Adopting new gamma cameras and reconstruction algorithms: Do we need to re-establish normal reference values?

The impact of adopting new single photon emission computed tomography (SPECT) cameras and new reconstruction algorithms on left ventricular (LV) volumes has not been well established. We sought to understand the impact of hardware and software changes on normal LV reference ranges. Consecutive patients who underwent stress Tc-99m tetrofosmin 8-frame gated SPECT MPI were screened. Patients with a history of myocardial infarction, coronary revascularization, abnormal MPI, or known LV dysfunction/reduced ejection fraction were excluded. A total of 1953 consecutive normal patients, with rest LV end-diastolic volume (EDV) and end-systolic volume (ESV) measurements were analyzed. After stratifying according to sex, LV volumes indexed to body surface area were compared across the different gamma cameras. In a normal population, measurements with CZT were different from those obtained by the dual-headed NaI gamma cameras for LV EDVi (men: 53.6 ± 10.4 vs 48.3 ± 10.2 mL/m2 and women: 43.3 ± 8.9 vs 37.8 ± 9.3 mL/m2; P < 0.001) and LV ESVi (men: 21.7 ± 7.0 vs 16.9 ± 6.2 mL/m2 and women: 13.4 ± 5.3 vs 10.6 ± 4.7 mL/m2; P < 0.001). Inter- and intra-observer reliability for all measures was excellent. These findings were verified in a prospectively collected cohort. A sub-analysis of LV measurements comparing new resolution recovery and standard reconstruction software revealed no significant differences in LV measures. This study demonstrates that normal references ranges of LV volumes vary between SPECT cameras and confirms the need for establishing reference values that are camera specific.