The Selfie Study- Assessing Novel Markers for Cervical Cancer Screening From Self-collected Samples (SELFIE)

Worldwide, cervical cancer remains the fourth most common cancer and fourth leading cause of cancer deaths among women, with the greatest burden occurring in low-resource settings that lack effective screening and treatment. Even in the United States (U.S.), where Pap cytology screening has resulted in dramatic declines in cervical cancer incidence and mortality, thousands of new cases and related-deaths still occur every year, most commonly among underserved women who face barriers to accessing screening and/or treatment. The recent implementation of human papillomavirus (HPV) DNA testing as a primary strategy for cervical cancer screening has the potential to alleviate these disparities by improving the sensitivity for cervical precancer detection compared to Pap cytology, while providing greater long term reassurance following a negative HPV test. Moreover, HPV testing can be successfully performed on self-collected specimens, offering the possibility of expanding access to cervical cancer screening among hard-to-reach, underserved women.

Despite the many advantages of primary HPV screening, the current challenge is optimizing triage testing to determine who among the many women testing HPV positive are at high-risk and require immediate colposcopy referral or treatment, while avoiding unnecessary harms among women at low-risk. A molecular triage test that can be conducted from the same primary screening sample (i.e., reflex testing) is particularly attractive for both high-and low-resource settings, particularly if it works from self-collected samples. Previously, DNA methylation of candidate host cell genes has been evaluated in self-collected samples and achieved acceptable performance for triage of HPV-positive women. In studies using clinician collected samples, there is a suggestion that methylation of carcinogenic HPV genotypes has better clinical performance (higher sensitivity/specificity) compared to host gene methylation; however, HPV methylation has not been evaluated in self-collected specimens. Because HPV DNA is present in only a small subset of infected cells in the lower genital tract, particularly in cervical precancers, it is likely that the signal-to noise ratio in self-collected samples is better for HPV compared to host gene methylation, resulting in improved specificity for cervical precancer detection. Evaluating the feasibility of HPV DNA methylation testing from self-collected samples is essential for determining the extent to which this assay can address the critical need for HPV triage in high- and low resource settings. In collaboration with the Cancer Genome Research Laboratory, a high-throughput, low-cost next-generation bisulfite sequencing assay that detects methylation of the 12 most carcinogenic HPV genotypes has been developed. Further, this assay includes a panel of host genes, including those that have been previously evaluated in self-collected specimens. The proposal is to test this assay in paired self and clinician-collected samples from 1,000 women enrolled in an ongoing prospective study of women undergoing cervical cancer screening and colposcopy at the George Washington University (GWU). The hypothesis is that HPV and host DNA methylation will show non-inferior sensitivity and specificity for detection of cervical precancer in self-collected compared to clinician-collected samples and that HPV methylation will have higher absolute specificity compared to host methylation in both sample types.