{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T15:18:32Z","timestamp":1787066312313,"version":"build-2736575974"},"posted":{"date-parts":[[2023,10,5]]},"group-title":"Cardiovascular Medicine","reference-count":49,"publisher":"openRxiv","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2023,10,5]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                <jats:sec>\n                  <jats:title>Background<\/jats:title>\n                  <jats:p>Extracting explainable flow metrics is a bottleneck to the clinical translation of advanced cardiac flow imaging modalities. We hypothesized that reduced-order models (ROMs) of intraventricular flow are a suitable strategy for deriving simple and interpretable clinical metrics suitable for further assessments. Combined with machine learning (ML) flow-based ROMs could provide new insight to help diagnose and risk-stratify patients.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>We analyzed 2D color-Doppler echocardiograms of 81 non-ischemic dilated cardiomyopathy (DCM) patients, 51 hypertrophic cardiomyopathy (HCM) patients, and 77 normal volunteers (Control). We applied proper orthogonal decomposition (POD) to build patient-specific and cohort-specific ROMs of LV flow. Each ROM aggregates a low number of components representing a spatially dependent velocity map modulated along the cardiac cycle by a time-dependent coefficient. We tested three classifiers using deliberately simple ML analyses of these ROMs with varying supervision levels. In supervised models, hyperparameter gridsearch was used to derive the ROMs that maximize classification power. The classifiers were blinded to LV chamber geometry and function. We ran vector flow mapping on the color-Doppler sequences to help visualize flow patterns and interpret the ML results.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>POD-based ROMs stably represented each cohort through 10-fold cross-validation. The principal POD mode captured &gt;80% of the flow kinetic energy (KE) in all cohorts and represented the LV filling\/emptying jets. Mode 2 represented the diastolic vortex and its KE contribution ranged from &lt;1% (HCM) to 13% (DCM). Semi-unsupervised classification using patient-specific ROMs revealed that the KE ratio of these two principal modes, the vortex-to-jet (V2J) energy ratio, is a simple, interpretable metric that discriminates DCM, HCM, and Control patients. Receiver operating characteristic curves using V2J as classifier had areas under the curve of 0.81, 0.91, and 0.95 for distinguishing HCM vs. Control, DCM vs. Control, and DCM vs. HCM, respectively.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusions<\/jats:title>\n                  <jats:p>Modal decomposition of cardiac flow can be used to create ROMs of normal and pathological flow patterns, uncovering simple interpretable flow metrics with power to discriminate disease states, and particularly suitable for further processing using ML.<\/jats:p>\n                <\/jats:sec>","DOI":"10.1101\/2023.10.03.23296524","type":"posted-content","created":{"date-parts":[[2023,10,5]],"date-time":"2023-10-05T21:10:21Z","timestamp":1696540221000},"source":"Crossref","is-referenced-by-count":2,"title":["Deriving Explainable Metrics of Left Ventricular Flow by Reduced-Order Modeling and Classification"],"prefix":"10.64898","author":[{"given":"Mar\u00eda Guadalupe","family":"Borja","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pablo","family":"Martinez-Legazpi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Cathleen","family":"Nguyen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Oscar","family":"Flores","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Andrew M.","family":"Kahn","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Javier","family":"Bermejo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5683-0239","authenticated-orcid":false,"given":"Juan C.","family":"del \u00c1lamo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2023101012550908000_2023.10.03.23296524v1.1","doi-asserted-by":"crossref","first-page":"2134","DOI":"10.1038\/s41467-018-04608-8","article-title":"Exploring