Comparative evaluation of junctional adhesion molecule 3 and paired box gene 1 methylation, cytology, E6/E7 messenger RNA, and high-risk human papillomavirus testing for detecting precancerous cervical lesions

INTRODUCTION

Cervical cancer ranks as the fourth most prevalent cancer in women, with an estimated 604,127 new cases and 341,831 deaths annually, representing a major public health and economic concern.^[1] In 2018, the World Health Organization (WHO) launched an initiative aiming to eradicate cervical cancer by 2030. Despite this, the disease remains a growing concern in developing countries, largely due to economic inequalities and limited access to healthcare resources.^[2] Effective and widespread screening plays a crucial role in lowering cervical cancer rates. At present, the primary approaches for detecting cervical cancer are cytological examination and high-risk human papillomavirus (hrHPV) testing.^[3] However, cervical cytology is influenced by the subjective factors and experience of pathologists, which leads to high subjectivity, poor reproducibility, and low sensitivity.^[4] hrHPV testing, while sensitive, suffers from low specificity, leading to increased colposcopy referrals and potential overtreatment, which impose unnecessary economic and psychological burdens on patients and their families.^[5] Therefore, it is crucial to develop a cervical cancer screening strategy that is highly accurate and cost-effective.

Among epigenetic mechanisms, DNA methylation is prominent for its ability to stably inactivate genes without altering the sequence itself. This modification governs gene regulation and is implicated in diverse processes, including organismal development and the onset of diseases.^[6] In cervical cancer, hypermethylation of gene promoters, which leads to gene silencing, is strongly linked to the development and progression of the disease. Among the more than 100 human methylation biomarkers investigated in cervical tissues, approximately 20 have been consistently reported in published studies.^[7,8] Multiple combinations of DNA methylation markers have demonstrated potential application value in cervical cancer screening. This underscores the critical role of epigenetic alterations in disease diagnosis and surveillance, suggesting these markers could enhance screening strategies alongside traditional methods.^[9,10] In addition, DNA methylation assays are relatively simple and can theoretically be automated, potentially improving detection efficiency while reducing costs. This would facilitate implementing these assays in economically disadvantaged areas.

Among these markers, paired box gene 1 (PAX1) and junctional adhesion molecule 3 (JAM3) have shown particularly strong diagnostic performance. PAX1, a transcription factor important for embryonic development, has been identified as a tumor suppressor in cervical carcinoma.^[11] Promoter hypermethylation of PAX1 leads to its silencing, which contributes to tumor development. Reactivation of PAX1 can inhibit the growth and motility of cervical carcinoma cells and enhance chemotherapy response by inhibiting the WNT/TIMELESS signaling pathway.^[12] Multiple studies have demonstrated a strong correlation between PAX1 methylation and the development and progression of cervical carcinoma, with higher methylation levels observed in more advanced disease stages.^[13,14] JAM3, an immunoglobulin superfamily protein that mediates cell– cell adhesion, regulates epithelial cell migration and immune responses.^[15] Within cervical carcinoma, JAM3 correlates with lymph node metastasis and may drive tumor progression through epithelial–mesenchymal transition and activation of the hypoxia-inducible factor 1-alpha/vascular endothelial growth factor A pathway.^[16] JAM3 has also demonstrated high specificity for detecting cervical intraepithelial neoplasia (CIN) grade 3 and higher lesions, especially among women positive for human papillomavirus (HPV) or those presenting with indeterminate cytology results.^[17] Taken together, recent evidence supports the combined use of PAX1 and JAM3 methylation testing as an effective triage strategy across various clinical contexts and HPV genotypes.^[18-23]

