Machine learning evaluation of LV outflow obstruction in hypertrophic cardiomyopathy using three-chamber cardiovascular magnetic resonance

The International Journal of Cardiovascular Imaging - Tập 38 - Trang 2695-2705 - 2022
Manisha Sahota1, Sepas Ryan Saraskani1, Hao Xu1, Liandong Li1, Abdul Wahab Majeed1, Uxio Hermida1, Stefan Neubauer2, Milind Desai3, William Weintraub4, Patrice Desvigne-Nickens5, Jeanette Schulz-Menger6, Raymond Y. Kwong7, Christopher M. Kramer8, Alistair A. Young1, Pablo Lamata1
1Department of Biomedical Engineering, King’s College London, London, UK
2Division of Cardiovascular Medicine, Radcliffe Department of Medicine, University of Oxford, Oxford, UK
3Cardiovascular Institute, Cleveland Clinic, Cleveland, USA
4MedStar Heart and Vascular Institute, Washington, USA
5National Heart, Lung, and Blood Institute, Bethesda, USA
6ECRC and Department of Cardiology, HELIOS Klinik Berlin-Buch, Clinic for Cardiology and Nephrology, DZHK Partnersite Berlin, Charité Medical University Berlin, Berlin, Germany
7Cardiovascular Division, Department of Medicine and Department of Radiology, Brigham and Women’s Hospital, Boston, USA
8Cardiovascular Division, University of Virginia Health, Charlottesville, USA

Tóm tắt

Left ventricular outflow tract obstruction (LVOTO) is common in hypertrophic cardiomyopathy (HCM), but relationships between anatomical metrics and obstruction are poorly understood. We aimed to develop machine learning methods to evaluate LVOTO in HCM patients and quantify relationships between anatomical metrics and obstruction. This retrospective analysis of 1905 participants of the HCM Registry quantified 11 anatomical metrics derived from 14 landmarks automatically detected on the three-chamber long axis cine CMR images. Linear and logistic regression was used to quantify strengths of relationships with the presence of LVOTO (defined by resting Doppler pressure drop of > 30 mmHg), using the area under the receiver operating characteristic (AUC). Intraclass correlation coefficients between the network predictions and three independent observers showed similar agreement to that between observers. The distance from anterior mitral valve leaflet tip to basal septum (AML-BS) was most highly correlated with Doppler pressure drop (R2 = 0.19, p < 10–5). Multivariate stepwise regression found the best predictive model included AML-BS, AML length to aortic valve diameter ratio, AML length to LV width ratio, and midventricular septal thickness metrics (AUC 0.84). Excluding AML-BS, metrics grouped according to septal hypertrophy, LV geometry, and AML anatomy each had similar associations with LVOTO (AUC 0.71, 0.71, 0.68 respectively, p = ns), significantly less than their combination (AUC 0.77, p < 0.05 for each). Anatomical metrics derived from a standard three-chamber CMR cine acquisition can be used to highlight risk of LVOTO, and suggest further investigation if necessary. A combination of geometric factors is required to provide the best risk prediction.

