Mayo Clinic Proceedings. Digital health最新文献

{"title":"Comparison of Participant and Site Perceptions of Decentralized Clinical Trials in the USA","authors":"Roland Barge PhD ,&nbsp;Patrick Floody MBA","doi":"10.1016/j.mcpdig.2025.100201","DOIUrl":"10.1016/j.mcpdig.2025.100201","url":null,"abstract":"<div><h3>Objective</h3><div>To define potential participant and site perceptions of decentralized clinical trials (DCTs).</div></div><div><h3>Participants and Methods</h3><div>Two qualitative surveys were conducted between January 2022 and August 2022 to assess current awareness of, and perceptions about, DCTs. The first survey received 141 responses from staff at our clinical trial sites; the second survey received 481 responses from US-based healthy individuals or those living with an illness.</div></div><div><h3>Results</h3><div>There was a difference in perceptions and willingness between participants and sites toward DCTs. Participants expressed more comfort with hybrid and fully remote trials than did the sites. Site staff were more concerned and less trusting than participants of DCTs; participants’ main concerns were regarding practicality and medical safety, whereas the focus for sites was on burden, trust, and security. Both sites and participants expressed confidence in fully remote clinical study activities when they have appropriate support; sites were less tolerant of fully remote clinical study activities if professional support was not provided. Overall, sites were more willing to manage the use of DCT-related technologies than were participants. It is highly likely that participants’ willingness to manage DCT technologies relates to the perceived burden of use (ie, willingness decreases as burden or impact on daily life increases). Sponsors, contract research organizations, and DCT vendors generally had positive views on DCTs. However, different stakeholders had different concerns.</div></div><div><h3>Conclusion</h3><div>These results highlight the need for collaborative research and development of DCTs, as well as a clear DCT framework and regulatory guidance.</div></div>","PeriodicalId":74127,"journal":{"name":"Mayo Clinic Proceedings. Digital health","volume":"3 2","pages":"Article 100201"},"PeriodicalIF":0.0,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143696654","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":"CF Tummy Tracker: A Cystic Fibrosis–Specific Patient-Reported Outcome Measure for Daily Gastrointestinal Symptom Burden","authors":"Rebecca J. Calthorpe BMBS ,&nbsp;Hisham A. Saumtally MBChB ,&nbsp;Laura M. Howells PhD ,&nbsp;Natalie J. Goodchild BA (Hons) ,&nbsp;Bethinn C. Evans MSc ,&nbsp;Zoe Elliott ,&nbsp;Bu’Hussain Hayee PhD ,&nbsp;Siobhán B. Carr MBBS ,&nbsp;Caroline M. Elston MBBS ,&nbsp;Alexander A.R. Horsley PhD ,&nbsp;Daniel G. Peckham DM ,&nbsp;Helen L. Barr PhD ,&nbsp;Giles A.D. Major PhD ,&nbsp;Iain D. Stewart PhD ,&nbsp;Kim S. Thomas ,&nbsp;Alan R. Smyth MD","doi":"10.1016/j.mcpdig.2025.100203","DOIUrl":"10.1016/j.mcpdig.2025.100203","url":null,"abstract":"<div><h3>Objective</h3><div>To develop a cystic fibrosis (CF)–specific patient-reported outcome measure (PROM) to measure the daily burden of gastrointestinal symptoms for people with cystic fibrosis (pwCF) aged 12 years and older and address the lack of validated outcome measures for gastrointestinal symptoms in CF.</div></div><div><h3>Patients and Methods</h3><div>CF Tummy Tracker was developed through a 5-stage approach in accordance with regulatory guidance. This included development and refinement of a conceptual framework; item generation; refinement; reduction; selection; and initial PROM testing. A mixed-methods approach, consisting of expert panel discussions, a focus group, interviews, and an online survey, was used. In initial testing, participants completed the PROM daily for 14 days via a smartphone application. This study was performed from March 14, 2022, December 12, 2023.</div></div><div><h3>Results</h3><div>The CF community were involved throughout the development via a focus group (n=7 pwCF), interviews (n=11 pwCF), and an online survey (n=180 pwCF). A formative model was confirmed for the PROM. The final PROM, CF Tummy Tracker, consists of 10 items capturing gastrointestinal symptom burden, tested in 151 pwCF. The PROM reported no floor or ceiling effects, high test–retest reliability (intra-class correlation coefficient=0.94), and strong correlation with the anchor question.