Currently, various novel cervical cancer screening strategies exist, including hrHPV testing, ThinPrep cytologic test (TCT), HPV integration testing, E6/E7 messenger RNA (mRNA) testing, and p16/Ki-67 immunohistochemical dual staining.^[24] These methods aim to optimize the performance of cervical carcinoma screening by incorporating advanced technologies and biomarkers. However, current detection methods still require optimization to achieve more efficient and cost-effective screening strategies. In addition, extensive clinical prospective studies are needed to explore their feasibility. The present study investigates the accuracy of PAX1/JAM3 methylation testing for cervical carcinoma screening. It compares the diagnostic performance of PAX1/JAM3 methylation testing, HPV testing, cytology, E6/E7 mRNA testing, and various combinations of these tests for identifying HSIL+ lesions. This research seeks to provide insights into the effectiveness of methylation markers as potential biomarkers for cervical carcinoma diagnosis.

MATERIAL AND METHODS

Methods

Liquid-based cytology testing

The TCT was used for diagnostic purposes. Cervical exfoliated cells were collected and suspended in a specialized liquid-based preservative solution (CellPreserve, Wuhan Heer Medical Technology Development Co., Ltd., Wuhan, Hubei, China) and sent to the pathology department. Two or more experienced pathologists assessed the cytological samples following the 2014 Bethesda System Classification.^[25]

Results description

The results are categorized as follows: Negative for intraepithelial lesion or malignancy (NILM); atypical squamous cells (ASC), including ASCs of undetermined significance (ASC-US) and atypical squamous cells that cannot exclude-high-grade squamous intraepithelial lesion (ASC-H); low-grade squamous intraepithelial lesion (LSIL); high-grade squamous intraepithelial lesion (HSIL); squamous cell carcinoma (SCC); atypical glandular cells; adenocarcinoma in situ (AIS); and adenocarcinoma. Representative cytological images are shown for (a) NILM [Figure 1a], (b) ASC-US [Figure 1b], (c) LSIL [Figure 1c], and (d) HSIL [Figure 1d] cases.

Close

Figure 1: Representative cytological images of cervical samples showing a spectrum of squamous epithelial abnormalities. (a) Negative for intraepithelial lesion or malignancy (NILM), showing normal squamous epithelial cells with regular nuclear contours and uniform chromatin distribution. (b) Atypical squamous cells of undetermined significance (ASC-US). Cells within the red circle show mild nuclear enlargement, slightly coarse or mildly uneven chromatin, and a mildly increased nuclear-to-cytoplasmic (N/C) ratio. (c) Low-grade squamous intraepithelial lesion (LSIL). The cell indicated by the red arrow shows koilocytotic features, including a prominent perinuclear halo, marked nuclear enlargement, irregular nuclear contours, coarsened but relatively evenly distributed chromatin, a mildly to moderately increased N/C ratio, and multinucleation. (d) High-grade squamous intraepithelial lesion (HSIL). Cells within the red rectangle show small cell size with markedly reduced cytoplasm, hyperchromatic nuclei with coarse and unevenly distributed chromatin, a significantly increased N/C ratio, and distinctly irregular nuclear outlines.All slides were prepared using the ThinPrep cytologic test and stained with the Papanicolaou stain. Original magnification ×40. Scale bar = 20 µm.

Export to PPT

hrHPV DNA testing

High-risk HPV genotyping was performed using a nucleic acid detection kit (fluorescent polymerase chain reaction [PCR] method; 20143402188, Chaozhou Kaipu Biochemical Co., Ltd., Chaozhou, Guangdong, China), following the manufacturer’s instructions. This kit detects 21 HPV genotypes, including 14 high-risk HPV types (HPV 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68), 6 low-risk HPV types (HPV 6, 11, 42, 43, 44, and CP8304), and 1 HPV type with unknown risk (HPV 53).

E6/E7 mRNA testing

Branched DNA (bDNA) technology was used to detect E6/E7 mRNA according to the manufacturer’s instructions. The kit (HPV E6/E7 detection kit [bDNA signal amplification method], 20163401261, Kodia, Zhenzhou, China) detects E6/E7 mRNA from 14 types of high-risk HPV (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68). The mRNA results for each sample are expressed in light units, which are converted to copy numbers using specialized calculation software. This conversion is based on the direct relationship between light emission and the amount of HPV mRNA. An E6/E7 mRNA result is interpreted as positive when the copy number is ≥1.0 and negative when it is <1.0.