patterns enriched in a dataset with contrastive principal component analysis","volume":"9","year":"2018","journal-title":"Nat Commun"},{"key":"2023101012550908000_2023.10.03.23296524v1.2","doi-asserted-by":"crossref","unstructured":"Agrawal, T. , 2020. Hyperparameter Optimization in Machine Learning. Apress Berkeley, CA.","DOI":"10.1007\/978-1-4842-6579-6"},{"key":"2023101012550908000_2023.10.03.23296524v1.3","first-page":"1405","article-title":"Vector flow mapping: A review from theory to practice","volume":"38","year":"2021","journal-title":"Echocardiography"},{"key":"2023101012550908000_2023.10.03.23296524v1.4","doi-asserted-by":"crossref","first-page":"485","DOI":"10.3389\/fphys.2019.00485","article-title":"Age-Dependence of Flow Homeostasis in the Left Ventricle","volume":"10","year":"2019","journal-title":"Front Physiol"},{"key":"2023101012550908000_2023.10.03.23296524v1.5","doi-asserted-by":"publisher","DOI":"10.1152\/ajpheart.00697.2013"},{"key":"2023101012550908000_2023.10.03.23296524v1.6","doi-asserted-by":"crossref","first-page":"315","DOI":"10.1146\/annurev-fluid-010814-014728","article-title":"The clinical assessment of intracardiac flows","volume":"47","year":"2015","journal-title":"Ann Rev Fluid Mech"},{"key":"2023101012550908000_2023.10.03.23296524v1.7","doi-asserted-by":"crossref","unstructured":"Brunton, S.L. , Noack, B.R. , Koumoutsakos, P. , 2020. Machine Learning for Fluid Mechanics. Ann Rev Fluid Mech 52.","DOI":"10.52843\/cassyni.27tbdb"},{"key":"2023101012550908000_2023.10.03.23296524v1.8","doi-asserted-by":"crossref","unstructured":"Celebi, M.E. , Aydin, K. , 2016. Unsupervised Learning Algorithms. Springer Cham.","DOI":"10.1007\/978-3-319-24211-8"},{"key":"2023101012550908000_2023.10.03.23296524v1.9","doi-asserted-by":"publisher","DOI":"10.1007\/s10439-013-0755-0"},{"key":"2023101012550908000_2023.10.03.23296524v1.10","doi-asserted-by":"crossref","first-page":"3501","DOI":"10.1016\/j.ultrasmedbio.2021.08.021","article-title":"Intraventricular Vector Flow Imaging with Blood Speckle Tracking in Adults: Feasibility, Normal Physiology and Mechanisms in Healthy Volunteers","volume":"47","year":"2021","journal-title":"Ultrasound Med Biol"},{"key":"2023101012550908000_2023.10.03.23296524v1.11","doi-asserted-by":"crossref","first-page":"286","DOI":"10.1007\/s13239-021-00519-w","article-title":"Proper Orthogonal Decomposition Analysis of the Flow Downstream of a Dysfunctional Bileaflet Mechanical Aortic Valve","volume":"12","year":"2021","journal-title":"Cardiovasc Eng Technol"},{"key":"2023101012550908000_2023.10.03.23296524v1.12","doi-asserted-by":"crossref","unstructured":"Di Labbio, G. , Kadem, L. , 2019. Reduced-order modeling of left ventricular flow subject to aortic valve regurgitation. Physics of Fluids 31.","DOI":"10.1063\/1.5083054"},{"key":"2023101012550908000_2023.10.03.23296524v1.13","first-page":"1","article-title":"4D flow cardiovascular magnetic resonance consensus statement","volume":"17","year":"2015","journal-title":"Journal of Cardiovascular Magnetic Resonance"},{"key":"2023101012550908000_2023.10.03.23296524v1.14","doi-asserted-by":"crossref","first-page":"78","DOI":"10.1186\/s12968-014-0078-9","article-title":"Vortex flow during early and late left ventricular filling in normal subjects: quantitative characterization using retrospectively-gated 4D flow cardiovascular magnetic resonance and threedimensional vortex core analysis","volume":"16","year":"2014","journal-title":"J Cardiovasc Magn Reson"},{"key":"2023101012550908000_2023.10.03.23296524v1.15","doi-asserted-by":"crossref","first-page":"794","DOI":"10.1002\/mrm.26129","article-title":"Assessment of viscous energy loss and the association with three-dimensional vortex ring formation in left ventricular inflow: In vivo evaluation using four-dimensional flow MRI","volume":"77","year":"2017","journal-title":"Magn Reson Med"},{"key":"2023101012550908000_2023.10.03.23296524v1.16","doi-asserted-by":"publisher","DOI":"10.1093\/ehjci\/jes159"},{"key":"2023101012550908000_2023.10.03.23296524v1.17","doi-asserted-by":"crossref","first-page":"e190032","DOI":"10.1148\/ryct.2020190032","article-title":"Fully