Pub Date : 2024-11-19DOI: 10.1038/s43856-024-00670-0
Sophie Starck, Vasiliki Sideri-Lampretsa, Jessica J M Ritter, Veronika A Zimmer, Rickmer Braren, Tamara T Mueller, Daniel Rueckert
Background: Reliable reference data in medical imaging is largely unavailable. Developing tools that allow for the comparison of individual patient data to reference data has a high potential to improve diagnostic imaging. Population atlases are a commonly used tool in medical imaging to facilitate this. Constructing such atlases becomes particularly challenging when working with highly heterogeneous datasets, such as whole-body images, which contain significant anatomical variations.
Method: In this work, we propose a pipeline for generating a standardised whole-body atlas for a highly heterogeneous population by partitioning the population into anatomically meaningful subgroups. Using magnetic resonance images from the UK Biobank dataset, we create six whole-body atlases representing a healthy population average. We furthermore unbias them, and this way obtain a realistic representation of the population. In addition to the anatomical atlases, we generate probabilistic atlases that capture the distributions of abdominal fat (visceral and subcutaneous) and five abdominal organs across the population (liver, spleen, pancreas, left and right kidneys).
Results: Our pipeline effectively generates high-quality, realistic whole-body atlases with clinical applicability. The probabilistic atlases show differences in fat distribution between subjects with medical conditions such as diabetes and cardiovascular diseases and healthy subjects in the atlas space.
Conclusions: With this work, we make the constructed anatomical and label atlases publically available, with the expectation that they will support medical research involving whole-body MR images.
{"title":"Using UK Biobank data to establish population-specific atlases from whole body MRI.","authors":"Sophie Starck, Vasiliki Sideri-Lampretsa, Jessica J M Ritter, Veronika A Zimmer, Rickmer Braren, Tamara T Mueller, Daniel Rueckert","doi":"10.1038/s43856-024-00670-0","DOIUrl":"https://doi.org/10.1038/s43856-024-00670-0","url":null,"abstract":"<p><strong>Background: </strong>Reliable reference data in medical imaging is largely unavailable. Developing tools that allow for the comparison of individual patient data to reference data has a high potential to improve diagnostic imaging. Population atlases are a commonly used tool in medical imaging to facilitate this. Constructing such atlases becomes particularly challenging when working with highly heterogeneous datasets, such as whole-body images, which contain significant anatomical variations.</p><p><strong>Method: </strong>In this work, we propose a pipeline for generating a standardised whole-body atlas for a highly heterogeneous population by partitioning the population into anatomically meaningful subgroups. Using magnetic resonance images from the UK Biobank dataset, we create six whole-body atlases representing a healthy population average. We furthermore unbias them, and this way obtain a realistic representation of the population. In addition to the anatomical atlases, we generate probabilistic atlases that capture the distributions of abdominal fat (visceral and subcutaneous) and five abdominal organs across the population (liver, spleen, pancreas, left and right kidneys).</p><p><strong>Results: </strong>Our pipeline effectively generates high-quality, realistic whole-body atlases with clinical applicability. The probabilistic atlases show differences in fat distribution between subjects with medical conditions such as diabetes and cardiovascular diseases and healthy subjects in the atlas space.</p><p><strong>Conclusions: </strong>With this work, we make the constructed anatomical and label atlases publically available, with the expectation that they will support medical research involving whole-body MR images.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":"4 1","pages":"237"},"PeriodicalIF":5.4,"publicationDate":"2024-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142677898","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-19DOI: 10.1038/s43856-024-00664-y
Ruth E Costello, Karen M J Waller, Rachel Smith, George F Mells, Angel Y S Wong, Anna Schultze, Viyaasan Mahalingasivam, Emily Herrett, Bang Zheng, Liang-Yu Lin, Brian MacKenna, Amir Mehrkar, Sebastian C J Bacon, Ben Goldacre, Laurie A Tomlinson, John Tazare, Christopher T Rentsch
Background: Biological evidence suggests ursodeoxycholic acid (UDCA)-a common treatment of cholestatic liver disease-may prevent severe COVID-19 outcomes. We aimed to compare the hazard of COVID-19 hospitalisation or death between UDCA users versus non-users in a population with primary biliary cholangitis (PBC) or primary sclerosing cholangitis (PSC).
Methods: With the approval of NHS England, we conducted a population-based cohort study using primary care records between 1 March 2020 and 31 December 2022, linked to death registration data and hospital records through the OpenSAFELY-TPP platform. Cox proportional hazards regression was used to estimate hazard ratios (HR) and 95% confidence intervals (CI) for the association between time-varying UDCA exposure and COVID-19 related hospitalisation or death, stratified by geographical region and considering models unadjusted and fully adjusted for pre-specified confounders.
Results: We identify 11,305 eligible individuals, 640 were hospitalised or died with COVID-19 during follow-up, 400 (63%) events among UDCA users. After confounder adjustment, UDCA is associated with a 21% relative reduction in the hazard of COVID-19 hospitalisation or death (HR 0.79, 95% CI 0.67-0.93), consistent with an absolute risk reduction of 1.35% (95% CI 1.07%-1.69%).