DNA methylation testing

The methylation status of PAX1 and JAM3 was assessed using bisulfite conversion combined with quantitative fluorescent PCR, employing the Human PAX1 and JAM3 gene methylation detection kit (PCR-fluorescent probe method; 20233400253; Beijing Origin Juhe Biotechnology Co., Ltd., Beijing, China) according to the manufacturer’s instructions. A 2-3 mL aliquot of each specimen was used for DNA extraction, followed by concentration and quality assessment.

After bisulfite conversion using 200–1000 ng of DNA, PCR amplification was conducted using the SLAN-96S automated medical PCR analysis system (Hongshi, Shanghai, China). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was selected as the internal control. The following conditions were used for the reaction: pre-denaturation at 96°C for 10 min; 1 cycle of denaturation at 94°C for 15 s; annealing at 64°C for 5 s; extension and fluorescence collection at 60°C for 30 s; followed by 45 cycles; and cooling to 25°C for 1 min. On completion of the reaction, the cycle threshold (Ct) values of the PAX1 gene, JAM3 gene, and the internal control gene (GAPDH) were recorded, and the ΔCt values of the target genes were calculated (ΔCt = Ct of the detected gene-Ct of the internal control gene). The JAM3 result is represented as ΔCtJ, and the PAX1 result as ΔCtP. The receiver operating characteristic (ROC) curve was generated to establish the quantitative diagnostic threshold for the methylation status of PAX1 and JAM3.

Colposcopy and pathology

Trained clinicians collected biopsies from regions exhibiting suspicious lesions. For patients without visible lesions, biopsies were taken from the cervical transformation zone at the 3, 6, 9, and 12 o’clock positions, along with an endocervical curettage. The biopsy samples were preserved in 10% formalin and submitted to the pathology department for analysis. Two experienced pathologists interpreted the results, and if discrepancies arose, a third pathologist was involved to establish a majority consensus diagnosis. The pathological findings were categorized following the 2020 WHO classification for cervical lesions: Chronic cervicitis, LSIL, HSIL, AIS, cancers including SCC and adenocarcinoma, and invasive cancer.^[26]

RESULTS

Cut-off value for DNA methylation

The cut-off points for PAX1 and JAM3 methylation were established at 7.31 and 10.0, respectively, according to sensitivity, specificity, and Youden index calculations, which yielded maximum Youden index values of 0.657 and 0.564. As shown in the ROC analysis [Figure 2], both markers effectively distinguished high-grade lesions from negative cases. The number of positive methylation cases was 68 (23.6%) for PAX1, 75 (26.0%) for JAM3, and 88 (30.6%) for the combined PAX1/JAM3.

Close

Figure 2: Receiver operating characteristic curves showing the performance of paired box gene 1 and junctional adhesion molecule 3 (PAX1 and JAM3) methylation in distinguishing high-grade squamous intraepithelial lesion + from negative cases. The area under the curve (AUC) reflects the diagnostic accuracy of each marker: PAX1 (AUC = 0.873, P < 0.001) and JAM3 (AUC = 0.839, P < 0.001). The difference between the two AUCs was not statistically significant (P = 0.29).

Export to PPT

Relationship between PAX1 and JAM3 gene methylation and cervical lesions

Representative methylation levels across pathological categories are shown for (a) JAM3 [Figure 3a] and (b) PAX1 [Figure 3b]. Methylation levels of PAX1 and JAM3 were comparable between the normal cervix and LSIL groups (P > 0.05). As the pathological grade increased, the ΔCt value of methylation significantly decreased. For PAX1, differences were significant between LSIL and HSIL (P < 0.0001) as well as between HSIL and cancer (P < 0.01). For JAM3, significant differences were observed between LSIL and HSIL (P < 0.0001) and between HSIL and cancer (P < 0.001). Compared with the PAX1^m/JAM3^m (-) group, patients with PAX1^m/JAM3^m (+) had a markedly higher risk of HSIL+ (odds ratio [OR] = 22.608, 95% CI: 10.787-47.380, P < 0.001).