</div></div><div><h3>Conclusion</h3><div>CF Tummy Tracker aims to address the gap in validated CF-specific PROMs for daily completion. Further testing of the psychometric properties of the PROM are planned in a new patient cohort to validate its use in clinical trials and support its use in both electronic and paper formats to increase accessibility.</div></div>","PeriodicalId":74127,"journal":{"name":"Mayo Clinic Proceedings. Digital health","volume":"3 2","pages":"Article 100203"},"PeriodicalIF":0.0,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143680721","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":"Retrospective Comparative Analysis of Prostate Cancer In-Basket Messages: Responses From Closed-Domain Large Language Models Versus Clinical Teams","authors":"Yuexing Hao MS ,&nbsp;Jason Holmes PhD ,&nbsp;Jared Hobson MD ,&nbsp;Alexandra Bennett MD ,&nbsp;Elizabeth L. McKone MD ,&nbsp;Daniel K. Ebner MD ,&nbsp;David M. Routman MD ,&nbsp;Satomi Shiraishi MD ,&nbsp;Samir H. Patel MD ,&nbsp;Nathan Y. Yu MD ,&nbsp;Chris L. Hallemeier MD ,&nbsp;Brooke E. Ball MSN ,&nbsp;Mark Waddle MD ,&nbsp;Wei Liu PhD","doi":"10.1016/j.mcpdig.2025.100198","DOIUrl":"10.1016/j.mcpdig.2025.100198","url":null,"abstract":"<div><h3>Objective</h3><div>To evaluate the effectiveness of RadOnc-generative pretrained transformer (GPT), a GPT-4 based large language model, in assisting with in-basket message response generation for prostate cancer treatment, with the goal of reducing the workload and time on clinical care teams while maintaining response quality.</div></div><div><h3>Patients and Methods</h3><div>RadOnc-GPT was integrated with electronic health records from both Mayo Clinic-wide databases and a radiation-oncology-specific database. The model was evaluated on 158 previously recorded in-basket message interactions, selected from 90 patients with nonmetastatic prostate cancer from the Mayo Clinic Department of Radiation Oncology in-basket message database in the calendar years 2022-2024. Quantitative natural language processing analysis and 2 grading studies, conducted by 5 clinicians and 4 nurses, were used to assess RadOnc-GPT’s responses. Three primary clinicians independently graded all messages, whereas a fourth senior clinician reviewed 41 responses with relevant discrepancies, and a fifth senior clinician evaluated 2 additional responses. The grading focused on 5 key areas: completeness, correctness, clarity, empathy, and editing time. The grading study was performed from July 20, 2024 to December 15, 2024.</div></div><div><h3>Results</h3><div>The RadOnc-GPT slightly outperformed the clinical care team in empathy, whereas achieving comparable scores with the clinical care team in completeness, correctness, and clarity. Five clinician graders identified key limitations in RadOnc-GPT’s responses, such as lack of context, insufficient domain-specific knowledge, inability to perform essential meta-tasks, and hallucination. It was estimated that RadOnc-GPT could save an average of 5.2 minutes per message for nurses and 2.4 minutes for clinicians, from reading the inquiry to sending the response.</div></div><div><h3>Conclusion</h3><div>RadOnc-GPT has the potential to considerably reduce the workload of clinical care teams by generating high-quality, timely responses for in-basket message interactions. This could lead to improved efficiency in health care workflows and reduced costs while maintaining or enhancing the quality of communication between patients and health care providers.</div></div>","PeriodicalId":74127,"journal":{"name":"Mayo Clinic Proceedings. Digital health","volume":"3 1","pages":"Article 100198"},"PeriodicalIF":0.0,"publicationDate":"2025-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143579568","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":"Optimizing Input Selection for Cardiac Model Training and Inference: An Efficient 3D Convolutional Neural Networks-Based Approach to Automate Coronary Angiogram Video Selection","authors":"Shih-Sheng Chang MD, PhD ,&nbsp;Behrouz Rostami PhD ,&nbsp;Gerardo LoRusso MD ,&nbsp;Chia-Hao Liu MD ,&nbsp;Mohamad Alkhouli MD","doi":"10.1016/j.mcpdig.2025.100195","DOIUrl":"10.1016/j.mcpdig.2025.100195","url":null,"abstract":"<div><h3>Objective</h3><div>To develop an efficient and automated method for selecting appropriate coronary angiography videos for training deep learning models, thereby improving the accuracy and efficiency of medical image analysis.