Meta-analysis of the diagnostic value of SOX1 methylation in different types of cervical cancer

Abstract

Objective

This meta-analysis evaluates the diagnostic value of SOX1 methylation across different cervical cancer types, including squamous cell carcinoma and adenocarcinoma, to assess its efficacy as a biomarker.

Methods

We reviewed studies published up to March 2024, employing a PICOS-based search strategy in databases like PubMed and Web of Science. We included clinical studies providing diagnostic performance indicators while excluding non-clinical and small-sample studies. Meta-Disc1.4 and Stata15.1 were used for statistical analyses focusing on SOX1 methylation’s sensitivity, specificity, and diagnostic odds ratio.

Results

Twelve articles encompassing 18 studies with 3,213 subjects were analyzed. The overall DOR for SOX1 methylation in cervical cancer diagnosis was 68.95 (95%CI: 27.63-172.07), with a Summary Receiver Operating Characteristic AUC of 0.92, indicating high diagnostic accuracy. Specifically, the DOR for adenocarcinoma was 87.57 (95%CI: 7.05-1087.44) with an AUC of 0.89, and for squamous cell carcinoma, it was 245.87 (95% CI: 26.49-2282.40) with an AUC of 0.93, reflecting significant diagnostic potential for both cancer types. No substantial publication bias was detected (P > 0.10).

Conclusion

SOX1 gene methylation demonstrates significant diagnostic value for both adenocarcinoma and squamous cell carcinoma of the cervix, particularly effective in large sample sizes and cervical exfoliated cell samples for early detection and screening, supporting its utility as a reliable biomarker.

Introduction

Cervical cancer has a high incidence in female malignant tumors worldwide, usually at the age of about 50 years, showing a higher tendency to develop, significantly lower than the average age of other common malignant tumors [1]. According to statistics, millions of women are diagnosed with high-grade cervical intraepithelial neoplasia (CIN) each year [2]. Although a variety of existing cervical cancer screening methods have been developed, the high rate of false negatives and false positives, combined with the limitations of health care resources, have led to the coverage of cervical cancer screening in a wide range of people far from ideal. Squamous cell carcinoma of cervix (SCC) and adenocarcinoma of cervix (ADC) constitute the two major histological subtypes of invasive cervical cancer, accounting for 75–90% and 10–25% of total cases, respectively [3, 4]. Compared with SCC, ADC often has no significant clinical symptoms and is prone to false negative results in routine examinations, resulting in the disease being mostly diagnosed at an advanced stage. In addition, ADC responds poorly to chemotherapy and radiotherapy, leading to a relatively poor prognosis. Therefore, the development of a simple, economical, highly accurate and acceptable screening technology is of great significance for improving screening participation rate and reducing the incidence of cervical cancer.

During human papillomavirus (HPV) infection and the development of associated precancerous lesions, human genes coexist with the methylation of HPV DNA, resulting in regulatory changes in the function of tumor suppressor genes [5]. This abnormal methylation is a molecular mechanism with potential value in distinguishing non-progressive from potentially carcinogenic HPV infections [6]. Studies have shown that increased DNA methylation is significantly associated with the persistence of high-risk HPV (HR-HPV) infection, the severity of CIN, and the risk of invasive cervical cancer [7,8,9]. DNA methylation is observed in most cases of cervical cancer (70-100%) and a significant proportion of precancerous lesions (30-80%) [10].

The SOX gene family, a group of transcription factors that play a key role in embryonic development and maintenance of stem cell function, is strongly associated with the onset and progression of tumors [11]. SOX1, a member of this family, is often considered as a cancer suppressor gene. The high methylation of its promoter region may lead to the silencing or inhibition of SOX1 gene expression, which in turn promotes the proliferation and migration of cancer cells, and ultimately promotes the occurrence and progression of cancer [12].

