Patient’s population 177 unique patient exams on dPET/CT: 52 with 3MBq (0,08mCi) protocol 1 and 125 with 4MBq (0,11mCi) protocol 2, and 74 patients on aPET/CT were analyzed.
[18F]FDG PET/CT indication was in 80 to 82 percent of patients oncological (initial or follow-up exam in proven malignancy), versus diagnostic (benign versus malignant pathology) or miscellaneous (inflammatory or infectious pathology).
Table 1:Patient and PET/CT characteristics
dPET/CT
|
aPET/CT
|
Protocole 1 Protocole 2
N=52 N=125
|
N=74
|
Sex (female %)
|
58%
|
60%
|
47%
|
Age (Y) mean±SD
[range]
|
62,3±14,7
[23-89]
|
60,9±13,6
[24-89]
|
62,8±13,0
[22-81]
|
Weight (kg) mean±SD
[range]
|
78±19
[41-127]
|
75±16
[35-110]
|
80±23
[38-130]
|
Height (m) mean±SD
[range]
|
1,67± 0,10
[1,46-1,83]
|
1,66 ±0,10
[1,51-1,85]
|
1,67±0,09
[1,48-1,84]
|
BMI (kg/m2) mean±SD
[range]
|
28±8
[15-44]
|
27±8
[15-42]
|
28±8
[14-52]
|
Fat (kg) mean±SD
[range]
|
27±12
[6-63]
|
25±13
[5-60]
|
27±14
[3-74]
|
Glycaemia (g/l) mean±SD
[range]
|
1,02 ± 0,18[0,76-1,5] |
1,01 ± 0,13
[0,70-1,38]
|
0,98 ± 0,11
[0,72-1,25]
|
Scan Delay (min)
mean±SD
[range]
|
58,5 ± 2,7
[55-65]
|
58,3 ±3,0
|
58,0 ± 3,0
idem
|
Injected Activity/kg
[18F]FDG mean±SD
range in %/patient)
|
3,00,1MBq/kg
0,080,003mCi (8%)
|
4,00,2 MBq/kg
0,110,015mCi
(8%)
|
4,00,2 MBq/kg
0,110,015mCi
(8%)
|
Bedposition scan duration
|
120sec
|
60sec
|
160sec BMI<25kg/m2
220sec BMI25kg/m2
|
Reconstruction
protocol
|
3D OSEM + PSF
2i10s
2mm voxel size
|
3D OSEM + PSF; 4i4s
2mm voxel size
|
3D OSEM + PSF
3i21s
4mm voxel size
|
without significant differences between groups in age, sex, weight, height, BMI, glycaemia, scan delay.
Image Quality Analysis
1.Semi- Quantitative Analysis
On dPET/CT (with both imaging protocols) and aPET/CT, CVliv was associated with weight, BMI and fatty mass in univariable and multivariable linear regression analyses (p0.0001 for dPET/CT and p0.009 for aPET/CT). There was also a significant, more moderate association
between sex and CVliv with higher CVliv in men on both camera’s, only for dPET/CT 4MBq (0,11mCi) protocol 2 in multivariable weight models (p=0.03), and except for dPET/CT 3MBq (0,08mCi) protocol in multivariable BMI models (p0.01).
Age, pathological exam, initial exam were not associated with CVliv.
The R2-fit association of CVliv was best described by a linear model and reached a good and same univariable and multivariable fit in weight model on dPET/CT 3MBq (0,08mCi) protocol 1; R2 =0.62 (versus multivariable R2=0,40 for 4MBq (0,11mCi) protocol 2 and R2=0,26 on aPET/CT). For CVliv versus fatty mass no differences in linear fitting were found. However not significantly different (p0,54), in both protocols on dPET/CT slightly less well linear fitting (-10%) was obtained between CVliv and BMI (versus weight) for dPET/CT (for dPET/CT protocol 1:R2 =0,54 versus 0,36 in protocol 2 and 0,26 on aPET/CT).
Figure1. Graph formula (univariable linear regression) was the following: CVliv = 6,2.10-4 ( 7,2.10-5 )* weight + 7,6.10-2 (5,5. 10-3). In this group no additional sex effect was found.
Better results with lower mean CVliv values were obtained with 4MBq (0,11mCi) protocol 2 versus 3MBq (0,08mCi) protocol 1 on dPET/CT. However slopes, curve coefficients of weight in the linear formula predicting CVliv were not significantly different (p=0,72).
