| UW (N = 51) | UTSW (N = 41) | JHU (N = 27) | Stanford (N = 38) | Overall (N = 157) | |
|---|---|---|---|---|---|
| Age | 50 (21.5, 59) | 46 (36, 56) | 41 (18.5, 58.5) | 17 (12.2, 21.8) | 39 (19, 55) |
| Sex - m | 33 (64.7%) | 24 (58.5%) | 14 (51.9%) | 20 (52.6%) | 91 (58%) |
| Sex - f | 18 (35.3%) | 17 (41.5%) | 13 (48.1%) | 18 (47.4%) | 66 (42%) |
| Weight | 75.3 (65.3, 97.3) | 83.6 (61.4, 103) | 66.2 (58.5, 80.5) | 52 (33, 67) | 69 (57.8, 90.1) |
| Height | 172.7 (162.6, 178) | 172.7 (162.3, 182.9) | 165.1 (158.8, 172.5) | 154 (139.6, 165) | 167.6 (157.8, 177.8) |
| Ferriscan | 3.4 (2, 6.6) | 1.3 (0.7, 2.5) | 3 (1.1, 6.8) | 3 (1.5, 8) | 2.6 (1.2, 6.5) |
Multi-Center Multi-Vendor Validation of Liver QSM in Patients with Iron Overload
Subject characteristics
Descriptive statistics for subject-level variables are tabulated below by site and overall. Quantitative variables are summarized by median (inter-quartile range; IQR) and categorical variables by N (%).
Linear regressions of QSM vs R2* and QSM vs LIC
Numbers of observations per site, field strength (FS), and test/retest are tabulated below.
| Site | FS | Retest | n |
|---|---|---|---|
| UW | 1.5 | test | 24 |
| UW | 1.5 | retest | 24 |
| UW | 3.0 | test | 27 |
| UW | 3.0 | retest | 30 |
| UTSW | 1.5 | test | 35 |
| UTSW | 1.5 | retest | 35 |
| UTSW | 3.0 | test | 31 |
| UTSW | 3.0 | retest | 30 |
| JHU | 1.5 | test | 14 |
| JHU | 1.5 | retest | 14 |
| JHU | 3.0 | test | 12 |
| JHU | 3.0 | retest | 14 |
| Stanford | 1.5 | test | 33 |
| Stanford | 3.0 | test | 27 |
QSM vs R2*
Site-specific regression lines of QSM vs R2* for each FS and test/retest scenario are plotted below. It appears that test data show stronger associations than retest data do.
We perform F-tests (with 2 degrees of freedom for intercept and slope) to compare the regression lines between pairs of sites. The p-values are summarized in Table 3a.
| UW vs UTSW | UW vs JHU | UW vs Stanford | UTSW vs JHU | UTSW vs Stanford | JHU vs Stanford | |
|---|---|---|---|---|---|---|
| 1.5T | 0.355 | <0.001 | 0.015 | 0.002 | 0.014 | <0.001 |
| 3.0T | 0.41 | <0.001 | 0.157 | <0.001 | 0.295 | <0.001 |
QSM vs LIC
Site-specific regression lines of QSM vs liver iron concentration (LIC) for each FS and test/retest scenario are plotted below. Test data also show stronger associations than retest data do.
Similarly, we perform F-tests to compare the regression lines between pairs of sites. The p-values are summarized in Table 3b.
| UW vs UTSW | UW vs JHU | UW vs Stanford | UTSW vs JHU | UTSW vs Stanford | JHU vs Stanford | |
|---|---|---|---|---|---|---|
| 1.5T | 0.182 | <0.001 | 0.09 | 0.003 | 0.004 | <0.001 |
| 3.0T | <0.001 | 0.089 | <0.001 | 0.003 | 0.003 | 0.013 |
Repeatability & reproducibility of QSM
Test-retest repeatability
Bland-Altman plots (difference vs mean) for test-retest QSM are plotted by FS and overall below. Data points are color-coded by site.