Automated Myocardial Strain Estimation from Cardiovascular MRI-tagged Images Using a Deep Learning Framework in the UK Biobank","volume":"2","year":"2020","journal-title":"Radiol Cardiothorac Imaging"},{"key":"2023101012550908000_2023.10.03.23296524v1.18","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2010.2049656"},{"key":"2023101012550908000_2023.10.03.23296524v1.19","doi-asserted-by":"crossref","first-page":"105384","DOI":"10.1016\/j.compbiomed.2022.105384","article-title":"Higher order dynamic mode decomposition: From fluid dynamics to heart disease analysis","volume":"144","year":"2022","journal-title":"Comput Biol Med"},{"key":"2023101012550908000_2023.10.03.23296524v1.20","doi-asserted-by":"publisher","DOI":"10.1023\/A:1010920819831"},{"key":"2023101012550908000_2023.10.03.23296524v1.21","first-page":"662","article-title":"The E-wave propagation index (EPI): A novel echocardiographic parameter for prediction of left ventricular thrombus","volume":"227","year":"2017","journal-title":"Derivation from computational fluid dynamic modeling and validation on human subjects. Int J Cardiol"},{"key":"2023101012550908000_2023.10.03.23296524v1.22","doi-asserted-by":"crossref","first-page":"489","DOI":"10.1007\/s13239-023-00666-2","article-title":"Intra-Cardiac Flow from Geometry Prescribed Computational Fluid Dynamics: Comparison with Ultrasound Vector Flow Imaging","volume":"14","year":"2023","journal-title":"Cardiovasc Eng Technol"},{"key":"2023101012550908000_2023.10.03.23296524v1.23","doi-asserted-by":"publisher","DOI":"10.1038\/35008075"},{"key":"2023101012550908000_2023.10.03.23296524v1.24","doi-asserted-by":"publisher","DOI":"10.1016\/j.echo.2014.10.003"},{"key":"2023101012550908000_2023.10.03.23296524v1.25","doi-asserted-by":"publisher","DOI":"10.1161\/01.CIR.92.7.1933"},{"key":"2023101012550908000_2023.10.03.23296524v1.26","doi-asserted-by":"publisher","DOI":"10.1016\/j.jacc.2014.06.1205"},{"key":"2023101012550908000_2023.10.03.23296524v1.27","doi-asserted-by":"publisher","DOI":"10.1016\/j.jcmg.2017.06.012"},{"key":"2023101012550908000_2023.10.03.23296524v1.28","doi-asserted-by":"crossref","first-page":"319","DOI":"10.1016\/j.echo.2018.10.018","article-title":"Intracardiac Flow Analysis: Techniques and Potential Clinical Applications","volume":"32","year":"2019","journal-title":"J Am Soc Echocardiogr"},{"key":"2023101012550908000_2023.10.03.23296524v1.29","doi-asserted-by":"crossref","first-page":"92","DOI":"10.1007\/s00380-020-01664-3","article-title":"Noninvasive assessment of intraventricular pressure difference in left ventricular dyssynchrony using vector flow mapping","volume":"36","year":"2021","journal-title":"Heart Vessels"},{"key":"2023101012550908000_2023.10.03.23296524v1.30","doi-asserted-by":"crossref","first-page":"197","DOI":"10.1007\/s11063-007-9069-2","article-title":"New Routes from Minimal Approximation Error to Principal Components","volume":"27","year":"2008","journal-title":"Neural Process Lett"},{"key":"2023101012550908000_2023.10.03.23296524v1.31","doi-asserted-by":"crossref","first-page":"111708","DOI":"10.1063\/5.0069560","article-title":"Flow examination in abdominal aortic aneurysms: Reduced-order models driven by in vitro data and spectral proper orthogonal decomposition","volume":"33","year":"2021","journal-title":"Physics of Fluids"},{"key":"2023101012550908000_2023.10.03.23296524v1.32","doi-asserted-by":"crossref","unstructured":"Ohh, C.Y. , Spedding, G.R. , 2022. Wake identification of stratified flows using dynamic mode decomposition. Physical Review Fluids 7.","DOI":"10.1103\/PhysRevFluids.7.024801"},{"key":"2023101012550908000_2023.10.03.23296524v1.33","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevLett.95.108101"},{"key":"2023101012550908000_2023.10.03.23296524v1.34","doi-asserted-by":"publisher","DOI":"10.1038\/nrcardio.2014.75"},{"key":"2023101012550908000_2023.10.03.23296524v1.35","doi-asserted-by":"crossref","unstructured":"Rodr\u00edguez-Gonz\u00e1lez, E. , Martinez-Legazpi, P. , Mombiela, T. , Gonz\u00e1lez-Mansilla, A. , Delgado-Montero, A. , Guzm\u00e1n De