Conclusions: We found evidence that UDCA is associated with a lower hazard of COVID-19 related hospitalisation and death, support calls for clinical trials investigating UDCA as a preventative measure for severe COVID-19 outcomes.
{"title":"Ursodeoxycholic acid and severe COVID-19 outcomes in a cohort study using the OpenSAFELY platform.","authors":"Ruth E Costello, Karen M J Waller, Rachel Smith, George F Mells, Angel Y S Wong, Anna Schultze, Viyaasan Mahalingasivam, Emily Herrett, Bang Zheng, Liang-Yu Lin, Brian MacKenna, Amir Mehrkar, Sebastian C J Bacon, Ben Goldacre, Laurie A Tomlinson, John Tazare, Christopher T Rentsch","doi":"10.1038/s43856-024-00664-y","DOIUrl":"https://doi.org/10.1038/s43856-024-00664-y","url":null,"abstract":"<p><strong>Background: </strong>Biological evidence suggests ursodeoxycholic acid (UDCA)-a common treatment of cholestatic liver disease-may prevent severe COVID-19 outcomes. We aimed to compare the hazard of COVID-19 hospitalisation or death between UDCA users versus non-users in a population with primary biliary cholangitis (PBC) or primary sclerosing cholangitis (PSC).</p><p><strong>Methods: </strong>With the approval of NHS England, we conducted a population-based cohort study using primary care records between 1 March 2020 and 31 December 2022, linked to death registration data and hospital records through the OpenSAFELY-TPP platform. Cox proportional hazards regression was used to estimate hazard ratios (HR) and 95% confidence intervals (CI) for the association between time-varying UDCA exposure and COVID-19 related hospitalisation or death, stratified by geographical region and considering models unadjusted and fully adjusted for pre-specified confounders.</p><p><strong>Results: </strong>We identify 11,305 eligible individuals, 640 were hospitalised or died with COVID-19 during follow-up, 400 (63%) events among UDCA users. After confounder adjustment, UDCA is associated with a 21% relative reduction in the hazard of COVID-19 hospitalisation or death (HR 0.79, 95% CI 0.67-0.93), consistent with an absolute risk reduction of 1.35% (95% CI 1.07%-1.69%).</p><p><strong>Conclusions: </strong>We found evidence that UDCA is associated with a lower hazard of COVID-19 related hospitalisation and death, support calls for clinical trials investigating UDCA as a preventative measure for severe COVID-19 outcomes.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":"4 1","pages":"238"},"PeriodicalIF":5.4,"publicationDate":"2024-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142677865","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-18DOI: 10.1038/s43856-024-00673-x
Jia Wei, Jiandong Zhou, Zizheng Zhang, Kevin Yuan, Qingze Gu, Augustine Luk, Andrew J Brent, David A Clifton, A Sarah Walker, David W Eyre
Background: Accurately predicting hospital discharge events could help improve patient flow and the efficiency of healthcare delivery. However, using machine learning and diverse electronic health record (EHR) data for this task remains incompletely explored.
Methods: We used EHR data from February-2017 to January-2020 from Oxfordshire, UK to predict hospital discharges in the next 24 h. We fitted separate extreme gradient boosting models for elective and emergency admissions, trained on the first two years of data and tested on the final year of data. We examined individual-level and hospital-level model performance and evaluated the impact of training data size and recency, prediction time, and performance in subgroups.
Results: Our models achieve AUROCs of 0.87 and 0.86, AUPRCs of 0.66 and 0.64, and F1 scores of 0.61 and 0.59 for elective and emergency admissions, respectively. These models outperform a logistic regression model using the same features and are substantially better than a baseline logistic regression model with more limited features. Notably, the relative performance increase from adding additional features is greater than the increase from using a sophisticated model. Aggregating individual probabilities, daily total discharge estimates are accurate with mean absolute errors of 8.9% (elective) and 4.9% (emergency). The most informative predictors include antibiotic prescriptions, medications, and hospital capacity factors. Performance remains robust across patient subgroups and different training strategies, but is lower in patients with longer admissions and those who died in hospital.
Conclusions: Our findings highlight the potential of machine learning in optimising hospital patient flow and facilitating patient care and recovery.
