102 Chapter 4 lavage (sensitivity 74%; specificity 79%) and brush (sensitivity 88%; specificity 81%) selfcollected samples from screening non-attendees 21. The ASCL1/LHX8 marker panel also demonstrated good triage performance on hrHPV-positive clinician-collected cervical samples 24, 25. Notably, in all studies, samples from women with cervical cancer were positive for the methylation marker panel. Recently, also the utility of methylation testing of ASCL1 and LHX8 for the detection of CIN3+ in urine has been demonstrated 36. To the best of our knowledge, this is the first study evaluating methylation markers in self-collected samples from women who were offered primary HPV self-sampling for cervical cancer screening. The relatively low sensitivity and specificity for CIN3+ of ASCL1/ LHX8 in primary HPV self-sampling may be related to the study population. Previous studies have mainly been performed on self-collected samples in under screened and never-screened women and referral populations 19-23, 34, which may not fully represent the women who attend population-based screening. In these populations, the incidence of CIN3 is higher than in a population of screening attendees 37. Moreover, due to screening at regular intervals, CIN3 lesions identified by population-based screening have a relatively short duration since onset, are relatively small, and are known to have lower methylation levels 14, 16, 22. Furthermore, cervicovaginal self-collected samples have a different cellular composition and proportion of cervical cells compared to cliniciancollected cervical samples. Technical advances in DNA methylation analysis to allow more sensitive and specific assessment in low-input samples will be of interest to optimise methylation-based triage of self-collected samples 38. A low-input may be circumvented in routine setting by use of a larger fraction of the self-sample in the methylation assay. This was not possible in our study given the setting of using restricted leftover sample material. Furthermore, direct cell conversion protocols and automated solutions for methylation analysis offer the ability to improve performance and provide a solution for high-throughput application of methylation assays in cervical cancer screening 39, 40. The strengths of this study include the large sample size and the use of samples from a primary self-sampling trial conducted within the setting of the Dutch cervical cancer screening. Although the study protocol slightly differed from the current national screening protocol for primary HPV self-sampling, our study demonstrates direct triage utility of host-cell DNA methylation markers with satisfactory performance on hrHPVpositive self-collected screening samples. These data are particularly relevant for settings where cytology is not routinely available. Another strength is the IMPROVE study design, which allowed the comparison of methylation data on hrHPV-positive self-collected
RkJQdWJsaXNoZXIy MjY0ODMw