SUVmax
Overall, the segmented brain VOIs showed higher average SUVmax values with TOF reconstruction than with non-TOF reconstruction (Fig. 2-A), with all VOIs showing a difference that was statistically significant at p < 0.05 (Table 2). The Highest %RAD was observed in CN (p < 0.0001; TOF-SUVmax: 10.54 ± 2.46, non-TOF-SUVmax: 9.68 ± 2.14; %RAD-SUVmax: 8.96%), and the lowest in SN (p = 0.003; TOF-SUVmax: 7.56 ± 1.72, non-TOF-SUVmax: 7.34 ± 1.58; %RAD-SUVmax: 3.06%), as shown in Table 2 and Fig. 2-A, B.
Table 2
Average SUVmax and SUVmean values in 18F-FDG (mean ± SD)
Brain VOIs
|
Mean ± SD
|
SUVmax
|
SUVmean
|
TOF
|
non-TOF
|
RAD%
|
p-value
|
TOF
|
non-TOF
|
RAD%
|
p-value
|
Frontal Lobe
|
10.27 ± 2.15
|
9.80 ± 2.01
|
4.75
|
< 0.0001
|
7.10 ± 1.44
|
6.82 ± 1.35
|
4.19
|
< 0.0001
|
Temporal Lobe
|
8.92 ± 1.84
|
8.38 ± 1.67
|
6.4
|
< 0.0001
|
6.12 ± 1.21
|
5.77 ± 1.1
|
5.96
|
< 0.0001
|
Parietal Lobe
|
12.33 ± 2.75
|
11.69 ± 2.51
|
5.41
|
< 0.0001
|
7.75 ± 1.64
|
7.41 ± 1.53
|
4.66
|
< 0.0001
|
Occipital Lobe
|
13.20 ± 2.95
|
12.76 ± 2.71
|
3.42
|
0.002
|
8.66 ± 1.79
|
8.36 ± 1.69
|
3.57
|
< 0.0001
|
Caudate Nuclei
|
10.54 ± 2.46
|
9.68 ± 2.14
|
8.97
|
< 0.0001
|
5.53 ± 1.91
|
5.21 ± 1.69
|
6.18
|
0.002
|
Nucleus Accumbens
|
8.96 ± 1.85
|
8.68 ± 1.75
|
3.22
|
0.0004
|
6.84 ± 1.43
|
6.71 ± 1.39
|
1.97
|
0.007
|
Putamen
|
12.33 ± 2.79
|
11.38 ± 2.52
|
8.54
|
< 0.0001
|
8.75 ± 1.93
|
8.28 ± 1.79
|
5.72
|
< 0.0001
|
Thalamus
|
11.80 ± 2.58
|
11.20 ± 2.46
|
5.42
|
< 0.0001
|
6.74 ± 1.80
|
6.48 ± 1.71
|
4.04
|
0.0003
|
Palladium
|
9.95 ± 2.15
|
9.45 ± 2.04
|
5.33
|
0.0001
|
6.22 ± 1.22
|
6.21 ± 1.23
|
0.13
|
0.848
|
Corpus callosum
|
9.29 ± 2.80
|
8.73 ± 2.50
|
6.44
|
0.008
|
3.15 ± 0.72
|
3.34 ± 0.76
|
-5.55
|
0.0002
|
Substantia Nigra
|
7.56 ± 1.72
|
7.34 ± 1.58
|
3.07
|
0.003
|
5.48 ± 1.25
|
5.58 ± 1.19
|
-1.68
|
0.058
|
Insulla
|
10.94 ± 2.49
|
10.27 ± 2.21
|
6.54
|
0.000
|
6.73 ± 1.38
|
6.49 ± 1.33
|
3.68
|
< 0.0001
|
Anterior cingulate cortex
|
10.52 ± 2.11
|
9.86 ± 1.9
|
6.65
|
< 0.0001
|
7.21 ± 1.36
|
6.9 ± 1.28
|
4.48
|
< 0.0001
|
Posterior cingulate cortex
|
12.36 ± 2.97
|
11.48 ± 2.67
|
7.61
|
< 0.0001
|
8.71 ± 2.11
|
8.32 ± 1.98
|
4.71
|
< 0.0001
|
Cerebellum Cortex
|
10.96 ± 2.49
|
10.29 ± 2.22
|
6.61
|
< 0.0001
|
6.44 ± 1.32
|
6.31 ± 1.26
|
2.1
|
< 0.0001
|
Brainstem
|
10.11 ± 2.16
|
9.29 ± 1.94
|
8.83
|
< 0.0001
|
4.96 ± 0.97
|
4.88 ± 0.91
|
1.65
|
0.002
|
White matter
|
13.53 ± 2.71
|
13.12 ± 2.47
|
3.09
|
0.003
|
5.54 ± 1.00
|
5.57 ± 1.02
|
-0.7
|
0.164
|
Cerebellum Medulla
|
10.02 ± 2.59