{"title":"Predicting individual patient and hospital-level discharge using machine learning.","authors":"Jia Wei, Jiandong Zhou, Zizheng Zhang, Kevin Yuan, Qingze Gu, Augustine Luk, Andrew J Brent, David A Clifton, A Sarah Walker, David W Eyre","doi":"10.1038/s43856-024-00673-x","DOIUrl":"10.1038/s43856-024-00673-x","url":null,"abstract":"<p><strong>Background: </strong>Accurately predicting hospital discharge events could help improve patient flow and the efficiency of healthcare delivery. However, using machine learning and diverse electronic health record (EHR) data for this task remains incompletely explored.</p><p><strong>Methods: </strong>We used EHR data from February-2017 to January-2020 from Oxfordshire, UK to predict hospital discharges in the next 24 h. We fitted separate extreme gradient boosting models for elective and emergency admissions, trained on the first two years of data and tested on the final year of data. We examined individual-level and hospital-level model performance and evaluated the impact of training data size and recency, prediction time, and performance in subgroups.</p><p><strong>Results: </strong>Our models achieve AUROCs of 0.87 and 0.86, AUPRCs of 0.66 and 0.64, and F1 scores of 0.61 and 0.59 for elective and emergency admissions, respectively. These models outperform a logistic regression model using the same features and are substantially better than a baseline logistic regression model with more limited features. Notably, the relative performance increase from adding additional features is greater than the increase from using a sophisticated model. Aggregating individual probabilities, daily total discharge estimates are accurate with mean absolute errors of 8.9% (elective) and 4.9% (emergency). The most informative predictors include antibiotic prescriptions, medications, and hospital capacity factors. Performance remains robust across patient subgroups and different training strategies, but is lower in patients with longer admissions and those who died in hospital.</p><p><strong>Conclusions: </strong>Our findings highlight the potential of machine learning in optimising hospital patient flow and facilitating patient care and recovery.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":"4 1","pages":"236"},"PeriodicalIF":5.4,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142670099","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-18DOI: 10.1038/s43856-024-00662-0
Thiago Bassi, Elizabeth Rohrs E, Melodie Parfait, Brett C Hannigan, Steven Reynolds, Julien Mayaux, Maxens Decavèle, Jose Herrero, Alexandre Demoule, Thomas Similowski, Martin Dres
Background: In critically ill patients, deep sedation and mechanical ventilation suppress the brain-diaphragm-lung axis and are associated with cognitive issues in survivors.
Methods: This exploratory crossover design study investigates whether phrenic nerve stimulation can enhance brain activity and connectivity in six deeply sedated, mechanically ventilated patients with acute respiratory distress syndrome.
Results: Our findings indicate that adding phrenic stimulation on top of invasive mechanical ventilation in deeply sedated, critically ill, moderate acute respiratory distress syndrome patients increases cortical activity, connectivity, and synchronization in the frontal-temporal-parietal cortices.
Conclusions: Adding phrenic stimulation on top of invasive mechanical ventilation in deeply sedated, critically ill, moderate acute respiratory distress syndrome patients increases cortical activity, connectivity, and synchronization. The observed changes resemble those during diaphragmatic breathing in awake humans. These results suggest that phrenic nerve stimulation has the potential to restore the brain-diaphragm-lung crosstalk when it has been shut down or impaired by mechanical ventilation and sedation. Further research should evaluate the clinical significance of these results.
{"title":"Restoring brain connectivity by phrenic nerve stimulation in sedated and mechanically ventilated patients.","authors":"Thiago Bassi, Elizabeth Rohrs E, Melodie Parfait, Brett C Hannigan, Steven Reynolds, Julien Mayaux, Maxens Decavèle, Jose Herrero, Alexandre Demoule, Thomas Similowski, Martin Dres","doi":"10.1038/s43856-024-00662-0","DOIUrl":"10.1038/s43856-024-00662-0","url":null,"abstract":"<p><strong>Background: </strong>In critically ill patients, deep sedation and mechanical ventilation suppress the brain-diaphragm-lung axis and are associated with cognitive issues in survivors.</p><p><strong>Methods: </strong>This exploratory crossover design study investigates whether phrenic nerve stimulation can enhance brain activity and connectivity in six deeply sedated, mechanically ventilated patients with acute respiratory distress syndrome.</p><p><strong>Results: </strong>Our findings indicate that adding phrenic stimulation on top of invasive mechanical ventilation in deeply sedated, critically ill, moderate acute respiratory distress syndrome patients increases cortical activity, connectivity, and synchronization in the frontal-temporal-parietal cortices.</p><p><strong>Conclusions: </strong>Adding phrenic stimulation on top of invasive mechanical ventilation in deeply sedated, critically ill, moderate acute respiratory distress syndrome patients increases cortical activity, connectivity, and synchronization. The observed changes resemble those during diaphragmatic breathing in awake humans. These results suggest that phrenic nerve stimulation has the potential to restore the brain-diaphragm-lung crosstalk when it has been shut down or impaired by mechanical ventilation and sedation. Further research should evaluate the clinical significance of these results.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":"4 1","pages":"235"},"PeriodicalIF":5.4,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142670101","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-17DOI: 10.1038/s43856-024-00671-z
Martin Smelik, Yelin Zhao, Dina Mansour Aly, AKM Firoj Mahmud, Oleg Sysoev, Xinxiu Li, Mikael Benson
Cancer screening tests are considered pivotal for early diagnosis and survival. However, the efficacy of these tests for improving survival has recently been questioned. This study aims to test if cancer screening could be improved by biomarkers in peripheral blood based on multi-omics data. We utilize multi-omics data from 500,000 participants in the UK Biobank. Machine learning is applied to search for proteins, metabolites, genetic variants, or clinical variables to diagnose cancers collectively and individually. Here we show that the overall performance of the potential blood biomarkers do not outperform clinical variables for collective diagnosis. However, we observe promising results for individual cancers in close proximity to peripheral blood, with an Area Under the Curve (AUC) greater than 0.8. Our findings suggest that the identification of blood biomarkers for cancer might be complicated by variable overlap between molecular changes in tumor tissues and peripheral blood. This explanation is supported by local proteomics analyses of different tumors, which all show high AUCs, greater than 0.9. Thus, multi-omics biomarkers for the diagnosis of individual cancers may potentially be effective, but not for groups of cancers. This study aimed to find out if we could improve cancer screening tests by looking for signs of cancer in blood samples. We used computer and mathematical models to analyze data from 500,000 people. We found that these blood tests were not better than existing methods for diagnosing multiple types of cancer at once. However, they did show promise for diagnosing individual types of cancer that are close to the bloodstream. This suggests that finding blood markers for cancer is complex and depends on how much the cancer affects the blood. These findings could help in the development of more effective tests for individual types of cancer in the future. Smelik et al. investigate the effectiveness of using multi-omics biomarkers in blood for cancer screening. The results indicate that while these biomarkers show promise for diagnosing individual cancers in close proximity to the blood stream, they do not surpass clinical variables for diagnosing multiple cancers.