SOX1 gene methylation has been identified as a key independent predictor of cervical cancer development and progression, particularly as a biomarker for high-grade precancerous lesions and early cervical cancer [13]. The hypermethylation status of its promoter region is particularly essential in the diagnosis and monitoring of cervical adenocarcinoma. The significant difference in the methylation level of SOX1 in ADC compared with SCC and normal cervical tissue indicates its potential as a specific diagnostic marker for cervical adenocarcinoma [14]. Further studies have shown that the methylation status of genes such as SOX1 and POU4F3 can effectively distinguish CIN3⁺ in atypical adenocyte patients, showing excellent diagnostic performance [15]. In addition, SOX1 and SOX14, as methylation biomarkers, exhibit similar sensitivity and greater specificity compared to HR-HPV testing for all cervical cancer types, including SCC and ADC [16].

The role of SOX1 methylation in the diagnosis of SCC has been widely recognized. However, its effectiveness in the early diagnosis of ADC remains to be determined and further clinical studies and systematic evaluation are warranted. This study systematically evaluated the clinical accuracy of SOX1 gene methylation in the diagnosis of different types of cervical cancer through meta-analysis and comprehensive literature, aiming to provide statistical evidence for its application in the diagnosis of SCC and ADC.

Methods

Search strategy

A comprehensive literature search was conducted using both MeSH terms and free words, including “Uterine Cervical Neoplasms,” “SOX1 (SRY-box transcription factor),” “diagnosis,” “sensitivity,” and “specificity.” This search was meticulously designed based on the PICOS (Population, Intervention, Comparison, Outcome, and Study Design) framework of evidence-based medicine to ensure thorough coverage. The databases searched included PubMed, Web of Science, Cochrane Library, China National Knowledge Infrastructure (CNKI), and China Wanfang databases. The search was not restricted by language or country and included all records available up until March 2024. This strategy was aimed at capturing a comprehensive dataset to provide robust diagnostic insights into the utility of SOX1 methylation in cervical cancer.

Inclusion and exclusion criteria

This study included published clinical research that investigated the relationship between SOX1 methylation and its diagnostic performance for cervical cancer from the inception of the database to March 2024. Eligible studies needed to provide result indicators such as true positives (TP), false positives (FP), true negatives (TN), and false negatives (FN) that could be directly or indirectly extracted to construct a 2 × 2 contingency table. Only studies involving histopathologically confirmed cervical cancer patients were considered, with a minimum sample size of five cases per study necessary to mitigate selection bias. Additionally, full texts and complete data sets had to be available for inclusion.

Excluded were studies lacking full text availability or complete data sets, those that did not support the construction of a 2 × 2 table, and studies with fewer than five participants in either the case or control group. Non-clinical studies such as reviews, conference abstracts, meta-analyses, case reports, basic research, or animal studies were also omitted. Publications from the same author or team that reported duplicate data and studies involving other types of malignancies were excluded to maintain data integrity and relevance.

Relevant data were organized using Endnote literature management software. Two researchers independently screened the collected studies based on the defined inclusion and exclusion criteria. The initial screening involved removing duplicate studies and excluding irrelevant publications by reviewing titles and abstracts. Subsequent in-depth full-text reviews excluded studies that failed to meet the inclusion standards or provide necessary diagnostic data. Discrepancies between reviewers were resolved through discussion.

Data extraction and quality assessment

Two independent researchers meticulously extracted essential data from each included study, compiling details such as the first author’s name, year of publication, true and false positives (TP/FP), true and false negatives (TN/FN), sample sizes for experimental and control groups, sample types, diagnostic methods, pathological types, and control group composition into a structured table. To evaluate the validity and applicability of the studies, the QUADAS tool was employed, assessing potential biases and clinical utility across four domains: Selection of Studies, Index Test, Reference Standard, and Flow and Timing. Each domain’s bias risk was classified as “Low,” “High,” or “Unclear,” based on responses to targeted questions that also considered the clinical implications of the findings. Quality assessments were independently conducted by the researchers, with any discrepancies resolved through discussion [17].

