Machine Learning. Health最新文献

{"title":"Machine-learning in optimization of CRISPR technology.","authors":"Risi Liyanage, Lei Jin, Shi-Jie Chen","doi":"10.1088/3049-477X/aea331","DOIUrl":"10.1088/3049-477X/aea331","url":null,"abstract":"<p><p>The clustered regularly interspaced short palindromic repeats (CRISPR)-Cas system has become an indispensable tool in modern gene-engineering applications over recent years. Despite its rapid adoption, further development and broader applicability are hindered by several inherent limitations. This review surveys a range of machine-learning based approaches that aim to address these challenges. In particular, we focus on the optimization of key components of the CRISPR system, including protospacer adjacent motif recognition, Cas-protein-engineering, guide RNA sequence design and extension of the approaches to alternative editing modalities. Applied machine-learning methodologies, their underlying rationale, and comparisons with experimental observations are critically discussed. Special emphasis is placed on the role of machine-learning frameworks in advancing biophysical research, where complex, high-dimensional data increasingly demand integrative computational approaches. Finally, we outline current limitations, draw overarching conclusions, and propose perspectives for future developments and applications in CRISPR-based technologies.</p>","PeriodicalId":521035,"journal":{"name":"Machine Learning. Health","volume":"2 2","pages":"021001"},"PeriodicalIF":0.0,"publicationDate":"2026-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13595359/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149027465","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Utilizing categorical boosting and SHAP to understand key predictors of frequency of use at discharge for completed substance abuse treatments in TEDS-D.","authors":"C Flynn, S Abad, B Wooton, C Jansen, B Johnson, B Harrison, J Nash, A J Stokes, H Turner","doi":"10.1088/3049-477X/ae9bc7","DOIUrl":"10.1088/3049-477X/ae9bc7","url":null,"abstract":"<p><p><i>Background:</i> Substance use disorder is a pressing US public health crisis, with 48.7 million people reporting past-year substance use and over 100 000 overdose deaths in 2022. Building on a growing body of machine learning research using the SAMHSA treatment episode Data set-discharges (TEDS-D), this study introduces <i>frequency of use at discharge</i> as an outcome variable and the inclusion of <i>U.S. state</i> as a predictor. <i>Research Design and Methods:</i> First, <i>reason for discharge</i> was evaluated as a dependent variable and compared to <i>frequency of use at discharge</i>. Utilizing 2411 351 completed treatment episodes from TEDS-D 2015-2019, a categorical boosting model was employed to predict <i>frequency of substance use at discharge</i> using 62 predictor variables. Shapley additive explanations (SHAP) calculated global feature importance and decomposed individual predictions. State-level models were compared to the national model using rank-biased overlap (RBO), with a Hawai'i case study demonstrating state-level decision support. <i>Results:</i> Although treatment completion yielded higher rates of <i>no use</i> at discharge than dropout (45% vs 28%), 25% of completed episodes reported active substance use, 9% reported homelessness, and 50% reported unemployment at discharge. The model demonstrated strong predictive performance (AUC = 0.96, sensitivity = 0.85, specificity = 0.95). Treatment and geographic factors outweighed patient attributes, with <i>frequency of use at admission, U.S state of treatment, length of stay</i>, and <i>service type</i> as the four most important features. Among all service types, long-term residential rehabilitation was associated with the lowest SHAP values and predicted probability of substance use at discharge. State-stratified models showed moderate feature importance consistency (mean RBO = 0.59, range: 0.30-0.73). The Hawai'i case study identified elevated probability of use at discharge for younger clients and revealed substantial underreporting of long-term residential and therapeutic community treatment modalities in federal data. <i>Discussion and Implications:</i> State-level variability in treatment outcomes reflects systemic inconsistencies in policy, reporting, and service access that complicate national benchmarking and may obscure the true drivers of patient recovery. Explainable ML provides actionable insights for tailoring treatment and informing data-driven policy to improve outcomes.