{"title":"Multiomics biomarkers were not superior to clinical variables for pan-cancer screening","authors":"Martin Smelik, Yelin Zhao, Dina Mansour Aly, AKM Firoj Mahmud, Oleg Sysoev, Xinxiu Li, Mikael Benson","doi":"10.1038/s43856-024-00671-z","DOIUrl":"10.1038/s43856-024-00671-z","url":null,"abstract":"Cancer screening tests are considered pivotal for early diagnosis and survival. However, the efficacy of these tests for improving survival has recently been questioned. This study aims to test if cancer screening could be improved by biomarkers in peripheral blood based on multi-omics data. We utilize multi-omics data from 500,000 participants in the UK Biobank. Machine learning is applied to search for proteins, metabolites, genetic variants, or clinical variables to diagnose cancers collectively and individually. Here we show that the overall performance of the potential blood biomarkers do not outperform clinical variables for collective diagnosis. However, we observe promising results for individual cancers in close proximity to peripheral blood, with an Area Under the Curve (AUC) greater than 0.8. Our findings suggest that the identification of blood biomarkers for cancer might be complicated by variable overlap between molecular changes in tumor tissues and peripheral blood. This explanation is supported by local proteomics analyses of different tumors, which all show high AUCs, greater than 0.9. Thus, multi-omics biomarkers for the diagnosis of individual cancers may potentially be effective, but not for groups of cancers. This study aimed to find out if we could improve cancer screening tests by looking for signs of cancer in blood samples. We used computer and mathematical models to analyze data from 500,000 people. We found that these blood tests were not better than existing methods for diagnosing multiple types of cancer at once. However, they did show promise for diagnosing individual types of cancer that are close to the bloodstream. This suggests that finding blood markers for cancer is complex and depends on how much the cancer affects the blood. These findings could help in the development of more effective tests for individual types of cancer in the future. Smelik et al. investigate the effectiveness of using multi-omics biomarkers in blood for cancer screening. The results indicate that while these biomarkers show promise for diagnosing individual cancers in close proximity to the blood stream, they do not surpass clinical variables for diagnosing multiple cancers.","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":"1-8"},"PeriodicalIF":5.4,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s43856-024-00671-z.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142645805","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}
Breathing patterns may inform on health. We note that the sites of earliest brain damage in Parkinson’s disease (PD) house the neural pace-makers of respiration. We therefore hypothesized that ongoing long-term temporal dynamics of respiration may be altered in PD. We applied a wearable device that precisely logs nasal airflow over time in 28 PD patients (mostly H&Y stage-II) and 33 matched healthy controls. Each participant wore the device for 24 h of otherwise routine daily living. We observe significantly altered temporal patterns of nasal airflow in PD, where inhalations are longer and less variable than in matched controls (mean PD = −1.22 ± 1.9 (combined respiratory features score), Control = 1.04 ± 2.16, Wilcoxon rank-sum test, z = −4.1, effect size Cliff’s δ = −0.61, 95% confidence interval = −0.79 – (−0.34), P = 4.3 × 10−5). The extent of alteration is such that using only 30 min of recording we detect PD at 87% accuracy (AUC = 0.85, 79% sensitivity (22 of 28), 94% specificity (31 of 33), z = 5.7, p = 3.5 × 10−9), and also predict disease severity (correlation with UPDRS-Total score: r = 0.49; P = 0.008). We conclude that breathing patterns are altered by H&Y stage-II in the disease cascade, and our methods may be further refined in the future to provide an indication with diagnostic and prognostic value. Andelman-Gur et al. use a nasal airflow monitoring device to detect alterations of respiratory dynamics in patients with Parkinson’s Disease. They reveal longer, but less variable, inhalations and show that changes in airflow dynamics are correlated with disease severity, plus 30 min of data is adequate to discriminate patients from controls. In its earliest stages, Parkinson’s disease damages the parts of the brain that control breathing. We built a small device that measures airflow patterns through the nose over time. People with Parkinson’s disease and healthy individuals wore this device for 24 h. We found that nasal inhalations in Parkinson’s patients were longer and less variable than in healthy individuals. This difference was so pronounced that, using only 30 min of recording, we could accurately determine most people who had Parkinson’s disease and how severe their disease was. Future studies will determine whether this tool can contribute to early diagnosis, and it may be useful to monitor disease progression.