</div></div><div><h3>Patients and Methods</h3><div>We developed deep learning models using 232 coronary angiographic studies from the Mayo Clinic. We utilized 2 state-of-the-art convolutional neural networks (CNN: ResNet and X3D) to identify low-quality angiograms through binary classification (satisfactory/unsatisfactory). Ground truth for the quality of the input angiogram was determined by 2 experienced cardiologists. We validated the developed model in an independent dataset of 3208 procedures from 3 Mayo sites.</div></div><div><h3>Results</h3><div>The 3D-CNN models outperformed their 2D counterparts, with the X3D-L model achieving superior performance across all metrics (AUC 0.98, accuracy 0.96, precision 0.87, and F1 score 0.92). Compared with 3D models, 2D architectures are smaller and less computationally complex. Despite having a 3D architecture, the X3D-L model had lower computational demand (19.34 Giga Multiply Accumulate Operation) and parameter count (5.34 M) than 2D models. When validating models on the independent dataset, slight decreases in all metrics were observed, but AUC and accuracy remained robust (0.95 and 0.92, respectively, for the X3D-L model).</div></div><div><h3>Conclusion</h3><div>We developed a rapid and effective method for automating the selection of coronary angiogram video clips using 3D-CNNs, potentially improving model accuracy and efficiency in clinical applications. The X3D-L model reports a balanced trade-off between computational efficiency and complexity, making it suitable for real-life clinical applications.</div></div>","PeriodicalId":74127,"journal":{"name":"Mayo Clinic Proceedings. Digital health","volume":"3 1","pages":"Article 100195"},"PeriodicalIF":0.0,"publicationDate":"2025-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143551149","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":"In Celebration of Mayo Clinic Proceedings: Digital Health","authors":"Fredric B. Meyer MD,&nbsp;John D. Poe MBA","doi":"10.1016/j.mcpdig.2025.100204","DOIUrl":"10.1016/j.mcpdig.2025.100204","url":null,"abstract":"","PeriodicalId":74127,"journal":{"name":"Mayo Clinic Proceedings. Digital health","volume":"3 1","pages":"Article 100204"},"PeriodicalIF":0.0,"publicationDate":"2025-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143551148","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":"External Validation of Persistent Severe Acute Kidney Injury Prediction With Machine Learning Model","authors":"Simone Zappalà PhD ,&nbsp;Francesca Alfieri MS ,&nbsp;Andrea Ancona PhD ,&nbsp;Antonio M. Dell’Anna MD ,&nbsp;Kianoush B. Kashani MD, MS","doi":"10.1016/j.mcpdig.2025.100200","DOIUrl":"10.1016/j.mcpdig.2025.100200","url":null,"abstract":"<div><h3>Objective</h3><div>To externally validate the persistent electronic alert (PersEA) machine learning model for predicting persistent severe acute kidney injury (psAKI), addressing the scarcity of validated prediction models.</div></div><div><h3>Patients and Methods</h3><div>We included adult patients (18 years or older) admitted to intensive care unit with at least stage 2 acute kidney injury (AKI) at a tertiary medical center, using retrospective data collected between January 1st, 2017 and December 31st, 2022. The data were accessed and analyzed during the period from March 1st, 2023, through July 28th, 2023. The psAKI was defined as AKI stage 3 lasting for ≥72 hours or AKI leading to death in 48 hours or kidney replacement therapy in 1 day. The performance of the PersEA model, a boosted tree algorithm fed by hourly patient data via electronic health records to provide real-time psAKI predictions, was evaluated using specific metrics that penalize late alarms. We measured the area under the receiver operating characteristic and the area under the precision-recall curves.