open-source LLMs for free text to structured data conversation in radiology, comparing their performance with GPT models. A training set of 3000 synthetic thyroid nodule dictations was generated to train six open-source models (Starcoderbase-1B, Starcoderbase-3B, Mistral-7B, Llama-3-8B, Llama-2-13B, and Yi-34B). ACR TI-RADS template was the target model output. The model performance was tested on 50 thyroid nodule dictations from MIMIC-III patient dataset and compared against 0-shot, 1-shot, and 5-shot performance of GPT-3.5 and GPT-4. GPT-4 5-shot and Yi-34B showed the highest performance with no statistically significant difference between the models. Various open models outperformed GPT models with statistical significance. Overall, models trained with synthetic data showed performance comparable to GPT models in structured text conversion in our study. Given privacy preserving advantages, open LLMs can be utilized as a viable alternative to proprietary GPT models.</p>","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"29 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144684651","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effectiveness of digital healthcare to improve clinical outcomes in discharged patients with coronary artery disease","authors":"Lanshu Yang, Zuoxiang Wang, Sheng Zhao, Mengyuan Liu, Yalin Zhu, Fenghuan Hu, Xiaojin Gao, Yongjian Wu","doi":"10.1038/s41746-025-01655-6","DOIUrl":"https://doi.org/10.1038/s41746-025-01655-6","url":null,"abstract":"<p>Post-discharge management of coronary artery disease (CAD) remains clinically challenging, with digital healthcare’s efficacy underexplored. This study analyzed 16,797 CAD patients enrolled in the HeartMed Digital Management System (June 2018–September 2022), comparing outcomes between a digital management (DM, n = 4,713) and conventional management (CM, n = 12,084) cohort over 12 months. Cox models adjusted for confounders revealed significantly reduced all-cause mortality in the DM group (1.6% vs. 2.7%; HR 0.58, 95% CI 0.45–0.75, p &lt; 0.001) and lower risks for major adverse cardiovascular events (MACCE: 6.4% vs. 9.2%; HR 0.67, 0.59–0.77, p &lt; 0.001), cardiovascular death (HR 0.70, 0.51–0.95), myocardial infarction (HR 0.38, 0.29–0.50), recurrent angina (HR 0.75, 0.65–0.87), revascularization (HR 0.84, 0.71–0.99), and readmissions (HR 0.76, 0.68–0.84) (p &lt; 0.05 for all). Digital healthcare demonstrates superior post-discharge optimization of CAD outcomes, significantly attenuating mortality and morbidity.</p>","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"16 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144684656","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}