Skip to main content

Advertisement

SpringerLink
Log in

Machine Learning Quantification of Pulmonary Regurgitation Fraction from Echocardiography

  • Research
  • Published:
Pediatric Cardiology Aims and scope Submit manuscript

Abstract

Assessment of pulmonary regurgitation (PR) guides treatment for patients with congenital heart disease. Quantitative assessment of PR fraction (PRF) by echocardiography is limited. Cardiac MRI (cMRI) is the reference-standard for PRF quantification. We created an algorithm to predict cMRI-quantified PRF from echocardiography using machine learning (ML). We retrospectively performed echocardiographic measurements paired to cMRI within 3 months in patients with ≥ mild PR from 2009 to 2022. Model inputs were vena contracta ratio, PR index, PR pressure half-time, main and branch pulmonary artery diastolic flow reversal (BPAFR), and transannular patch repair. A gradient boosted trees ML algorithm was trained using k-fold cross-validation to predict cMRI PRF by phase contrast imaging as a continuous number and at > mild (PRF ≥ 20%) and severe (PRF ≥ 40%) thresholds. Regression performance was evaluated with mean absolute error (MAE), and at clinical thresholds with area-under-the-receiver-operating-characteristic curve (AUROC). Prediction accuracy was compared to historical clinician accuracy. We externally validated prior reported studies for comparison. We included 243 subjects (median age 21 years, 58% repaired tetralogy of Fallot). The regression MAE = 7.0%. For prediction of > mild PR, AUROC = 0.96, but BPAFR alone outperformed the ML model (sensitivity 94%, specificity 97%). The ML model detection of severe PR had AUROC = 0.86, but in the subgroup with BPAFR, performance dropped (AUROC = 0.73). Accuracy between clinicians and the ML model was similar (70% vs. 69%). There was decrement in performance of prior reported algorithms on external validation in our dataset. A novel ML model for echocardiographic quantification of PRF outperforms prior studies and has comparable overall accuracy to clinicians. BPAFR is an excellent marker for > mild PRF, and has moderate capacity to detect severe PR, but more work is required to distinguish moderate from severe PR. Poor external validation of prior works highlights reproducibility challenges.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5

Similar content being viewed by others

Abbreviations

AUROC:

Area under the receiver operating characteristic curve

BPA:

Branch pulmonary artery

BPAFR:

Branch pulmonary artery flow reversal

cMRI:

Cardiac magnetic resonance imaging

ICC:

Intraclass correlation coefficient

MAE:

Mean absolute error

ML:

Machine learning

MPA:

Main pulmonary artery

OLS:

Ordinary least squares

PHT:

Pressure halftime

PR:

Pulmonary regurgitation

PRF:

Pulmonary regurgitation fraction

TAP:

Transannular patch

VCR:

Vena contracta ratio

References

  1. Stout KK, Daniels CJ, Aboulhosn JA, Bozkurt B, Broberg CS, Colman JM, Crumb SR, Dearani JA, Fuller S, Gurvitz M (2018) AHA/ACC guideline for the management of adults with congenital heart disease: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. J Am Coll Cardiol. https://doi.org/10.1161/CIR.0000000000000602

    Article  PubMed  Google Scholar 

  2. Geva T (2011) Repaired tetralogy of Fallot: the roles of cardiovascular magnetic resonance in evaluating pathophysiology and for pulmonary valve replacement decision support. J Cardiovasc Magn Reson 13:1–24

    Article  Google Scholar 

  3. Rebergen SA, van der Wall EE, Doornbos J, de Roos A (1993) Magnetic resonance measurement of velocity and flow: technique, validation, and cardiovascular applications. Am Heart J 126:1439–1456

    Article  CAS  PubMed  Google Scholar 

  4. Redington AN (2006) Determinants and assessment of pulmonary regurgitation in tetralogy of Fallot: practice and pitfalls. Cardiol Clin 24:631–639

    Article  PubMed  Google Scholar 

  5. Pennell DJ, Sechtem UP, Higgins CB, Manning WJ, Pohost GM, Rademakers FE, van Rossum AC, Shaw LJ, Yucel EK (2004) Clinical indications for cardiovascular magnetic resonance (CMR): Consensus Panel report. Eur Heart J 25:1940–1965

