BMC Womens Health · 2026;26:108 · Colposcopy-Referred Study · China

PAX1/JAM3 Methylation & Vaginal Dysbiosis in High-Grade Cervical Lesions

PAX1/JAM3 甲基化与阴道微生态失调在宫颈高级别病变中的关联
264 colposcopy-referred women · CISCER AUC 0.826 (CIN2+) / 0.852 (CIN3+) · specificity 89.1% vs hrHPV 7.3% · dysbiosis (Lactobacillus↓, Gardnerella↑) predicts CIN2+
Luo L, Wang L, Liu Z, Cai Y, Zhang J (correspondence)
Beijing Anzhen Hospital, Capital Medical University · CISCER® PAX1/JAM3 · doi: 10.1186/s12905-026-04285-6
Methylation signal · CIN2+ vs CIN1−
ΔCt PAX1
4.25 vs 9.35
ΔCt JAM3
9.34 vs 13.2
lower ΔCt = higher methylation
Vaginal dysbiosis · correlates with CIN2+
Lactobacillus
r = −0.316
Gardnerella
r = +0.354
Nugent score
r = +0.499

1Background & Objective

  • Gap: hrHPV testing is sensitive but very low specificity (7.3% in this cohort) → excessive colposcopy referral.
  • Angle: link PAX1/JAM3 methylation with clinically assessed vaginal microecology in colposcopy-referred women.
  • Objective: assess CISCER® as molecular triage and its association with vaginal dysbiosis for CIN2+.

2Study Design & Cohort

264
colposcopy-referred women
42y
median age
71
CIN2+
CISCER methylation+ hrHPV / LBC+ vaginal microecology
  • Setting: Beijing Anzhen Hospital, Feb–Dec 2024; TZ III 66.7%.
  • CIN2+ (n=71): CIN2 42.3% · CIN3 49.3% · SCC 7.0% · AD 1.4%.
  • HPV in CIN2+: 93% hrHPV+; HPV16/18 50.7% vs 33.7% in CIN1− (P<0.05).
  • Microecology: clinician-evaluated in 172 women (Lactobacillus, Gardnerella, Nugent, pH, AV).
  • Baseline: median 42 y; TZ III 66.7%.
  • Dysbiosis scores: Nugent, AV, pH, H₂O₂, sialidase, polyamine.
  • Scoring: clinician per microecology scoring method (Table S3).
30
CIN2
35
CIN3
6
cancer (5 SCC + 1 AD)
CIN2
CIN3
SCC
AD
CISCER+ = ΔCt PAX1 ≤ 6.6 or ΔCt JAM3 ≤ 10.0 (Kruskal-Wallis ***P<0.001).

3Diagnostic Performance

0.826
AUC · CIN2+
0.852
AUC · CIN3+
76.1%
Se · CIN2+
89.1%
Sp · CIN2+
91%
NPV · CIN2+
CISCER Sp 89.1% vs hrHPV 7.3% (CIN1−) · CIN3+ Se 87.8% / Sp 82.5% · all SCC & AD positive · normal women: 92.9% hrHPV+ but only 7.1% CISCER+.

4Methylation & Clinical Efficacy

ΔCt distribution by pathological groups
Fig. 2 ΔCt falls with severity — higher methylation in high-grade lesions.
Clinical efficacy of different screening methods
Fig. 3 Se (CIN2+/CIN3+), Sp (CIN1−) & ROC — CISCER best specificity.
  • Specificity wins: 89.1% vs hrHPV 7.3%; all SCC & AD methylation positive (cytology missed 4).

5Vaginal Microecology Association

Correlation heatmap of clinical indexes
Fig. 4 Correlation heatmap (n=172) — dysbiosis markers track high-grade lesions.
Correlation with CIN2+ (r) Nugent Gardnerella Sialidase pH Lactobacillus 0.499 0.354 0.321 0.283 −0.316
Fig. Dysbiosis markers — positive with CIN2+; Lactobacillus negative (depletion).
  • Dysbiosis pattern: Lactobacillus depletion (r=−0.316) & Gardnerella enrichment (r=0.354).
Nugent r=0.499 · ΔCt JAM3 r=−0.607 — both predict high-grade lesions.
Clinical Significance