</p>","PeriodicalId":521035,"journal":{"name":"Machine Learning. Health","volume":"2 2","pages":"025017"},"PeriodicalIF":0.0,"publicationDate":"2026-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13643250/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149140314","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Autonomous radiotherapy planning via agentic orchestration using a multimodal TPS-integrated compound AI platform.","authors":"Austen Matthew Maniscalco, Yang Kyun Park, Sean J Domal, Mu-Han Lin, Sanda Harabagiu, Steve B Jiang, Dan Nguyen","doi":"10.1088/3049-477X/ae7978","DOIUrl":"10.1088/3049-477X/ae7978","url":null,"abstract":"&lt;p&gt;&lt;p&gt;&lt;i&gt;Purpose.&lt;/i&gt; Radiation therapy (RT) treatment planning requires iterative, multi-day optimization workflows in which subjective planning strategies produce inter-planner variability in plan quality. Existing computational approaches automate isolated aspects of this workflow, yet none orchestrates an end-to-end pipeline from physician directive to deliverable plan. We developed a compound artificial intelligence (AI) platform for fully autonomous RT treatment planning that combines multi-agent large language model (LLM) orchestration with directive-conditioned three-dimensional (3D) dose prediction, natively integrated with a commercial treatment planning system (TPS). &lt;i&gt;Methods.&lt;/i&gt; Seven specialized agents navigated the multi-objective optimization landscape through structured clinical reasoning, iteratively analyzing dose-volume histogram (DVH) metrics and spatial dose patterns, formulating trade-off strategies, and executing validated modifications through the TPS across five fully autonomous iterations per case. A directive-conditioned 3D dose prediction model supplied patient-specific DVH values from which initial optimization objectives were autonomously derived, eliminating the need for curated templates or manual initialization. A retrieval-augmented generation (RAG) system encoded institutional knowledge into the planning workflow. We evaluated 60 retrospective cases across brain, lung, and prostate sites, with 10 intensity-modulated RT (IMRT) and 10 volumetric modulated arc therapy (VMAT) plans per site spanning 20.0-79.2 Gy in 3-44 fractions, scored by the proportion of dosimetric criteria satisfied. &lt;i&gt;Results.&lt;/i&gt; Across all 60 cases, AI plans achieved 89.8 &lt;math&gt;&lt;mrow&gt;&lt;mo&gt;±&lt;/mo&gt;&lt;/mrow&gt; &lt;/math&gt; 9.4% of dosimetric criteria versus 85.2 &lt;math&gt;&lt;mrow&gt;&lt;mo&gt;±&lt;/mo&gt;&lt;/mrow&gt; &lt;/math&gt; 10.8% for clinical reference plans ( &lt;math&gt;&lt;mrow&gt;&lt;mi&gt;p&lt;/mi&gt; &lt;mo&gt;&lt;&lt;/mo&gt; &lt;mn&gt;0.001&lt;/mn&gt;&lt;/mrow&gt; &lt;/math&gt; ). IMRT plans improved in 25 of 30 cases with none worsened (94.1 &lt;math&gt;&lt;mrow&gt;&lt;mo&gt;±&lt;/mo&gt;&lt;/mrow&gt; &lt;/math&gt; 6.7% vs 84.3 &lt;math&gt;&lt;mrow&gt;&lt;mo&gt;±&lt;/mo&gt;&lt;/mrow&gt; &lt;/math&gt; 11.8%, &lt;math&gt;&lt;mrow&gt;&lt;mi&gt;p&lt;/mi&gt; &lt;mo&gt;&lt;&lt;/mo&gt; &lt;mn&gt;0.001&lt;/mn&gt;&lt;/mrow&gt; &lt;/math&gt; ); VMAT plans showed no significant difference (85.6 &lt;math&gt;&lt;mrow&gt;&lt;mo&gt;±&lt;/mo&gt;&lt;/mrow&gt; &lt;/math&gt; 9.9% vs 86.1 &lt;math&gt;&lt;mrow&gt;&lt;mo&gt;±&lt;/mo&gt;&lt;/mrow&gt; &lt;/math&gt; 9.7%, &lt;math&gt;&lt;mrow&gt;&lt;mi&gt;p&lt;/mi&gt; &lt;mo&gt;=&lt;/mo&gt; &lt;mn&gt;0.770&lt;/mn&gt;&lt;/mrow&gt; &lt;/math&gt; ). Each plan iteration completed in &lt;math&gt;&lt;mrow&gt;&lt;mn&gt;20.2&lt;/mn&gt; &lt;mo&gt;±&lt;/mo&gt; &lt;mn&gt;12.7&lt;/mn&gt;&lt;/mrow&gt; &lt;/math&gt; min, of which agent reasoning consumed &lt;math&gt;&lt;mrow&gt;&lt;mn&gt;5.2&lt;/mn&gt; &lt;mo&gt;±&lt;/mo&gt; &lt;mn&gt;1.7&lt;/mn&gt;&lt;/mrow&gt; &lt;/math&gt; min ( &lt;math&gt;&lt;mrow&gt;&lt;mn&gt;114,429&lt;/mn&gt; &lt;mo&gt;±&lt;/mo&gt; &lt;mn&gt;11,798&lt;/mn&gt;&lt;/mrow&gt; &lt;/math&gt; tokens, &lt;math&gt;&lt;mrow&gt;&lt;mn&gt;0.43&lt;/mn&gt; &lt;mo&gt;±&lt;/mo&gt; &lt;mn&gt;0.04&lt;/mn&gt;&lt;/mrow&gt; &lt;/math&gt; ). &lt;i&gt;Conclusions.