</div></div><div><h3>Results</h3><div>After screening, 4479 patients from the Mayo Clinic cohort were included in the current external validation study, with 234 (5.22%) having psAKI. Results from the Amsterdam UMCdb (531 patients, 59 [11.11%] positive) and MIMIC-III (495 patients, 57 [11.52%] positive) cohorts were obtained in a prior development study. The model demonstrated an area under the receiver operating characteristic curve of 0.98 (95% CI, 0.97-0.98) and an area under the precision-recall curve of 0.67 (95% CI, 0.60-0.73), and when applying the threshold that reached 0.80 sensitivity on the internal cohort, PersEA achieved 0.88 sensitivity, 0.94 specificity, and 0.47 precision, all based on Mayo Clinic data.</div></div><div><h3>Conclusion</h3><div>The PersEA model performed excellently on an external cohort, showing that it is scalable on high-quality data with little to no tuning once a noisy training set is chosen.</div></div>","PeriodicalId":74127,"journal":{"name":"Mayo Clinic Proceedings. Digital health","volume":"3 2","pages":"Article 100200"},"PeriodicalIF":0.0,"publicationDate":"2025-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143748494","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":"Topological Data Analysis in the Assessment of Coronary Atherosclerosis: A Comprehensive Narrative Review","authors":"Yashbir Singh ME, PhD ,&nbsp;Quincy A. Hathaway MD, PhD ,&nbsp;Colleen Farrelly MS ,&nbsp;Matthew J. Budoff MD ,&nbsp;Bradley Erickson MD, PhD ,&nbsp;Jeremy D. Collins MD ,&nbsp;Michael J. Blaha MD ,&nbsp;Tim Leiner MD, PhD ,&nbsp;Francisco Lopez-Jimenez MD, MBA ,&nbsp;Jennifer Rozenblit PhD ,&nbsp;Deepa Sarkar PhD ,&nbsp;Gunnar Carlsson PhD","doi":"10.1016/j.mcpdig.2025.100199","DOIUrl":"10.1016/j.mcpdig.2025.100199","url":null,"abstract":"<div><div>Plaque build-up in the coronary arteries can restrict blood flow or become unstable and contribute to cardiovascular events. Evaluating these plaques is essential for cardiovascular risk stratification and determining the most appropriate treatment strategies. This review aimed to investigate the potential of topological data analysis (TDA), a novel technique, in assessing coronary atherosclerosis. We perform a narrative review of the emerging TDA field, discussing its basic principles and advantages over traditional methods, which often struggle to capture the intricate patterns and relationships within coronary plaques. Topological data analysis can evaluate coronary atherosclerosis in novel ways using variations in the cover and filtration functions of the algorithm, highlighting new ways to characterize calcified and noncalcified plaques. It also highlights the potential of TDA in enhancing the understanding and management of this complex disease. Despite the challenges and future directions, which need to be addressed, TDA shows potential as a valuable tool in assessing coronary atherosclerosis, specifically in noncalcified plaque detection, possibly surpassing traditional techniques in capturing the disease’s intricate patterns. A Pubmed, Google Scholar, Scopus - based literature review was undertaken using the following search terms: <em>Topological data analysis in Coronary atherosclerosis, Topological data analysis in Cardiovascular</em>.</div></div>","PeriodicalId":74127,"journal":{"name":"Mayo Clinic Proceedings. Digital health","volume":"3 2","pages":"Article 100199"},"PeriodicalIF":0.0,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143696681","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":"Predicting Inpatient Admissions From Emergency Department Triage Using Machine Learning: A Systematic Review","authors":"Ethan L. Williams MD ,&nbsp;Daniel Huynh MD ,&nbsp;Mohamed Estai MBBS, PhD ,&nbsp;Toshi Sinha PhD ,&nbsp;Matthew Summerscales MBBS ,&nbsp;Yogesan Kanagasingam PhD","doi":"10.1016/j.mcpdig.2025.100197","DOIUrl":"10.1016/j.mcpdig.2025.100197","url":null,"abstract":"<div><div>This study aimed to evaluate the quality of evidence for using machine learning models to predict inpatient admissions from emergency department triage data, ultimately aiming to improve patient flow management. A comprehensive literature search was conducted according to the PRISMA guidelines across 5 databases, PubMed, Embase, Web of Science, Scopus, and CINAHL, on August 1, 2024, for English-language studies published between August 1, 2014, and August 1, 2024. This yielded 700 articles, of which 66 were screened in full, and 31 met the inclusion and exclusion criteria. Model quality was assessed using the PROBAST appraisal tool and a modified TRIPOD+AI framework, alongside reported model performance metrics. Seven studies demonstrated rigorous methodology