Prevention of Cervical Cancer Using the Genotyping Screening and Same-day Self-sampling

In the fight against cancer, cervical cancer prevention strategies in high-resource settings have been extremely successful. Secondary prevention with cervical cytology, linked with timely treatment of precancer, has led to declines in invasive cervical cancer (ICC) rates by more than 75% over the past half century in the United States and other high-resource nations. In stark contrast, low- and middle-income countries (LMIC), where ICC is the leading cause of cancer death in women in 42 countries, carry the global burden with about 85% of new cases per year. In 2018, the WHO called for action to achieve the global elimination of ICC. An elimination strategy was ratified by member states in August 2020 and launched in November 2020. If the proposed WHO strategies are followed it would take a century to truly make an impact on disease burden. Thus, it is imperative to speed up this timeline. Low-cost and effective strategies that increase access to screening and treatment of cervical precancer will enable these goals to be met sooner. The purpose of this study is to clinically validate the modified AmpFire® which provides a platform for effective, simple, rapid and affordable cervical cancer screening in low resource settings.

Aim 1. To estimate the test performance of self- and provider- collected modified AmpFire® screening platform in Honduras.

Activities: The investigators will test 4,000 women with both self- and provider-collected AmpFire® and quantify the ability of each collection modality to detect CIN2+. To do this, all HPV+ women by either modality and 5% randomly chosen HPV negative (HPV-) women on both modalities will undergo a pelvic exam where visual inspection with acetic acid (VIA), Automated Visual Evaluation (AVE), colposcopy, and biopsy will be conducted. VIA and AVE will be done to conduct retrospective, secondary analyses comparing screening and triage efficacy of these methods vs. AmpFire® genotyping. Colposcopy and biopsy will be done to ascertain true disease prevalence. All patients who are HPV+ will be offered immediate ablation treatment, if eligible. Women will be deemed ineligible for treatment if they have lesions that cover more than 75% of the cervix, extend into the endocervical canal or onto the vagina, have evidence of ICC, and/or have obstructing or unusual anatomy that makes it difficult to visualize the entire cervix. Participants with suspected ICC will be referred per Honduran guidelines.

Pathologist readings: Cervical biopsies and ECC specimens will be sent to a private lab with extensive experience with cervical cancer pathology. Additionally, an expert pathologist will evaluate slides with CIN2+ diagnoses and a 5% random sample of cervical intraepithelial neoplasia grade 1(CIN1)/normal pathology slides every six months to ensure quality and for expert confirmation of study endpoints. In cases of disagreement between the local and expert pathology analyses, a third expert pathologist will adjudicate the discrepancy.

Analysis: The primary analysis will separately assess the test performance (sensitivity, specificity, PPV, NPV) of provider collected modified AmpFire® and self-collected modified AmpFire® for the detection of high grade cervical precancer (CIN2+). All women who are positive with either test and 5% of women who test negative on both tests will undergo colposcopy and biopsy to ascertain true disease prevalence. Estimates of sensitivity and specificity and corresponding 95% confidence intervals will be calculated using the approach of Begg and Greenes to correct for the verification biased sampling of true disease status. A study of 4,000 women will result in 751 women receiving colposcopy and biopsy-580 women who screen positive on at least one modified AmpFire® test assuming 14.5% AmpFire® test positivity and 171 women who correspond to the 5% random sample of the 3,420 women who screen negative on both tests. Of the women receiving colposcopy and biopsy, it estimated that 56 women will have CIN2+. PPV and NPV will be estimated as proportions along with corresponding exact 95% confidence intervals. If 580 screen positive on the modified AmpFire® test, PPV can be estimated with a precision of 0.029 for a PPV of 0.145). NPV can be estimated with a precision of 0.033 if there are 171 women who screen for AmpFire and receive colposcopy and biopsy and a true NPV of 0.95. VIA and VAT results will be retrospectively correlated to biopsy pathology and used as data points for the cost-effectiveness analysis.

Aim 2: To evaluate the feasibility of a single visit approach using a self-sampled modified AmpFire® screening platform.

Activities: An additional 1,000 women will self-collect samples with the modified AmpFire® platform. These tests will be processed the same day. All HPV+ women and 5% of women who tested negative will return in one year for repeat HPV testing, colposcopy and biopsy. A mixed-methods feasibility evaluation will be conducted to identify potential barriers and facilitators of HPV testing with the modified AmpFire® platform.

Baseline visit: As women arrive at the clinic, a research assistant will explain the study and, if the woman agrees, conduct informed consent procedures and administer a brief sociodemographic and health history questionnaire. Once this is done, the woman will proceed to self-collection. Women who are HPV+ will be triaged based on their HPV genotype. All women who test HPV+ will be given appointments to return in one year for repeat HPV testing, AVE, and colposcopy/biopsy to ascertain disease status. Women with CIN2+ at one year will be referred per Honduran guidelines.

Follow-up visit (12mos): All HPV+ patients regardless of risk group will be scheduled to return for HPV re-testing, colposcopy, and biopsy. This will ensure all cases of CIN2+ or invasive cancer among the HPV+ women are detected. Using an approach explicitly grounded in implementation science, the investigators will utilize a mixed-methods approach to evaluate the feasibility of the same-day screen and treat (ST) strategy. Data will be collected on clinical indicators and implementation outcomes. The primary feasibility outcome will be patient throughput, while the main implementation assessment will be provider and patient acceptability. Secondary outcomes are also described below.

Throughput. Patient attendance will be logged at the screening appointment and wait times between steps (screening to test processing, processing to result delivery, result delivery to speculum exam, etc.) will be recorded. Screening or treatment refusals will also be tracked and recorded. Outcomes will include the percentage of women screened, triaged and treated in the target time frames, and test processing time.

