CLINICAL RESEARCH: CARDIAC IMAGING
A Novel Two-Dimensional Echocardiographic Image Analysis System Using Artificial Intelligence-Learned Pattern Recognition for Rapid Automated Ejection Fraction
Maxime Cannesson, MD*,2,
Masaki Tanabe, MD*,
Matthew S. Suffoletto, MD*,
Dennis M. McNamara, MD, FACC*,
Shobhit Madan, MD ,
Joan M. Lacomis, MD and
John Gorcsan, III, MD, FACC*,1,*
* Cardiovascular Institute, University of Pittsburgh, Pittsburgh, Pennsylvania
Department of Radiology, University of Pittsburgh, Pittsburgh, Pennsylvania.
Manuscript received April 10, 2006;
revised manuscript received August 17, 2006,
accepted August 21, 2006.
* Reprint requests and correspondence: Dr. John Gorcsan III, University of Pittsburgh, Scaife 564, 200 Lothrop Street, Pittsburgh, Pennsylvania 15213-2582. (Email: gorcsanj{at}upmc.edu).
OBJECTIVES: We sought to test the hypothesis that a novel 2-dimensional echocardiographic image analysis system using artificial intelligence-learned pattern recognition can rapidly and reproducibly calculate ejection fraction (EF).
BACKGROUND: Echocardiographic EF by manual tracing is time consuming, and visual assessment is inherently subjective.
METHODS: We studied 218 patients (72 female), including 165 with abnormal left ventricular (LV) function. Auto EF incorporated a database trained on >10,000 human EF tracings to automatically locate and track the LV endocardium from routine grayscale digital cineloops and calculate EF in 15 s. Auto EF results were independently compared with manually traced biplane Simpson’s rule, visual EF, and magnetic resonance imaging (MRI) in a subset.
RESULTS: Auto EF was possible in 200 (92%) of consecutive patients, of which 77% were completely automated and 23% required manual editing. Auto EF correlated well with manual EF (r = 0.98; 6% limits of agreement) and required less time per patient (48 ± 26 s vs. 102 ± 21 s; p < 0.01). Auto EF correlated well with visual EF by expert readers (r = 0.96; p < 0.001), but interobserver variability was greater (3.4 ± 2.9% vs. 9.8 ± 5.7%, respectively; p < 0.001). Visual EF was less accurate by novice readers (r = 0.82; 19% limits of agreement) and improved with trainee-operated Auto EF (r = 0.96; 7% limits of agreement). Auto EF also correlated with MRI EF (n = 21) (r = 0.95; 12% limits of agreement), but underestimated absolute volumes (r = 0.95; bias of –36 ± 27 ml overall).
CONCLUSIONS: Auto EF can automatically calculate EF similarly to results by manual biplane Simpson’s rule and MRI, with less variability than visual EF, and has clinical potential.
|
Abbreviations and Acronyms
| | 2D = 2-dimensional | | EF = ejection fraction | | LV = left ventricle/ventricular | | MRI = magnetic resonance imaging |
|
This article has been cited by other articles:
|
|
|
|
 
M. Hravnak, J. M. Frangiskakis, E. A. Crago, Y. Chang, M. Tanabe, J. Gorcsan III, and M. B. Horowitz
Elevated Cardiac Troponin I and Relationship to Persistence of Electrocardiographic and Echocardiographic Abnormalities After Aneurysmal Subarachnoid Hemorrhage
Stroke,
November 1, 2009;
40(11):
3478 - 3484.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
B. R.J. H. Bird and S. M. Swain
Cardiac Toxicity in Breast Cancer Survivors: Review of Potential Cardiac Problems
Clin. Cancer Res.,
January 1, 2008;
14(1):
14 - 24.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. St John Sutton and T. Plappert
Core lab, no core lab or automated LVEF?
Eur J Echocardiogr,
August 1, 2007;
8(4):
239 - 240.
[Full Text]
[PDF]
|
|
|
|