|
9.63 ± 2.52
|
4.02
|
0.01
|
5.61 ± 0.95
|
5.59 ± 0.93
|
0.31
|
0.126
|
* Difference if significant at the level of p < 0.05 |
SUV
mean
Higher average SUVmean was observed with TOF reconstruction than with non-TOF reconstruction in all segmented brain VOIs except the SN, CoC, and WM regions, as illustrated in Fig. 3A, B. Statistically significant differences (p < 0.05) were seen among all the segmented VOIs except the PA, SN, WM, and CM (Table 2). The highest %RAD was observed in CN (p = 0.002; TOF-SUVmean: 5.53 ± 1.91, non-TOF-SUVmean: 5.21 ± 1.69; %RAD-SUVmean: 6.18%), and the lowest in PA (p = 0.848; TOF-SUVmean: 6.22 ± 1.22, non-TOF-SUVmean: 6.21 ± 1.23; %RAD-SUVmean: 0.13%; Table 2). Higher SUVmean enhancement was noted in the non-TOF reconstructions in the CoC (p < 0.001; TOF-SUVmean: 3.15 ± 0.72, non-TOF-SUVmean: 3.34 ± 0.76), SN (p = 0.058; TOF-SUVmean: 5.48 ± 1.25, non-TOF-SUVmean: 5.58 ± 1.19) and WM (p = 0.164; TOF-SUVmean: 5.54 ± 1.00, non-TOF-SUVmean: 5.57 ± 1.02) VOIs (Table 2 and Fig. 3-A, B).
Percentage of relative average difference of SUVmax and SUVmean (%RAD)
The %RAD values (TOF vs. non-TOF) of SUVmax and SUVmean were calculated for the segmented brain VOIs according to Eq. 3. The %RAD-SUVmax differences for all segmented brain VOIs were positive, indicating that TOF reconstruction increased SUVmax compared with non-TOF reconstruction (Table 2-B). The %RAD-SUVmean difference was positive for all segmented VOIs except the following regions: CoC (− 5.55%), SN (− 1.68%), and WM (− 0.7%) (Fig. 3-B).
SNR and Contrast
The SNR gain was measured on TOF and non-TOF reconstructions according to Eq. 1. Higher average SNR was observed with TOF than with non-TOF reconstruction in 9 of the 17 regions, as illustrated in Fig. 4-A. However, significant improvement in SNR for TOF reconstruction was identified only in TL, CN, and PU. In contrast, 8/17 regions showed higher SNR with non-TOF reconstruction, with the difference being significant for NA (p = 0.008; TOF-SNR:1.221 ± 0.36, non-TOF SNR: 1.290 ± 0.38), PA (p < 0.0001; TOF-SNR: 0.830 ± 0.34, non-TOF-SNR: 0.949 ± 0.35), CoC (p = 0.003; TOF-SNR: −1.214 ± 0.33, non-TOF-SNR: −1.125 ± 0.35), SN (p < 0.0001; TOF-SNR: 0.342 ± 0.35, non-TOF-SNR: 0.492 ± 0.37), CC (p = 0.005; TOF-SNR: 1.014 ± 0.33, non-TOF-SNR: 1.057 ± 0.34), WM (p < 0.0001; TOF-SNR: 0.383 ± 0.28, non-TOF-SNR: 0.499 ± 0.29), and CM (p < 0.0001; TOF-SNR: 0.383 ± 0.22, non-TOF-SNR: 0.457 ± 0.21) (Table 3). The corpus callosum showed less signal than the background VOI (Fig. 4-A).