&lt;/i&gt; These results established the feasibility of end-to-end, fully autonomous, universal RT treatment planning through compound AI. Integrating dose prediction as an agent-invoked tool for objective initialization resolved the dependency on curated templates a","PeriodicalId":521035,"journal":{"name":"Machine Learning. Health","volume":"2 2","pages":"025005"},"PeriodicalIF":0.0,"publicationDate":"2026-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13374601/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148475796","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Detecting cognitive impairment and psychological well-being among older adults.","authors":"Xiaofan Mu, Merna Bibars, Salman Seyedi, Iris Zheng, Zifan Jiang, Liu Chen, Bolaji Omofojoye, Rachel Hershenberg, Allan I Levey, Gari D Clifford, Hiroko H Dodge, Hyeokhyen Kwon","doi":"10.1088/3049-477X/ae250c","DOIUrl":"10.1088/3049-477X/ae250c","url":null,"abstract":"<p><p>The aging society urgently requires scalable methods to monitor cognitive decline and identify social and psychological factors indicative of dementia risk in older adults. Our machine learning models captured facial, acoustic, linguistic, and cardiovascular features from 39 older adults with normal cognition or mild cognitive impairment, derived from remote video conversations and quantified their cognitive status, social isolation, neuroticism, and psychological well-being. Our model could distinguish Clinical Dementia Rating Scale of 0.5 (vs 0) with 0.77 area under the receiver operating characteristic curve (AUC), social isolation with 0.74 AUC, social satisfaction with 0.75 AUC, psychological well-being with 0.72 AUC, and negative affect with 0.74 AUC. Our feature importance analysis showed that speech and language patterns were useful for quantifying cognitive impairment, whereas facial expressions and cardiovascular patterns were useful for quantifying social and psychological well-being. Our bias analysis showed that the best-performing models for quantifying psychological well-being and cognitive states in older adults exhibited significant biases concerning their age, sex, disease condition, and education levels. Our comprehensive analysis shows the feasibility of monitoring the cognitive and psychological health of older adults, as well as the need for collecting large-scale interview datasets of older adults to benefit from the latest advances in deep learning technologies to develop generalizable models across older adults with diverse demographic backgrounds and disease conditions.</p>","PeriodicalId":521035,"journal":{"name":"Machine Learning. Health","volume":"2 1","pages":"015005"},"PeriodicalIF":0.0,"publicationDate":"2026-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12820936/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146032700","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A systematic evaluation of uncertainty quantification techniques in deep learning: a case study in photoplethysmography signal analysis.","authors":"Ciaran Bench, Oskar Pfeffer, Vivek Desai, Mohammad Moulaeifard, Loic Coquelin, Peter H Charlton, Nils Strodthoff, Nando Hegemann, Philip J Aston, Andrew Thompson","doi":"10.1088/3049-477X/ae4c8e","DOIUrl":"10.1088/3049-477X/ae4c8e","url":null,"abstract":"<p><p>In principle, deep learning models trained on medical time-series, including wearable photoplethysmography sensor data, can provide a means to continuously monitor physiological parameters outside of clinical settings. However, there is considerable risk of poor performance when deployed in practical measurement scenarios leading to negative patient outcomes. Reliable uncertainties accompanying predictions can provide guidance to clinicians in their interpretation of the trustworthiness of model outputs. It is therefore of interest to compare the effectiveness of different approaches. Here we implement an unprecedented set of eight uncertainty quantification (UQ) techniques to models trained on two clinically relevant prediction tasks: atrial fibrillation (AF) detection (classification), and two variants of blood pressure regression. We formulate a comprehensive evaluation procedure to enable a rigorous comparison of these approaches. We observe a complex picture of uncertainty reliability across the different techniques, where the most optimal for a given task depends on the chosen expression of uncertainty, evaluation metric, and scale of reliability assessed. We find that assessing local calibration and adaptivity provides practically relevant insights about model behaviour that otherwise cannot be acquired using more commonly implemented global reliability metrics. We emphasise that criteria for evaluating UQ techniques should cater to the model's practical use case, where the use of a small number of measurements per patient places a premium on achieving small-scale reliability for the chosen expression of uncertainty, while preserving as much predictive performance as possible.