Provider Acceptability. At the end of the 4-month enrollment period, the investigators will conduct a semi-structured questionnaire with providers at each of the participating health centers. These questionnaires will be adapted to different roles (physicians, nurses, laboratory technicians, etc.) Topics covered will include perceptions of the screening test and its processing, the self-sampling component, AVE as a triage method, and the treatment provided, attitudes toward patient throughout and workflow. A sub-sample of 5-10 providers will be asked to participate in an ethnographic interview to further understand perceived barriers and facilitators of the strategy.

Patient Acceptability. All women will be asked to complete a brief acceptability assessment of self-sampling and associated instructions after completing this procedure. In addition, HPV+ women will be administered a more detailed questionnaire to understand perceptions of the same-day strategy. This will be done after the speculum exam is completed for women in Groups 1 and 2, or after result delivery for those in Group 3. Topics will include wait times and opportunity costs, comfort and/or pain associated with different steps, comprehension and perception of the counseling sessions, and overall satisfaction with the strategy. Women who receive treatment will be contacted via telephone at 4-6 weeks for a brief survey regarding side-effects. A sub-sample of \~10-15 HPV+ women will be asked to participate in an ethnographic interview to further elucidate these topics.

Aim 3: Evaluate the cost-effectiveness of the same-day screen-and-treat approach using the modified AmpFire® test followed by triage with Automated Visual Evaluation (AVE) versus the current Visual Inspection with Acetic Acid (VIA) strategy.

The investigators will conduct assessment of the cost and cost-effectiveness of implementing a screening strategy for CIN2+ based on the modified AmpFire® test followed by triage with AVE scores compared to VIA. Measurements of costs will be conducted using a micro-costing (ingredients) approach in which resource use required for each screening strategy will be tracked, and unit costs for each of the resources will be applied to generate an average screening cost per woman. Costs will include 1) direct medical costs (e.g., personnel time, supplies, screening and treatment equipment); 2) direct non-medical costs (e.g., patient transportation to/from study facility); 3) patient time costs (e.g., time spent traveling to, waiting for, and receiving care); and 4) programmatic costs (e.g., costs incurred at the administrative levels rather than the point-of-care delivery, such as training of local providers). Analyses will distinguish financial costs, which reflect actual expenditures of the program, from economic costs, which also include the value of donated and shared resources to more fully assess opportunity costs. Projections of economic cost implications over time will be made under varying assumptions of screening and treatment uptake, follow-up compliance, disease distribution by age, and treatment efficacy. Because the ultimate outcomes of interest associated with different screening strategies (i.e., cancer and mortality reduction) are not observable by the time decisions for adoption must be made, decision-makers have increasingly relied on projections from mathematical disease simulation models. For evaluation of the cost-effectiveness of the strategy based on AmpFire® combined with AVE scores versus VIA exclusively, the investigators will employ and adapt an individual-based microsimulation model of cervical cancer disease progression calibrated to HPV and cervical disease in Honduras [ref]. The female-only model includes health states representing HPV infection stratified by group according to risk of developing cervical cancer in the future (groups 1 to 4 defined in specific aims 1 and 2); high-grade cervical precancer (i.e., CIN2, CIN3); invasive cervical cancer (i.e., stages I-IV), and death. Women enter the model prior to sexual initiation and transition between health states as a function of age, HPV persistence by genotype, history of prior HPV infection, previously treated CIN, and screening patterns. Based on results from aim 1, women will receive treatment given accuracy and timing of screening, as well as availability of treatment. Given the individual-level simulation, the model can closely adhere to complex screening, triage, and treatment algorithms and keep track of each individual woman's health status and resource use over time. In order to contextualize the model to reflect the burden of cervical disease in the study population, the model will undergo an extensive likelihood-based calibration process to fit epidemiological data on the prevalence of different HPV types and precancer in Honduras [ref]. Once the model is calibrated to Honduras, the progression of disease will be simulated to generate a theoretical cohort of women before sexual initiation. The model will use monthly transitions through the health states described above. Data from the clinical test performance (e.g., sensitivity, specificity) and feasibility (e.g., loss to follow-up), along with the cost data described above, will be integrated into the model to determine the health and economic consequences of each strategy. Importantly, the model is able to project the impact of uncertainty in the data (e.g., CIN regression, invasion, and type-specific HPV reinfection), which can allow for expressions of confidence bounds around the cost-effectiveness results. Furthermore, one-way and multi-way (i.e., probabilistic) sensitivity analysis will be performed to determine the univariate and joint effects of other uncertain parameters and assumptions in the model.

Analysis: The comparative performance of the strategies will be measured using the incremental cost-effectiveness ratio (ICER), defined as the additional cost of modified AmpFire® and with AVE compared to VIA, divided by the additional health benefit as measured by disability-adjusted life-years (DALYs). Consistent with guidelines for health economic evaluations, future health gains and costs will be discounted at an annual rate of 3% per year to reflect time preference, but the investigators will vary the discount rate across a range of common values across different settings (e.g., 0% to 6%). The ICER provides a measure of value for money, which can be compared to other uses of health care resources. Analyses will be conducted from the societal and payer perspectives over both short- and long-term time horizons to appeal to different stakeholders and decision-makers. Given the lack of consensus for cost-effectiveness ("willingness-to-pay") thresholds in Honduras, the investigators will compare cost-effectiveness over a range of willingness-to-pay. How the investigators present and interpret results from cost-effectiveness analyses will be critical to our objective to ultimately inform adoption decisions. In this regard, the investigators will make substantive efforts to measure and report other outcomes viewed as relevant by decision-makers and key stakeholders.