Table 3
Average Contrast and SNR values in 18F-FDG (mean ± SD)
Brain VOI
|
Mean ± SD
|
Contrast
|
SNR
|
TOF
|
Non-TOF
|
p-value
|
TOF
|
Non-TOF
|
p-value
|
Frontal Lobe
|
1.437 ± 0.14
|
1.400 ± 0.13
|
0.0593
|
1.420 ± 0.40
|
1.389 ± 0.38
|
0.258
|
Temporal Lobe
|
1.238 ± 0.12
|
1.187 ± 0.10
|
0.0001
|
0.763 ± 0.31
|
0.640 ± 0.27
|
0.000
|
Parietal Lobe
|
1.570 ± 0.19
|
1.523 ± 0.18
|
0.0000
|
1.858 ± 0.51
|
1.821 ± 0.51
|
0.061
|
Occipital Lobe
|
1.754 ± 0.20
|
1.719 ± 0.19
|
0.0000
|
2.481 ± 0.55
|
2.524 ± 0.56
|
0.084
|
Caudate Nuclei
|
1.102 ± 0.30
|
1.057 ± 0.26
|
0.0000
|
0.329 ± 0.93
|
0.181 ± 0.88
|
0.012
|
Nucleus Accumbens
|
1.379 ± 0.14
|
1.373 ± 0.14
|
0.0128
|
1.221 ± 0.36
|
1.290 ± 0.38
|
0.008
|
Putamen
|
1.764 ± 0.19
|
1.694 ± 0.17
|
0.4110
|
2.490 ± 0.45
|
2.409 ± 0.44
|
0.061
|
Thalamus
|
1.351 ± 0.23
|
1.320 ± 0.22
|
0.0000
|
1.141 ± 0.65
|
1.102 ± 0.67
|
0.264
|
Pallidum
|
1.261 ± 0.13
|
1.277 ± 0.12
|
0.0036
|
0.830 ± 0.34
|
0.949 ± 0.35
|
0.000
|
Corpus Callosum
|
0.638 ± 0.09
|
0.685 ± 0.10
|
0.0410
|
-1.214 ± 0.33
|
-1.125 ± 0.35
|
0.003
|
Substantia Nigra
|
1.108 ± 0.13
|
1.145 ± 0.13
|
0.0000
|
0.342 ± 0.35
|
0.492 ± 0.37
|
0.000
|
Insulla
|
1.359 ± 0.12
|
1.330 ± 0.11
|
0.0000
|
1.166 ± 0.33
|
1.144 ± 0.32
|
0.356
|
Anterior Cingulate Cortex
|
1.462 ± 0.13
|
1.421 ± 0.12
|
0.0004
|
1.510 ± 0.38
|
1.465 ± 0.37
|
0.089
|
Posterior Cingulate Cortex
|
1.761 ± 0.29
|
1.707 ± 0.27
|
0.0000
|
2.477 ± 0.80
|
2.455 ± 0.80
|
0.626
|
CC
|
1.301 ± 0.11
|
1.294 ± 0.10
|
0.0002
|
1.014 ± 0.33
|
1.057 ± 0.34
|
0.005
|
White matter
|
1.124 ± 0.11
|
1.149 ± 0.11
|
0.0933
|
0.383 ± 0.28
|
0.499 ± 0.29
|
0.000
|
Cerebellar Medulla
|
1.109 ± 0.08
|
1.120 ± 0.09
|
0.0000
|
0.383 ± 0.22
|
0.457 ± 0.21
|
0.000
|
* Difference if significant at the level of p < 0.05 |
Image contrast in all brain VOIs was calculated with TOF and non-TOF reconstruction using
Eq. 2. TOF reconstruction led to higher average contrast in 12 brain regions compared with non-TOF reconstruction. A significant difference in contrast was observed in TL (
p < 0.001; TOF-Contrat:1.238 ± 0.12, non-TOF-Contrast:1.187 ± 0.10), PL (
p < 0.0001; TOF-Contrast:1.570 ± 0.19, non-TOF-Contrast: 1.523 ± 0.18), OL (
p < 0.0001; TOF-Contrast: 1.754 ± 0.20, non-TOF-Contrast: 1.719 ± 0.19), CN (
p < 0.0001; TOF-Contrast:1.102 ± 0.30, non-TOF-Contrast: 1.057 ± 0.26), TH (
p < 0.0001; TOF-Contrast:1.351 ± 0.23, non-TOF-Contrast: 1.320 ± 0.22), IN (
p < 0.0001; TOF-Contrast: 1.359 ± 0.12, non-TOF-Contrast: 1.330 ± 0.11), ACC (
p < 0.0001; TOF-Contrast: 1.462 ± 0.13, non-TOF-Contrast: 1.421 ± 0.12), PCC (
p < 0.001; 1.761 ± 0.29, 1.707 ± 0.27), and CC (
p < 0.001; TOF-Contrast: 1.301 ± 0.11,non-TOF-Contrast: 1.294 ± 0.10) (Fig.
4-B). However, PA, CoC, SN, WM, and CM showed higher average contrast with non-TOF reconstruction, with the difference being significant for PA (
p = 0.0036; TOF-Contrast:1.261 ± 0.13, non-TOF-Contrast: 1.277 ± 0.12), CoC (
p < 0.041; TOF-Contrast: 0.638 ± 0.09, non-TOF-Contrast: 0.685 ± 0.10), and SN (
p < 0.0001; TOF-Contrast:1.108 ± 0.13, non-TOF-Contrast: 1.145 ± 0.13)
(Table
3).