</p>","PeriodicalId":521035,"journal":{"name":"Machine Learning. Health","volume":"2 1","pages":"015011"},"PeriodicalIF":0.0,"publicationDate":"2026-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12997033/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147489303","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"'See' through the surface: surface-derived three-dimensional AI-driven real-time imaging solution for intra-treatment image guidance.","authors":"Tingliang Zhuang, Hua-Chieh Shao, Ruiqi Li, Tsuicheng Chiu, Kenneth Westover, Steve Jiang, You Zhang","doi":"10.1088/3049-477X/ae3321","DOIUrl":"10.1088/3049-477X/ae3321","url":null,"abstract":"<p><p>Respiratory motion is a long-standing challenge for lung stereotactic body radiotherapy (SBRT), particularly for centrally located lung tumors where increased toxicity demands more precise motion management during treatment. Current two-dimensional imaging approaches are insufficient for 3D tumor deformable motion tracking. In this study, we developed and evaluated a Surface-derived Three-dimensional (3D) AI-driven Real-time (STAR) imaging system by transforming a surface imaging system into a 3D real-time imaging solution. STAR integrated two key components: (1) prior-model-free spatiotemporal implicit neural representation (PMF-STINR): a machine-learning sub-system for pretreatment dynamic cone-beam CT (CBCT) reconstruction and motion modeling; and (2) surface-to-deformation network (Surf2DefNet): a deep-learning model that correlates intra-treatment body surface images with internal 3D anatomy and motion fields, trained based on the dynamic CBCT and motion model output of PMF-STINR. Specifically, PMF-STINR reconstructs a reference CBCT and solves an eigenvector-based motion model from a pretreatment CBCT, while Surf2DefNet predicts the motion eigen-vector weightings from surface images, enabling it to infer real-time CBCTs and motion vector fields (MVFs) using intra-treatment surface maps acquired later. We evaluated STAR imaging using both a digital extended cardiac torso (XCAT) phantom with regular and irregular motion patterns and ten patient datasets. The relative error (RE), center-of-mass error (COME), 95th percentile Hausdorff distance (HD95), Dice coefficients (DICE), and Pearson correlation coefficient (PCC) metrics between STAR images and the 'GT' were evaluated. For the XCAT phantom and the ten patients, the mean COME values are within 1 mm for all but one patient (1.3 mm). The RE values were consistently low, and the DICE and PCC values exceeded 0.89 in all cases. The HD95 are all within 2 mm except for one patient (2. 78 mm). These results demonstrate that the STAR imaging system can achieve accurate spatiotemporal reconstructions from surface images, providing CBCTs and MVFs for intra-treatment real-time image-guidance, and has the potential to improve safety and efficacy of SBRT for lung cancer.</p>","PeriodicalId":521035,"journal":{"name":"Machine Learning. Health","volume":"2 1","pages":"015006"},"PeriodicalIF":0.0,"publicationDate":"2026-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12834327/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146069726","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine learning models enhance detection of arrhythmogenic right ventricular cardiomyopathy.","authors":"Kwaku K Quansah, Sean A Murphy, Esther Kwon, Emma Anderson, Crystal Tichnell, Brittney Murray, Sean Gaines, Alessio Gasperetti, Cynthia A James, Hugh Calkins, Richard T Carrick, Chulan Kwon","doi":"10.1088/3049-477X/ae451a","DOIUrl":"10.1088/3049-477X/ae451a","url":null,"abstract":"<p><p>Arrhythmogenic right ventricular cardiomyopathy (ARVC) is a heritable cardiac disorder associated with sudden cardiac death, yet its diagnosis remains slow, resource-intensive, and dependent on expert interpretation of multimodal tests. Machine learning approaches may enable earlier and more standardized