Original Contributions
Determination of normal regional left ventricular function from cine-MR images using a semi-automated edge detection method

https://doi.org/10.1016/S0730-725X(98)00158-1Get rights and content

Abstract

A semi-automated edge detection method for the delineation of the endo- and epicardial borders of the left ventricle from cine MR images has been developed. The feasibility of this was demonstrated by processing end diastolic and end systolic ECG-gated images of four short axis images in 10 healthy subjects. The first derivative method combined with a 2D weighted polynomial fitting procedure was used to determine the endo- and epicardial borders, which then allowed determination of the wall motion, wall thickening, and ejection fraction, of the left ventricle. The results show that the end-systolic radial wall motion varies from (32 ± 8)% to (76 ± 12)%, and wall thickening from (0.60 ± 0.46) cm to (1.26 ± 0.50) cm. An average ejection fraction of (69 ± 6)% was found which agrees well with literature values. The method described, for the delineation of the borders, reduces considerably the long and tedious operator time inherent in manual measurement and greatly increases the reproducibility of the measurements.

Introduction

Detailed information on cardiac dimensions and function is of great interest for cardiac diagnosis1, 2 and treatment monitoring,3 as well as for basic physiological4 and biomechanical5, 6 studies of cardiac wall motion. MRI has the capability of providing an accurate and precise determination of cardiac geometry. Several attempts have been made to use MRI as a non-invasive technique for the assessment of cardiac function, specifically, the determination of left ventricular (LV) volumes, ejection fractions7, 8, 9, 10 and myocardial wall thickness, and thickening and motion,11, 12, 13, 14 in patients with congenital or acquired heart disease. These prior studies used manual tracing of epicardial and endocardial borders and manual segmentation of the myocardium in multiple, parallel images for subsequent determination of functional parameters.

Manual tracing and segmentation, although reasonably accurate, is time consuming and subjective, potentially leading to substantial inter- and intra-observer variability. To reduce the variability and time constraints inherent in observer tracing of biomedical images, digital image processing methods have been widely used in other imaging modalities. Those methods published rely on complicated techniques such as closed loop B-spline curve fitting,15 segmentation by region growing,16 graph searching,17 deformable contour models18 and others.19, 20 We have developed a method that relies on first derivative edge detection and weighted polynomial fitting, both of which are elementary mathematical techniques, to determine the epicardial and endocardial borders. The epicardial and endocardial borders were then used to determine the LV end diastolic and end systolic volumes, and to calculate segmental wall motion and thickening.

Section snippets

Imaging techniques

After giving informed consent, ten healthy male volunteers, aged 24 to 30 years, (mean 27 years) with normal physical examination results and no history of heart or lung disease underwent MR imaging using a 0.95 Tesla Siemens (Erlangen, Germany) Impact imager, with a standard body transmit and receive RF coil. Multiple temporal cine-MR images were obtained in the short axis, positioned using the end-systolic long axis view as the frame of reference, using a FISP pulse sequence with gradient

Results

Evaluation of the wall thickening, wall motion, and ejection fraction from the ED and ES images of the four SA slices took less than 3 min, including manual positioning of the initial circles. Figure 5 shows an example of both the endo- and epicardial borders obtained automatically superimposed onto four original end systolic images. The automatic software successfully detected the LV borders in all of the images where good contrast was present. In the cases where cardiac activity was too

Discussion

Cine-MR imaging has been proved to be highly accurate in assessment of cardiac volumes and function. In the present study we used a cardiac gated cine-MR imaging technique with non-contiguous slices and a semi-automated edge detection method. The use of these two techniques allows the endo- and epicardial borders of the LV to be delineated in a time efficient manner.

Evaluation of left ventricular function is of great importance in the diagnosis and management of patients with known or suspected

References (39)

  • J Wynne et al.

    Estimation of left ventricular volumes in man from biplane cineangiograms filmed in oblique projections

    Am J Cardiol

    (1978)
  • W.P Hood et al.

    Wall stress in the normal and hypertrophied left ventricle

    Am J Cardiol

    (1968)
  • F.H Sheehan et al.

