Systematic Review & Meta-Analysis · 论著

PAX1/JAM3 Methylation for Diagnosis of Cervical High-Grade Intraepithelial Neoplasia: A Meta-Analysis

PAX1、JAM3 甲基化对子宫颈高级别上皮内瘤变诊断的 meta 分析
7 diagnostic studies · QUADAS-2 · Meta-Disc 1.4 · CIN2+ & CIN3+ pooled accuracy
肖金芬、张旭梅、刘懿、丁尚玮、曾倩、李芹 · 昆明医科大学附属延安医院
现代妇产科进展 2025;24(9):680-685 · DOI:10.13283/j.cnki.xdfckjz.2025.09.004 · PROSPERO CRD420251008484
7
diagnostic studies
0.9694
CIN2+ pooled AUC
77.36
CIN2+ DOR
2024.2
search cutoff

1Background & Objective

13.83/10万
China incidence 2022
HPV
low specificity
TCT
low sensitivity
  • Single tests fall short: HPV DNA cannot distinguish transient from persistent infection; cytology is operator-dependent.
  • Methylation rationale: promoter hypermethylation silences tumor-suppressor genes — an early, quantitative molecular signal.
  • Aim: systematically evaluate combined PAX1/JAM3 methylation accuracy for HSIL (CIN2+/CIN3+).

2Methods

8
databases
7
studies included
QUADAS-2
quality tool
  • Search: CNKI, Wanfang, VIP, CBM, PubMed, Embase, Web of Science, Cochrane — from inception to Feb 10, 2024; PROSPERO-registered.
  • Inclusion: diagnostic studies of PAX1/JAM3 dual methylation with 2×2 table data (TP/FP/TN/FN), histopathology reference standard.
  • Quality: QUADAS-2 double-blind appraisal; heterogeneity via Spearman threshold analysis; Meta-Disc 1.4 pooling.
  • Outcomes: pooled Se, Sp, PLR, NLR, AUC, DOR for CIN2+ and CIN3+.
2015–2024
publication window
2×2
table data required
CN
+ EN databases
  • Language: Chinese (CNKI/Wanfang/VIP/CBM) and English (PubMed/Embase/WoS/Cochrane) sources combined.
  • Threshold effect: Spearman correlation evaluated before pooling; Meta-Disc used for bivariate accuracy synthesis.
  • Reference standard: internationally accepted histopathology diagnosis; duplicate reports excluded.
Se 0.78
CIN2+ pooled
Sp 0.95
CIN2+ pooled
AUC .97
CIN2+
方法要点：双人独立提取、QUADAS-2 质量评价、阈值效应检验与 Meta-Disc 1.4 合并——保证诊断准确性综合的稳健性。
检索覆盖 8 个中英文数据库至 2024 年 2 月，纳入 7 项诊断试验，共提取四格表数据 7 组。
Meta-analysis follows PRISMA framework; independent dual data extraction with third-party arbitration.

3CIN2+ — Pooled Diagnostic Accuracy

CIN2+ forest plot
0.78
Se [0.75–0.80]
0.95
Sp [0.94–0.95]
  • Rule-in power: PLR 14.61 — positive methylation markedly raises HSIL probability.
  • Rule-out power: NLR 0.202 — negative result strongly reduces high-grade risk.
  • Overall DOR: 77.36 [43.1–139.0] — excellent discriminative ability.
7 studies pooled · AUC 0.9694 · consistency supports triage application in screening follow-up.

4CIN3+ — Pooled Diagnostic Accuracy

CIN3+ forest plot
0.84
Se [0.80–0.87]
0.87
Sp [0.86–0.88]
  • Higher Se: pooled sensitivity rises to 0.84 for CIN3+ — strongest for clinically important lesions.
  • Stronger exclusion: NLR 0.172 — negative result nearly rules out CIN3+.
  • DOR: 34.76 [20.1–60.1] · AUC 0.9251 — high overall accuracy.
CIN3+ is the clinical action threshold — high sensitivity supports safe deferral of colposcopy when negative.

5Likelihood Ratios & Interpretation

PLR · CIN2+
14.61
14.61
PLR · CIN3+
5.911
5.911
NLR · CIN2+
0.202
0.202
NLR · CIN3+
0.172
0.172
0.9694
CIN2+ AUC
0.9251
CIN3+ AUC
  • Strong rule-in: PLR 14.61 (CIN2+) — a positive PAX1/JAM3 methylation result markedly increases HSIL probability.
  • Strong rule-out: NLR 0.202 — a negative result substantially reduces the chance of high-grade disease.
  • Clinical role: supports methylation as an objective triage tool for cervical screening follow-up.
  • Pre-test caution: interpret alongside local prevalence & cytology/HPV context — pooled values span heterogeneous settings.
QUADAS-2 quality assessment; threshold-effect analysis performed; Meta-Disc 1.4 bivariate pooling.
Clinical Significance