Background and rational: Cervical cancer ranks twelfth in terms of cancer mortality in Canadian women, but remains the second most frequent cancer in Canadian women aged 20-44. Fortunately, its natural history makes it an ideal candidate for successful prevention. Cervical cancer and its precursors are caused by a persistent infection of the cervical epithelium by one of the 15 oncogenic (or high risk) types of human papillomaviruses (HPVs). The preinvasive changes of cervical intra-epithelial neoplasia (CIN) can be identified through cervical cytology (also known as the Pap test). When cellular abnormalities are identified, women are referred for diagnostic testing. High-grade precursor lesions (CIN2/3) carry a high risk of progression to invasive cancer and for this reason their treatment is recommended. Treatment success rates of CIN2/3 are 85%. However, those who fail treatment need to be identified promptly because their risk of invasive disease is increased more then 10 fold. The current strategy used in Canada to identify treatment failures consists of a follow-up every 6 months in colposcopy clinics for 2 years. Recent research underlines the fact that routine colposcopy is unreliable and may miss significant lesions. The identification of oncogenic HPV DNA in the cervical secretions is known as HPV testing. A few studies have investigated the use of HPV testing in the identification of treatment failures. Although they point to a very good sensitivity (90-95%) there were methodological limitations: studies were small; the HPV test used was often not suitable for clinical laboratory use; endpoints were not assessed by histological confirmation. Most importantly, none compared HPV testing to colposcopy, the strategy used in Canada. Research question: Is HPV testing more sensitive than routine colposcopy to identify CIN2/3 treatment failures? Methodology: Design: We propose a parallel randomized controlled trial, where participants will be randomized 1:1 to routine colposcopy vs. HPV testing after treatment of CIN2/3 lesions. Study Population: Women treated for high-grade precursor lesions. Exclusion criteria are: pregnancy, less than 18 years of age, known immunosuppression or immunodeficiency, unable to provide informed consent. Study interventions: (1) HPV testing with Hybrid Capture 2. (2) Routine colposcopy. In order to evaluate the performance of the current follow-up strategy, there will be no standardized protocol for the colposcopy intervention, but documentation of procedures will be done. Patients will receive the study interventions 6 months after treatment. All participants will undergo expert diagnostic assessment at 12 and 24. Sample size calculation and recruitment rate. The sample size was calculated to have 80% power to detect an increase in sensitivity of 15%, with an alpha set at 0.05 (double sided test). Assuming a conservative baseline sensitivity of 70% for colposcopy, 134 cases of recurrent/persistent disease will be necessary in each study group. Assuming a 15% failure rate of treatment and a 20% drop out rate, a total of 2250 participants will need to be accrued. Recruitment will take place in high volume colposcopy clinics across Canada over 2 years. Analysis: The primary analysis will consist of the comparison of the diagnostic indices of the 2 follow-up strategies: sensitivity, specificity, positive and negative predictive values (with their 95% asymptotic confidence intervals). Cumulative persistent/recurrent cases identified during the 2 year follow-up will make up the case group. Since all participants will undergo expert diagnostic assessment. No subgroup analysis will be performed. 4. Expected contribution: The result from this trial will provide high quality data on which to base management recommendation. If HPV testing is found to be more sensitive than routine colposcopy, follow-up of treated lesion may occur mainly in primary care physician's office with HPV testing, reducing costs, and making it possible to focus colposcopy activities on a group of women truly at risk of significant disease.