    Article  PubMed  Google Scholar 

  6. Bonello B, Kilner PJ (2012) Review of the role of cardiovascular magnetic resonance in congenital heart disease, with a focus on right ventricle assessment. Arch Cardiovasc Dis 105:605–613

    Article  PubMed  Google Scholar 

  7. Marwick TH, Neubauer S, Petersen SE (2013) Use of cardiac magnetic resonance and echocardiography in population-based studies: why, where, and when? Circ: Cardiovasc Imaging 6:590–596

    PubMed  Google Scholar 

  8. Bansal N, Gupta P, Joshi A, Zerin JM, Aggarwal S (2019) Utility of Doppler echocardiography to estimate the severity of pulmonary valve regurgitation fraction in patients with repaired tetralogy of Fallot. Pediatr Cardiol 40:404–411

    Article  PubMed  Google Scholar 

  9. Beurskens NE, Gorter TM, Pieper PG, Hoendermis ES, Bartelds B, Ebels T, Berger RM, Willems TP, van Melle JP (2017) Diagnostic value of Doppler echocardiography for identifying hemodynamic significant pulmonary valve regurgitation in tetralogy of Fallot: comparison with cardiac MRI. Int J Cardiovasc Imaging 33:1723–1730

    Article  PubMed  PubMed Central  Google Scholar 

  10. Li W, Davlouros PA, Kilner PJ, Pennell DJ, Gibson D, Henein MY, Gatzoulis MA (2004) Doppler-echocardiographic assessment of pulmonary regurgitation in adults with repaired tetralogy of Fallot: comparison with cardiovascular magnetic resonance imaging. Am Heart J 147:165–172

    Article  PubMed  Google Scholar 

  11. Mercer-Rosa L, Yang W, Kutty S, Rychik J, Fogel M, Goldmuntz E (2012) Quantifying pulmonary regurgitation and right ventricular function in surgically repaired tetralogy of Fallot: a comparative analysis of echocardiography and magnetic resonance imaging. Circ: Cardiovasc Imaging 5:637–643

    PubMed  Google Scholar 

  12. Puchalski MD, Askovich B, Sower CT, Williams RV, Minich LL, Tani LY (2008) Pulmonary regurgitation: determining severity by echocardiography and magnetic resonance imaging. Congenit Heart Dis 3:168–175

    Article  PubMed  Google Scholar 

  13. Silversides CK, Veldtman GR, Crossin J, Merchant N, Webb GD, McCrindle BW, Siu SC, Therrien J (2003) Pressure half-time predicts hemodynamically significant pulmonary regurgitation in adult patients with repaired tetralogy of Fallot. J Am Soc Echocardiogr 16:1057–1062

    Article  PubMed  Google Scholar 

  14. Renella P, Aboulhosn J, Lohan DG, Jonnala P, Finn JP, Satou GM, Williams RJ, Child JS (2010) Two-dimensional and Doppler echocardiography reliably predict severe pulmonary regurgitation as quantified by cardiac magnetic resonance. J Am Soc Echocardiogr 23:880–886

    Article  PubMed  Google Scholar 

  15. Deo RC (2015) Machine learning in medicine. Circulation 132:1920–1930

    Article  PubMed  PubMed Central  Google Scholar 

  16. Jone P-N, Gearhart A, Lei H, Xing F, Nahar J, Lopez-Jimenez F, Diller G-P, Marelli A, Wilson L, Saidi A (2022) Artificial intelligence in congenital heart disease: current state and prospects. JACC: Adv 1:100153

    Google Scholar 

  17. Gearhart A, Gaffar S, Chang AC (2020) A primer on artificial intelligence for the paediatric cardiologist. Cardiol Young 30:934–945

    Article  PubMed  Google Scholar 

  18. Chen T, Guestrin C (2016) XGBoost: a scalable tree boosting system. Paper/Poster presented 2016/08/13/, https://arxiv.org/abs/1603.02754. Accessed 04 June 2023

  19. Scikit-Learn Developers. Iterative imputer. https://scikit-learn.org/stable/modules/generated/sklearn.impute.IterativeImputer.html. Accessed 04 June 2023

