Introduction: Cervical cancer remains a leading cause of cancer mortality worldwide. Infection with human papillomavirus (HPV) is necessary for the development of cervical cancer but the factors related to persistent infection are poorly understood. Many potential environmental and behavioural exposures related to the development of cervical cancer have been studied, but randomised control trials are uncommon, and the reliability of evidence gained from observational studies remains unclear. Meanwhile heritability of cervical cancer has been estimated to approach 36% but genome-wide association exploration has been limited. There is increasing evidence for an association between DNA methylation and cervical carcinogenesis, with potential for use as a molecular marker for early detection of high-grade cervical lesions.

Methods: We performed an umbrella review of the literature regarding any environmental or behavioural exposure relating to cervical outcomes. We did a genome-wide association study (GWAS) in women with invasive cervical cancer (ICC) and CIN3 (N=4769) and female controls (N=145,545) in the UK Biobank. Epigenetic signatures were examined in an epigenome-wide study on the Illumina 850K array in liquid-based cytology samples from a cohort of 241 women. Through a systematic review and meta-analysis, we examined the diagnostic test accuracy of HPV DNA methylation as a test for cervical cancer.

Results: From 171 meta-analyses we found strong evidence relating to HIV positivity (RR=2.20, (95%CI=1.89-2.54)), vaginal dysbiosis (RR=1.59(1.40-1.81)) and chronic immunosuppressive therapy (RR=1.33(1.89-2.54)) for cervical cancer outcomes. Novel genetic loci rs1017546 (PAX8;P=1.07x10-9) and rs27069 (CLPTM1L;P=1.51x10-9), and previously reported rs9272050 (HLA-DQA1;P=2.51x10-28) were strongly associated to cervical cancer. EWAS further identified two independent CpG sites in PAX1 (P=3.90x10-12) and NREP-AS1 (P=2.45x10-10) genes that were able to detect CIN3 or Cancer with an AUC=0.92(0.92-0.92) in the discovery cohort and AUC=0.77(0.76-0.79) in an independent cohort. The HPV16 L1 gene had the strongest association with CIN2 or worse, with a pooled sensitivity of 77%(63-87%), specificity 64%(55–71%), and AUC of 0.73(0.69–0.77).

Conclusions: This work evaluates and presents the strongest environmental risk factors and substantial new evidence for genetic susceptibility to cervical cancer, mediated through disruption in immune function and apoptotic pathways. PAX1, NREP-AS1 and HPV16 L1 are identified as promising methylation markers for triage of women identified as hrHPV positive at cervical screening.