{"title":"Discriminating Parkinson’s disease patients from healthy controls using nasal respiratory airflow","authors":"Michal Andelman-Gur, Kobi Snitz, Danielle Honigstein, Aharon Weissbrod, Timna Soroka, Aharon Ravia, Lior Gorodisky, Liron Pinchover, Adi Ezra, Neomi Hezi, Tanya Gurevich, Noam Sobel","doi":"10.1038/s43856-024-00660-2","DOIUrl":"10.1038/s43856-024-00660-2","url":null,"abstract":"Breathing patterns may inform on health. We note that the sites of earliest brain damage in Parkinson’s disease (PD) house the neural pace-makers of respiration. We therefore hypothesized that ongoing long-term temporal dynamics of respiration may be altered in PD. We applied a wearable device that precisely logs nasal airflow over time in 28 PD patients (mostly H&Y stage-II) and 33 matched healthy controls. Each participant wore the device for 24 h of otherwise routine daily living. We observe significantly altered temporal patterns of nasal airflow in PD, where inhalations are longer and less variable than in matched controls (mean PD = −1.22 ± 1.9 (combined respiratory features score), Control = 1.04 ± 2.16, Wilcoxon rank-sum test, z = −4.1, effect size Cliff’s δ = −0.61, 95% confidence interval = −0.79 – (−0.34), P = 4.3 × 10−5). The extent of alteration is such that using only 30 min of recording we detect PD at 87% accuracy (AUC = 0.85, 79% sensitivity (22 of 28), 94% specificity (31 of 33), z = 5.7, p = 3.5 × 10−9), and also predict disease severity (correlation with UPDRS-Total score: r = 0.49; P = 0.008). We conclude that breathing patterns are altered by H&Y stage-II in the disease cascade, and our methods may be further refined in the future to provide an indication with diagnostic and prognostic value. Andelman-Gur et al. use a nasal airflow monitoring device to detect alterations of respiratory dynamics in patients with Parkinson’s Disease. They reveal longer, but less variable, inhalations and show that changes in airflow dynamics are correlated with disease severity, plus 30 min of data is adequate to discriminate patients from controls. In its earliest stages, Parkinson’s disease damages the parts of the brain that control breathing. We built a small device that measures airflow patterns through the nose over time. People with Parkinson’s disease and healthy individuals wore this device for 24 h. We found that nasal inhalations in Parkinson’s patients were longer and less variable than in healthy individuals. This difference was so pronounced that, using only 30 min of recording, we could accurately determine most people who had Parkinson’s disease and how severe their disease was. Future studies will determine whether this tool can contribute to early diagnosis, and it may be useful to monitor disease progression.","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":"1-9"},"PeriodicalIF":5.4,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11564766/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142633483","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}
Pub Date : 2024-11-13DOI: 10.1038/s43856-024-00666-w
Julian Klug, Guillaume Leclerc, Elisabeth Dirren, Emmanuel Carrera
Prediction of outcome after stroke is critical for treatment planning and resource allocation but is complicated by fluctuations during the first days after onset. We propose a machine learning model that can provide hourly predictions based on the integration of continuous variables acquired within 72 h of hospital admission. We analyzed 2492 admissions for ischemic stroke in the Geneva University Hospital from 01.01.2018 to 31.12.2021, amounting to 2’131’752 unique data points. We developed a transformer model that continuously included clinical, physiological, imaging, and biological data recorded within 72 h of admission. This model was trained to generate hourly predictions of mortality and morbidity. Shapley additive explanations were used to identify the most relevant predictors to explain outcomes for each patient. The MIMIC-III database was used for external validation. Our transformer model predicts mortality, with an area under the receiver operating characteristic curve of 0.830 (95% CI 0.763–0.885) on admission, reaching 0.893 (95% CI 0.839–0.933) 72 h later for a 3-month outcome. Validated in an independent cohort, it outperforms all static models. Based on their mean explanatory weights, the top predictors included continuous clinical evaluation, baseline patient characteristics, timing from admission to acute treatment, and markers of inflammation and organ dysfunction. The performance of our transformer model demonstrates the potential of machine learning models integrating clinical, physiological, imaging, and biological variables over time after stroke. The clinical applicability of our model is further strengthened by access to hourly updated predictions along with accompanying explanations. Stroke is the most frequent cause of disability in industrialized countries. To determine the best treatment and allocate resources, an early and accurate prediction of outcome is essential. Although modern stroke units gather a continuous stream of data, existing tools for outcome prediction are rarely used as they are static and fail to adapt to the evolving condition of the patient. We developed a machine learning model, a computer system learning from existing data, to provide real-time predictions of in-hospital mortality and 3-month outcomes. Our model was able to provide accurate hourly prediction of outcome based on regularly updated clinical data obtained from the patient. This study demonstrates the potential of integrating the continuous data stream recorded in the electronic health record after stroke. Similar predictive models could help personalize treatment planning, empower patients and their families through counseling, and facilitate resource allocation. Klug et al. present a machine learning model for continuous monitoring and prediction of functional outcome after acute ischemic stroke. Integrating clinical, physiological, and biological variables over time, the system detects patients at risk as well as potential cause
