Main Article Content
Background: Left Ventricular Ejection Fraction (LVEF) has been used in evaluating cardiac function. It is calculated generally by the difference between end-diastolic volumes (EDV) and end-systolic volumes (ESV). In order to obtain the EDV and ESV from magnetic resonance imaging (MRI), an experienced cardiologist is required to select the smallest and largest axial images from 30 phases in each slice that represent the ESV and EDV, respectively. This process is time consuming and could have individual-dependent variability.
Objectives: To develop an algorithm that determines the EDV and ESV, and automatically calculates the LVEF as an output in MRI in order to reduce operation time and interobserver variability for the user.
Materials and methods: Fifteen thalassemia patients were recruited into this study. Image data were acquired using a 1.5 Tesla MRI scanner. Scanning protocol included field of view (FOV) of 300 mm, matrix size of 256×256, slice thickness of 8 mm, pixel spacing of 1.25 mm, and repetition time (TR) and echo time (TE) of 3.83 and 1.69 ms, respectively. The proposed algorithm included 3 steps; left ventricle (LV) segmentation, end diastolic (ED) and end systolic (ES) phase selection, and LVEF calculation. The LV region was segmented by using the patch-based clustering method. Shape of the LV was tuned by mathematical morphologies and the Gaussian filter. The processes were repeated for all phases. The ED and ES phases were selected from those that had the maximum and minimum number of pixels, respectively, in the LV area. The EDV and ESV were the sum of the number of pixels in the segmented LV areas of the ED and ES phases, which were multiplied by pixel spacing and slice thickness from all slices. Finally, the LVEF was calculated and reported.
Results: Precision and recall were used to evaluate segmentation performance, which was good in the experimental results of the proposed method. Precision and recall were on average approximately 0.9 and 0.7 for the ED and ES phase, respectively. The different percentage values of the LVEF was 3.32% between the proposed and manual segmentation method.
Conclusion: This work proposed automatic LVEF evaluation from MRI. Patch-based clustering techniques, mathematical morphologies and the Gaussian filter were used to segment the LV area automatically. Experimental results showed that the proposed method could evaluate LVEF values that were close to those from two experts who used the manual segmentation method.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Personal views expressed by the contributors in their articles are not necessarily those of the Journal of Associated Medical Sciences, Faculty of Associated Medical Sciences, Chiang Mai University.
Foley TA, Mankad SV, Anavekar NS, Bonnichsen CR, Morris MF, Miller TD, et al. Measuring left ventricular ejection fraction-techniques and potential pitfalls. European Cardiology. 2012; 8(2): 108-14.
Lang RM, Badano LP, Mor-Avi V, Afilalo J, Armstrong A, Ernande L, et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr. 2015; 28(1): 1-39.
Petitjean C, Dacher J-N. A review of segmentation methods in short axis cardiac MR images. Med Image Anal. 2011; 15(2): 169-84.
Wang L, Pei M, Codella NCF, Kochar M, Weinsaft JW, Li J, et al. Left ventricle: fully automated segmentation based on spatiotemporal continuity and myocardium information in cine cardiac magnetic resonance imaging (LV-FAST). Biomed Res Int. 2015; 367583.
Wantanajittikul K, Theera-Umpon N, Saekho S, Auephanwiriyakul S, Phrommintikul A, Leemasawat K. Automatic cardiac T2* relaxation time estimation from magnetic resonance images using region growing method with automatically initialized seed points. Comput Methods Programs Biomed. 2016; 130: 76-86.
Hu H, Gao Z, Liu L, Liu H, Gao J, Xu S, et al. Automatic Segmentation of the Left Ventricle in Cardiac MRI Using Local Binary Fitting Model and Dynamic Programming Techniques. PLOS ONE. 2014; 9(12): e114760.
Uzunbaş MG, Zhang S, Pohl KM, Metaxas D, Axel L. Segmentation of myocardium using deformable regions and graph cuts. In: 2012 9th IEEE International Symposium on Biomedical Imaging (ISBI). 2012; p. 254-7.
Grosgeorge D, Petitjean C, Caudron J, Fares J, Dacher J-N. Automatic cardiac ventricle segmentation in MR images: a validation study. Int J Comput Assist Radiol Surg. 2011; 6(5): 573-81.
Pieciak T. Segmentation of the Left Ventricle Using Active Contour Method with Gradient Vector Flow Forces in Short-Axis MRI. Information Technologies in Biomedicine. Berlin, Heidelberg: Springer; 2012; p. 24-35.
Ngo TA, Carneiro G. Left ventricle segmentation from cardiac MRI combining level set methods with deep belief networks. 2013 IEEE International Conference on Image Processing. 2013; p. 695-9.
Zheng Q, Lu Z, Zhang M, Xu L, Ma H, Song S, et al. Automatic Segmentation of Myocardium from Black-Blood MR Images Using Entropy and Local Neighborhood Information. PLOS ONE. 2015; 10(3): e0120018.
Hadhoud MMA, Eladawy MI, Farag A, Montevecchi FM, Morbiducci U. Left Ventricle Segmentation in Cardiac MRI Images. American Journal of Biomedical Engineering. 2012; 2(3): 131-5.
Nambakhsh CMS, Yuan J, Punithakumar K, Goela A, Rajchl M, Peters TM, et al. Left ventricle segmentation in MRI via convex relaxed distribution matching. Medical Image Analysis. 2013; 17(8): 1010-24.
Dreijer JF, Herbst BM, du Preez JA. Left ventricular segmentation from MRI datasets with edge modelling conditional random fields. BMC Med Imaging 13. 2013; 13(1): 1-24.
Hu H, Liu H, Gao Z, Huang L. Hybrid segmentation of left ventricle in cardiac MRI using Gaussian-mixture model and region restricted dynamic programming. Magn Reson Imaging. 2013; 31(4): 575-84.
Wantanajittikul K, Saekho S, Phrommintikul A, Theera-Umpon N, Auephanwiriyakul S, Tapanya M, et al. Fully automatic cardiac T2* relaxation time estimation using marker-controlled watershed. 2015 IEEE International Conference on Control System, Computing and Engineering (ICCSCE). 2015; p. 377-82.
Tan LK, Liew YM, Lim E, McLaughlin RA. Convolutional neural network regression for short-axis left ventricle segmentation in cardiac cine MR sequences. Med Image Anal. 2017; 39:78-86.
Lalande A, Salvé N, Comte A, Jaulent MC, Legrand L, Walker PM, et al. Left ventricular ejection fraction calculation from automatically selected and processed diastolic and systolic frames in short-axis cine-MRI. J Cardiovasc Magn Reson. 2004; 6(4): 817-27.
Lu Y-L, Connelly KA, Dick AJ, Wright GA, Radau PE. Automatic functional analysis of left ventricle in cardiac cine MRI. Quant Imaging Med Surg. 2013; 3(4): 200-9.
Theera-Umpon N. White Blood Cell Segmentation and Classification in Microscopic Bone Marrow Images. Fuzzy Systems and Knowledge Discovery. Berlin, Heidelberg: Springer; 2005; p. 787-96.
Blinchikoff HJ, Zverev AI. Filtering in the Time and Frequency Domains. Revised ed. edition. Raleigh, NC: SciTech Publishing; 2001.
Bezdek JC, Keller J, Krisnapuram R, Pal N. Fuzzy Models and Algorithms for Pattern Recognition and Image Processing. 1999 edition. New York: Springer; 2005.
Gonzalez RC, Woods RE. Digital Image Processing. 3rd ed. Upper Saddle River, N.J: Pearson; 2007.