Close

Figure 3: Methylation levels of paired box gene 1/junctional adhesion molecule 3 (PAX1/JAM3) (ΔCt) in the study group. (a) Distribution of JAM3 methylation ΔCt based on pathological results. (b) Distribution of PAX1 methylation ΔCt based on pathological results. ^✶✶P < 0.01, ^✶✶✶P < 0.001, ^✶✶✶✶P < 0.0001, ns: Not significant.

Export to PPT

Positive rates of different pathological grades detected by various testing methods

Among the normal cervical and LSIL groups, hrHPV positivity rates were 76.7% and 95.8%, respectively. Similarly, the LSIL group exhibited a high E6/E7 mRNA positivity rate of 75.0%. In the HSIL group, the positivity rates for E6/E7 mRNA, HPV 16/18, JAM3^m, PAX1^m, and PAX1^m/JAM3^m were 75.5%, 56.6%, 66%, 71.7%, and 79.2%, respectively. For cervical cancer detection, E6/E7 mRNA positivity was 87.5%, while all other tests detected cervical cancer with 100% accuracy [Figure 4]. Using histopathological results as the reference standard, with HSIL+designated as the outcome measure, the positive rates of different tests for cervical lesion screening are presented in Table 2.

Close

Figure 4: Bar chart showing the positive rates of each test within each histological category. LSIL: Low-grade squamous intraepithelial lesion; HSIL: High-grade squamous intraepithelial lesion; cancer, including cervical squamous cell carcinoma and cervical adenocarcinoma.

Export to PPT

Table 2: Positive rates of different testing methods for cervical lesion screening with HSIL+as the Endpoint.

Testing method	LSIL (n, %)	HSIL+ (n, %)	Summary	χ2	P
TCT (%)
Positive	68 (30.0)	51 (83.6)	119 (41.3)	57.1	<0.001
Negative	159 (70.0)	10 (16.4)	169 (58.7)
hrHPV (%)
Positive	188 (82.8)	60 (98.4)	248 (86.1)	9.7	=0.002
Negative	39 (17.2)	1 (1.6)	40 (13.9)
E6/E7 mRNA (%)
Positive	103 (45.4)	48 (78.7)	151 (52.4)	21.4	<0.001
Negative	124 (54.6)	13 (21.3)	137 (47.6)
PAX1m
Positive	22 (9.7)	46 (75.4)	68 (23.6)	115.1	<0.001
Negative	205 (90.3)	15 (24.6)	220 (76.4)
JAM3m
Positive	32 (14.1)	43 (70.5)	75 (26.0)	79.4	<0.001
Negative	195 (85.9)	18 (29.5)	213 (74.0)
PAX1m/JAM3m
Positive	38 (16.7)	50 (82.0)	88 (30.6)	96.4	<0.001
Negative	189 (83.3)	11 (18.0)	200 (69.4)

LSIL: Low-grade squamous intraepithelial lesion, HSIL: High-grade squamous intraepithelial lesion, TCT: ThinPrep cytologic test, hrHPV: High-risk human papillomavirus, mRNA: Messenger RNA, PAX1: Paired box gene 1, JAM3: Junctional adhesion molecule 3. HSIL+, the diagnosis of HSIL and cancer, is collectively referred to as HSIL+; LSIL, the diagnosis of cervical inflammation and LSIL, is collectively referred to as LSIL; liquid-based cytology results≥ASCUS are considered positive, while the rest are negative. PAX1^m refers to PAX1 methylation, JAM3^m refers to JAM3 methylation, and PAX1^m/JAM3^m refers to either PAX1 or JAM3 methylation