{"title":"Machine learning for early dynamic prediction of functional outcome after stroke","authors":"Julian Klug, Guillaume Leclerc, Elisabeth Dirren, Emmanuel Carrera","doi":"10.1038/s43856-024-00666-w","DOIUrl":"10.1038/s43856-024-00666-w","url":null,"abstract":"Prediction of outcome after stroke is critical for treatment planning and resource allocation but is complicated by fluctuations during the first days after onset. We propose a machine learning model that can provide hourly predictions based on the integration of continuous variables acquired within 72 h of hospital admission. We analyzed 2492 admissions for ischemic stroke in the Geneva University Hospital from 01.01.2018 to 31.12.2021, amounting to 2’131’752 unique data points. We developed a transformer model that continuously included clinical, physiological, imaging, and biological data recorded within 72 h of admission. This model was trained to generate hourly predictions of mortality and morbidity. Shapley additive explanations were used to identify the most relevant predictors to explain outcomes for each patient. The MIMIC-III database was used for external validation. Our transformer model predicts mortality, with an area under the receiver operating characteristic curve of 0.830 (95% CI 0.763–0.885) on admission, reaching 0.893 (95% CI 0.839–0.933) 72 h later for a 3-month outcome. Validated in an independent cohort, it outperforms all static models. Based on their mean explanatory weights, the top predictors included continuous clinical evaluation, baseline patient characteristics, timing from admission to acute treatment, and markers of inflammation and organ dysfunction. The performance of our transformer model demonstrates the potential of machine learning models integrating clinical, physiological, imaging, and biological variables over time after stroke. The clinical applicability of our model is further strengthened by access to hourly updated predictions along with accompanying explanations. Stroke is the most frequent cause of disability in industrialized countries. To determine the best treatment and allocate resources, an early and accurate prediction of outcome is essential. Although modern stroke units gather a continuous stream of data, existing tools for outcome prediction are rarely used as they are static and fail to adapt to the evolving condition of the patient. We developed a machine learning model, a computer system learning from existing data, to provide real-time predictions of in-hospital mortality and 3-month outcomes. Our model was able to provide accurate hourly prediction of outcome based on regularly updated clinical data obtained from the patient. This study demonstrates the potential of integrating the continuous data stream recorded in the electronic health record after stroke. Similar predictive models could help personalize treatment planning, empower patients and their families through counseling, and facilitate resource allocation. Klug et al. present a machine learning model for continuous monitoring and prediction of functional outcome after acute ischemic stroke. Integrating clinical, physiological, and biological variables over time, the system detects patients at risk as well as potential cause","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":"1-13"},"PeriodicalIF":5.4,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11561255/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142633489","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}
Pub Date : 2024-11-12DOI: 10.1038/s43856-024-00650-4
Thale D. J. Hovdun Patrick-Brown, Andreas Barratt-Due, Marius Trøseid, Anne Ma Dyrhol-Riise, Katerina Nezvalova-Henriksen, Trine Kåsine, Pål Aukrust, Inge C. Olsen, NOR Solidarity consortium
There is an unmet need for treatment of long-term symptoms following COVID-19. Remdesivir is currently the only antiviral approved by the European Medicines Agency for hospitalised patients. Here, we report on the effect of remdesivir in addition to standard of care on long-term symptoms and quality of life in hospitalised patients with COVID-19 as part of the open-label randomised NOR-Solidarity trial (NCT04321616). A total of 185 patients were included in the main trial, of which 118 (60%) were randomised to either remdesivir (n = 42; 36%) or a post-hoc defined control group composed of patients who received standard of care alone or standard of care with hydroxychloroquine (n = 76; 64%). Participants were given quality of life surveys to fill out to gauge their self-reported health over time (the COPD assessment test, the EQ-5D-5L and the RAND SF-36). Here we show that after three months, patients treated with remdesivir do not show significant improvements in stated health compared to those who were not. There are self-reported symptoms of fatigue [mean remdesivir group 2.6 (standard deviation 1.5) v control 2.1 (1.6), 95% confidence interval(CI) −1.17 to 0.15, p = 0.129], shortness of breath [3.0 (1.7) v 2.1 (1.8), 95% CI −1.53 to 0.16, p = 0.110] and coughing [1.8 (1.6) v 1.2 (1.5), 95% CI −1.3 to 0.33, p = 0.237] 3 months after randomisation assessed via the COPD Assessment Test. Our findings indicate that treatment with remdesivir during hospitalisation does not provide any clinically relevant long-term benefit. Remdesivir is a medicine that is used to treat people with COVID-19. It has been found to help people get better faster, but we did not know whether it also relieved them of long-term symptoms such as persistent coughing, fatigue, or shortness of breath. To research this, we randomly assigned hospitalised patients with COVID-19 to either remdesivir on top of their normal care, or only normal care, with or without hydroxychloroquine (a drug later found to have no effect on COVID-19). We then compared participant’s symptoms after 3 months. Our results show that there is probably no benefit of using remdesivir during hospitalisation for long-term symptom relief. Patrick-Brown et al report the findings of a secondary study adjunct to the Nor-Solidarity trial that evaluated remdesivir versus standard of care for the treatment of COVID-19. While remdesivir appears to be safe for use in these patients, there does not appear to be any long-term clinical benefit to its use in terms of long-COVID symptoms.