Statistical analysis

In this study, Meta-Disc1.4 and Stata15.1 software were used for statistical analysis [18, 19]. Threshold effect is a key source of heterogeneity in diagnostic tests. First, Meta-Disc1.4 software was used to plot and analyze the receiver operating characteristic (ROC) curve, with Spearman’s correlation coefficient to evaluate the threshold effect. Threshold effect existed when the ROC curve was “one-arm” and P < 0.05. If there was no significant threshold effect, Stata15.1 would be used to combine the effect size, including sensitivity, specificity, positive/negative likelihood ratio (PLR/NLR), and diagnostic odds ratio (DOR), with the area under the curve (AUC) of ROC calculated to evaluate diagnostic performance. Heterogeneity was determined by Cochran Q test and I² statistics. When P > 0.05 and I²<50%, the heterogeneity was low, and the fixed-effects model was used to combine the effect size. Otherwise, the random-effects model was applied. Fagan plot was used to evaluate the clinical application value of SOX1 methylation, with Deek’s funnel plot to evaluate publication bias. P < 0.1 indicated the presence of publication bias. In the face of significant non-threshold effect heterogeneity, the potential causes would be explored by subgroup analysis.

Results

Literature screening, characteristics of included studies, and quality assessment

According to PRISMA’s statement [20], a total of 172 articles were obtained in the preliminary search, and 96 were left after being de-duplicated by Endnote software. By reading the title, abstract and keywords in detail, we excluded 61 papers that were not relevant to the study, such as reviews, meta-analyses, and conference abstracts. After reviewing the full text of the remaining 35 records, 18 records that could not extract 2 × 2 tables, 4 non-diagnostic studies and 1 study with a sample size of less than 5 were excluded (Fig. 1). A total of 12 articles were finally identified [16, 21,22,23,24,25,26,27,28,29,30,31,32], including 18 studies with a total sample size of 3213 patients, of which 1024 were experimental and 2189 were control. In these included studies, the pathologic types of cervical cancer consisted of SCC, ADC, and mixed types of SCC and ADC, or not explicitly specified. Specifically, seven studies focused on SCC, four studies explored ADC, and the other seven studies involved cervical cancer that contained multiple pathological types or did not specify the pathological type. In terms of sample types, 9 studies used cervical tissue samples, while 9 studies were analyzed based on exfoliated cervical cells. These studies covered different cervical cancer pathological types and sample types, using methylation detection methods such as quantitative methylation-specific PCR (qMSP), bisulfite sequencing PCR (BSP) and pyrosequencing (Table 1).

Fig. 1

Study selection flowchart

Table 1 Baseline characteristics of included studies for meta-analysis

The QUADAS tool quality assessment was conducted for all the included studies. Evaluation results showed that the quality of most of the literature was medium to low risk, especially in case selection and evaluation trials, which was related to the fact that some studies did not clearly indicate whether cases were selected continuously or case-control study designs were not avoided. The risk of clinical practicability was mainly due to case selection, which was limited to patients undergoing surgery or colposcopic biopsy (Table 2).

Table 2 Details of quality assessment with QUADAS scale diagnostic quality evaluation form

Heterogeneity test

In this review, Meta-Disc1.4 software was used to perform a threshold effect analysis to examine the Spearman correlation coefficient between sensitivity and (1-specificity) of SOX1 methylation in the diagnosis of cervical cancer. The correlation coefficient was 0.227, with a P value of 0.349, exceeding the threshold of 0.05. The SROC curve did not present a “one-arm” pattern, suggesting that heterogeneity in diagnostic tests was not caused by threshold effects, but might result from non-threshold effects. In addition, Cochran-Q test for the DOR showed 56.43, with 18 degrees of freedom (P < 0.01, and I² = 68.1% > 50%), further confirming the existence of significant heterogeneity due to non-threshold effects. Based on these findings, it is appropriate to apply a random-effects model to merge the effect sizes.

Combination of effect size of diagnostic value

In this meta-analysis, SOX1 gene methylation exhibited a high pooled sensitivity of 0.88 (95%CI: 0.84–0.91) and specificity of 0.90 (95%CI: 0.80–0.96) in the diagnosis of cervical cancer (Fig. 2A and B). The PLR and NLR were 9.23 (95%CI: 4.18–20.37 and 0.13 (95% CI: 0.10–0.18), respectively (Fig. 3A and B). The DOR was 68.95 (95%CI: 27.63-172.07), with an AUC of 0.92, indicating that SOX1 gene methylation has high accuracy in the diagnosis of cervical cancer (Figs. 4 and 5). These data highlight the potential value of SOX1 methylation as a diagnostic tool for cervical cancer.

