The proposed study seeks to compare the diagnostic performance of Human Papillomavirus (HPV) testing in self-collected samples via the Eve Medical self-collection system© (Eve) with standard physician-collected samples for the detection of cervical intraepithelial neoplasia grade 1 or worse (CIN1+) and cervical cancer among women referred for colposcopy. The performance of the Eve sample will also be compared with that of a second self-sample via a cobas® PCR Female swab.

Approximately 1000 adult women with an abnormal Pap test at the level of an atypical squamous cells of undetermined significance or worse squamous or glandular abnormality (i.e., ASCUS+) or an abnormal co-test (ASCUS+ and HPV-positive) result will be recruited over a period of 12 months via colposcopy clinics located at the Jewish General Hospital, St-Mary's Hospital, and the McGill University Health Centers (Royal Victoria Hospital).

Participating women will undergo three cervical or cervicovaginal sampling techniques: 1) self-sampling using the Eve Medical self-collection system©; 2) self-sampling using a cobas® PCR Female swab; and 3) physician-collected sampling. The participants will also fill in a questionnaire on their experience with the convenience and acceptability of the Eve system, relative to the other two sampling approaches. The decision as to which self-sample is to be collected first will be dependent on randomization

HPV testing will be done using the cobas® 4800 HPV Test. The liquid medium of within the cobas® PCR CELL Collection Media with the provider collected sample and the cobas® PCR media with the two self-collected samples will be used to suspend the cellular material prior to HPV testing. We have made collaborative arrangements with Dr. Marcel Behr, Chief of the Department of Clinical Microbiology at the McGill University Health Centre for the HPV genotyping work.

Histology-confirmed CIN1+ will form the study outcome or case definition. Sensitivity, specificity, and predictive values (along with their respective 95% confidence intervals) will be calculated for each sample type to evaluate the clinical performance of the various sampling techniques. We will use CIN1+ as definition of disease but analyses will also be performed for more stringent definitions, e.g. CIN2+ or CIN3/cancer.

After the discovery that Human Papillomavirus (HPV) infection is the necessary cause of cervical cancer there have been much interest in the introduction of novel screening strategies involving HPV DNA testing [1]. HPV DNA testing was shown to have higher sensitivity in detecting cervical cancer and its precursors and greater reproducibility than Papanicolaou (Pap) cytology screening and has been therefore proposed as the primary screening method for the future. In high-income countries, cases of cervical cancer continue to develop among unscreened and under-screened women [2]. Factors discouraging women from attending routine screening include fear of pain or discomfort, lack of time or perceived inconvenience, cultural or religious concerns, and poor socioeconomic status [3]. For a screening program to be effective high coverage is essential and thus attendance is a critical requirement.

Because of the high sensitivity of molecular HPV testing to detect cervical cancer precursors there is great potential for screening coverage to be increased by inviting non-attendant to provide self-collected vaginal samples. Providing these women with a simple, convenient, and inexpensive means of self-testing may improve cervical screening participation [3]. Self-sampling is also an attractive approach to assist screening in poor countries, which typically do not have structured cervical cancer screening programs while having to bear the greatest burden of cervical cancer morbidity and mortality [4]. Self-sampling can also enhance the value of post-HPV vaccination surveillance by allowing a more efficient monitoring of HPV type distribution in populations. Finally, self-sampling represents a good research tool to assist prospective studies of genital HPV infection and natural history of cervical neoplasia.

Specifically for cervical cancer screening, it has been shown that self-collected cervicovaginal specimens from women who received proper instruction for collecting them yield HPV test results that are comparable to those in specimens collected by physicians [5-8]. Women also prefer self-sampling relative to clinician-provided samples, provided that they can be assured that they are told how to properly collect the sample [8, 9]. Although there is conceivably some loss of sensitivity and specificity in screening for cervical cancer in a specimen that is not directly collected from the ecto- and endo-cervix the overall accuracy of the HPV testing results (to identify presence of cervical precancerous lesions) is still superior to that of physician-collected Pap smears. Therefore, replacing a more anatomically-correct specimen (the one collected with direct visualization of the cervix by a primary healthcare provider) with one collected from the vagina (and thus diluted with exfoliated cells from a wider epithelial surface area) is compensated by the high sensitivity of the molecular screening approach [6, 10, 11]. Performance of screening seems also to be unaffected by storage and transport of the self-sample swabs, irrespective of whether they have been kept dry or transferred and resuspended into a liquid transport medium [10, 12].

A critical feature in enhancing women's acceptance and adherence with the self-sampling approach for cervical cancer screening and HPV surveillance studies is the convenience of the device used for collection. Dacron and polyester swabs are simple, common, and inexpensive. However, they are inconvenient and not isolated within a sheath that prevents contact of the sampling area with mucosal surfaces in the labia and vaginal opening. Ideally, sampling devices should be anatomically correct to facilitate insertion and the sampling surface that will retain the exfoliated cells should be protected from contact with the labia and lower vaginal mucosa while the device is inserted, with the objective of sampling cells that are mostly from the cervix and upper vaginal area.

HPV DNA testing using self-collected cervicovaginal specimens represents a promising strategy to increase cervical screening participation and thus reduce rates of cervical cancer in countries with established cervical screening programs. Furthermore, this approach can also greatly improve the coverage and quality of cervical cancer screening in developing countries, as well as in remote regions in developed countries, e.g., aboriginal populations in Northern Quebec and in First Nations territories. Facilitating and improving uptake of cervical cancer screening would save lives, reduce costs of treating invasive cancer and potentially reduce inequalities in avoidable mortality caused from cervical cancer.

(full protocol available upon request)

Inclusion Criteria:

• Women of all ages are eligible to the study if they have been referred to the participating colposcopy clinic because of an abnormal Pap test at the level of an atypical squamous cells of undetermined significance or worse squamous or glandular abnormality (i.e., ASCUS+) or an abnormal co-test (ASCUS+ and HPV-positive) result.

Exclusion Criteria:

• none