Clinical performance of different methods for detecting HSIL + lesions

Using histopathological findings as the benchmark and considering HSIL + as the target outcome, PAX1 and JAM3 methylation markers showed favorable diagnostic performance. The combined PAX1^m/JAM3^m test achieved the highest sensitivity (82.0%) and a high NPV (94.5%), with an AUC of 0.826. Individually, PAX1^m had higher sensitivity (75.4%) and specificity (90.3%) than JAM3^m. Detailed diagnostic metrics for all methods are summarized in Table 3.

Table 3: Clinical performance of different methods for detecting HSIL+lesions (% [95% CI]).

Testing method	Sensitivity (95% CI)	Specificity (95% CI)	PPV (95% CI)	NPV (95% CI)	AUC (95% CI)
Single test screening
TCT (%)	83.6 (71.9–91.8)	70.0 (63.6–75.9)	42.9 (37.4–48.5)	94.1 (90.0–96.6)	0.768 (0.703–0.833)
HPV16/18	62.3 (49.0–74.4)	64.3 (57.7–70.5)	31.9 (26.5–37.9)	86.4 (81.9–89.9)	0.633 (0.554–0.712)
hrHPV	98.4 (91.2–100.0)	17.2 (12.5–22.7)	24.2 (23.0–25.5)	97.5 (84.5–99.6)	0.578 (0.503–0.652)
E6/E7 mRNA	78.7 (66.3–88.1)	54.6 (47.9–61.2)	31.8 (27.7–36.1)	90.5 (85.3–94.0)	0.667 (0.594–0.740)
PAX1m	75.4 (62.7–85.5)	90.3 (85.7–93.8)	67.6 (57.8–76.1)	93.2 (89.8–95.5)	0.829 (0.761–0.896)
JAM3m	70.5 (57.4–81.5)	85.9 (80.7–90.2)	57.3 (48.4–65.8)	91.5 (88.0–94.1)	0.782 (0.710–0.854)
PAX1m/JAM3m	82.0 (70.0–90.6)	83.3 (77.8–87.9)	56.8 (49.0–64.3)	94.5 (90.9–96.7)	0.826 (0.764–0.889)
Combined Screening
PAX1m/JAM3mand hrHPV	80.3 (68.2–89.4)	85.0 (79.7–89.4)	59.0 (50.8–66.8)	94.2 (90.1–96.4)	0.827 (0.778–0.869)
PAX1m/JAM3m and E6/E7 mRNA	67.2 (54.0–78.7)	88.6 (83.7–92.4)	61.2 (51.3–70.2)	91.0 (87.5–93.5)	0.779 (0.726–0.825)
PAX1m/JAM3m and TCT	73.8 (60.9–84.2)	92.5 (88.3–95.6)	72.6 (62.1–81.1)	92.9 (89.6–95.2)	0.831 (0.783–0.873)
PAX1m/JAM3mand HPV16/18	52.5 (39.3–65.4)	95.6 (92.0–97.9)	76.2 (62.5–86.0)	88.2 (85.2–91.0)	0.740 (0.686–0.790)

TCT: ThinPrep cytologic test, hrHPV: High-risk human papillomavirus, mRNA: Messenger RNA PAX1: Paired box gene 1, JAM3: Junctional adhesion molecule 3, AUC: Area under the curve, PPV: Positive predictive value, NPV: Negative predictive value, CI: Confidence interval. The AUC values for PAX1 and JAM3 in this table are calculated based on a binary classification. Differences from AUC values shown in Figure 2 are due to distinct classification methods and grouping criteria.

The diagnostic value of combining PAX1^m/JAM3^m with other screening strategies was further evaluated. Combining PAX1^m/JAM3^m with cytology (≥ ASC-US), E6/E7 mRNA, HPV 16/18, or overall hrHPV positivity yielded AUCs ranging from 0.779 to 0.831, indicating modest improvements in diagnostic accuracy compared with individual tests. The diagnostic performance of these test combinations and the corresponding sensitivity–specificity trade-offs are shown in Figure 5a and 5b.