Fig. 2

Pooled results of meta-analysis of the diagnostic value of SOX1 methylation of cervical cancer: (A) sensitivity; (B) specificity

Fig. 3

Pooled results of meta-analysis of the diagnostic value of SOX1 methylation of cervical cancer: (A) positive diagnostic likelihood ratio; (B) negative diagnostic likelihood ratio

Fig. 4

Pooled diagnostic odds ratio of meta-analysis of the diagnostic value of SOX1 methylation of cervical cancer

Fig. 5

Area under the ROC curve of meta-analysis of the diagnostic value of SOX1 methylation of cervical cancer

Subgroup analysis

In this meta-analysis, subgroup analyses based on total sample size, sample type, and control group composition were performed for included studies to gain insight into potential sources of heterogeneity. In subgroup analysis of total sample size, we observed that studies with sample size ≥ 200 showed higher diagnostic accuracy (AUC: 0.97) compared to studies with sample size < 200 (AUC: 0.90), indicating that studies with larger sample sizes were able to provide more accurate cervical cancer diagnosis information. Regarding sample types, the diagnostic accuracy of exfoliated cervical cell samples (AUC: 0.94) exceeded that of cervical tissue samples (AUC: 0.91), suggesting that SOX1 gene methylation analysis using exfoliated cervical cells is a non-invasive and highly effective method for cervical cancer screening. Subgroup analysis of the control group showed that the control group without precancerous lesions had better sensitivity, specificity, PLR/NLR, DOR and AUC than the control group with precancerous lesions (Table 3).

Table 3 Subgroup analysis results of meta-analysis

Although subgroup analysis did not explicitly identify the source of heterogeneity, I² values of DOR were lower in the cervical tissue group (38.20%) and the control group containing precancerous lesions (3.60%), suggesting relatively little intra-group heterogeneity in these two subgroups. These findings highlight the significance of considering sample size, sample type, and control group composition when designing and interpreting cervical cancer diagnosis studies. Detailed data of sensitivity, specificity, PLR/NLR, DOR, AUC and I² of heterogeneity are listed in Table 3.

Publication bias analysis

In this study, Stata 15.1 software was applied to detect publication bias, specifically using Deek’s funnel plot method. The results of the analysis showed a P-value of 0.98 (P > 0.10), indicating that there was no evident publication bias in the included studies (Fig. 6).

Fig. 6

Deeks’ funnel plot for publication assessment of the diagnostic value of SOX1 methylation of cervical cancer

Evaluation of clinical efficacy of SOX1 gene methylation in the diagnosis of cervical cancer

Using Fagan plot (Fig. 7), a positive SOX1 methylation result increased the posterior probability of cervical cancer to 70%, while a negative result reduced the probability to 3%, when the prior probability of cervical cancer diagnosis was set at 20%. This result highlights the significant clinical application value of SOX1 gene methylation detection in the diagnosis of cervical cancer, indicating its high utility as a diagnostic tool for cervical cancer.

Fig. 7

Fagan’s nomogram in detecting diagnostic probability of the diagnostic value of SOX1 methylation of cervical cancer

Meta-analysis of SOX1 gene methylation in cervical adenocarcinoma and cervical squamous cell carcinoma

SOX1 gene methylation had a pooled sensitivity of 0.88 (95%CI: 0.83–0.92), and specificity of 0.92 (95%CI: 0.50–0.99) in the diagnosis of ADC. The PLR and NLR were 11.00 (95%CI: 1.17-103.78) and 0.13 (95%CI: 0.08–0.20), respectively. The DOR was 87.57 (95%CI: 7.05-1087.44), with an AUC under the SROC of 0.89, indicating that the SOX1 gene methylation test is highly accurate for the diagnosis of ADC. In addition, Deek’s funnel plot (P = 0.86, P > 0.10) presented no significant publication bias in the included studies concerning ADC. The above results added credibility and validity to the findings (Supplemental Figs. 1–5).