On dPET/CT compared to weight category < 70kg as a reference, patient category 70 <90kg had on
average a significantly higher CVliv (+7%, range [-9% to 24%]) for 3MBq (0,08mCi) protocol 1 and +9% range [-13% to 36%]) for 4MBq (0,11mCi) protocol 2; p0,02). More pronounced relative CVliv increase was observed for patients 90 kg (+22% [-3%-50%] for 3MBq (0,08mCi) protocol 1 + 26% [2%-56%] for 4MBq (0,11mCi) protocol 2).
Table 2: Mean CVliv in weight categories
Mean
CVliv
(SD)
range
|
dPET/CT
|
aPET/CT
|
<70kg
|
70-89kg
|
90kg
|
<70kg
|
70-89kg
|
90kg
|
P1 3MBq/kg
|
P2
4MBq/kg
|
P1
|
P2
|
P1
|
P2
|
P3
|
N=19
|
N=48
|
N=17
|
N=53
|
N=16
|
N=24
|
N=30
|
N=17
|
N=27
|
0,122
0,008
0,104-
0,138
|
0,113
0,013
0,084-0,147
|
0,130
0,011
0,116-
0,147
|
0,123
0,013
0,089-
0,161
|
0,149
0,022
0,126-0,187
|
0,142
0,014
0,111-0,178
|
0,121
0,014
0,099 -0,155
|
0,124
0,015
0,098 -0,154
|
0,133
0,010
0,114-
0,154
|
With significant lower CVliv values on dPET/CT with 4MBq protocol 2 versus 3MBq protocol 1 in each weight category (p=0,01 <70kg and p=0,049 in 70-89kg) except for 90kg (p=0,2)
Between 4MBq (0,11mCi) protocol 2 on dPET/CT and aPET/CT lower CVliv in patients <70kg on dPET/CT and no significantly different CVliv in intermediate weight category and Pearson’s correlation coefficients (R) were found (p≥0,26), despite longer mean scanning time and per bed position for aPET/CT. However curve slope of CVliv versus weight was significantly steeper for dPET/CT (p=0,00005), although flattened and not comparable on aPET/CT in particular by its BMI adaptive protocol.
2.Visual IQ analysis
Table 3: Visual IQG per reader, camera, and weight category
|
dPET/CT
protocol 1
N=52
|
aPET/CT
N=74
|
ALL
N=126
|
Reader 1
(mean score ±SD)
|
3,39/5 ±0,63
|
3,53/5 ±0,57
|
3,48/5 ±0,60
|
Reader 2
(mean score ±SD)
|
3,61/5 ±0,67
|
3,81/5 ±0,51
|
3,73/5 ±0,58
|
Total mean score ±SD
§ <70kg
§ 70 <90kg
§ 90 kg
|
3,50/5 ±0,66
3,90 ± 0,53 (N=21)3,30 ± 0,60 (N=15)3,13 ± 0,57 (N=16) |
3,67/5 ±0,56
3,95 ± 0,44 (N=29)3,69 ± 0,52 (N=21)3,38 ± 0,56 (N=24) |
3,60/5 ±0,60
|
Lower scores if 70kg versus <70kg for dPET/CT (p0,003) and between each higher weight category for aPET/CT (p0,008). Higher overall scores and in categories 70 kg on aPET/CT versus dPET/CT (p0,03).
Significant higher scores by reader 2 (all patients summed; p=0,0009).
Interreader Cohen’s kappa was only moderate between both readers, identical on both camera’s (=0,4).
For both readers there was a significant and similar negative relation of visual image quality score and BMI and weight (p0001) on both camera’s.
Table 4: Relation between visual IQ, weight and BMI
Visual
IQ
|
dPET/CT protocol 1
|
aPET/CT
|
Weight
Reader 1 Reader 2
|
BMI
Reader 1 Reader 2
|
Weight
Reader 1 Reader 2
|
BMI
Reader 1 Reader 2
-0,50 -0,49
|
Spearman rho
|
-0,60 -0,63
|
-0,66 -0,68
|
-0,47 -0,45
|
Figure 2. Graph 2
Lower global IQ was found in women on dPET/CT only (3,41/5 versus 3,60/5; p=0,04), without sex difference in hepatic homogeneity scores.
For both readers there was a significant and similar negative relation of global visual IQ and CVliv for both camera’s (spearman rho = -0,4; p0,05), higher when scoring only visual hepatic homogeneity in a pair-wise comparison (= -0,65; p=0,001).
Visual analysis was not performed for 4MBq (0,11mCi)/kg protocol 2 on dPET/CT as semi-quantitative data were comparable in both protocols.
3.Case examples (figure 3)