Skip to main content

Table 3 Clinical image scoring from two readers of 20 whole-body FDG scans acquired on a DMI scanner. p-values were calculated using the Wilcoxon signed-rank test. The values in bold represent the best (highest) score for each metric

From: Sequential deep learning image enhancement models improve diagnostic confidence, lesion detectability, and image reconstruction time in PET

Reconstruction

Lesion detectability

Diagnostic confidence

Image quality

ToF-BSREM

3.62 ± 0.11

3.55 ± 0.12

2.77 ± 0.13

ToF-OSEM + DLE

3.48 ± 0.09

(p = 1.00)

3.52 ± 0.09

(p = 1.00)

3.65 ± 0.12

(p = 0.005)

OSEM + DLE + DLT

2.55 ± 0.13

(p = 0.002)

2.50 ± 0.12

(p = 0.001)

2.85 ± 0.15

(p = 1.00)

ToF-OSEM + DLE + DLT

4.05 ± 0.13

(p = 0.27)

4.10 ± 0.11

(p = 0.05)

4.03 ± 0.10

(p = 0.001)

ToF-BSREM + DLT

4.15 ± 0.07

(p = 0.004)

4.12 ± 0.07

(p = 0.004)

3.40 ± 0.12

(p = 0.001)

ICC

0.41

0.33

0.51