Close

Figure 5: Comparative receiver operating characteristic (ROC) curves for multiple cervical cancer screening methods. Methylation was defined as positive for either paired box gene 1 or junctional adhesion molecule 3 (PAX1 or JAM3). (a) ROC curves displaying the diagnostic performance of different combinations of tests in detecting cervical lesions. (b) ROC curves illustrating the trade-off between sensitivity and specificity for different screening methods in detecting cervical cancer and its precursors. Each curve represents a test: ThinPrep cytologic test (liquid-based cytology), high-risk human papillomavirus testing, E6/E7 messenger RNA, and PAX1/JAM3 methylation testing. The area under the curve quantifies the diagnostic performance of each method, with values closer to 1 indicating better performance.

Export to PPT

DISCUSSION

This retrospective study employed a cohort design to evaluate the classification efficacy of PAX1^m/JAM3^m in cervical cancer screening for high-grade cervical lesions, comprehensively comparing the effectiveness of different screening methods. It also explored the diagnostic performance of combined screening with PAX1^m/JAM3^m and various methods for cervical cancer.

A previous meta-analysis including 1,055 patients demonstrated that PAX1 methylation was significantly associated with the progression of cervical lesions, with an OR of 0.09 (95% CI: 0.04-0.19) for the transition from CIN I to CIN II/III and an OR of 0.16 (95% CI: 0.05-0.46) for progression from CIN II/III to cervical cancer, suggesting PAX1 methylation may serve as an auxiliary biomarker for estimating CIN progression risk.^[27] Consistent with previous findings, within our study population, higher PAX1 methylation levels correlated with increasing severity of cervical lesions.

Previous studies have found that PAX1 demonstrates clinical performance comparable to cytology in hrHPV+ patients and offers better accuracy and specificity than HPV 16/18 as a triage tool for detecting CIN3+ in hrHPV+ women.^[28] Detecting HSIL in patients with ASC-US is crucial. Chen et al. (2024) found that the methylation assay achieved a sensitivity of 83.8% and a specificity of 95.8%, surpassing HPV-DNA testing in distinguishing high-grade cervical lesions in women with ASC-US.^[20] Similarly, in our study, PAX1 methylation also showed strong performance in detecting HSIL+, with detection performance superior to that of cytology and hrHPV (AUC 0.829 vs. 0.768 vs. 0.578).

Research indicates that increased JAM3 methylation is linked to higher Fédération Internationale de Gynécologie et d’Obstétrique (FIGO) stage and lymph node involvement, suggesting its potential as an independent biomarker.^[29] Boers et al. used quantitative methods to detect JAM3 gene methylation levels in hrHPV+ women, reporting a sensitivity of 68% and specificity of 94% for diagnosing CIN2 and above, and a sensitivity and specificity of 80% and 76% for diagnosing CIN3+ lesions.^[30] In this study, the sensitivity of JAM3 gene methylation for diagnosing HSIL+ was 70.5%, and the specificity was 85.9%.

Moreover, in our study, PAX1^m/JAM3^m positivity progressively rose with increasing severity of cervical lesions. The positive rate was low at 11.6% in the normal cervix group, increased to 27.8% in LSIL, and significantly rose to 79.2% in HSIL, with a 100% positive rate in cervical cancer. This trend suggests PAX1^m/JAM3^m could serve as a triage tool for patients with abnormal cervical screening results. In contrast, this progressive pattern was less pronounced in other tests used in our study, including hrHPV DNA, E6/E7 mRNA, and HPV 16/18 genotyping.

