Machine learning based automated dynamic quantification of left heart chamber volumes

Akhil Narang, Victor Mor-Avi, Aldo Prado, Valentina Volpato, David Prater, Gloria Tamborini, Laura Fusini, Mauro Pepi, Neha Goyal, Karima Addetia, Alexandra Gonçalves, Amit R Patel, Roberto M Lang

Research output: Contribution to journalArticle

Abstract

Aims: Studies have demonstrated the ability of a new automated algorithm for volumetric analysis of 3D echocardiographic (3DE) datasets to provide accurate and reproducible measurements of left ventricular and left atrial (LV, LA) volumes at end-systole and end-diastole. Recently, this methodology was expanded using a machine learning (ML) approach to automatically measure chamber volumes throughout the cardiac cycle, resulting in LV and LA volume-time curves. We aimed to validate ejection and filling parameters obtained from these curves by comparing them to independent well-validated reference techniques.

Methods and results: We studied 20 patients referred for cardiac magnetic resonance (CMR) examinations, who underwent 3DE imaging the same day. Volume-time curves were obtained for both LV and LA chambers using the ML algorithm (Philips HeartModel), and independently conventional 3DE volumetric analysis (TomTec), and CMR images (slice-by-slice, frame-by-frame manual tracing). Automatically derived LV and LA volumes and ejection/filling parameters were compared against both reference techniques. Minor manual correction of the automatically detected LV and LA borders was needed in 4/20 and 5/20 cases, respectively. Time required to generate volume-time curves was 35 ± 17 s using ML algorithm, 3.6 ± 0.9 min using conventional 3DE analysis, and 96 ± 14 min using CMR. Volume-time curves obtained by all three techniques were similar in shape and magnitude. In both comparisons, ejection/filling parameters showed no significant inter-technique differences. Bland-Altman analysis confirmed small biases, despite wide limits of agreement.

Conclusion: The automated ML algorithm can quickly measure dynamic LV and LA volumes and accurately analyse ejection/filling parameters. Incorporation of this algorithm into the clinical workflow may increase the utilization of 3DE imaging.

Original languageEnglish
JournalEuropean Heart Journal Cardiovascular Imaging
DOIs
Publication statusE-pub ahead of print - Oct 9 2018

Fingerprint

Cardiac Volume
Magnetic Resonance Spectroscopy
Diastole
Workflow
Systole
Machine Learning

Cite this

Machine learning based automated dynamic quantification of left heart chamber volumes. / Narang, Akhil; Mor-Avi, Victor; Prado, Aldo; Volpato, Valentina; Prater, David; Tamborini, Gloria; Fusini, Laura; Pepi, Mauro; Goyal, Neha; Addetia, Karima; Gonçalves, Alexandra; Patel, Amit R; Lang, Roberto M.

In: European Heart Journal Cardiovascular Imaging, 09.10.2018.

Research output: Contribution to journalArticle

Narang, Akhil ; Mor-Avi, Victor ; Prado, Aldo ; Volpato, Valentina ; Prater, David ; Tamborini, Gloria ; Fusini, Laura ; Pepi, Mauro ; Goyal, Neha ; Addetia, Karima ; Gonçalves, Alexandra ; Patel, Amit R ; Lang, Roberto M. / Machine learning based automated dynamic quantification of left heart chamber volumes. In: European Heart Journal Cardiovascular Imaging. 2018.
@article{019b6a07a1944788b43a11eaa351ca59,
title = "Machine learning based automated dynamic quantification of left heart chamber volumes",
abstract = "Aims: Studies have demonstrated the ability of a new automated algorithm for volumetric analysis of 3D echocardiographic (3DE) datasets to provide accurate and reproducible measurements of left ventricular and left atrial (LV, LA) volumes at end-systole and end-diastole. Recently, this methodology was expanded using a machine learning (ML) approach to automatically measure chamber volumes throughout the cardiac cycle, resulting in LV and LA volume-time curves. We aimed to validate ejection and filling parameters obtained from these curves by comparing them to independent well-validated reference techniques.Methods and results: We studied 20 patients referred for cardiac magnetic resonance (CMR) examinations, who underwent 3DE imaging the same day. Volume-time curves were obtained for both LV and LA chambers using the ML algorithm (Philips HeartModel), and independently conventional 3DE volumetric analysis (TomTec), and CMR images (slice-by-slice, frame-by-frame manual tracing). Automatically derived LV and LA volumes and ejection/filling parameters were compared against both reference techniques. Minor manual correction of the automatically detected LV and LA borders was needed in 4/20 and 5/20 cases, respectively. Time required to generate volume-time curves was 35 ± 17 s using ML algorithm, 3.6 ± 0.9 min using conventional 3DE analysis, and 96 ± 14 min using CMR. Volume-time curves obtained by all three techniques were similar in shape and magnitude. In both comparisons, ejection/filling parameters showed no significant inter-technique differences. Bland-Altman analysis confirmed small biases, despite wide limits of agreement.Conclusion: The automated ML algorithm can quickly measure dynamic LV and LA volumes and accurately analyse ejection/filling parameters. Incorporation of this algorithm into the clinical workflow may increase the utilization of 3DE imaging.",
author = "Akhil Narang and Victor Mor-Avi and Aldo Prado and Valentina Volpato and David Prater and Gloria Tamborini and Laura Fusini and Mauro Pepi and Neha Goyal and Karima Addetia and Alexandra Gon{\cc}alves and Patel, {Amit R} and Lang, {Roberto M}",
year = "2018",
month = "10",
day = "9",
doi = "10.1093/ehjci/jey137",
language = "English",
journal = "European Heart Journal Cardiovascular Imaging",
issn = "2047-2404",
publisher = "Oxford University Press",

