Shifted Helical Computed Tomography to Optimize Cardiac Positron Emission Tomography–Computed Tomography Coregistration: Quantitative Improvement and Limitations

Abstract

Positron emission tomography-computed tomography (PET-CT) uses CT attenuation correction but suffers from misregistration artifacts. However, the quantitative accuracy of helical versus cine CT in the same patient after optimized coregistration by shifting both CT data as needed for each patient is unknown. We studied 293 patients undergoing cardiac perfusion PET-CT using helical CT attenuation correction for comparison to cine CT. Objective, quantitative criteria identified perfusion abnormalities that were associated visually with PET-CT misregistration. Custom software shifted CT data to optimize coregistration with quantitative artifact improvement. The majority (58.1%) of cases with both helical and shifted helical CT data (n = 93) had artifacts that improved or resolved by software shifting helical CT data. Translation of shifted helical CT was greatest in the x-direction (8.8 ± 3.3 mm) and less in the y- and z-directions (approximately 3.5 mm). The magnitude of differences in quantitative end points was greatest for helical (p = .0001, n = 177 studies), less for shifted helical but significant (p = .0001, n = 93 studies), and least for cine (not significant, n = 161 studies) CT compared to optimal attenuation correction for each patient. Frequent artifacts owing to attenuation-emission misregistration are substantially corrected by software shifting helical CT scans to achieve proper coregistration that, however, remains on average significantly inferior to cine CT attenuation quantitatively.