DNA methylation as a triage marker for colposcopy referral in HPV-based cervical cancer screening: a systematic review and meta-analysis

Table 3 Meta-analysis of the performance of DNA methylation assays for the detection of CIN2+ and CIN3+ according to the study design

	N datasets	N samples	Pooled prevalence	Pooled sensitivity (95% CI)	Pooled specificity (95% CI)	Q	I2	Pooled AUC,
	N datasets	N samples	Pooled prevalence	Pooled sensitivity (95% CI)	Pooled specificity (95% CI)	(p-value)	(95% CI)	(95% CI)
CIN2 + detection ^a,b,c
Population Referral studies [14, 18,19,20,21,22, 27]	8	3627	0.28	0.60 (0.52–0.66)	0.70 (0.62–0.77)	91.489 (p < 0.001)	98 (96–99)	0.68 (0.64–0.72)
Cohort studies [17, 24, 28, 29, 31, 33]	6	1293	0.42	0.70 (0.62–0.76)	0.78 (0.69–0.86)	35.657 (p < 0.001)	94 (90–99)	0.79 (0.75–0.82)
Convenience studies [12, 13, 15, 16]	8	3234	0.20	0.69 (0.64–0.73)	0.79 (0.77–0.81)	0.478 (p = 0.394)	0 (0–100)	0.82 (0.78–0.85)
CIN3 + detection ^a,b,d
Population referral studies [14, 18,19,20,21,22,23, 27]	9	4601	0.16	0.69 (0.64–0.74)	0.71 (0.64–0.78)	28.0.19 (p < 0.001)	93 (86–99)	0.75 (0.71–0.78)
Cohort studies [17, 24, 26, 29,30,31]	6	1331	0.36	0.81 (0.72–0.87)	0.72 (0.65–0.79)	27.134 (p < 0.001)	93 (86–99)	0.83 (0.80–0.86)
Convenience studies [12, 13, 15, 16]	8	3234	0.14	0.80 (0.75–0.84)	0.77 (0.75–0.79)	0.646 (p = 0.362)	0 (0–100)	0.81 (0.78–0.84)

^aDue to the limited number of case–control studies (n = 3), it was not possible to perform statistical analysis. ^bThe sensitivity and specificity were estimated as reported by authors. When multiple thresholds reported, 70% specificity was selected. Only one entrance per study was considered. The best combination reported by the authors was selected. ^cOnly one entrance per study was considered. The best combination reported by the authors was selected. For [16] was considered JAM3; for [29] was considered C13orf18/EPB41L3/JAM3. ^dOnly one entrance per study was considered. The best combination reported by the authors was selected. For H5[30] was considered PAX1/ZNF582; for K3 [29] was considered C13orf18/EPB41L3/JAM3; for J7 [27] was considered SOX1/ZSCAN1 selection bias

ISSN: 1868-7083