Determining a minimum set of variables for machine learning cardiovascular event prediction: results from REFINE SPECT registry
Rios R, Miller RJH, Hu LH, Otaki Y, Singh A, Diniz M, Sharir T, Einstein AJ, Fish MB, Ruddy TD, Kaufmann PA, Sinusas AJ, Miller EJ, Bateman TM, Dorbala S, DiCarli M, Van Kriekinge S, Kavanagh P, Parekh T, Liang JX, Dey D, Berman DS, Slomka P. Determining a minimum set of variables for machine learning cardiovascular event prediction: results from REFINE SPECT registry. Cardiovascular Research 2021, 118: 2152-2164. PMID: 34259870, PMCID: PMC9302886, DOI: 10.1093/cvr/cvab236.Peer-Reviewed Original ResearchConceptsMajor adverse cardiac eventsMyocardial perfusion imagingTotal perfusion deficitPrognostic accuracyREFINE SPECT registryImaging variablesTraditional multivariable modelsMultivariable modelStress total perfusion deficitAdverse cardiac eventsReceiver-operating characteristic curveOptimal risk stratificationTomography (SPECT) MPIComparable prognostic accuracyComparable prognostic performanceHigher prognostic accuracySingle photon emissionCardiovascular event predictionCardiac eventsRisk stratificationClinical variablesPerfusion deficitsPrognostic performancePerfusion imagingCollected variables