背景：治疗 COVID-19 后长期症状的需求尚未得到满足。雷米替韦是目前欧洲药品管理局批准用于住院患者的唯一抗病毒药物。在此，我们报告了雷米替韦在标准护理基础上对 COVID-19 住院患者长期症状和生活质量的影响，这是开放标签随机 NOR-Solidarity 试验（NCT04321616）的一部分：主要试验共纳入了185名患者，其中118人（60%）被随机分配到雷米地韦组（n = 42；36%）或由单独接受标准护理或标准护理加羟氯喹的患者组成的事后定义对照组（n = 76；64%）。参试者需填写生活质量调查表，以评估其自我健康状况（慢性阻塞性肺病评估测试、EQ-5D-5L 和 RAND SF-36）：结果：我们在此表明，与未接受治疗的患者相比，接受雷米替韦治疗三个月后，患者的健康状况并没有明显改善。自我报告的症状包括疲劳[平均雷米替韦组 2.6（标准差 1.5）v 对照组 2.1（1.6），95% 置信区间（CI）-1.17 至 0.15，p = 0.129]、气短[3.0（1.7) v 2.1 (1.8), 95% CI -1.53 to 0.16, p = 0.110]和咳嗽[1.8 (1.6) v 1.2 (1.5), 95% CI -1.3 to 0.33, p = 0.237]：我们的研究结果表明，住院期间使用雷米替韦治疗不会带来任何临床相关的长期益处。
{"title":"The effects of remdesivir on long-term symptoms in patients hospitalised for COVID-19: a pre-specified exploratory analysis","authors":"Thale D. J. Hovdun Patrick-Brown, Andreas Barratt-Due, Marius Trøseid, Anne Ma Dyrhol-Riise, Katerina Nezvalova-Henriksen, Trine Kåsine, Pål Aukrust, Inge C. Olsen, NOR Solidarity consortium","doi":"10.1038/s43856-024-00650-4","DOIUrl":"10.1038/s43856-024-00650-4","url":null,"abstract":"There is an unmet need for treatment of long-term symptoms following COVID-19. Remdesivir is currently the only antiviral approved by the European Medicines Agency for hospitalised patients. Here, we report on the effect of remdesivir in addition to standard of care on long-term symptoms and quality of life in hospitalised patients with COVID-19 as part of the open-label randomised NOR-Solidarity trial (NCT04321616). A total of 185 patients were included in the main trial, of which 118 (60%) were randomised to either remdesivir (n = 42; 36%) or a post-hoc defined control group composed of patients who received standard of care alone or standard of care with hydroxychloroquine (n = 76; 64%). Participants were given quality of life surveys to fill out to gauge their self-reported health over time (the COPD assessment test, the EQ-5D-5L and the RAND SF-36). Here we show that after three months, patients treated with remdesivir do not show significant improvements in stated health compared to those who were not. There are self-reported symptoms of fatigue [mean remdesivir group 2.6 (standard deviation 1.5) v control 2.1 (1.6), 95% confidence interval(CI) −1.17 to 0.15, p = 0.129], shortness of breath [3.0 (1.7) v 2.1 (1.8), 95% CI −1.53 to 0.16, p = 0.110] and coughing [1.8 (1.6) v 1.2 (1.5), 95% CI −1.3 to 0.33, p = 0.237] 3 months after randomisation assessed via the COPD Assessment Test. Our findings indicate that treatment with remdesivir during hospitalisation does not provide any clinically relevant long-term benefit. Remdesivir is a medicine that is used to treat people with COVID-19. It has been found to help people get better faster, but we did not know whether it also relieved them of long-term symptoms such as persistent coughing, fatigue, or shortness of breath. To research this, we randomly assigned hospitalised patients with COVID-19 to either remdesivir on top of their normal care, or only normal care, with or without hydroxychloroquine (a drug later found to have no effect on COVID-19). We then compared participant’s symptoms after 3 months. Our results show that there is probably no benefit of using remdesivir during hospitalisation for long-term symptom relief. Patrick-Brown et al report the findings of a secondary study adjunct to the Nor-Solidarity trial that evaluated remdesivir versus standard of care for the treatment of COVID-19. While remdesivir appears to be safe for use in these patients, there does not appear to be any long-term clinical benefit to its use in terms of long-COVID symptoms.","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":"1-7"},"PeriodicalIF":5.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11557865/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142633500","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}
Pub Date : 2024-11-11DOI: 10.1038/s43856-024-00661-1
Dominik J. Ose, Elena Gardner, Morgan Millar, Andrew Curtin, Jiqiang Wu, Mingyuan Zhang, Camie Schaefer, Jing Wang, Jennifer Leiser, Kirsten Stoesser, Bernadette Kiraly
{"title":"Author Correction: A cross-sectional and population-based study from primary care on post-COVID-19 conditions in non-hospitalized patients","authors":"Dominik J. Ose, Elena Gardner, Morgan Millar, Andrew Curtin, Jiqiang Wu, Mingyuan Zhang, Camie Schaefer, Jing Wang, Jennifer Leiser, Kirsten Stoesser, Bernadette Kiraly","doi":"10.1038/s43856-024-00661-1","DOIUrl":"10.1038/s43856-024-00661-1","url":null,"abstract":"","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":"1-1"},"PeriodicalIF":5.4,"publicationDate":"2024-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s43856-024-00661-1.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142599013","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}
Pub Date : 2024-11-09DOI: 10.1038/s43856-024-00651-3
Aaron R. Glick, Colin Jones, Lisa Martignetti, Lisa Blanchette, Theresa Tova, Allen Henderson, Marc D. Pell, Nicole Y. K. Li-Jessen