  20. Dai W, Nazzari H, Namasivayam M, Hung J, Stultz CM (2023) Identifying aortic stenosis with a single parasternal long-axis video using deep learning. J Am Soc Echocardiogr 36:116–118. https://doi.org/10.1016/j.echo.2022.10.014

    Article  PubMed  Google Scholar 

Download references

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Author information

Author notes

  1. Jennifer Cohen and Son Q. Duong contributed equally to this work.

Authors and Affiliations

  1. Department of Pediatrics, Icahn School of Medicine at Mount Sinai, New York, NY, USA

    Jennifer Cohen, Son Q. Duong, David M. Barris, Surkhay Bebiya, Marjorie Gayanilo, Kali Hopkins, Maya Kailas, Grace Kong, Molly Marshall, Erin A. Paul & David Ezon

  2. Mount Sinai Kravis Children’s Heart Center, The Mount Sinai Hospital, 1468 Madison Ave, Annenberg 3rd Floor, New York, NY, 10029, USA

    Jennifer Cohen, Son Q. Duong, David M. Barris, Surkhay Bebiya, Rosalie Castaldo, Marjorie Gayanilo, Maya Kailas, Grace Kong, Xiye Ma, Molly Marshall, Erin A. Paul, Melanie Tan, Jen Lie Yau & David Ezon

  3. Charles Bronfman Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA

    Son Q. Duong, Naveen Arivazhagan & Girish N. Nadkarni

  4. Mindich Child Health and Development Institute, Icahn School of Medicine at Mount Sinai, New York, NY, USA

    Son Q. Duong & Kali Hopkins

  5. Adult Congenital Heart Disease, Mount Sinai Heart, The Mount Sinai Hospital, New York, NY, USA

    Kali Hopkins

Authors
  1. Jennifer Cohen
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Son Q. Duong
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Naveen Arivazhagan
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. David M. Barris
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Surkhay Bebiya
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Rosalie Castaldo
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Marjorie Gayanilo
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Kali Hopkins
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Maya Kailas
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Grace Kong
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Xiye Ma
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Molly Marshall
    View author publications

    You can also search for this author in PubMed Google Scholar

  13. Erin A. Paul
    View author publications

    You can also search for this author in PubMed Google Scholar

  14. Melanie Tan
    View author publications

    You can also search for this author in PubMed Google Scholar

  15. Jen Lie Yau
    View author publications

    You can also search for this author in PubMed Google Scholar

  16. Girish N. Nadkarni
    View author publications

    You can also search for this author in PubMed Google Scholar

  17. David Ezon
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Conception and design of the study:JC, SQD, NA, GNN, DE; Acquisition of data: JC, SQD, DMB, SB, RC, MG, KH, MK, GK, XM, MM, EAP, MT, JLY, DE; Analysis and interpretation of data: JC, SQD, NA, GNN, DE; Drafting the article or revising it critically for important intellectual content: JC, SQD, NA, DMB, SB, RC, MG, KH, MK, GK, XM, MM, EAP, MT, JLY, GNN, DE; final approval of the version to be submitted: All authors have approved the final article.

Corresponding author

Correspondence to Son Q. Duong.

Ethics declarations

Conflict of interest

David Ezon is CEO of Tapestry Insight, Inc. Girish Nadkarni reports consultancy agreements with AstraZeneca, BioVie, GLG Consulting, Pensieve Health, Reata, Renalytix, Siemens Healthineers and Variant Bio; research funding from Goldfinch Bio and Renalytix; honoraria from AstraZeneca, BioVie, Lexicon, Daiichi Sankyo, Menarini Health and Reata; patents or royalties with Renalytix; owns equity and stock options in Pensieve Health and Renalytix as a scientific cofounder; owns equity in Verici Dx; has received financial compensation as a scientific board member and advisor to Renalytix; serves on the advisory board of Neurona Health; and serves in an advisory or leadership role for Pensieve Health and Renalytix. No other authors report competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (DOCX 13 KB)

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Cohen, J., Duong, S.Q., Arivazhagan, N. et al. Machine Learning Quantification of Pulmonary Regurgitation Fraction from Echocardiography. Pediatr Cardiol (2024). https://doi.org/10.1007/s00246-024-03511-y

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s00246-024-03511-y

Keywords

Search

Navigation