Clinically, the strong correlation between PAX1^m/JAM3^m positivity and lesion severity highlights its utility in improving detection accuracy and reducing unnecessary referrals. Although hrHPV testing demonstrates high sensitivity, its specificity remains low (17.2%), leading to many false positives. In contrast, PAX1^m/JAM3^m showed a balanced diagnostic profile with 82.0% sensitivity and 83.3% specificity, which may help reduce overdiagnosis and overtreatment inherent in current hrHPV-based screening. In the LSIL group, hrHPV and E6/E7 mRNA positivity rates were high (95.8% and 75.0%, respectively), but PAX1^m/JAM3^m positivity remained substantially lower (27.8%). This suggests that methylation testing can differentiate transient HPV infections from lesions at higher risk requiring clinical intervention. Since many HPV infections are transient and asymptomatic, the high positivity rates of HPV-based tests in low-grade lesions may represent infections unlikely to progress.

Currently, hrHPV nucleic acid testing is endorsed by Chinese guidelines as the primary approach for cervical cancer screening.^[31] Based on previous studies, hrHPV gene testing is characterized by high sensitivity but relatively lower specificity as a preliminary screening test.^[32,33] This study is consistent with prior research, showing the sensitivity of hrHPV gene testing for HSIL+ is 98.4%, while its specificity is only 17.2%.^[34] Therefore, effective triage of hrHPV+ cases is essential. TCT testing remains the mainstream triage method for hrHPV+ patients; however, TCT results can be influenced by subjective interpretation, particularly in developing countries where cytology testing may lack stringent quality control, potentially leading to lower sensitivity and higher false-negative rates.^[35] In our study, cytology showed relatively high sensitivity (83.6%), possibly due to pathologists’ awareness of patients’ HPV test results, as hrHPV+ cases are more likely to be interpreted as abnormal.^[36] However, cytology as a triage tool did not significantly improve specificity for hrHPV+ patients.^[37]

In this study, PAX1^m/JAM3^m exhibited sensitivity similar to that of TCT (82% vs. 83.6%) but showed higher specificity than TCT (83.3% vs. 70%). Furthermore, PAX1^m/JAM3^m identified all cervical cancer cases, while TCT missed two cases, classifying one as NILM and the other as LSIL. Thus, PAX1^m/JAM3^m offers a viable auxiliary method for triaging hrHPV+ patients in cervical cancer screening. Analyzing the combination of cytology and PAX1^m/JAM3^m testing revealed a trade-off: While sensitivity decreased slightly (83.6% vs. 73.8%), specificity improved considerably (70.0% vs. 92.5%). Compared with hrHPV testing, PAX1^m/JAM3^m demonstrates acceptable sensitivity with higher specificity. In addition, the positive rate of PAX1^m/JAM3^m in the HSIL+ group was significantly higher than in the benign/inflammatory and LSIL groups (82.0% vs. 16.7%), suggesting PAX1^m/JAM3^m may aid in stratifying patients with abnormal cervical screening results into different management pathways, combining hrHPV testing with PAX1^m/JAM3^m also significantly improved specificity (17.2% vs. 85.0%).

Prior research indicates a positive association between E6/E7 mRNA detection and the severity of cytological and histological lesions.^[38] E6/E7 mRNA testing may serve as an effective tool for triaging patients at risk of high-grade CIN.^[39,40] In this research, we compared the performance of E6/E7 mRNA with PAX1^m/JAM3^m in detecting HSIL+ (AUC 0.667 vs. 0.826) and found that PAX1^m/JAM3^m performed better, consistent with previous studies.^[41] A large-scale cross-sectional study in China found that HPV 16/18 genotyping could be an effective approach for screening for cervical cancer in Chinese women.^[42] Combining cytology with HPV 16/18 genotyping improves sensitivity, though at a slight cost to specificity.^[43] By comparison, in our study, combining HPV 16/18 genotyping with PAX1^m/JAM3^m testing slightly reduced sensitivity (62.3% vs. 52.5%) but significantly improved specificity (64.3% vs. 95.6%). This suggests the combination may be particularly useful in reducing false positives and avoiding overtreatment. Among all combined tests, hrHPV(+) and PAX1^m/JAM3^m(+), as well as TCT(+) and PAX1^m/JAM3^m(+), demonstrated superior specificity, sensitivity, and diagnostic performance in detecting HSIL+ compared with other combinations [Table 3]. In conclusion, PAX1^m/JAM3^m performs effectively in cervical cancer screening and holds promise as a triage tool for both hrHPV+ and TCT-abnormal patients.

