Table 1 Demographic and Clinical Characteristics of People Screened with Chest X-ray and Xpert MTB/RIF.

From: Using artificial intelligence to read chest radiographs for tuberculosis detection: A multi-site evaluation of the diagnostic accuracy of three deep learning systems

 

Overall

Nepal

Cameroon

Xpert Positive

Xpert Negative

Total

Xpert Positive

Xpert Negative

Total

Xpert Positive

Xpert Negative

Total

N

109

1087

1196

94

421

515

15

666

681

Age (median [IQR])

36 [25, 47]

48 [31, 61]

46 [30, 61]

38 [25, 48.75]

47 [29, 62]

45 [28, 61]

33 [25.50, 38]

48 [32, 61]

47 [32, 61]

Sex = F/M (%)

32/77 (29.4/70.6)

614/473 (56.5/43.5)

646/550 (54.0/46.0)

27/67 (28.7/71.3)

200/221 (47.5/52.5)

227/288 (44.1/55.9)

5/10 (33.3/66.7)

414/252 (62.2/37.8)

419/262 (61.5/38.5)

Cough (%)

105 (96.3)

632 (58.1)

737 (61.6)

93 (98.9)

368 (87.4)

461 (89.5)

12 (80)

264 (39.6)

276 (40.5)

Fever (%)

76 (69.7)

444 (40.8)

520 (43.5)

68 (72.3)

218 (51.8)

286 (55.5)

8 (53.3)

226 (33.9)

234 (34.4)

Weight loss (%)

74 (67.9)

550 (50.6)

624 (52.2)

62 (66)

150 (35.6)

212 (41.2)

12 (80)

400 (60.1)

412 (60.5)

Night Sweats (%)

37 (33.9)

269 (24.7)

306 (25.6)

29 (30.9)

83 (19.7)

112 (21.7)

8 (53.3)

186 (27.9)

194 (28.5)

TB treatment history (%)

Yes

89 (81.7)

982 (90.3)

1071 (89.5)

20 (21.3)

76 (18.1)

96 (18.6)

0 (0)

22 (3.3)

22 (3.2)

No

0 (0.0)

7 (0.6)

7 (0.6)

74 (78.7)

345 (81.9)

419 (81.4)

15 (100)

637 (95.6)

652 (95.7)

Unknown

20 (18.3)

98 (9.0)

118 (9.9)

   

0 (0)

7 (1.1)

7 (1)

HIV status (%)

Negative

17 (15.6)

254 (23.4)

271 (22.7)

9 (9.6)

50 (11.9)

59 (11.5)

8 (53.3)

204 (30.6)

212 (31.1)

Positive

10 (9.2)

28 (2.6)

38 (3.2)

4 (4.3)

3 (0.7)

7 (1.4)

6 (40)

25 (3.8)

31 (4.6)

Unknown

82 (75.2)

805 (74.1)

887 (74.2)

81 (86.2)

368 (87.4)

449 (87.2)

1 (6.7)

437 (65.6)

438 (64.3)

AFB positive (%)

76 (70.4)

2 (0.2)

78 (6.6)

67 (71.3)

1 (0.2)

68 (13.2)

9 (64.3)

1 (0.2)

10 (1.5)

Xpert positive (%)

109 (100.0)

0 (0.0)

109 (9.1)

94 (100)

0 (0)

94 (18.3)

15 (100)

0 (0)

15 (2.2)

CAD4TB (median [IQR])

86 [75, 99]

45 [25, 54]

46 [28, 59]

86 [77, 99]

48 [25, 69]

54 [40.5, 79.5]

77 [58, 99]

44 [25, 49]

44 [25, 50]

Lunit (median [IQR])

0.99 [0.97, 0.99]

0.08 [0.01, 0.51]

0.10 [0.01, 0.80]

0.99 [0.97, 0.99]

0.37 [0.05, 0.90]

0.69 [0.09, 0.96]

0.97 [0.90, 0.98]

0.03 [0.01, 0.15]

0.03 [0.01, 0.18]

qXR (median [IQR])

0.90 [0.82, 0.94]

0.17 [0.10, 0.36]

0.19 [0.11, 0.50]

0.92 [0.84, 0.94]

0.30 [0.12, 0.65]

0.44 [0.14, 0.82]

0.78 [0.64, 0.86]

0.14 [0.09, 0.23]

0.15 [0.10, 0.24]

Senior Radiologist (Nepal) = Normal CXR (%)

   

4 (4.3)

203 (48.2)

207 (40.2)

   

Junior Radiologist & Residents (Nepal) = Normal CXR (%)

   

12 (12.8)

290 (68.9)

302 (58.6)

   

Radiologist (Cameroon) = Normal CXR (%)

      

3 (20)

495 (74.3)

498 (73.1)

Teleradiology Company (Cameroon) = Normal CXR (%)

      

3 (20)

494 (74.2)

497 (73.0)