Purpose: Following the transarterial injection of Y-90 microspheres for liver radioembolization, PET/CT images can be acquired for dosimetric analysis. Unfortunately, Y-90 emits positrons with a very low abundance (32 ppm) so long acquisition times (>15 min/bed) are required for acceptable image quality. Therefore, imaging is often limited to the liver (~15 cm axial extent) and excludes other organs of interest (i.e. lungs). This study investigated the improvement of Y-90 PET/CT image quality using a digital PET system with higher sensitivity and longer axial extent.
Methods: Using a modified IEC NEMA image quality phantom, 2 sets of 3 spheres (22, 28, 37 mm diameter) were prepared with 6.6:1 and 12.5:1 sphere-to-background activity ratios. Images were acquired at 4 time points (0.50, 0.41, 0.26, and 0.14 MBq/mL background activity concentrations) on both a DMI (GE Discovery MI, 25 cm axial extent) and a D710 (GE Discovery 710, 15 cm axial extent) using a single 30 min/bed acquisition. Images were reconstructed following the QUEST phantom study recommendations. Contrast-to-noise ratios (CNR) and errors in activity quantification were calculated for all spheres and time points.
Results: Sphere CNRs were, on average, 1.51 times higher on the DMI than D710 images. Conversely, Y-90 DMI single-bed acquisition times can be reduced to 44% to achieve comparable CNRs to D710 acquisitions, which facilitates a two-bed acquisition to cover the lungs. In terms of quantitation, DMI was more accurate than D710 (<20% versus ~25% for 37 mm 12.5:1 sphere); however, accuracy degraded for both scanners with lower sphere uptake and diameters.
Conclusion: Post-therapy Y-90 PET/CT acquisition times can be greatly reduced, without sacrificing image quality, using digital PET/CT with longer axial extent compared to standard PET/CT systems. Beyond improving hepatic dosimetry accuracy, the digital PET/CT with longer axial extent would also permit imaging the thorax for lung dosimetry.