From an implementation perspective, PAX1^m/JAM3^m testing can be readily incorporated into existing cervical cancer screening workflows. For hrHPV+ women, reflex methylation testing on the same cervical sample collected for HPV or cytology testing allows for efficient, risk-based stratification without requiring additional procedures. This molecular method is objective, automatable, and well-suited for centralized laboratories, offering advantages in quality control and reproducibility over subjective cytology interpretation. In particular, in low-resource settings where cytology infrastructure and trained personnel may be limited, methylation testing provides a scalable alternative that may enhance screening coverage and accuracy. Integrating PAX1^m/JAM3^m as a triage step may thus reduce unnecessary colposcopies while ensuring timely identification of women at true risk for CIN2+ or cancer.

SUMMARY

This study demonstrates that PAX1 and JAM3 gene methylation serve as a valuable supplementary method for the detection of HSIL+, showing higher specificity than cytology and hrHPV testing. When integrated with conventional screening strategies, it can enhance diagnostic accuracy and limit avoidable colposcopy procedures. These results underscore the clinical utility of PAX1^m/JAM3^m as an effective triage tool, particularly for hrHPV+ women, and highlight its potential to improve patient management and resource allocation in cervical cancer screening programs. Future large-scale, multicenter, and longitudinal studies are needed to validate these results and explore their role in routine clinical implementation, especially in low-resource settings.

AVAILABILITY OF DATA AND MATERIALS

The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.

ABBREVIATIONS

AGC: Atypical glandular cells

AIS: Adenocarcinoma in situ

ASC: Atypical squamous cells

ASC-H: Atypical squamous cells that cannot exclude high-grade squamous intraepithelial lesion

ASC-US: Atypical squamous cells of undetermined significance

AUC: Area under the curve

bDNA: Branched DNA

CI: Confidence interval

CIN: Cervical intraepithelial neoplasia

Ct: Cycle threshold

ECC: Endocervical curettage

FIGO: Fédération Internationale de Gynécologie et d’Obstétrique

GAPDH: Glyceraldehyde-3-phosphate dehydrogenase

GCP: Good Clinical Practice

HIV: Human immunodeficiency virus

HPV: Human papillomavirus

hrHPV: High-risk human papillomavirus

HSIL: High-grade squamous intraepithelial lesion

JAM3: Junctional adhesion molecule 3

LSIL: Low-grade squamous intraepithelial lesion

mRNA: Messenger RNA

NILM: Negative for intraepithelial lesion or malignancy

NPV: Negative predictive value

OR: Odds ratio

PAX1: Paired box gene 1

PCR: Polymerase chain reaction

PPV: Positive predictive value

ROC: Receiver operating characteristic

SCC: Squamous cell carcinoma

TCT: ThinPrep cytologic test

WHO: World Health Organization

ΔCt: Delta cycle threshold

AUTHOR CONTRIBUTIONS

SL and CT: Made substantial contributions to the conception and design of the study, development of experimental protocol, acquisition and analysis of data, and interpretation of results and drafted the manuscript and critically revised it for important intellectual content; FW: Contributed to data acquisition and analysis and participated in drafting and critically revising the manuscript; TS and RZ: Supervised the study and contributed to conception and design and critically revised the manuscript for important intellectual content. All authors gave final approval of the version to be published and agreed to be accountable for all aspects of the work. All authors meet the criteria for authorship recommended by the ICMJE.