Background: Gated single photon emission computed tomography imaging allows simultaneous determination of myocardial perfusion and function. Quantitative perfusion measurements can be based on regional tracer uptake, but function measurements ordinarily require endocardial and epicardial edge detection, which is problematic because of the inherently low spatial resolution and image noise in single photon emission computed tomography images. This article presents methods for quantification of both function and perfusion that do not require edge detection.
Methods and results: In SPECT imaging the partial volume effect causes changes in myocardial wall thickness to be reflected as changes in pixel counts in pixels representing the myocardial wall. This effect allows an estimation of changes in myocardial wall thickness by comparing pixel counts in end-systolic images with corresponding pixel counts in end-diastolic images. This article first describes a standard method to quantify myocardial perfusion by sampling myocardial tracer activity at rest and stress. The same method is then used to sample tracer activity in diastolic and systolic images. A new method is developed to convert the diastolic and systolic samples into quantitative estimates of regional wall thickening. A method is then developed to convert the regional wall thickening fractions into a global left ventricular ejection fraction. A normal database is presented. Receiver operating characteristic analysis is used to establish normal limits.
Conclusion: This method requires no edge detection or geometric boundary estimates. Computer results are presented in a simple and intuitive format, which is uniform for parameters of both perfusion and function. The method is robust and produces relatively few false-positive results.