detection. Here, we sought to identify an optimal machine learning strategy for ARVC detection, and to define its role within the diagnostic pathway. A composite dataset of 688 patients from the Johns Hopkins ARVC Registry was used to train and evaluate eight models based on 31 clinical, electrocardiogram (ECG), imaging, and genetic variables. Following an 80/20 train-test split, models were developed on the train set using five-fold stratified cross-validation and evaluated on the hold-out test cohort. Performance was assessed using area under the curve (AUC), accuracy, sensitivity, specificity, and 95% confidence intervals. Model comparison was performed using the Friedman test with Nemenyi post-hoc analysis. Gradient boosted trees (GBT) achieved the highest overall performance (AUC = 0.943, 95% CI: 0.902-0.99; accuracy = 0.847, 95% CI: 0.786-0.907). Nemenyi testing showed GBT significantly outperformed Decision Tree and TabNet (<i>p</i> < 0.05), while differences vs the remaining five models were not statistically significant. The next-best model (Random Forest) showed a minimal performance gap (ΔAUC = 0.008), whereas low-ranked models showed larger deficits (ΔAUC = 0.040-0.042). An ECG-only version of the GBT model achieved an AUC of 0.884, exceeding the previously reported ECG-deep learning waveform model (AUC = 0.87). GBT performs best among evaluated algorithms and offers clinically interpretable feature relevance consistent with task force criteria for ARVC diagnosis. An ECG-only deployment supports early triage, while the multimodal model functions as confirmatory decision support after advanced testing. These findings support a tiered ML-assisted diagnostic strategy for ARVC and justify prospective external validation in broader clinical settings.</p>","PeriodicalId":521035,"journal":{"name":"Machine Learning. Health","volume":"2 1","pages":"010501"},"PeriodicalIF":0.0,"publicationDate":"2026-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13012209/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147518358","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial intelligence (AI)-based multi-organ contour quality assurance with uncertainty estimation for online adaptive radiotherapy (oART).","authors":"Shunyu Yan, Jiacheng Xie, Na Chen, Dan Nguyen, Fan-Chi Su, Daniel Yang, You Zhang, Steve Jiang, Chenyang Shen","doi":"10.1088/3049-477X/ae3320","DOIUrl":"10.1088/3049-477X/ae3320","url":null,"abstract":"<p><p>Accurate delineation of treatment targets and organs at risk (OARs) is essential to the success of radiotherapy (RT). Although artificial intelligence (AI)-based segmentation methods have successfully automated the delineation process, a reliable and efficient quality assurance (QA) mechanism is still missing, particularly in the time-critical setting of online adaptive RT (oART). This study aims to address this unmet clinical need by introducing an AI-driven multi-organ contour QA framework with uncertainty quantification. Using MR-guided oART for prostate cancer as the testbed, we developed an automatic multi-structure QA framework by employing a carefully designed contour quality estimation (ConQuE) model. Built on a ResNet34 backbone, ConQuE processes binary segmentation masks with corresponding MR images to classify contours as 'acceptable' or 'revision required' in a slice-by-slice manner. To ensure scalability, we proposed a unified framework for multi-organ QA by embedding structure-specific code into the fully connected layers. To quantify classification uncertainty, Monte Carlo (MC) dropout was employed, enabling confidence assessment of model predictions for informative decision-making in oART. ConQuE was trained to assess the segmentation quality of prostate and six OARs, including rectum, bladder, urethra, penile bulb, cauda equina, and femoral heads. Training was performed using 249 cases for training, 31 for validation, and another 31 cases were saved for testing. With 30 MC dropout passes, ConQuE can complete QA for one image slice in ∼13.1 ms. Overall, ConQuE achieved accuracy of 93.9% on all structures with structure-specific performance consistently above 90%, demonstrating its strong capability in identifying <i>revision required</i> contours. Moreover, incorrect predictions were strongly correlated with high uncertainty scores, underscoring the effectiveness in improving QA reliability. The proposed AI-based framework holds the potential to automate contour QA with quantified uncertainty. It is well-suited for oART with heavy time constraints given its high accuracy, efficiency, and reliability.