The pooled sensitivity and specificity of SOX1 methylation in the diagnosis of SCC were 0.85 (95%CI: 0.77–0.91) and 0.98 (95%CI: 0.87-1.00). The PLR and NLR were 37.44 (95% CI: 5.77–242.80) and 0.15 (95% CI: 0.10–0.24), respectively. The diagnostic odds ratio reached 245.87 (95% CI: 26.49-2282.40). With an AUC under SROC of 0.93, these results highlight the significance of SOX1 gene methylation in the diagnosis of SCC. The analysis of Deek’s funnel plot showed a P-value of 0.34 (P > 0.10), indicating no significant publication bias in the SCC studies analyzed. This enhanced the reliability and validity of the study results, and provided a solid scientific basis for the application of SOX1 gene methylation in the diagnosis of cervical squamous cell carcinoma (Supplemental Figs. 6–10).

Discussion

In this meta-analysis, a total of 12 articles involving 18 studies were included. Quality assessment by applying the QUADAS tool for diagnostic test accuracy showed that the majority of studies were of moderate to low-risk quality, indicating high study quality overall. Deek’s funnel plot analysis was used to test publication bias, and the P-value was 0.98 (P > 0.10), indicating no publication bias in the included literature.

The combined sensitivity of SOX1 gene methylation in the diagnosis of cervical cancer was 0.88 (95%CI: 0.84–0.91), with the combined specificity being 0.90 (95%CI: 0.80–0.96), indicating that SOX1 gene methylation has high sensitivity and specificity in the diagnosis of cervical cancer. Sensitivity and specificity are the key indicators of diagnostic accuracy, with their values close to 1 reflecting a good diagnostic effect. PLR and NLR were considered as indicators reflecting both sensitivity and specificity. The PLR was 9.23 (95%CI: 4.18–20.37 in this meta-analysis, suggesting that the probability of cervical cancer was increased by 5.85 times when SOX1 gene methylation was positive [33]. An NLR of 0.13 (95% CI: 0.10–0.18) indicates that when the test is negative, it reduces the likelihood of cervical cancer to 13%, showing a low false negative rate and a good exclusion effect [33].

DOR and AUC under the SROC are indicators to evaluate the overall accuracy of diagnostic tests. The DOR of this study was 68.95 (95%CI: 27.63-172.07) with an AUC of 0.92, further confirming the high accuracy and excellent discriminative efficacy of SOX1 methylation in the diagnosis of cervical cancer [34, 35]. Fagan plot analysis emphasizes the significance of SOX1 methylation detection in clinical diagnosis of cervical cancer and provides strong support for clinical decision-making. These findings provide a solid scientific basis for using SOX1 gene methylation as a biomarker for the early diagnosis of cervical cancer.

Subgroup analysis revealed that studies with large sample sizes (AUC = 0.97, DOR = 62.08) were superior to those with small sample sizes (AUC = 0.90, DOR = 41.03) in cervical cancer diagnosis accuracy, highlighting the importance of SOX1 gene methylation in extensive screening. The diagnostic performance of exfoliated cervical cell samples (AUC = 0.94, DOR = 72.24) exceeded that of cervical tissue samples (AUC = 0.91, DOR = 23.01), indicating its advantages in the diagnosis of non-invasive cervical cancer. Analysis of the composition of the control group showed that the control group without precancerous lesions was superior to the control group with precancerous lesions on various diagnostic indicators, which emphasized the high diagnostic value of SOX1 methylation in distinguishing cervical cancer from non-cancerous states. These findings provide a scientific basis for using SOX1 methylation to improve cervical cancer screening and diagnosis.