Why so few data points for 3T? Many subjects do not have full pairs of test-retest data:
# A tibble: 137 × 7
UID Site FS test retest mean diff
<chr> <fct> <chr> <dbl> <dbl> <dbl> <dbl>
1 J007 JHU 3T 1.16 NA NA NA
2 J013 JHU 3T -0.124 NA NA NA
3 J016 JHU 3T 0.946 NA NA NA
4 J017 JHU 3T -0.370 NA NA NA
5 J020 JHU 3T 1.07 NA NA NA
6 J021 JHU 3T -0.101 NA NA NA
7 J026 JHU 3T 0.536 0.524 0.530 -0.0127
8 J029 JHU 3T 0.690 0.822 0.756 0.132
9 J031 JHU 3T -0.339 -0.138 -0.238 0.201
10 J034 JHU 3T 0.0373 0.0251 0.0312 -0.0121
11 J035 JHU 3T -0.0526 -0.0501 -0.0513 0.00244
12 J036 JHU 3T 1.39 1.94 1.66 0.547
13 J039 JHU 3T NA -0.278 NA NA
14 J040 JHU 3T NA 3.16 NA NA
15 J041 JHU 3T NA 0.0605 NA NA
16 J042 JHU 3T NA -0.140 NA NA
17 J043 JHU 3T NA -0.356 NA NA
18 J045 JHU 3T NA -0.344 NA NA
19 J054 JHU 3T NA -0.00491 NA NA
20 J055 JHU 3T NA 1.20 NA NA
21 S004 Stanford 3T -0.320 NA NA NA
22 S005 Stanford 3T -0.174 NA NA NA
23 S007 Stanford 3T -0.403 NA NA NA
24 S008 Stanford 3T 0.829 NA NA NA
25 S010 Stanford 3T 1.40 NA NA NA
26 S011 Stanford 3T -0.131 NA NA NA
27 S012 Stanford 3T 0.206 NA NA NA
28 S013 Stanford 3T -0.207 NA NA NA
29 S015 Stanford 3T 1.39 NA NA NA
30 S017 Stanford 3T 0.137 NA NA NA
31 S018 Stanford 3T -0.313 NA NA NA
32 S019 Stanford 3T -0.276 NA NA NA
33 S020 Stanford 3T 0.448 NA NA NA
34 S022 Stanford 3T 0.450 NA NA NA
35 S023 Stanford 3T 0.715 NA NA NA
36 S024 Stanford 3T 0.0279 NA NA NA
37 S026 Stanford 3T 1.68 NA NA NA
38 S027 Stanford 3T 1.42 NA NA NA
39 S028 Stanford 3T 0.956 NA NA NA
40 S029 Stanford 3T 0.329 NA NA NA
41 S030 Stanford 3T 2.12 NA NA NA
42 S031 Stanford 3T -0.364 NA NA NA
43 S032 Stanford 3T -0.404 NA NA NA
44 S033 Stanford 3T -0.263 NA NA NA
45 S035 Stanford 3T -0.371 NA NA NA
46 S036 Stanford 3T 0.716 NA NA NA
47 S037 Stanford 3T -0.142 NA NA NA
48 T001 UTSW 3T 0.239 NA NA NA
49 T002 UTSW 3T -0.263 NA NA NA
50 T005 UTSW 3T -0.304 NA NA NA
51 T006 UTSW 3T 0.0606 NA NA NA
52 T007 UTSW 3T 0.240 NA NA NA
53 T008 UTSW 3T -0.158 NA NA NA
54 T009 UTSW 3T -0.207 NA NA NA
55 T010 UTSW 3T -0.469 NA NA NA
56 T012 UTSW 3T 0.450 NA NA NA
57 T013 UTSW 3T -0.429 -0.375 -0.402 0.0548
58 T014 UTSW 3T -0.298 NA NA NA
59 T015 UTSW 3T -0.185 NA NA NA
60 T016 UTSW 3T 1.71 1.63 1.67 -0.0793
61 T017 UTSW 3T -0.147 -0.127 -0.137 0.0200
62 T019 UTSW 3T -0.418 -0.422 -0.420 -0.00383
63 T020 UTSW 3T -0.351 -0.173 -0.262 0.178
64 T021 UTSW 3T -0.260 -0.292 -0.276 -0.0319
65 T023 UTSW 3T -0.296 -0.201 -0.249 0.0953
66 T024 UTSW 3T 0.230 0.236 0.233 0.00636
67 T025 UTSW 3T -0.209 -0.200 -0.205 0.00898
68 T026 UTSW 3T -0.202 -0.221 -0.212 -0.0188
69 T027 UTSW 3T 0.137 -0.0791 0.0287 -0.216
70 T028 UTSW 3T -0.0669 -0.0939 -0.0804 -0.0270
71 T029 UTSW 3T -0.341 -0.370 -0.356 -0.0282
72 T030 UTSW 3T -0.115 -0.104 -0.109 0.0103
73 T031 UTSW 3T -0.173 -0.143 -0.158 0.0298
74 T032 UTSW 3T -0.0813 -0.0872 -0.0843 -0.00591
75 T033 UTSW 3T -0.148 -0.158 -0.153 -0.00940
76 T034 UTSW 3T -0.201 -0.190 -0.196 0.0106
77 T035 UTSW 3T 0.00189 -0.0365 -0.0173 -0.0384
78 T036 UTSW 3T -0.0922 -0.106 -0.0990 -0.0135
79 T038 UTSW 3T NA 0.0763 NA NA
80 T039 UTSW 3T NA -0.419 NA NA
81 T040 UTSW 3T NA -0.321 NA NA
82 T041 UTSW 3T NA -0.146 NA NA
83 T042 UTSW 3T NA 1.21 NA NA
84 T043 UTSW 3T NA 1.36 NA NA
