Abstract
Though myocardial perfusion imaging (MPI) with single-photon emission computed tomography (SPECT) is an established diagnostic method, equivocal studies are commonly encountered. New software has been introduced that incorporates resolution recovery (RR) and noise regulation into the reconstruction algorithm and has been used to facilitate "half-dose" and "half-time" studies. Its utility with "full-time, full-dose" acquisition has not been well studied.
We sought to understand the potential benefit of incorporating RR software in equivocal SPECT studies.
Patients with full-time, full-dose SPECT MPI were reviewed and those with equivocal results, who subsequently underwent cardiac Rb-82 positron emission tomography (PET) scan were identified. Image reconstruction was performed with iterative reconstruction (IR), attenuation correction (IR + AC), and RR software (IR + AC + RR). Images were anonymized and read blindly by consensus of two experienced readers. All images were qualitatively assessed and semi-quantitatively graded using summed stress and summed rest scores.
45 patients were included (28 males, age = 59.6 +/- A 9.9 years) and the diagnostic accuracy of each of the reconstruction algorithms (IR, IR + AC, IR + AC + RR) was compared to Rb-82 PET. Agreement of clinical diagnosis of each SPECT reconstruction with Rb-PET showed incremental improvement. The agreement with PET for IR + AC + RR (kappa = 0.66, CI 0.454-0.875) is significantly better than for IR (kappa = 0.22, CI 0.0-0.450, P = .005) and for IR + AC (kappa = 0.32, CI 0.077-0.563, P = .03). Also, IR + AC + RR improved the clinical diagnosis in 14 cases and with overall improvement of reclassification proportion of 23.5% compared to IR (P = .01). Using PET as a reference standard, ROC curves were created for IR + AC + RR, IR + AC, and IR which showed incremental value of the area under the curve of IR + AC + RR (AUC: 0.87; CI 0.76-0.98) over IR + AC (AUC: 0.75; CI 0.61-0.89, P = .078), and over IR (AUC 0.68; CI 0.52-0.84, P = .025).
The addition of RR may help in the diagnosis of patients with equivocal SPECT MPI without the need for additional testing. Further prospective studies are needed to define the role of this new software.