Gender differences in normal myocardial SPECT distributions using a hybrid SPECT-CT system for attenuation correction

Abstract

199

Objectives: SPECT-CT systems with advanced iterative reconstruction algorithms correcting for attenuation correction (AC), scatter and collimator response are expected to improve the quantitative accuracy of the tracer distribution in the myocardium using SPECT imaging. In particular, these corrections are expected to eliminate differences in the normal distribution due to body habitus and gender. The purpose of this study was to evaluate gender differences in the tracer distribution in normal myocardium from SPECT imaging with and without attenuation correction (AC) using CT based attenuation maps.

Methods: Using a multi-slice SPECT-CT imaging system (SYMBIA-T6), 40 low likelihood normal volunteers (20F:47±7yo, 20M:50±11yo) were studied. All subjects were imaged using a stress 99mTc-Sestamibi protocol with CT based AC. Free breathing CT acquisitions were acquired with 1.5 sec rotational speed, 70 mAs, 130 kV, pitch 0.45. Prior to CT based AC of SPECT images, co-registration was verified and the position of the CT data was manually adjusted as necessary. Three sets of SPECT images were reconstructed using (1) filtered backprojection (FBP) with no corrections, (2) scatter and collimator response without CT AC (IT_NOAC), and (3) CT AC with scatter and collimator response (AC_ALL). Each set of images were resampled into polar map format and gender specific databases were constructed using the 4D-MSPECT software. Regional differences using a 9 segment overlay were computed between gender distributions using the paired Student’s t-test.

Results: As expected, the FBP gender distributions were statistically different (4 of 9 segments) in both inferior segments, and the anterior-distal and lateral-proximal segments. The same differences were statistically evident with IT_NOAC. When AC_ALL was applied, the normal distributions for both male and female distributions were statistically equivalent. To investigate differences in uniformity for the various corrections applied, the intensity ratio between the lateral-distal segment and the inferior-proximal segment were computed. For males, the ratios were FBP:1.29, IT_NOAC: 1.27, and AC_ALL: 0.98, while for females the ratios were FBP:1.10, IT_NOAC: 1.08, and AC_ALL: 0.96.

Conclusions: Uniformity in the normal distribution is clearly improved using AC. While image quality was visually improved with IT_NOAC, regional nonuniformity is relatively unchanged from that using only FBP. Most importantly, the dependence of the normal intensity distribution on body habitus (i.e. gender) is removed using CT AC which previous data has demonstrated will improved detection accuracy for coronary disease using AC myocardial SPECT imaging.