Subsequently, we conducted an analysis of studies focusing exclusively on ADC and SCC. In the case of ADC, SOX1 methylation demonstrated that sensitivity and specificity are of 0.88 and 0.92, respectively. Conversely, for SCC, the sensitivity was 0.85 and the specificity reached 0.98. These findings underscore the robust diagnostic performance of SOX1 gene methylation in identifying both ADC and SCC, aligning with the findings reported by Wang R et al. [16]. In particular, the diagnostic sensitivity of ADC was better than that of SCC. While the specificity for adenocarcinoma (ADC) was marginally lower, SOX1 methylation still exhibited potential to enhance ADC detection rates. Furthermore, the diagnostic accuracy for ADC and squamous cell carcinoma (SCC) was evident from their respective Area Under the Curve (AUC) values of 0.89 and 0.93, demonstrating high diagnostic efficacy of SOX1 gene methylation for these cervical cancer types. However, the limited number of studies specifically addressing ADC poses a potential risk of statistical bias, highlighting the need for further research to ascertain the relationship between SOX1 methylation and cervical adenocarcinoma more reliably and to enhance the statistical robustness of these analyses. Current findings suggest that the diagnostic performance of SOX1 methylation in ADC is comparable to that in SCC. Its clinical application could significantly reduce the rates of missed diagnoses in early ADC screening, thereby serving as a critical molecular tool for enhancing the effectiveness of cervical cancer screening and early diagnosis programs.

The application of DNA methylation technology in cervical cancer diagnosis has been widely concerned. A meta-analysis by Helen Kelly et al. indicated that DNA methylation levels were significantly higher in high-grade cervical endothelial lesions (CIN2⁺ and CIN3⁺) compared to samples below or equal to Grade 1 cervical endothelial lesions (CIN1) [6]. In addition, a set of DNA methylation markers (including CADM1, MAL, MIR-124-2, FAM19A4, POU4F3, EPB41L3, PAX1, SOX1) demonstrated greater specificity in diagnostic grading than cytological tests for ASCUS+, and its sensitivity exceeded that of HPV16/18 genotyping, which further confirmed the critical role of DNA methylation in cervical cancer screening, diagnosis and triage [6]. The existing meta-analysis showed that when SOX1 methylation was used to distinguish high-grade squamous intraepithelial lesions (HSIL), CIN3⁺ or cervical cancer, the AUC was 0.82, among which the combined sensitivity was 0.71 with a specificity of 0.64. Compared with the diagnostic efficacy of SOX1 methylation in distinguishing cervical cancer samples from non-cancer samples, the results of this study showed higher accuracy in our study [36].

A meta-analysis by Huang et al. evaluated the association between SOX1 promoter hypermethylation and cervical cancer or squamous intraepithelial lesions (SIL), finding a 4.20-fold increased risk of SOX1 promoter hypermethylation in HSIL and 41.26-fold increased risk in cervical cancer compared to controls [37]. The combined sensitivity of SOX1 methylation in distinguishing patients with cervical cancer was 0.85 (95%CI: 0.81–0.88), and the specificity was 0.72 (95%CI: 0.69–0.75), with an AUC of 0.925 [37]. Compared with the results of Huang et al., the sensitivity, specificity and AUC indicators obtained in this study all presented advantages, which might be caused by the influence of factors such as the larger sample size of cervical cancer included, the geographical distribution of the study objects mainly in Asia and the large sample size of ADC. This study is the first systematic meta-analysis of the role of SOX1 methylation in the diagnosis of ADC, and a satisfactory conclusion has been obtained, which provides strong scientific support for the use of SOX1 methylation as a biomarker for the early diagnosis of cervical cancer.

One significant area for future research is the potential integration of SOX1 methylation with other biomarkers, particularly HPV status. HPV infection, especially with high-risk types, is the primary cause of cervical cancer, and combining SOX1 methylation with HPV testing could significantly enhance diagnostic sensitivity [38]. The interplay between SOX1 methylation and HPV status has been explored in some studies, but further research is needed to understand how these biomarkers interact and how their combined use could improve early detection [39]. SOX1 methylation may help identify high-risk HPV infections that are more likely to progress to cancer, making it a valuable tool in the triage of HPV-positive women. Future studies should focus on combining SOX1 methylation with HPV testing and cytological findings to optimize the diagnostic accuracy for cervical cancer.