    Effect of interventions in salvaging left ventricular function in acute myocardial infarction: A study of intracoronary streptokinase

    Am J Cardiol

    (1983)
  • Y Kong et al.

    Assessment of regional myocardial performance from biplane coronary cineangiograms

    Am J Cardiol

    (1971)
  • O Risum et al.

    Mortality and morbidity after coronary artery bypass surgery related to pre-operative left ventricular ejection fraction: A follow up study

    Eur Heart J

    (1996)
  • Editorial. Left ventricular ejection fraction: An important but incomplete determinant of long-term outcome after coronary bypass surgery

    Eur Heart J

    (1996)
  • C.B Higgins

    MR of the heart: Anatomy, physiology and metabolism

    AJR

    (1988)
  • A Young et al.

    Three-dimensional left ventricular deformation in hypertrophic cardiomyopathy

    Circulation

    (1994)
  • R.C Semelka et al.

    Normal left ventricular dimensions and function: interstudy reproducibility of measurements with cine-MR imaging

    Radiology

    (1990)
  • Cited by (47)

    • Assessment of left ventricular ejection fraction using the wall motion score index in cardiac magnetic resonance imaging

      2012, Archives of Cardiovascular Diseases
      Citation Excerpt :

      We did not grade hyperkinesias (our maximal LVEF was 64%) where planimetric methods would be more appropriate. Our range of normal LVEF values is similar to previously published MRI studies (mean of 66 ± 6%) [20–24] as well as RNA studies (mean of 63 ± 6%) [25–28]. Finally, estimation of wall motion abnormalities is subjective and the WMSI results are dependent on expertise and image quality.

    • Unsupervised 4D myocardium segmentation with a Markov Random Field based deformable model

      2011, Medical Image Analysis
      Citation Excerpt :

      At that time, segmentation techniques were usually general purpose methods (Clarke et al., 1995) applied to the cardiac case. Some references of interest are Goshtasby and Turner (1995), Nachtomy et al. (1998), Waiter et al. (1999), which constitute good entry points into the cardiac segmentation problem. In these techniques, some sort of manual intervention is needed to let the algorithm start or to guide it through the segmentation process.

    • A review of segmentation methods in short axis cardiac MR images

      2011, Medical Image Analysis
      Citation Excerpt :

      Contours in each individual image are then recovered using minimum surface segmentation. Several solutions are proposed, in order not to include papillary muscles: computation of the convex hull of the contour (van der Geest et al., 1994; Lin et al., 2006), applications of mathematical morphology (MM) operations such as opening and closing on the contour (Nachtomy et al., 1998; Cousty et al., 2010), or fitting a parametric curve to the contour, allowing it to be smooth (Waiter et al., 1999). The epicardium is found during a second step, often relying on the endocardial contour, that makes use of a spatial model incorporating myocardial thickness or MM operators, applied on the endocardial contour.

    • Blood oxygen level-dependent (BOLD) MRI: A novel technique for the assessment of myocardial ischemia as identified by nuclear imaging SPECT

      2007, European Journal of Internal Medicine
      Citation Excerpt :

      Using the standard two-dimensional (2D) Fourier transformation algorithm, the raw data was reconstructed to 256 × 256 pixel images. MR images were transferred to an offline workstation and processed using a home-written package [16]. In each patient, two short-axis tomograms from mid-LV with 16 segments each, corresponding to SPECT slices showing areas of reversible ischemia, were analysed for signal change on BOLD MRI.

    • Blood oxygen level-dependent (BOLD) magnetic resonance imaging in patients with dypiridamole induced ischaemia; a PET comparative study

      2007, International Journal of Cardiology
      Citation Excerpt :

      In each patient, two short axis tomograms from mid LV were analysed for signal change on BOLD MRI. The myocardium in each of the slices was masked out, from the rest of the image, using semi-automated edge detection method [18]. These slices, in turn, were then aligned with the corresponding short axis tomographic images of myocardial 13NH3 images from the PET study.

    View all citing articles on Scopus
    1

    Present address of Roger J. Trent: Medical Cardiology, Western Infirmary, Glasgow, UK.

    View full text