Original article
Quantitative gated blood pool SPECT: Analysis of 3-dimensional models for the assessment of regional myocardial wall motion

https://doi.org/10.1067/mnc.2002.121448Get rights and content

Abstract

Background. Gated blood pool single photon emission computed tomography (SPECT) (GBPS) uses truly 3-dimensional (3D) data, requiring attention to appropriate reference systems and segmentation models for proper quantification. To date, optimal 3D reference models have not been evaluated. In this study several techniques for 3D GBPS were evaluated. Methods and Results. Static and dynamic cardiac phantom evaluations were performed, and GBPS studies for 3 healthy subjects and 9 patients were processed by a variety of 3D analysis techniques to determine the optimum parameters for identification of abnormal segments when compared with coronary arteriography and left ventriculography. Left ventricular wall motion was quantified by calculation of regional ejection fraction (rEF) through use of count, volume, and cord length changes from end diastole to end systole. Three contractile models were evaluated: (1) fixed center of mass (COM), (2) floating COM, and (3) a modification of the method developed by Slager et al (J Am Coll Cardiol 1986;7:317-26), based on the motion of implanted endocardial markers. Eight, twelve, and eighteen 3D segments were analyzed by means of the 3 contractile models and correlated against coronary artery disease assessed by coronary arteriography. Single-head gamma-camera acquisition provided adequate counting statistics to reliably compute rEF for up to 18 left ventricular segments. Using count changes the overall results were able to identify myocardium supplied by diseased coronary arteries when compared with coronary arteriography. Cord length and, to a lesser degree, volume changes provided somewhat poorer sensitivities and specificities when compared with rEF computed from regional count changes, as compared with coronary arteriography. Conclusions. Three-dimensional quantitative GBPS appears to be a sensitive method for assessing wall motion defects due to coronary artery disease. (J Nucl Cardiol 2002;9:271-84.)

References (0)

Cited by (17)

  • Evaluation of ventricular function in Nuclear Medicine

    2008, Revista Espanola de Medicina Nuclear
  • SPECT versus planar gated blood pool imaging for left ventricular evaluation

    2007, Journal of Nuclear Cardiology
    Citation Excerpt :

    IL has a more stable bias, not correlated with the value of the ejection fraction. Groch et al8,14,15 previously studied other algorithms for cardiac blood pool segmentation. They reported a significant correlation between their 3D method and planar ventriculography (r = 0.92), but unlike our study, their method generated a bias that correlated with the magnitude of the LVEF.

  • Nuclear Cardiology: Present and Future

    2006, Current Problems in Cardiology
  • Medical Imaging Technology and Imaging Agents

    2023, Advances in Experimental Medicine and Biology
  • Left-Ventricle Segmentation of SPECT Images of Rats

    2015, IEEE Transactions on Biomedical Engineering
View all citing articles on Scopus

This work was supported, in part, by a grant from Siemens Medical Systems, Hoffman Estates, Ill.

View full text