A Multicenter Cohort Study on <scp>DNA</scp> Methylation for Endometrial Cancer Detection in Cervical Scrapings

Xiao Ma, Xiaojun Chen, Jing Liang, Jingbo Zhang, Qixi Wu, Dong Wang, Xianghua Huang, Dan Zi, Dexin Chen, Hua Wan, Li Qu, Zhaoyun Jiang, Wenyu Shao, Jie Sun, Luyuan Chang, Yunchao Liu, Qin Zhang, Y. Li, Yani Ding, Biao Tang, Fang‐Jie Zhao, Hanqing Zhao, Dongyan CaoCancer 2024 · DOI

TL;DR

This paper introduces EMPap, a novel DNA methylation classifier using cervical scrapings, to improve early detection of endometrial cancer (EC). EMPap demonstrates high accuracy, outperforming traditional methods like transvaginal ultrasonography (TVS) in certain scenarios, and shows promise for identifying aggressive EC subtypes and premalignant lesions, potentially reducing unnecessary invasive procedures.

What Problem Does This Paper Address?

Endometrial cancer (EC) is a common gynecological malignancy with a rising incidence. Current diagnostic approaches for suspected EC, particularly in cases of abnormal uterine bleeding (AUB), often rely on invasive procedures like hysteroscopy or dilation and curettage (D&C). Transvaginal ultrasonography (TVS) is used to measure endometrial thickness (ET), but its accuracy can be unsatisfactory, leading to a significant number of women undergoing unnecessary invasive procedures. There's a clear need for a non-invasive, accurate, and reliable method for early detection and triage of patients at risk for EC.

Methods

The researchers developed a DNA methylation classifier called EMPap. This classifier integrates the methylation status of two key genes, BHLHE22 and CDO1, with patient demographic data (age and body mass index - BMI). The methylation levels were measured using quantitative methylation-specific PCR (qMSP). A logistic regression model was used to construct the classifier, generating an "endometrial cancer methylation (EM) score." This model was trained on a cohort of women with suspected EC and then validated on an independent cohort. The performance of EMPap was evaluated using metrics like sensitivity, specificity, area under the receiver operating characteristic curve (AUC), positive predictive value (PPV), and negative predictive value (NPV). They also investigated the impact of cervical scraping timing relative to biopsy procedures on EMPap's accuracy.

Datasets

The study utilized data from a multicenter cohort of 1297 women aged 18 or older with a high suspicion of EC or newly diagnosed EC from seven hospitals in China. The cohort was divided into a training set (n=630) and a validation set (n=772). The training set was used to develop the EMPap model, and the validation set was used to assess its performance.

Key Insights and Ideas

  • Cervical scrapings as a viable sample source: The study reinforces that cervical scrapings, a minimally invasive sample, can effectively capture DNA methylation markers indicative of EC, offering a convenient alternative to more invasive procedures.
  • Timing matters for sample collection: The timing of cervical scraping relative to intrauterine procedures like biopsies significantly impacts EMPap's accuracy. Collecting samples before a biopsy yielded more accurate results, suggesting that intrauterine procedures might influence the presence of exfoliated cells or DNA in cervical samples.
  • EMPap's potential for aggressive subtypes: EMPap showed promise in detecting Type II ECs (non-endometrioid) and premalignant lesions (AH/EIN), which are often more aggressive and harder to detect. This is a crucial advancement as current methods may struggle with these subtypes.
  • EMPap as a superior alternative to TVS for PMB: For patients with postmenopausal bleeding (PMB), EMPap demonstrated significantly better diagnostic performance (higher AUC, sensitivity, and specificity) compared to TVS, suggesting it could be a more reliable initial screening tool in this population.
  • EMPap reflects disease progression: The EM scores progressively increased from normal endometrium through endometrial hyperplasia (EH), atypical hyperplasia (AH/EIN), and finally to EC, indicating that EMPap can potentially reflect the continuum of endometrial lesion progression.

Key Results

  • Overall Performance: EMPap achieved an overall AUC of 0.93, with a sensitivity of 90.3% and specificity of 89.3% in the entire cohort.
  • Type II EC Detection: EMPap correctly identified 87.5% of Type II EC cases.
  • Premalignant Lesion Detection: EMPap detected over 50% of premalignant lesions (specifically, 66.7% of AH/EIN cases).
  • Comparison with TVS for PMB: In patients with PMB, EMPap showed superior sensitivity (100% vs. 82.0%) and specificity (85.2% vs. 38.5%) compared to TVS.
  • Impact of Scraping Timing: Excluding samples collected within 10 days after biopsy increased EMPap's accuracy to 92.3% in the overall cohort.
  • Subgroup Performance: EMPap performed consistently well across various patient demographics, symptoms (like AUB), and clinical characteristics (age, BMI, hypertension, diabetes).

Limitations and Future Work

  • Sample Size for Specific Subtypes: While promising, the study acknowledges a limited sample size for certain subtypes, particularly Type II EC and premalignant lesions, suggesting that further validation with larger cohorts is needed.
  • Impact of Intrauterine Procedures: The study identified that intrauterine procedures can affect EMPap accuracy, recommending sample collection before such procedures. Future work could explore standardized protocols to mitigate this impact.
  • Molecular Subtyping: The authors note that due to economic and logistical challenges in their multicenter study, molecular subtyping was not included. They recognize its importance and plan further research to incorporate it for a deeper understanding and improved detection methods.
  • Cost and Laboratory Standardization: Variations in medical conditions, facilities, and laboratory standards across the participating hospitals were acknowledged as a factor, suggesting a need for standardization in future large-scale implementations.

Summary updated

Abstract

This study demonstrated the potential of EMPap as an effective tool for EC detection. Despite the limited sample size, EMPap showed promise for identifying type II EC and detecting over 50% of premalignant lesions. As a DNA methylation classifier, EMPap can reduce unnecessary uterine interventions and improve diagnosis and outcomes.
BibTeX
@inproceedings{ma2024multicenter,
  author = {Xiao Ma and Xiaojun Chen and Jing Liang and Jingbo Zhang and Qixi Wu and Dong Wang and Xianghua Huang and Dan Zi and Dexin Chen and Hua Wan and Li Qu and Zhaoyun Jiang and Wenyu Shao and Jie Sun and Luyuan Chang and Yunchao Liu and Qin Zhang and Y. Li and Yani Ding and Biao Tang and Fang‐Jie Zhao and Hanqing Zhao and Dongyan Cao},
  title = {{A Multicenter Cohort Study on <scp>DNA</scp> Methylation for Endometrial Cancer Detection in Cervical Scrapings}},
  booktitle = {Cancer},
  year = {2024},
  doi = {10.1002/cam4.70361},
}
Figure