Tài liệu tham khảo

Maron BJ, Maron MS (2013) Hypertrophic cardiomyopathy. Lancet 381:242–255 Hypertropic cardiomyopathy | British Heart Foundation. https://www.bhf.org.uk/informationsupport/conditions/cardiomyopathy/hypertrophic-cardiomyopathy. Accessed 30 Nov 2021 Maron BJ (2018) Clinical course and management of hypertrophic cardiomyopathy. N Engl J Med 379:655–668 Authors/Task Force members, ElliottAnastasakis PMA et al (2014) 2014 ESC Guidelines on diagnosis and management of hypertrophic cardiomyopathy. Eur Heart J 35:2733–2779 Autore C, Bernabò P, Barillà CS, Bruzzi P, Spirito P (2005) The prognostic importance of left ventricular outflow obstruction in hypertrophic cardiomyopathy varies in relation to the severity of symptoms. J Am Coll Cardiol 45:1076–1080 Nishimura RA, Holmes DR (2004) Hypertrophic obstructive cardiomyopathy. N Engl J Med 350:1320–1327 Donati F, Myerson S, Bissell MM, Smith NP, Neubauer S, Monaghan MJ, Nordsletten DA, Lamata P (2017) Beyond Bernoulli: improving the accuracy and precision of noninvasive estimation of peak pressure drops. Circ Cardiovasc Imaging. https://doi.org/10.1161/CIRCIMAGING.116.005207 Gill H, Fernandes J, Chehab O, Prendergast B, Redwood S, Chiribiri A, Nordsletten D, Rajani R, Lamata P (2021) Evaluation of aortic stenosis: from Bernoulli and Doppler to Navier-stokes. Trends Cardiovasc Med. https://doi.org/10.1016/j.tcm.2021.12.003 Bois JP, Geske JB, Foley TA, Ommen SR, Pellikka PA (2017) Comparison of maximal wall thickness in hypertrophic cardiomyopathy differs between magnetic resonance imaging and transthoracic echocardiography. Am J Cardiol 119:643–650 Hindieh W, Weissler-Snir A, Hammer H, Adler A, Rakowski H, Chan RH (2017) Discrepant measurements of maximal left ventricular wall thickness between cardiac magnetic resonance imaging and echocardiography in patients with hypertrophic cardiomyopathy. Circ Cardiovasc Imaging. https://doi.org/10.1161/CIRCIMAGING.117.006309 Phelan D, Sperry BW, Thavendiranathan P, Collier P, Popović ZB, Lever HM, Smedira NG, Desai MY (2017) Comparison of ventricular septal measurements in hypertrophic cardiomyopathy patients who underwent surgical myectomy using multimodality imaging and implications for diagnosis and management. Am J Cardiol 119:1656–1662 Rickers C, Wilke NM, Jerosch-Herold M, Casey SA, Panse P, Panse N, Weil J, Zenovich AG, Maron BJ (2005) Utility of cardiac magnetic resonance imaging in the diagnosis of hypertrophic cardiomyopathy. Circulation 112:855–861 Posma JL, Blanksma PK, van der Wall EE, Hamer HP, Mooyaart EL, Lie KI (1996) Assessment of quantitative hypertrophy scores in hypertrophic cardiomyopathy: magnetic resonance imaging versus echocardiography. Am Heart J 132:1020–1027 Schulz-Menger J, Abdel-Aty H, Busjahn A, Wassmuth R, Pilz B, Dietz R, Friedrich M (2006) Left ventricular outflow tract planimetry by cardiovascular magnetic resonance differentiates obstructive from non-obstructive hypertrophic cardiomyopathy. J Cardiovasc Magn Reson 8:741–746 Ibrahim M, Rao C, Ashrafian H, Chaudhry U, Darzi A, Athanasiou T (2012) Modern management of systolic anterior motion of the mitral valve. Eur J Cardiothorac Surg 41:1260–1270 Raut M, Maheshwari A, Swain B (2018) Awareness of “systolic anterior motion” in different conditions. Clin Med Insights Cardiol 12:1179546817751921 Patel P, Dhillon A, Popovic ZB, Smedira NG, Rizzo J, Thamilarasan M, Agler D, Lytle BW, Lever HM, Desai MY (2015) Left ventricular outflow tract obstruction in hypertrophic cardiomyopathy patients without severe septal hypertrophy: implications of mitral valve and papillary muscle abnormalities assessed using cardiac magnetic resonance and echocardiography. Circ Cardiovasc Imaging 8:e003132 Nara I, Iino T, Watanabe H, Sato W, Watanabe K, Shimbo M, Umeta Y, Ito H (2018) Morphological determinants of obstructive hypertrophic cardiomyopathy obtained using echocardiography. Int Heart J 59:339–346 Kramer CM, Appelbaum E, Desai MY et al (2015) Hypertrophic cardiomyopathy registry: the rationale and design of an international, observational study of hypertrophic cardiomyopathy. Am Heart J 170:223–230 Neubauer S, Kolm P, Ho CY et al (2019) Distinct subgroups in hypertrophic cardiomyopathy in the NHLBI HCM registry. J Am Coll Cardiol 74:2333–2345 Hatle L, Brubakk A, Tromsdal A, Angelsen B (1978) Noninvasive assessment of pressure drop in mitral stenosis by Doppler ultrasound. Br Heart J 40:131–140 Yushkevich PA, Piven J, Hazlett HC, Smith RG, Ho S, Gee JC, Gerig G (2006) User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. Neuroimage 31:1116–1128 Cicchetti DV (1994) Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychol Assess 6:284–290 MathWorks MATLAB—MathWorks, Natick, MA, USA. https://uk.mathworks.com/products/matlab.html. Accessed 13 Mar 2021 Sun X, Xu W (2014) Fast implementation of DeLong’s algorithm for comparing the areas under correlated receiver operating characteristic curves. IEEE Signal Process Lett 21:1389–1393 YOUDEN WJ, (1950) Index for rating diagnostic tests. Cancer 3:32–35 Koo TK, Li MY (2016) A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med 15:155–163 Maron MS, Olivotto I, Harrigan C, Appelbaum E, Gibson CM, Lesser JR, Haas TS, Udelson JE, Manning WJ, Maron BJ (2011) Mitral valve abnormalities identified by cardiovascular magnetic resonance represent a primary phenotypic expression of hypertrophic cardiomyopathy. Circulation 124:40–47 Brownlee J. (2021) Impact of dataset size on deep learning model skill and performance estimates. In: Machine learning mastery. https://machinelearningmastery.com/impact-of-dataset-size-on-deep-learning-model-skill-and-performance-estimates/. Accessed 10 Apr 2021 Luo Y, Yang D, Liu H, Wan K, Sun J, Zhang T, Chen Y (2016) Mitral valve leaflet length as an important factor to differentiate hypertrophic cardiomyopathy from other causes of left ventricular hypertrophy. J Cardiovasc Magn Reson 18:P272 Healio (2021) Venturi effect. In: Healio. https://www.healio.com/cardiology/learn-the-heart/cardiology-review/topic-reviews/venturi-effect. Accessed 14 Apr 2021 Critoph CH, Pantazis A, Tome Esteban MT, Salazar-Mendiguchía J, Pagourelias ED, Moon JC, Elliott PM (2014) The influence of aortoseptal angulation on provocable left ventricular outflow tract obstruction in hypertrophic cardiomyopathy. Open Heart 1:e000176 Doddamani S, Bello R, Friedman MA et al (2007) Demonstration of left ventricular outflow tract eccentricity by real time 3D echocardiography: implications for the determination of aortic valve area. Echocardiography 24:860–866 Ferreira PF, Gatehouse PD, Mohiaddin RH, Firmin DN (2013) Cardiovascular magnetic resonance artefacts. J Cardiovasc Magn Reson 15:41 Maron BJ, Desai MY, Nishimura RA, Spirito P, Rakowski H, Towbin JA, Rowin EJ, Maron MS, Sherrid MV (2022) Diagnosis and evaluation of hypertrophic cardiomyopathy: JACC state-of-the-art review. J Am Coll Cardiol 79:372–389