Division of Image Processing / LKEB

Department of Radiology, Leiden University Medical Center

Quantification of Cardiovascular Magnetic Resonance Imaging (CMR)

Background

In this project, methods are developed for quantification of left ventricular function from magnetic resonance imaging (MRI) of the heart. A software package (MASS) has been developed providing automated contour detection and manual contour tracing facilities as well as dedicated global and regional LV function quantification tools. A short-axis MRI study of the heart consists of multiple slices covering the whole heart and multiple phases within the cardiac cycle. Left ventricular volumes can be measured from these images without using any geometrical assumptions from contours describing the endocardial (inner) and epicardial (outer) boundaries of the myocardium. Regional function analysis is also possible by measuring the left ventricular wall motion or wall thickening.

Goals

Development of automated contour detection techniques to be able to reduce the total analysis time and to reduce the inter- and intraobserver variabilities associated with manual contour tracing.

Approach

The contour detection that has been developed follows a number of steps. In the first step the center of the left ventricular cavity is detected using the Hough transform. By using this method for all slices through the left ventricle the long-axis of the LV is detected, which results in center points for the individual slice locations. Using this center point, epicardial contours are found in the first phase and subsequently in the remaining phases using a frame-to-frame contour detection procedure. This frame-to-frame epicardial contour detection procedure is based on matching of line profiles which are positioned perpendicularly to the model contour (derived from the first phase) and then automatically positioned at the corresponding tissue transitions in other phases within the same slice level. By this approach the algorithm is able to deal with the fact that the epicardial boundary of the myocardium is adjacent to regions having different gray value characteristics. A first estimate of the endocardial contour is found using an optimal thresholding technique within the region described by the epicardial contour. The final endocardial contour is found by using a model-based edge-detection technique using dynamic programming.

Status

The developed contour detection algorithms are integrated into the software package QMass-MR ®, which is commercially available through MEDIS medical imaging systems.

Contact

Rob J. van der Geest, PhD Division of Image Processing Department of Radiology, 1-C2S Leiden University Medical Center P.O. Box 9600 2300 RC Leiden The Netherlands Tel. +31 (0)71 526 2138 Fax. +31 (0)71 526 6801 e-mail: R.J.van_der_Geest@lumc.nl

Publications

1995

Helbing WA, Bosch JG, Maliepaard C, Rebergen SA, van der Geest RJ, Hansen B, Ottenkamp J, Reiber JHC, de Roos A. Comparison of echocardiographic methods with magnetic resonance imaging for assessment of right ventricular function in children. Am J Cardiol 1995; 76:589-594.