ABSTRACT Endometrial cancer is the sixth most common cancer in women globally and has a rising incidence, accounting for 417,000 new diagnoses and more than 97,000 deaths in 2020. Early detection is critical because of a poor advanced-stage prognosis and high relapse risk. Postmenopausal bleeding precedes endometrial cancer in 90% of cases, yet only 5% to 10% of patients with this symptom have an underlying malignancy. Recent cytology research has demonstrated endometrial cancer cells in vaginal samples, suggesting urine and cervicovaginal self-samples may allow convenient and minimally invasive screening. In addition, screening samples for DNA methylation in tumor suppressor genes can increase testing efficacy of triaging patients with postmenopausal bleeding as it does not require intact tumor cells. This study aimed to evaluate the diagnostic performance of endometrial cancer detection using DNA methylation analysis in paired urine, cervicovaginal self-samples, and clinician-taken cervical scrapes. Paired samples from participants of the SOLUTION1 study with endometrial cancer were collected between October 2016 and August 2020. A complete urine void and cervicovaginal self-sample was collected at home, whereas the cervical scrape was taken by a clinician in the operating room before surgery. Each patient had a paired urine, cervicovaginal self-sample, and clinician-taken cervical scrape available for methylation analysis. Controls were collected from the Urine Controls (URIC) biobank and Dutch national cervical cancer screening program. Promoter hypermethylation of the ADCYAP1, BHLHE22, CDH13, CDO1, GALR1, GHSR, HAND2, SST, and ZIC1 genes was tested using quantitative methylation-specific polymerase chain reaction. Optimal 3-marker combinations were determined for each sample type using multivariable logistic regression analysis, and diagnostic performances of each marker and 3-marker panels were assessed by leave-one-out cross validation. A total of 103 endometrial cancer patients were included in this study, with 317 unpaired samples of control women. Methylation levels of all markers were significantly higher in all 3 sample types of endometrial cancer patients compared with healthy control women. In urine, the non–cross-validated area under the curve of the DNA methylation markers ranged between 0.61 and 0.93, in cervicovaginal self-samples between 0.62 and 0.91, and in clinician-taken cervical scrapes between 0.61 and 0.95. Most markers (7/9) showed the highest performance in urine, with the remaining 2 showing best performance in clinician-taken cervical scrapes. The optimal 3-marker combination panels showed area under the curve values of 0.95 (05% confidence interval [CI], 0.92–0.98) for urine, 0.94 (95% CI, 0.90–0.97) for cervicovaginal self-samples, and 0.97 (95% CI, 0.96–0.99) for clinician-taken cervical scrapes. The sensitivity and specificity of the optimal urine testing panel were both 90%, were 89% sensitivity and 92% specificity for cervicovaginal self-samples, and were 93% sensitivity and 90% specificity for clinician-taken cervical scrapes. The leave-one-out cross-validation analysis revealed virtually equal performances in marker panels for each of the 3 testing modalities. The results of this study demonstrate high diagnostic potential of DNA methylation testing in minimally and noninvasive samples for endometrial cancer detection, emphasizing the potential for patient-friendly home-based sample collection.
Read full abstract