Villoria, J. , D\u00edaz-Otero, F. , Prieto, R. , Juarez, M. , del Rey MdC , G., 2023. Stasis Imaging Predicts the Risk of Cardioembolic Stroke Related to Acute Myocardial Infarction.","DOI":"10.1101\/2023.09.15.23295650"},{"key":"2023101012550908000_2023.10.03.23296524v1.36","doi-asserted-by":"crossref","first-page":"929","DOI":"10.1038\/s42256-021-00399-8","article-title":"Designing clinically translatable artificial intelligence systems for high-dimensional medical imaging","volume":"3","year":"2021","journal-title":"Nature Machine Intelligence"},{"key":"2023101012550908000_2023.10.03.23296524v1.37","doi-asserted-by":"crossref","first-page":"e008130","DOI":"10.1161\/CIRCIMAGING.118.008130","article-title":"Left Ventricular Flow Analysis","volume":"12","year":"2019","journal-title":"Circ Cardiovasc Imaging"},{"key":"2023101012550908000_2023.10.03.23296524v1.38","doi-asserted-by":"crossref","unstructured":"Sundin, J. , Engvall, J. , Nylander, E. , Ebbers, T. , Bolger, A.F. , Carlh\u00e4ll, C.-J. , 2020. Improved Efficiency of Intraventricular Blood Flow Transit Under Cardiac Stress: A 4D Flow Dobutamine CMR Study. Frontiers in Cardiovascular Medicine 7.","DOI":"10.3389\/fcvm.2020.581495"},{"key":"2023101012550908000_2023.10.03.23296524v1.39","doi-asserted-by":"crossref","first-page":"4013","DOI":"10.2514\/1.J056060","article-title":"Modal analysis of fluid flows","volume":"55","year":"2017","journal-title":"AIAA Journal"},{"key":"2023101012550908000_2023.10.03.23296524v1.40","doi-asserted-by":"crossref","first-page":"6776","DOI":"10.1038\/s41467-022-34245-1","article-title":"A formal validation of a deep learning-based automated workflow for the interpretation of the echocardiogram","volume":"13","year":"2022","journal-title":"Nature Communications"},{"key":"2023101012550908000_2023.10.03.23296524v1.41","doi-asserted-by":"crossref","first-page":"e023417","DOI":"10.1161\/JAHA.121.023417","article-title":"Introduction to Hemodynamic Forces Analysis: Moving Into the New Frontier of Cardiac Deformation Analysis","volume":"10","year":"2021","journal-title":"J Am Heart Assoc"},{"key":"2023101012550908000_2023.10.03.23296524v1.42","doi-asserted-by":"crossref","first-page":"640","DOI":"10.1148\/radiol.2018180339","article-title":"Left Ventricular Strain Is Abnormal in Preclinical and Overt Hypertrophic Cardiomyopathy: Cardiac MR Feature Tracking","volume":"290","year":"2019","journal-title":"Radiology"},{"key":"2023101012550908000_2023.10.03.23296524v1.43","doi-asserted-by":"publisher","DOI":"10.1152\/ajpheart.00041.2008"},{"key":"2023101012550908000_2023.10.03.23296524v1.44","doi-asserted-by":"publisher","DOI":"10.1016\/S0140-6736(09)60040-4"},{"key":"2023101012550908000_2023.10.03.23296524v1.45","doi-asserted-by":"crossref","unstructured":"Wehbe, R.M. , Katsaggleos, A.K. , Hammond, K.J. , Hong, H. , Ahmad, F.S. , Ouyang, D. , Shah, S.J. , McCarthy, P.M. , Thomas, J.D. , 2023. Deep Learning for Cardiovascular Imaging: A Review. JAMA Cardiology.","DOI":"10.1001\/jamacardio.2023.3142"},{"key":"2023101012550908000_2023.10.03.23296524v1.46","doi-asserted-by":"publisher","DOI":"10.1161\/CIRCIMAGING.113.000722"},{"key":"2023101012550908000_2023.10.03.23296524v1.47","doi-asserted-by":"crossref","unstructured":"Yu, P. , Durgesh, V. , 2022. Application of Dynamic Mode Decomposition to Study Temporal Flow Behavior in a Saccular Aneurysm. J Biomech Eng 144.","DOI":"10.1115\/1.4052999"},{"key":"2023101012550908000_2023.10.03.23296524v1.48","doi-asserted-by":"publisher","DOI":"10.1161\/CIRCULATIONAHA.118.034338"},{"key":"2023101012550908000_2023.10.03.23296524v1.49","doi-asserted-by":"publisher","DOI":"10.1109\/jproc.2021.3054390"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2023.10.03.23296524","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T19:42:27Z","timestamp":1768506147000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2023.10.03.23296524"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,10,5]]},"references-count":49,"URL":"https:\/\/doi.org\/10.1101\/2023.10.03.23296524","relation":{},"subject":[],"published":{"date-parts":[[2023,10,5]]},"subtype":"preprint"}}