A multicenter evaluation of a new post-processing method with depth-dependent collimator resolution applied to full-time and half-time acquisitions without and with simultaneously acquired attenuation correction.

Background: The field of nuclear cardiology is limited by image quality and length of procedure. The use of depth-dependent resolution recovery algorithms in conjunction with iterative reconstruction holds promise to improve image quality and reduce acquisition time. This study compared the Astonish algorithm employing depth-dependent resolution recovery and iterative reconstruction to filtered backprojection (FBP) using both full-time (FTA) and half-time (HTA) data. Attenuation correction including scatter correction in conjunction with the Astonish algorithm was also evaluated.

Methods: We studied 187 consecutive patients (132 with cardiac catheterization and 55 with low likelihood for CAD) from three nuclear cardiology laboratories who had previously undergone clinically indicated rest/stress Tc-99m sestamibi or tetrofosmin SPECT. Acquisition followed ASNC guidelines (64 projections, 20-25 seconds). Processing of the full-time data sets included FBP and Astonish (FTA). A total of 32 projection data sets were created by stripping the full-time data sets and processing with Astonish (HTA). Attenuation correction was applied to both full-time and half-time Astonish-processed images (FTA-AC and HTA-AC, respectively). A consensus interpretation of three blinded readers was performed for image quality, interpretative certainty, and diagnostic accuracy, as well as severity and reversibility of perfusion and functional parameters.

Results: Full-time and half-time Astonish processing resulted in a significant improvement in image quality in comparison with FBP. Stress and rest perfusion image quality (excellent or good) were 85%/80% (FBP), 98%/95% (FTA), and 95%/92% (HTA), respectively (p < 0.001). Interpretative certainty and diagnostic accuracy were similar with FBP, FTA, and HTA. Left ventricular functional data were not different despite a slight reduction in half-time gated image quality. Application of attenuation correction resulted in similar image quality and improved normalcy (FTA vs. FTA-AC: 76% vs. 95%; HTA vs. HTA-AC: 76% vs. 100%) and specificity (FTA vs. FTA-AC: 62% vs. 78%; HTA vs. HTA-AC: 63% vs. 84%) (p < 0.01 for all comparisons).

Conclusion: Astonish processing, which incorporates depth-dependent resolution recovery, improves image quality without sacrificing interpretative certainty or diagnostic accuracy. Application of simultaneously acquired attenuation correction, which includes scatter correction, to full-time and half-time images processed with this method, improves specificity and normalcy while maintaining high image quality.