Skip to main content

Table 2 Univariate analyses for prediction of missed diagnosis HSIL+ in patients with LSIL diagnosed by colposcopy biopsy in the training set

From: Assessing the risk of high-grade squamous intraepithelial lesions (HSIL+) in women with LSIL biopsies: a machine learning-based study

Parameter

Numbers

OR (95%CI)

p-values

Total

Missed diagnosis

(%)

Age (years)

  

0.654 (0.370–1.156)

0.143

 < 50

211

29 (13.7%)

  

 ≥ 50

143

28 (19.6%)

  

Menopause

  

0.883 (0.490–1.592)

0.68

 No

232

36 (15.5%)

  

 Yes

122

21 (17.2%)

  

HPV-16/18 infection

  

2.813 (1.569–5.042)

0

 No

255

30 (11.8%)

  

 Yes

99

27 (27.3%)

  

Multiple HR-HPV infection

  

1.992 (1.118–3.549)

0.018

 No

240

31 (12.9%)

  

 Yes

114

26 (22.8%)

  

TZ types

  

2.368 (1.305–4.296)

0.004

 TZ1/2

265

34 (13.3%)

  

 TZ3

89

23 (26.1%)

  

Lesion range

  

2.369 (1.285–4.366)

0.005

 < 3

166

17 (10.2%)

  

 ≥ 3

188

40 (21.3%)

  

TCT ≥ ASC-H

  

11.200 (4.762–26.340)

 < 0.001

 NO

328

41 (12.5%)

  

 Yes

26

16 (61.5%)

  

Colposcopic impression

  

9.743 (4.765–19.921)

 < 0.001

 Normal/G1

314

35 (11.2%)

  

 G2

40

22 (55%)

  

Immune disease

  

0.806 (0.367–1.772)

0.591

 No

306

48 (15.7%)

  

 Yes

48

9 (18.8%)

  

Smoking

  

0.571 (0.232–1.409)

0.219

 No

325

50 (15.4%)

  

 Yes

29

7 (24.1%)

  

Gravidity

  

1.185 (0.657–2.137)

0.573

 < 3

119

21 (17.6%)

  

 ≥ 3

235

36 (15.3%)

  

Parity

  

0.921 (0.516–1.642)

0.78

 < 2

217

34 (15.7%)

  

 ≥ 2

137

23 (16.8%)

  

Hormone use

  

0.556(0.163–1.897)

0.342

No

324

54 (16.7)

  

Yes

30

10 (10%)