Assessment of six markers for cervical (pre)cancer detection 53 2 other, non-cervix gynaecologic cancers. This concept has been corroborated for other methylation markers by recent studies 46-48. Methylation analysis has shown before to particularly detect high-grade CIN lesions associated with a persistent HPV infection of at least 5 years, as a surrogate for a more advanced lesion 27, 49. In a recent methylation study on HPV-positive cervical samples, including host-cell genes GHSR, SST and ZIC1, it was found that 72% of CIN3 showed a methylation profile similar to cervical cancer 26. These methylation-positive CIN3 lesions are therefore considered to have a substantial higher short-term risk of progression to cancer. The sensitivity of 80% for CIN3 as found in our study corresponds to these previous findings, and suggests that a methylation-negative result may be indicative of lesions with a low chance of progression to cervical cancer. This hypothesis is further supported by a recent study evaluating a methylation panel of host-cell and viral genes that showed the ability to differentiate between regressive and progressive CIN2 in young women 50. Hence, methylation analysis may stratify between immediate treatment and close surveillance, preventing overtreatment and the associated cervical morbidity, which is especially relevant for women of childbearing age 51. This association between methylation status and regression probability is subject of further studies 52. Self-sampling is increasingly offered in cervical cancer screening programs as an alternative or additive to clinician-taken cervical samples to reduce screening barriers and increase screening participation 14, 53, 54. The candidate host-cell methylation markers ASCL1 and LHX8 were originally identified on self-collected cervico-vaginal samples and validated in independent series of HPV-positive self-samples with AUCs between 0.88 and 0.90 for CIN3 detection 35. Previous research demonstrated that the performance of methylation markers on cervical samples and self-collected samples is not necessarily similar 55, as these two sample types can differ in cellular composition. Of interest, this study validated ASCL1 and LHX8 on HPV-positive cervical samples with an equally well performance compared with the previous results from HPV-positive self-samples. This highlights compatibility of this marker panel with both self-collected and cliniciancollected cervical samples, which is highly beneficial for a triage test. In addition, ASCL1 and LHX8 have been evaluated in a South African, HIV-positive study cohort as a primary screening tool for the detection of CIN3+ 56. Both markers showed a good performance with an AUC of 0.79 for ASCL1 and 0.81 for LHX8. These data support the diagnostic potential of ASCL1 and LHX8, although direct extrapolation is not feasible due to the HIV status and the absence of a well-organised screening program in South Africa.
RkJQdWJsaXNoZXIy MjY0ODMw