Skip to main content

Table 3 Reading results

From: Impact of attenuation correction on clinical [18F]FDG brain PET in combined PET/MRI

Reader   PETPET/CT PETCTderived PETDixon PETPseudoCT PETBoneDixon
1 Hypometabolic areas (visual) 1.1 ± 1.7 1.3 ± 2.1 1.6 ± 2.0 1.3 ± 1.7 1.7 ± 1.7
  Hypometabolic areas (NEUROSTAT) 1.5 ± 2.1 1.8 ± 2.0 2.4 ± 2.1 2.3 ± 2.3 1.9 ± 2.0
  Hypometabolic areas (BRASS) 1.1 ± 1.6 1.9 ± 2.6 2.2 ± 2.3 2.1 ± 2.1 2.1 ± 2.1
  Correctly classified brain regions compared to PETPET/CT (visual)   10.8 ± 1.7 10.6 ± 1.5 10.5 ± 1.9 9.1 ± 2.3
  Correctly classified brain regions compared to PETPET/CT (NEUROSTAT)   10.9 ± 1.7 10.6 ± 1.8 10.7 ± 2 9.6 ± 2.5
  Correctly classified brain regions compared to PETPET/CT (BRASS)   10.9 ± 1.7 10.5 ± 1.7 10.7 ± 1.7 9.6 ± 1.7
  Correct imaging diagnosis compared to PETPET/CT   11/13 11/13 11/13 7/7
2 Hypometabolic areas (visual) 1.4 ± 2.0 1.4 ± 1.9 1.3 ± 1.7 1.5 ± 2.2 1.7 ± 2.1
  Hypometabolic areas (NEUROSTAT) 2.8 ± 3.1 2.1 ± 2.4 2.5 ± 2.8 1.8 ± 2.3 3.0 ± 2.7
  Hypometabolic areas (BRASS) 2.2 ± 2.9 1.9 ± 2.4 2.2 ± 2.8 1.1 ± 2.0 1.3 ± 2.1
  Correctly classified brain regions compared to PETPET/CT (visual)   11.5 ± 0.7 11.6 ± 0.9 10.8 ± 1.8 11.1 ± 1.1
  Correctly classified brain regions compared to PETPET/CT (NEUROSTAT)   11.2 ± 1.1 11.2 ± 1.4 10.4 ± 2.4 10.7 ± 1.7
  Correctly classified brain regions compared to PETPET/CT (BRASS)   11.0 ± 1.2 11.4 ± 0.7* 10.4 ± 1.5* 10.1 ± 1.5
  Correct imaging diagnosis compared to PETPET/CT   13/13 12/13 12/13 6/7
3 Hypometabolic areas (visual) 1.8 ± 2.5 2.1 ± 1.8 2.6 ± 1.9 2.5 ± 1.9 2.7 ± 1.9
  Hypometabolic areas (NEUROSTAT) 2.2 ± 2.5 1.9 ± 1.6 2.5 ± 1.9 2.5 ± 1.8 2.1 ± 1.6
  Hypometabolic areas (BRASS) 2.2 ± 2.5 1.9 ± 1.7 2.4 ± 1.7 2.6 ± 1.9 2.1 ± 1.6
  Correctly classified brain regions compared to PETPET/CT (visual)   10.8 ± 1.7 10.6 ± 1.5 10.5 ± 1.9 9.1 ± 2.3
  Correctly classified brain regions compared to PETPET/CT (NEUROSTAT)   10.7 ± 1.6 10.8 ± 1.3 10.5 ± 1.9 9.1 ± 2.3
  Correctly classified brain regions compared to PETPET/CT (BRASS)   10.9 ± 1.7 10.6 ± 2.8 10.7 ± 2.0 9.6 ± 2.5
  Correct imaging diagnosis compared to PETPET/CT   11/13 11/13 11/13 7/7
  1. PETPET/CT, PETCTderived, PETDixon, PETPseudoCT, and PETBoneDixon from 13 patients were evaluated by three experienced readers. For each PET, 12 brain areas were classified as either hypometabolic or normal according to the visual impression and according to the semiquantitiative approaches NEUROSTAT and BRASS. For PETCTderived, PETDixon, PETPseudoCT, and PETBoneDixon, the number of correctly classified brain regions was determined according to the standard of truth PETPETCT. Moreover, the readers were asked to formulate an imaging diagnosis on the basis of the PET data without clinical information. The imaging diagnosis was also compared to the standard of truth PETPETCT. For PETPETCT, PETCTderived, PETDixon, and PETPseudoCT (PETBonedixon), paired (two-sampled) t tests (two-sided; α = 0.05) were calculated. α was Bonferroni corrected for multiple comparisons. The number of hypometabolic areas across the patients did not differ significantly between PETCTderived, PETDixon, PETPseudoCT, and PETBoneDixon and the reference: PETPETCT in either of the readers. *Reader 2 classified significantly more brain regions correctly using BRASS from PETDixon as compared to PETPseudoCT (p = 0.012). Apart from that, the number of correctly classified brain regions did not differ between PETCTderived, PETDixon, PETPseudoCT, and PETBoneDixon across patients. No systematic inferiority of any reconstruction was observed