85 T045 UTSW 3T NA -0.185 NA NA
86 T046 UTSW 3T NA -0.101 NA NA
87 T048 UTSW 3T NA -0.219 NA NA
88 T049 UTSW 3T NA 0.0357 NA NA
89 W001 UW 3T 0.350 NA NA NA
90 W003 UW 3T 0.118 NA NA NA
91 W004 UW 3T -0.213 NA NA NA
92 W005 UW 3T 0.462 NA NA NA
93 W007 UW 3T -0.252 NA NA NA
94 W009 UW 3T -0.374 NA NA NA
95 W010 UW 3T -0.0729 NA NA NA
96 W011 UW 3T 1.06 NA NA NA
97 W012 UW 3T -0.277 NA NA NA
98 W013 UW 3T 0.233 NA NA NA
99 W014 UW 3T 0.675 NA NA NA
100 W015 UW 3T -0.329 NA NA NA
101 W016 UW 3T 0.337 NA NA NA
102 W018 UW 3T -0.170 NA NA NA
103 W019 UW 3T 1.31 NA NA NA
104 W021 UW 3T 0.753 NA NA NA
105 W023 UW 3T 0.703 NA NA NA
106 W024 UW 3T 1.09 NA NA NA
107 W025 UW 3T 0.495 0.339 0.417 -0.156
108 W026 UW 3T 0.362 0.317 0.340 -0.0452
109 W027 UW 3T -0.291 -0.257 -0.274 0.0337
110 W029 UW 3T 0.0145 0.281 0.148 0.266
111 W030 UW 3T -0.107 -0.0512 -0.0793 0.0563
112 W032 UW 3T 0.0155 -0.000904 0.00732 -0.0165
113 W033 UW 3T -0.148 0.165 0.00832 0.313
114 W034 UW 3T -0.107 -0.202 -0.154 -0.0946
115 W035 UW 3T 0.619 NA NA NA
116 W038 UW 3T NA 0.671 NA NA
117 W039 UW 3T NA -0.474 NA NA
118 W040 UW 3T NA 0.729 NA NA
119 W041 UW 3T NA 0.957 NA NA
120 W042 UW 3T NA -0.00725 NA NA
121 W043 UW 3T NA 0.0606 NA NA
122 W044 UW 3T NA -0.0710 NA NA
123 W045 UW 3T NA -0.187 NA NA
124 W046 UW 3T NA -0.246 NA NA
125 W047 UW 3T NA -0.0294 NA NA
126 W048 UW 3T NA -0.184 NA NA
127 W049 UW 3T NA 0.178 NA NA
128 W050 UW 3T NA 0.491 NA NA
129 W051 UW 3T NA 0.185 NA NA
130 W052 UW 3T NA -0.0851 NA NA
131 W053 UW 3T NA -0.208 NA NA
132 W054 UW 3T NA -0.170 NA NA
133 W055 UW 3T NA 0.271 NA NA
134 W056 UW 3T NA -0.0747 NA NA
135 W057 UW 3T NA 0.692 NA NA
136 W058 UW 3T NA 0.318 NA NA
137 W059 UW 3T NA 0.194 NA NA The bias (mean test-retest difference), repeatability coefficient (RC; range covering 95% test-retest differences), intraclass correlation coefficient (ICC) with 95% confidence interval (CI) and p-value (for testing ICC = 0) are presented in Table 4 below.
| FS | Bias | RC | ICC | P |
|---|---|---|---|---|
| 1.5T | -0.076 | 1.024 | 0.693 (0.552, 0.795) | <0.001 |
| 3T | 0.034 | 0.271 | 0.959 (0.92, 0.979) | <0.001 |
| Overall | -0.041 | 0.863 | 0.746 (0.648, 0.819) | <0.001 |
Field strength reproducibility
The Bland-Altman plot for reproducibility between 1.5T and 3T is shown below, with bias, ICC (95% CI), and p-value indicated on the figure.
Sex differences in LIC
LIC values are compared between females and males by site and overall in Table 5 below. The p-values are based on the Wilcoxon rank sum test.
| Site | Female | Male | P |
|---|---|---|---|
| UW | 3.6 (2.2, 6.3) | 3.4 (1.8, 7.5) | 0.805 |
| UTSW | 2.4 (1.4, 5.8) | 0.9 (0.7, 1.6) | 0.019 |
| JHU | 3.8 (1.2, 7.6) | 2.7 (1.1, 3.7) | 0.409 |
| Stanford | 6.7 (2.4, 9.9) | 2.5 (1.2, 3.9) | 0.023 |
| Overall | 3.8 (1.8, 7.5) | 2.1 (1, 4) | 0.009 |
The following boxplot visualizes the comparisons.
Sex differences in QSM vs (R2*, LIC)
Sex-specific regression lines of QSM vs R2* for each FS and test/retest scenario are plotted below, with F-test p-values comparing the two lines indicated on the plots.
Similarly, sex-specific regression lines of QSM vs LIC for each FS and test/retest scenario are plotted below, with F-test p-values comparing the two lines indicated on the plots.
