Skip to main content

Table 5 Image quality assessment, lesion detectability, and SUV values for three different PET scanner models

From: Artificial intelligence for reduced dose 18F-FDG PET examinations: a real-world deployment through a standardized framework and business case assessment

Comparison between three different PET scanner models
PET scanner Dataset Dataset identification Image quality
(1 to 5)
Number of lesions SUVmax of target lesion SUVmean of the liver
    Mean ± SD Mean ± SD Mean ± SD Mean ± SD
1 (n=20) PET-native 96.2% 4.55 ± 0.61 1.91 ± 3.07 15.80 ± 11.64 1.87 ± 0.84
PET-processed 97.5% 4.57 ± 0.76 1.91 ± 3.07 13.84 ± 9.89 1.99 ± 0.86
2 (n=21) PET-native 98.8% 4.54 ± 0.59 3.58 ± 3.48 13.21 ± 10.8 1.81 ± 0.37
PET-processed 100% 4.70 ± 0.51 3.55 ± 3.50 11.35 ± 9.96 2.00 ± 0.46
3 (n=20) PET-native 96.2% 4.53 ± 0.62 3.53 ± 3.63 7.07 ± 4.63 2.19 ± 0.77
PET-processed 96.2% 4.60 ± 0.74 3.52 ± 3.65 6.05 ± 4.39 2.36 ± 0.81
  1. PET scanner 1= Discovery ST-4; PET scanner 2= Discovery ST-16; PET scanner 3= Discovery IQ