Novel triple-target panels utilizing methylation-sensitive restriction enzyme-based quantitative PCR for detecting advanced cervical precancers and cancers among high-risk HPV-positive women.

BACKGROUND

Optimal triage option for high-risk HPV-positive (hrHPV) women remains uncertain. We aimed to utilize methylation-sensitive restriction enzyme-based quantitative PCR (MSRE-qPCR) technique and develop triple-target human gene methylation panels to improve detection of advanced cervical precancers and cancers (CIN3 +) among hrHPV women.

METHODS

Eighteen candidate genes were detected by MSRE-qPCR in cervical samples from hrHPV women. All possible triple-target panels from these genes were generated by logistic regression models with repeated ten-fold cross-validation on a training set of 1223 women (180 CIN3 +; 1043 < CIN3). Panels with the top two AUCs for CIN3 + on a validation set of 937 women (69 CIN3 +; 868 < CIN3) were ultimately selected for qPCR reconstruction, retesting, and retraining. Triage performance, screening efficiency and risk stratification of the selected panels were then compared with traditional triage strategies (cytology [ASCUS as the threshold, atypical squamous cells of undetermined significance], HPV16/18 genotyping, HPV16/18 genotyping combined with cytology [ASCUS]) within the validation set.

RESULTS

Two panels were finally identified and validated for CIN3 + detection. Panel 1 includes JAM3, PCDHGB7, and SORCS1; while Panel 2 consists of PAX1, ZNF671, and ASCL1. Compared to traditional triage strategies, both panels demonstrated superior AUCs (Panel 1: 0.799; Panel 2: 0.790; Cytology: 0.532; HPV16/18 genotyping: 0.589; HPV16/18 + Cytology: 0.515; all P < 0.001), with substantially higher specificities (83.06%; 86.98%; 29.49%; 72.93%; 14.52%), comparable sensitivities (76.81%; 71.01%; 76.81%; 44.93%; 88.41%), and requiring fewer colposcopies per CIN3 + case (3.77; 3.31; 12.55; 8.58; 13.16). However, Panel 1 and Panel 2 were statistically indistinguishable (P = 0.603; Ratio: 0.92 [0.76-1.13]; Ratio: 1.05 [1.01-1.08]). Unlike traditional strategies (CIN3 + risks: 5.59% ~ 17.97%), women tested negative on either panel had immediate CIN3 + risks below 3% (Panel 1: 2.17%; Panel 2: 2.58%), supporting safe deferral for at least one year.

CONCLUSIONS

Two triple-target human gene methylation panels were successfully developed, each integrated into a single MSRE-qPCR system for one-tube detection. Both panels outperformed current triage strategies, indicating their potential as alternatives, though external validation among diverse settings is needed before clinical application.