Simulation of lung volume and SPECT count errors due to mismatch between SPECT and CT during free-breathing in lung...

Respiratory phase mismatch between single-photon emission computed tomography (SPECT) and computed tomography (CT) acquisition phases presents a challenge in lung perfusion scintigraphy using SPECT/CT. This study simulated lung volume and SPECT counts changes under free-breathing and breath-hold CT conditions compared to respiratory-synchronized acquisition. Chest 4D-CT images, divided into 10 respiratory phases, were used to generate lung, soft tissue, liver, and bone regions for each phase. A digital phantom was constructed via image processing using ImageJ. SPECT images were generated from these phantoms by employing the Prominence Processor to simulate projection data and reconstruct images. Simulations included a 'synchronized image,' where both SPECT and μMAP for attenuation correction were created in the same phase; a 'free-breathing image,' combining a free-breathing SPECT and μMAP; and a 'CT breath-hold image,' using phase-specific μMAPs with the free-breathing SPECT image for attenuation correction. Lung volumes and SPECT counts in the free-breathing and CT breath-hold images were compared with those in the synchronized image. By analyzing the relative errors caused by differences in the μMAPs, the study evaluated the impact of mismatch between SPECT and CT phases. Results indicated that lung volumes appeared reduced during inspiration and increased during expiration compared with synchronized images. No significant difference in the relative error was observed between the free-breathing and CT breath-hold images. Our findings revealed that in the quantitative evaluation of lung perfusion SPECT, varying the μ-map phase during free-breathing acquisition did not result in a significant improvement, suggesting that the mismatch between SPECT and CT had no statistically significant effect on quantitative accuracy. Compared with respiratory-gated SPECT, free-breathing acquisitions introduced potential errors of approximately 2.5% in lung volume measurement and 1.2% in SPECT counts. However, these errors were within acceptable tolerance limits for clinical diagnosis, indicating that free-breathing acquisition had minimal effects on diagnostic capability.