NPJ Digital Medicine最新文献

Eviction is a significant yet understudied social determinants of health (SDoH), linked to housing instability, unemployment, and mental health. While eviction appears in unstructured electronic health records (EHRs), it is rarely coded in structured fields, limiting downstream applications. We introduce SynthEHR-Eviction, a scalable pipeline that adapts and integrates human-in-the-loop annotation, automated prompt optimization (APO), and reasoning-augmented fine-tuning for low-resource eviction-related SDoH extraction from clinical notes. Using this pipeline, we created a large public eviction-related SDoH dataset to date, comprising 14 fine-grained categories. Fine-tuned LLMs (e.g., Qwen2.5, LLaMA3) trained on SynthEHR-Eviction achieved Macro-F1 scores of 88.8% (eviction) and 90.3% (other SDoH) on human validated data, outperforming GPT-4o-APO (87.8%, 87.3%), GPT-4o-mini-APO (69.1%, 78.1%), and BioBERT (60.7%, 68.3%), while enabling cost-effective deployment across various model sizes. The pipeline reduces annotation effort by over 80%, accelerates dataset creation, enables scalable eviction detection, and generalizes to other information extraction tasks.

Advances in AI offer significant opportunities to enhance drug development. While several regulatory agencies have begun issuing guidance on AI adoption, its application to causal inference-a critical piece to understand treatment effects and inform regulatory decisions-remains limited. This paper reviews regulatory activities and examines statistical methodologies for AI-driven causal inference. We discuss key regulatory challenges and illustrate how AI adds value across diverse data sources and studies.

Federated learning (FL) and travelling model (TM) allow privacy-preserving model training across sites without sharing patient-sensitive data. While both approaches have shown success, they face unique challenges related to distribution shifts between sites. To address this, we propose FedTM, a hybrid framework combining the strengths of FL and TM. FedTM begins with FL warmup training at sites with larger datasets, followed by sequential refinement through TM across all sites. We evaluated FedTM for Parkinson's disease classification using 1817 brain scans from 83 international sites. Model performance, misclassification disparities, and communication costs were computed and compared to standard FL and TM approaches. Our results reveal that FedTM improves AUROC from 77 ± 0.01% to 82 ± 0.01%, reduces misclassification disparities from 34 ± 0.01% to 26 ± 0.01%, and decreases training load for smaller sites from 22 to 12 cycles. These advancements mark an important step toward promoting global healthcare equity and advancing responsible AI development.

Alzheimer's disease (AD) is a progressive neurodegenerative disorder increasingly associated with peripheral inflammatory conditions such as chronic periodontitis (CP); however, the molecular mechanisms linking these conditions remain poorly understood. Here, we investigated the therapeutic effects of Huanglian Jieddu Decoction (HLJDD) on CP-induced AD using an integrative machine learning-guided multi-omics approach. Analysis of public single-cell RNA-sequencing data revealed pronounced inflammatory activation in microglia from AD samples. We further established a CP-induced AD rat model and performed hippocampal transcriptomic profiling. Multiple complementary machine learning strategies, including Random Forest-based feature selection, support vector machine-based refinement, network modeling, and interpretable model analysis, were applied to prioritize disease-relevant pathways from high-dimensional transcriptomic data. Across models, components of the cGAS-STING signaling pathway consistently exhibited strong and directional contributions to CP-AD pathology, indicating a central inflammatory axis linking peripheral infection to neurodegeneration. Guided by these data-driven insights, in vivo and in vitro experiments demonstrated that HLJDD suppressed cGAS-STING activation, attenuated neuroinflammation, and improved cognitive function in CP-induced AD models. Collectively, this study highlights the value of machine learning-assisted transcriptomic interpretation for mechanistic prioritization and identifies HLJDD as a multitarget therapeutic strategy for CP-induced AD.

Recent advances in cough sound analysis using deep learning techniques enable smartphone-based respiratory disease screening suitable for self-management care in a home setting, yet their utility is limited by device heterogeneity, population diversity, and challenges in multimodal integration. We propose a device-invariant, multimodal deep learning framework that jointly models cough acoustics, demographic data, and symptom descriptions for multi-label classification of adult respiratory diseases. To address the issues of device effect, an adversarial branch is embedded in the audio encoder to enforce device-invariant feature learning, while an invariant risk minimization-augmented loss enhances robustness to non-structural shifts. To evaluate the effectiveness of our proposed method, a real-world, multi-center dataset containing over 10,000 cases spanning seven major respiratory conditions was curated. On the tasks of individual respiratory disease identification for chronic obstructive pulmonary disease (COPD), lower respiratory tract infection (LRTI) and pulmonary shadows (PS), our method achieves superior performance with the area under the receiver operating characteristic curve (AUROC) of 0.9698, 0.8483 and 0.8720, respectively. It also shows promising results in identifying the presence of comorbidities for 7 respiratory diseases with an overall AUROC of 0.8907. More importantly, extensive experimental results demonstrate our method mitigates the issues of device effect and facilitates the cross-device generalization for cough-based respiratory disease diagnoses. This work demonstrates a scalable and transferable AI-based approach for cough-driven respiratory screening, emphasizing the importance of multimodal fusion and robust representation learning in advancing clinical applicability.

Biological sex is closely linked to patterns embedded within the electrocardiogram (ECG) with essential health and disease implications. We report multicenter verification of an AI-enabled ECG model to predict biological sex across pediatric development. A previously published Mayo Clinic model confirmed puberty-linked AUROC gradient during external validation at Texas Children's Hospital (pre-puberty AUROC 0.64, peri-puberty AUROC 0.84, post-puberty AUROC 0.94). This phenomenon was replicated at Boston Children's Hospital. Saliency mapping revealed established sex-related electrophysiologic patterns.

Most ambient AI medical scribes process audio only, omitting clinically important visual details. We developed a vision-enabled AI scribe using Google's Gemini model and Ray-Ban Meta smart glasses to document medication histories-a task requiring both audio and visual input. Ten clinical pharmacists video-recorded 110 simulated medication history interviews. Following iterative prompt engineering on 10 training recordings, the scribe was evaluated on 100 test recordings (2160 data points) across patient details and medication-specific fields. The vision-enabled scribe achieved 98% overall accuracy (2114/2,160 data points), ranging from 96% for patient details to 99% for dosing directions and indication. Video input significantly outperformed audio-only processing (98% vs 81%, P < 0.001), primarily through reduced omissions (10 vs 358 errors). Vision-enabled AI scribes substantially improved documentation accuracy for tasks requiring visual input, demonstrating potential to markedly reduce omission errors in clinical documentation.

Pulse wave velocity (PWV) is a marker of vascular aging and cardiovascular risk. Obstructive sleep apnea (OSA) may accelerate vascular decline, but evidence from single-night assessments is inconsistent. We examined associations of multi-night OSA severity, night-to-night variability, and snoring with arterial stiffness in a real-world setting. Adults used two in-home digital devices over a ~ 4 y period: an under-mattress sleep sensor to quantify nightly OSA severity and snoring, and a smart scale to measure aortic-leg PWV. Among 29,653 participants from 20 countries (52 ± 12 years; 84% male; BMI 27.3 ± 4.9 kg/m²), increasing OSA severity was associated with higher PWV in a dose-response manner, independent of age, sex, and BMI. Participants with mild OSA but high variability had PWV levels comparable to severe OSA. Higher snoring burden independently predicted higher PWV across OSA severity categories. Multi-night in-home assessments of OSA and snoring may better reflect cardiovascular risk with potential to inform personalized management.