Moreover, integrating SOX1 methylation with liquid-based cytology and HPV testing could lead to a more comprehensive cervical cancer screening protocol [40]. Current screening methods rely on HPV testing and cytology, both of which have limitations in certain clinical scenarios. Adding SOX1 methylation to these screening protocols could provide a more accurate and reliable means of detecting cervical cancer, particularly in women with atypical cytology or HPV-negative results [40].

Beyond its diagnostic utility, SOX1 methylation may have applications in monitoring disease progression and treatment response. As a stable biomarker, SOX1 methylation could be used for non-invasive monitoring of treatment efficacy, allowing clinicians to track tumor dynamics and detect recurrence at an early stage [24]. Current cervical cancer management strategies often rely on clinical assessments and imaging, which may not always detect early recurrence [37]. Therefore, integrating SOX1 methylation testing into routine follow-up care could provide additional value by enabling timely interventions in cases of treatment failure or recurrence.

Furthermore, SOX1 methylation has the potential to be used as a prognostic tool in cervical cancer. Monitoring the methylation status of SOX1 over time could help assess the risk of disease progression, providing insights into tumor behavior and patient prognosis [41]. Future studies should explore the role of SOX1 methylation in assessing tumor burden, predicting relapse, and tailoring individualized treatment plans.

The variability in methylation detection methods across studies presents a challenge in comparing results and assessing the generalizability of findings. Different studies employed a variety of techniques, including quantitative methylation-specific PCR (qMSP), bisulfite sequencing PCR (BSP), and pyrosequencing, each with different levels of sensitivity and specificity. Standardizing methylation detection methods would not only improve the comparability of results but also enhance the reproducibility of findings across laboratories. We recommend that future studies explore the comparative diagnostic accuracy of different methylation detection methods and work toward establishing standardized protocols for SOX1 methylation testing in clinical settings. This will facilitate the broader adoption of SOX1 methylation as a routine biomarker for cervical cancer screening and diagnosis.

This meta-analysis provides valuable insights into the diagnostic potential of SOX1 methylation in cervical cancer; however, several limitations should be considered. Firstly, significant heterogeneity was observed across studies due to variations in study design, sample types, methodologies, and methylation detection techniques. The absence of standardized detection methods hampers the ability to effectively compare results and assess the most reliable diagnostic approach. To improve consistency, future studies should focus on using homogeneous data and standardized techniques, incorporating multi-center trials to control for these variables. Another limitation is the inclusion of small sample studies, which may have reduced statistical power and affected the robustness of the results. While these studies enhance diversity, it is recommended that future research prioritize large-scale, well-designed studies to validate SOX1 methylation’s diagnostic potential across different populations and clinical settings. Furthermore, the exclusion of unpublished studies or gray literature may have introduced publication bias, and including these sources could provide a more comprehensive view of the evidence. The lack of detailed HPV status and cytology data in several studies limits the ability to directly compare SOX1 methylation with these established diagnostic methods. Integrating HPV and cytology findings with SOX1 methylation in future studies will help assess their combined diagnostic value and improve cervical cancer detection accuracy. Additionally, the limited exploration of the correlation between SOX1 methylation and clinicopathological features, such as tumor stage and lymph node metastasis, restricts understanding of its prognostic value. Investigating these relationships in future studies will provide a better understanding of SOX1 methylation’s role in disease progression and prognosis. Finally, the scarcity of literature exploring SOX1 methylation in adenocarcinoma (ADC) limits the stability of the meta-analysis and its overall diagnostic test accuracy for cervical cancer. Future studies should expand on this area to provide more robust evidence for ADC detection.

Conclusion

In conclusion, this meta-analysis indicates that SOX1 gene methylation is of high value in the diagnosis of SCC and ADC, especially in large sample sizes and in cervical exfoliated cell samples, showing higher sensitivity and specificity. These findings provide an available biomarker for cervical cancer screening and diagnosis, being expected to optimize the early detection and management of cervical cancer.