Platform-specific impact of CT attenuation correction on CZT myocardial perfusion SPECT: per-vessel detection of obstructive coronary artery disease

Attenuation artifacts on myocardial perfusion SPECT differ across camera geometries. This study evaluated whether CT attenuation correction (CTAC) improves per-vessel detection of obstructive coronary artery disease (CAD) on a stationary 19-pinhole CZT system. Rest–stress ⁹⁹ᵐTc-tetrofosmin CZT SPECT studies from 48 patients with angiographically confirmed CAD (185 vessels) were retrospectively reconstructed with and without CTAC. The reference standard was ≥ 70% diameter stenosis on coronary CT angiography or invasive angiography, with no perfusion input. Disease-negative participants were excluded by design. Perfusion was quantified as the percent defect averaged within each coronary territory, assigned per patient on hybrid SPECT/CT fusion images. Per-vessel discrimination of the reversible (stress − rest) defect rate was compared between reconstructions. Compared with non-attenuation-corrected (NAC) reconstructions, CTAC lowered the global rest defect rate ( P < 0.0001) and raised the reversible defect rate (median 5.9% vs. 1.5%; P < 0.0001). The reversible defect rate discriminated territories subtended by obstructive stenosis better with CTAC than with NAC (AUC 0.717 vs. 0.614; ΔAUC 0.103; DeLong P = 0.0164; bootstrap P = 0.010; within-patient permutation P = 0.0474). At the Youden point, sensitivity rose to 75.5% from 52.8%, with specificity 64.4% versus 68.9%. The gain was greatest in the LCX (0.803 vs. 0.658; P = 0.043) and RCA territories. On this 19-pinhole CZT platform, CTAC improves per-vessel discrimination of obstructive CAD, most prominently in the LCX and RCA territories. The benefit is sensitivity-led, opposite to the specificity-led gain reported on Anger cameras, at a modest cost in specificity.