Table 2 Meta-analysis of the performance of DNA methylation assays for detection of CIN2+ and CIN3+

From: Performance of DNA methylation assays for detection of high-grade cervical intraepithelial neoplasia (CIN2+): a systematic review and meta-analysis

 

N studies

N women

Pooled CIN2+/CIN3+ prevalence

Pooled sensitivity, % (95% CI)

I2

Pooled specificity, % (95% CI)

I2

CIN2+ detection

       

All studies, irrespective of threshold level10,24,25,26,27,28,29,30,31,33,34,35,36,37,38,39,40,4243,46,47,48,49,52,53,54,55,56,57,58,59,60,61,62,63,65 a,b

38

12,552

36.7%

63.2 (56.4–69.5)

90.5 (88.2–92.8)

75.9 (71.9–79.5)

85.7 (81.9–89.6)

 Population-based screening studies

8

6589

27.2%

66.5 (56.8–75.0)

93.0 (89.6–96.5)

70.8 (69.4–72.2)

42.9 (0.0–89.5)

 Cohort studies

7

2227

24.3%

68.0 (57.0–77.3)

83.0 (71.5–94.6)

76.7 (66.3–84.6)

95.3 (93.1–97.5)

 Case-control studiesc

5

1054

33.0%

58.7 (48.1–68.5)

34.2 (0.0–98.4)

78.2 (62.8–88.4)

76.5 (55.7–97.4)

 Convenience studiesc

9

1468

36.4%

45.9 (31.7–60.9)

87.2 (79.7–94.7)

84.1 (74.8–90.4)

83.5 (73.0–93.9)

 HPV16-positive samplesd

9

1214

58.3%

71.8 (54.5–84.4)

95.3 (93.3–97.4)

73.5 (66.6–79.4)

57.1 (23.3–90.9)

Set threshold to achieve 70% specificity10,24,25,26,27,28,29,31,33,35,37,39,42,46,48,55,56,57,58,59,60,63,65 e

24

9646

35.9%

68.6 (62.9–73.8)

86.3 (81.7–90.9)

70.5 (69.3–71.6)

0.0 (0.0–59.1)

Set threshold to achieve 50% specificity10,25,27,28,33,34,35,37,39,42,48,49,55,57,58,60 e

17

7225

34.9%

80.3 (75.6–84.4)

83.5 (76.2–90.7)

50.1 (48.7–51.5)

0.0 (0.0–74.0)

CIN3+ detection

       

All studies, irrespective of threshold level10,25,26,27,27,29,30,31,32,34,35,38,40,41,42,44,46,47,48,49,50,51,52,53,54,56,58,62,63,65 a, f

30

7393

21.5%

70.5 (64.8–75.6)

80.0 (73.3–86.6)

74.7 (70.8–78.1)

86.2 (82.1–90.4)

 Referral-population-based studies

6

2708

17.4%

67.6 (60.4–74.0)

67.2 (38.8–95.6)

70.1 (67.8–72.3)

38.2 (0.0–95.2)

 Cohort studies

5

1331

17.7%

78.2 (68.3–85.7)

77.7 (58.2–97.3)

76.4 (61.9–86.6)

97.0 (95.5–98.5)

 Case-control studies

5

1306

14.8%

71.0 (60.0–80.0)

61.6 (24.0–99.2)

70.0 (67.0–73.0)

0.0 (0.0–100.0)

 Convenience studies

10

1544

21.5%

72.2 (64.5–78.8)

54.0 (21.1–86.8)

76.6 (71.1–81.2)

73.1 (56.0–90.2)

 HPV16-positive samples

4

504

43.1%

58.8 (34.7–79.3)

92.6 (86.9–98.2)

80.0 (59.4–91.6)

83.5 (68.2–98.9)

Set threshold to achieve 70% specificity10,25,26,27,29,31,32,35,41,42,44,46,48,51,53,58,63,65 e

19

5197

19.8%

71.1 (65.7–76.0)

64.0 (45.9–82.1)

69.8 (68.3–71.3)

0.0 (0.0–69.3)

Set threshold to achieve 50% specificity10,25,27,32,34,35,42,44,48,49,51,58,65 e

14

4216

19.0%

82.3 (77.8–86.1)

53.6 (25.6–81.7)

49.7 (48.1–51.3)

0.0 (0.0–79.9)

  1. aThe sensitivity estimate derived based on 70% specificity is used in this analysis when multiple estimates are given and for all other studies, the sensitivity and specificity estimates as reported by authors are used
  2. bOne study36 provides estimates for two genes (EPB41L3 and SOX1); in this analysis only data for EPB41L3 are included so that data from the same population of women are considered only once; similarly for ref. 38—data for PAX1 and not SOX1 are included
  3. cFor analysis of case-control studies, ref. 10 was removed from the analysis; and for convenience studies ref. 49 was removed from the analysis to allow best model fit in stata
  4. dAmong women with HPV16 infection; e24 studies allowed standardization of threshold-level for methylation positivity, which corresponded to a specificity of 70% for CIN2+ either by providing the raw data or a ROC curve; 17 studies allowed estimation at a specificity of 50% for CIN2+; 19 studies allowed estimation at a specificity of 70% for CIN3+
  5. e14 studies allowed estimation at a specificity of 50% for CIN3+
  6. fOne study32 provides estimates for two gene combinations (CADM1/MAL and MAL/MiR-124-2) and only data for MAL/MiR-124-2 combination were used in the analysis (as authors conclude this was the best combination of markers)
Back to article page