Objectives: DNA methylation analysis of cancer-related genes is a promising tool for HPV-positive women to identify those with cervical (pre)cancer (CIN3+) in need of treatment. However, clinical performance of methylation markers can be influenced by the sample type utilized. We describe a multiplex quantitative methylation-specific PCR that targets FAM19A4 and mir124-2 loci, to detect CIN3+ using both HPV-positive lavage- and brush self-samples.
Methods: We determined methylation thresholds for clinical classification using HPV-positive training sets comprising lavage self-samples of 182 women (including 40 with CIN3+) and brush self-samples of 224 women (including 61 with CIN3+). Subsequently, independent HPV-positive validation sets of 389 lavage self-samples (including 78 with CIN3+), and 254 brush self-samples (including 72 with CIN3+) were tested using the preset thresholds. Furthermore, the clinical performance of combined methylation analysis and HPV16/18 genotyping was determined.
Results: Training set analysis revealed similar FAM19A4 and mir124-2 thresholds for both self-sample types to yield highest CIN3+ sensitivity at 70% specificity. Validation set analysis resulted in a CIN3+ sensitivity of 70.5% (95%CI: 60.4-80.6) at a specificity of 67.8% (95%CI: 62.7-73.0) for lavage self-samples, and a CIN3+ sensitivity of 69.4% (95%CI: 58.8-80.1) at a 76.4% (95%CI: 70.2-82.6) specificity for brush self-samples. In combination with HPV16/18 genotyping, CIN3+ sensitivity and specificity were 88.5% (95%CI: 81.4-95.6) and 46.0% (95%CI: 40.4-51.5) for lavage self-samples, and 84.7% (95%CI: 76.4-93.0) and 54.9% (95%CI: 47.7-62.2) for brush self-samples.
Conclusions: FAM19A4/mir124-2 methylation analysis performs equally well in HPV-positive lavage- and brush self-samples to identify women with CIN3+. In combination with HPV16/18 genotyping, significantly higher CIN3+ sensitivities are obtained, at decreased specificity.
Keywords: Cervical cancer; DNA methylation; HPV16/18 genotyping; Human papillomavirus; Molecular screening; Reflex test; Self-sampling.
Copyright © 2016 The Authors. Published by Elsevier Inc. All rights reserved.