and promising in silico performance, with an area under the receiver operating characteristic ranging from 0.81 to 0.93. However, further performance analysis was limited by heterogeneity in model development and an unclear-to-high risk of bias and applicability concerns in the remaining 24 articles, as evaluated by the PROBAST tool. The current literature demonstrates a good degree of in silico accuracy in predicting inpatient admission from triage data alone. Future research should emphasize transparent model development and reporting, temporal validation, concept drift analysis, exploration of emerging artificial intelligence techniques, and analysis of real-world patient flow metrics to comprehensively assess the usefulness of these models.</div></div>","PeriodicalId":74127,"journal":{"name":"Mayo Clinic Proceedings. Digital health","volume":"3 1","pages":"Article 100197"},"PeriodicalIF":0.0,"publicationDate":"2025-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143444616","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":"Policing the Boundary Between Responsible and Irresponsible Placing on the Market of Large Language Model Health Applications","authors":"Oscar Freyer ,&nbsp;Isabella C. Wiest Dr. med ,&nbsp;Stephen Gilbert PhD","doi":"10.1016/j.mcpdig.2025.100196","DOIUrl":"10.1016/j.mcpdig.2025.100196","url":null,"abstract":"","PeriodicalId":74127,"journal":{"name":"Mayo Clinic Proceedings. Digital health","volume":"3 1","pages":"Article 100196"},"PeriodicalIF":0.0,"publicationDate":"2025-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143402575","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":"Leveraging Comprehensive Echo Data to Power Artificial Intelligence Models for Handheld Cardiac Ultrasound","authors":"D.M. Anisuzzaman PhD,&nbsp;Jeffrey G. Malins PhD,&nbsp;John I. Jackson PhD,&nbsp;Eunjung Lee PhD,&nbsp;Jwan A. Naser MBBS,&nbsp;Behrouz Rostami PhD,&nbsp;Grace Greason BA,&nbsp;Jared G. Bird MD,&nbsp;Paul A. Friedman MD,&nbsp;Jae K. Oh MD,&nbsp;Patricia A. Pellikka MD,&nbsp;Jeremy J. Thaden MD,&nbsp;Francisco Lopez-Jimenez MD, MSc, MBA,&nbsp;Zachi I. Attia PhD,&nbsp;Sorin V. Pislaru MD, PhD,&nbsp;Garvan C. Kane MD, PhD","doi":"10.1016/j.mcpdig.2025.100194","DOIUrl":"10.1016/j.mcpdig.2025.100194","url":null,"abstract":"<div><h3>Objective</h3><div>To develop a fully end-to-end deep learning framework capable of estimating left ventricular ejection fraction (LVEF), estimating patient age, and classifying patient sex from echocardiographic videos, including videos collected using handheld cardiac ultrasound (HCU).</div></div><div><h3>Patients and Methods</h3><div>Deep learning models were trained using retrospective transthoracic echocardiography (TTE) data collected in Mayo Clinic Rochester and surrounding Mayo Clinic Health System sites (training: 6432 studies and internal validation: 1369 studies). Models were then evaluated using retrospective TTE data from the 3 Mayo Clinic sites (Rochester, n=1970; Arizona, n=1367; Florida, n=1562) before being applied to a prospective dataset of handheld ultrasound and TTE videos collected from 625 patients. Study data were collected between January 1, 2018 and February 29, 2024.</div></div><div><h3>Results</h3><div>Models showed strong performance on the retrospective TTE datasets (LVEF regression: root mean squared error (RMSE)=6.83%, 6.53%, and 6.95% for Rochester, Arizona, and Florida cohorts, respectively; classification of LVEF ≤40% versus LVEF &gt; 40%: area under curve (AUC)=0.962, 0.967, and 0.980 for Rochester, Arizona, and Florida, respectively; age: RMSE=9.44% for Rochester; sex: AUC=0.882 for Rochester), and performed comparably for prospective HCU versus TTE data (LVEF regression: RMSE=6.37% for HCU vs 5.57% for TTE; LVEF classification: AUC=0.974 vs 0.981; age: RMSE=10.35% vs 9.32%; sex: AUC=0.896 vs 0.933).</div></div><div><h3>Conclusion</h3><div>Robust TTE datasets can be used to effectively power HCU deep learning models, which in turn demonstrates focused diagnostic images can be obtained with handheld devices.</div></div>","PeriodicalId":74127,"journal":{"name":"Mayo Clinic Proceedings. Digital health","volume":"3 1","pages":"Article 100194"},"PeriodicalIF":0.0,"publicationDate":"2025-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143349121","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}