Skip to main content

Table 1 Characteristics of the study populations

From: Retinal photograph-based deep learning system for detection of hyperthyroidism: a multicenter, diagnostic study

 

20 medical centers

BTH

BFH

CZMC

SYMC

SZMC

WJMC

Development dataset

Internal validation dataset

External validation dataset 1

External validation dataset 2

External validation dataset 3

External validation dataset 4

External validation dataset 5

External validation dataset 6

Model 1

Model 2

Model 1

Model 2

Participants

5080

5080

717

717

227

191

363

363

363

363

 Hyperthyroidism

1804

1804

251

251

105

101

116

128

133

129

 Non-hyperthyroidism

3276

3276

466

466

122

90

247

235

230

234

Age, years

46 (20–88)

40 (15–85)

45 (19–88)

41 (16–88)

50 (40–60)

43 (34–56)

47 (24–88)

46 (20–73)

47 (24–88)

46 (24–82)

Gender

 Male

2072

1938

344

327

114

73

188

166

168

150

 Female

3008

3142

373

390

113

118

178

197

195

213

Body Mass Index, kg/m2

22.8 (17.0–28.2)

23.6 (17.2–28.9)

23.3 (16.8–28.9)

23.6 (16.9–28.1)

23.2 (16.8–27.2)

22.5 (16.2–28.0)

23.3 (16.4–28.7)

23.3 (17.2–28.5)

23.9 (15.9–27.8)

23.2 (16.5–28.3)

Systolic blood pressure, mm Hg

130 (88–175)

128 (86–173)

129 (88–173)

128 (83–175)

126 (86–173)

124 (83–175)

130 (87.4–171)

129 (88–179)

130 (86–173)

130 (88–175)

Diastolic blood pressure, mm Hg

78 (49–107)

78 (47–106)

77 (47–105)

78 (49–106)

80 (48–105)

75

(47–106)

78 (50–102)

77 (49–106)

78 (48–106)

79 (48–105)

Retinal photographs

10,160

10,160

1434

1434

592

366

726

726

726

726

 Hyperthyroidism

3608

3608

502

502

296

183

232

256

266

258

 Non-hyperthyroidism

6552

6552

932

932

296

183

494

470

460

468

  1. Model 1: the control group were randomly selected and unmatched to the hyperthyroidism subjects; Model 2: age and gender of control group were matched to the hyperthyroidism subjects. Hyperthyroidism participants were overlapped in Model 1 and Model 2, whereas non-hyperthyroidism participants were different between Model 1 and Model 2. Data are presented as n, or median values (range)
  2. BTH Beijing tongren hospital, BFH Beijing friendship hospital, CZMC Chongqing Zhuoyue medical centre of ikang healthcare group, SYMC Shanghai Yuanhua medical centre of ikang guobin healthcare group, SZMC Shenzhen Zhuoyue Medical Centre of iKang Healthcare Group, WJMC Wuhan Jindun road medical centre of iKang healthcare group