</p>","PeriodicalId":521035,"journal":{"name":"Machine Learning. Health","volume":"2 1","pages":"015001"},"PeriodicalIF":0.0,"publicationDate":"2026-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12813885/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146014163","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"TSMS-SAM2: Multi-scale Temporal Sampling Augmentation and Memory-Splitting Pruning for Promptable Video Object Segmentation and Tracking in Surgical Scenarios.","authors":"Guoping Xu, Hua-Chieh Shao, You Zhang","doi":"10.1088/3049-477x/ae291a","DOIUrl":"10.1088/3049-477x/ae291a","url":null,"abstract":"<p><p>Promptable video object segmentation and tracking (VOST) has seen significant advances with the emergence of foundation models like Segment Anything Model 2 (SAM2); however, their application in surgical video analysis remains challenging due to complex motion dynamics and the redundancy of memory that impedes effective learning. In this work, we propose TSMS-SAM2, a novel framework that enhances promptable VOST in surgical videos by addressing challenges of rapid object motion and memory redundancy in SAM2. TSMS-SAM2 introduces two key strategies: multi-temporal-scale video sampling augmentation to improve robustness against motion variability, and a memory splitting and pruning mechanism that organizes and filters past frame features for more efficient and accurate segmentation. Evaluated on EndoVis2017 and EndoVis2018 datasets, TSMS-SAM2 achieved the highest mean (± s.d.) Dice scores of 95.24±0.96% and 86.73±15.46%, respectively, outperforming prior SAM-based and task-specific methods. Extensive ablation studies confirm the effectiveness of multiscale temporal augmentation and memory splitting, highlighting the framework's potential for robust, efficient segmentation in complex surgical scenarios. Our source code will be made available at https://github.com/apple1986/TSMS-SAM2.</p>","PeriodicalId":521035,"journal":{"name":"Machine Learning. Health","volume":"2 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12830041/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146055939","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Generalizable deep learning for photoplethysmography-based blood pressure estimation-A benchmarking study.","authors":"Mohammad Moulaeifard, Peter H Charlton, Nils Strodthoff","doi":"10.1088/3049-477X/ae01a8","DOIUrl":"10.1088/3049-477X/ae01a8","url":null,"abstract":"<p><p>Photoplethysmography (PPG)-based blood pressure (BP) estimation represents a promising alternative to cuff-based BP measurements. Recently, an increasing number of deep learning (DL) models have been proposed to infer BP from the raw PPG waveform. However, these models have been predominantly evaluated on in-distribution (ID) test sets, which immediately raises the question of the generalizability of these models to external datasets. To investigate this question, we trained five DL models on the recently released PulseDB dataset, provided ID benchmarking results on this dataset, and then assessed their out-of-distribution (OOD) performance on several external datasets. The best model (XResNet1d101) achieved ID mean absolute errors (MAEs) of 9.0 and 5.8 mmHg for systolic and diastolic BP, respectively, on PulseDB with subject-specific calibration, and 13.9 and 8.5 mmHg, respectively, without calibration. The equivalent MAEs on external test datasets without calibration ranged from 10.0 to 18.6 mmHg (SBP) and 5.9 to 10.3 mmHg (DBP). Our results indicate that performance is strongly influenced by the differences in BP distributions between datasets. We investigated a simple way of improving performance through sample-based domain adaptation and put forward recommendations for training models with good generalization properties. With this work, we hope to educate more researchers about the importance and challenges of OOD generalization.</p>","PeriodicalId":521035,"journal":{"name":"Machine Learning. Health","volume":"1 1","pages":"010501"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12435175/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145077146","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}