}

TY - JOUR

T1 - Machine learning based automated dynamic quantification of left heart chamber volumes

AU - Narang, Akhil

AU - Mor-Avi, Victor

AU - Prado, Aldo

AU - Volpato, Valentina

AU - Prater, David

AU - Tamborini, Gloria

AU - Fusini, Laura

AU - Pepi, Mauro

AU - Goyal, Neha

AU - Addetia, Karima

AU - Gonçalves, Alexandra

AU - Patel, Amit R

AU - Lang, Roberto M

PY - 2018/10/9

Y1 - 2018/10/9

N2 - Aims: Studies have demonstrated the ability of a new automated algorithm for volumetric analysis of 3D echocardiographic (3DE) datasets to provide accurate and reproducible measurements of left ventricular and left atrial (LV, LA) volumes at end-systole and end-diastole. Recently, this methodology was expanded using a machine learning (ML) approach to automatically measure chamber volumes throughout the cardiac cycle, resulting in LV and LA volume-time curves. We aimed to validate ejection and filling parameters obtained from these curves by comparing them to independent well-validated reference techniques.Methods and results: We studied 20 patients referred for cardiac magnetic resonance (CMR) examinations, who underwent 3DE imaging the same day. Volume-time curves were obtained for both LV and LA chambers using the ML algorithm (Philips HeartModel), and independently conventional 3DE volumetric analysis (TomTec), and CMR images (slice-by-slice, frame-by-frame manual tracing). Automatically derived LV and LA volumes and ejection/filling parameters were compared against both reference techniques. Minor manual correction of the automatically detected LV and LA borders was needed in 4/20 and 5/20 cases, respectively. Time required to generate volume-time curves was 35 ± 17 s using ML algorithm, 3.6 ± 0.9 min using conventional 3DE analysis, and 96 ± 14 min using CMR. Volume-time curves obtained by all three techniques were similar in shape and magnitude. In both comparisons, ejection/filling parameters showed no significant inter-technique differences. Bland-Altman analysis confirmed small biases, despite wide limits of agreement.Conclusion: The automated ML algorithm can quickly measure dynamic LV and LA volumes and accurately analyse ejection/filling parameters. Incorporation of this algorithm into the clinical workflow may increase the utilization of 3DE imaging.

AB - Aims: Studies have demonstrated the ability of a new automated algorithm for volumetric analysis of 3D echocardiographic (3DE) datasets to provide accurate and reproducible measurements of left ventricular and left atrial (LV, LA) volumes at end-systole and end-diastole. Recently, this methodology was expanded using a machine learning (ML) approach to automatically measure chamber volumes throughout the cardiac cycle, resulting in LV and LA volume-time curves. We aimed to validate ejection and filling parameters obtained from these curves by comparing them to independent well-validated reference techniques.Methods and results: We studied 20 patients referred for cardiac magnetic resonance (CMR) examinations, who underwent 3DE imaging the same day. Volume-time curves were obtained for both LV and LA chambers using the ML algorithm (Philips HeartModel), and independently conventional 3DE volumetric analysis (TomTec), and CMR images (slice-by-slice, frame-by-frame manual tracing). Automatically derived LV and LA volumes and ejection/filling parameters were compared against both reference techniques. Minor manual correction of the automatically detected LV and LA borders was needed in 4/20 and 5/20 cases, respectively. Time required to generate volume-time curves was 35 ± 17 s using ML algorithm, 3.6 ± 0.9 min using conventional 3DE analysis, and 96 ± 14 min using CMR. Volume-time curves obtained by all three techniques were similar in shape and magnitude. In both comparisons, ejection/filling parameters showed no significant inter-technique differences. Bland-Altman analysis confirmed small biases, despite wide limits of agreement.Conclusion: The automated ML algorithm can quickly measure dynamic LV and LA volumes and accurately analyse ejection/filling parameters. Incorporation of this algorithm into the clinical workflow may increase the utilization of 3DE imaging.

U2 - 10.1093/ehjci/jey137

DO - 10.1093/ehjci/jey137

M3 - Article

C2 - 30304500

JO - European Heart Journal Cardiovascular Imaging

JF - European Heart Journal Cardiovascular Imaging

SN - 2047-2404

ER -