Professional voice users often experience stigma associated with voice disorders and are reluctant to seek medical help. This study deployed empirical and computational tools to (1) quantify the experience of vocal stigma and help-seeking behaviors in performers; and (2) predict their modulations with peer influences in social networks. Experience of vocal stigma and information-motivation-behavioral (IMB) skills were prospectively profiled using online surveys from a total of 403 Canadians (200 singers and actors and 203 controls). Data were used to formulate an agent-based network model of social interactions on vocal stigma (self-stigma and social-stigma) and help-seeking behaviors. Network analysis was performed to evaluate the effect of social network structure on the flow of IMB among virtual agents. Larger social networks are more likely to contribute to an increase in vocal stigma. For small social networks, total stigma is reduced with higher total IMB but not much so for large networks. For agents with high social-stigma and risk for voice disorder, their vocal stigma is resistant to large changes in IMB ( > 2 standard deviations). Agents with extreme IMB and stigma values are likely to polarize their networks faster in larger social groups. We integrated empirical surveys and computational techniques to contextualize vocal stigma and IMB beyond theory and to quantify the interaction among stigma, health-seeking behavior and influence of social interactions. This work establishes an effective, predictable experimental platform to provide scientific evidence in developing interventions to reduce health stigma in voice disorders and other medical conditions. Voice professionals such as singers and actors can experience stigma if they have a voice disorder. This stigma can result from their personal experience and knowledge (internalized) or be based on input from their peers, employment, and healthcare providers (externalized). To understand how negative vocal stigma spreads, we surveyed the stigma experience of voice professionals and developed computational models. We find that people tend to have more polarized stigma experiences when they are in larger social groups. Vocal stigma is not changed by a person’s knowledge, beliefs, and tendency to seek help. Our method could be used to study other stigmatized health conditions. Our research could also be used to reduce stigma and promote more equitable health care for vocal professionals with a voice disorder. Glick et al. investigate the stigma experience and help-seeking behavior in professional singers and actors using de novo data and social simulation. They find that vocal performers experience greater discrimination against their vocal injury with simulation data also predicting that vocal stigma could be worsened with larger social groups.
{"title":"An integrated empirical and computational study to decipher help-seeking behaviors and vocal stigma","authors":"Aaron R. Glick, Colin Jones, Lisa Martignetti, Lisa Blanchette, Theresa Tova, Allen Henderson, Marc D. Pell, Nicole Y. K. Li-Jessen","doi":"10.1038/s43856-024-00651-3","DOIUrl":"10.1038/s43856-024-00651-3","url":null,"abstract":"Professional voice users often experience stigma associated with voice disorders and are reluctant to seek medical help. This study deployed empirical and computational tools to (1) quantify the experience of vocal stigma and help-seeking behaviors in performers; and (2) predict their modulations with peer influences in social networks. Experience of vocal stigma and information-motivation-behavioral (IMB) skills were prospectively profiled using online surveys from a total of 403 Canadians (200 singers and actors and 203 controls). Data were used to formulate an agent-based network model of social interactions on vocal stigma (self-stigma and social-stigma) and help-seeking behaviors. Network analysis was performed to evaluate the effect of social network structure on the flow of IMB among virtual agents. Larger social networks are more likely to contribute to an increase in vocal stigma. For small social networks, total stigma is reduced with higher total IMB but not much so for large networks. For agents with high social-stigma and risk for voice disorder, their vocal stigma is resistant to large changes in IMB ( > 2 standard deviations). Agents with extreme IMB and stigma values are likely to polarize their networks faster in larger social groups. We integrated empirical surveys and computational techniques to contextualize vocal stigma and IMB beyond theory and to quantify the interaction among stigma, health-seeking behavior and influence of social interactions. This work establishes an effective, predictable experimental platform to provide scientific evidence in developing interventions to reduce health stigma in voice disorders and other medical conditions. Voice professionals such as singers and actors can experience stigma if they have a voice disorder. This stigma can result from their personal experience and knowledge (internalized) or be based on input from their peers, employment, and healthcare providers (externalized). To understand how negative vocal stigma spreads, we surveyed the stigma experience of voice professionals and developed computational models. We find that people tend to have more polarized stigma experiences when they are in larger social groups. Vocal stigma is not changed by a person’s knowledge, beliefs, and tendency to seek help. Our method could be used to study other stigmatized health conditions. Our research could also be used to reduce stigma and promote more equitable health care for vocal professionals with a voice disorder. Glick et al. investigate the stigma experience and help-seeking behavior in professional singers and actors using de novo data and social simulation. They find that vocal performers experience greater discrimination against their vocal injury with simulation data also predicting that vocal stigma could be worsened with larger social groups.","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":"1-13"},"PeriodicalIF":5.4,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s43856-024-00651-3.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142599016","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}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号