Management algorithms for cervical cancer screening and precancer treatment for resource‐limited settings

Management algorithms for screen‐positive women in cervical cancer prevention programs have undergone substantial changes in recent years. The WHO strongly recommends human papillomavirus (HPV) testing for primary screening, if affordable, or if not, then visual inspection with acetic acid (VIA), and promotes treatment directly following screening through the screen‐and‐treat approach (one or two clinic visits). While VIA‐positive women can be offered immediate ablative treatment based on certain eligibility criteria, HPV‐positive women need to undergo subsequent VIA to determine their eligibility. Simpler ablative methods of treatment such as cryotherapy and thermal coagulation have been demonstrated to be effective and to have excellent safety profiles, and these have become integral parts of new management algorithms. The challenges faced by low‐resource countries are many and include, from the management perspective, identifying an affordable point‐of‐care HPV detection test, minimizing over‐treatment, and installing an effective information system to ensure high compliance to treatment and follow‐up.


| INTRODUCTION
During the late 20th century, considerable reduction in cervical cancer incidence and mortality was achieved in high-resource countries owing to the systematic implementation of cytology-based cervical cancer screening programs (Pap smears), using a population-based approach. 1 These programs rely on frequently repeated cytology screening because of the low sensitivity of the method, and multiple visits are required for disease confirmation (by colposcopy and/or histopathology), treatment, and follow-up. Low-and middle-income countries (LMICs) have not been able to implement such a logistically complex model because of under-developed health systems-including a lack of laboratory infrastructure and human resources-and consequently are burdened with 86.5% of the deaths from cervical cancer worldwide. 2 Human papillomavirus (HPV) DNA testing has been recommended recently by the WHO as the first choice for primary screening for cervical cancer because of the objective nature of the test, its highthroughput capability, excellent reproducibility, and high negative predictive value, which allows extension of the screening interval to beyond 5 years. 3 A few high-resource countries in Europe have already replaced cytology with HPV testing in their screening programs. 4 In LMICs the use of HPV testing remains limited to small-scale demonstration projects because of high costs and the need for at least modest laboratory facilities. 5 acid (VIA) as the second-best screening test for low-resource settings, and more than 25 countries have introduced VIA in national screening programs, while many more are conducting pilot programs. 8 With introduction of these newer screening tests, novel management algorithms for screen-positive women have also been investigated and recommended. Key goals are to limit the number of visits to health facilities and to ensure high compliance with treatment for women with cervical lesions. This is most relevant in low-resource settings, where women must overcome huge social and economic barriers to reach screening or treatment clinics and are likely to have only a once-in-a-lifetime opportunity to access services. 9 Algorithms include the following, which are discussed in the next sections-screen-positive women can be: (1) referred for diagnosis confirmation by colposcopy (a traditional practice in cytology-based programs); (2) triaged by a second test before referral; or (3) treated immediately for suspected premalignant lesions (Fig. 1). The comparison of the different algorithms with regard to their referral rates and efficacies to detect or prevent cervical intraepithelial neoplasia (CIN) 3+ disease is shown in Table 1.

| REFERRAL TO CONFIRMATORY COLPOSCOPY
The standard of care for cytology-based programs in high-resource countries has been colposcopic verification and localization of disease in screen-positive women. However, facilities for colposcopy are limited in low-resource settings because the specialized and expensive equipment is difficult to procure and maintain, the training requirements for providers are high, and the necessary histopathology services are rarely available. As a consequence, the new management algorithms of screen-positive women in LMICs aim to minimize the use of colposcopy, but some discussion of the method is included in this review.
In addition to the logistical challenges, colposcopy in noncytologybased programs is more challenging, as the colposcopist often relies on the cytology result to diagnose the morphological abnormality. 10 Highly-sensitive HPV tests can detect potential CIN 2 or CIN 3 at very early stages, when the lesions are too small or subtle to be recognized visually. 11 In the atypical cells of undetermined significance, low-grade squamous intraepithelial lesion (ASCUS-LSIL) Triage Study (ALTS), the sensitivity of baseline colposcopy for the subsequent detection of CIN 3+ was only 53%. 12 In a randomized controlled trial in India, the risk of invasive cancer among VIA-positive women with apparently normal colposcopy and histopathology during 12 years of follow-up was much higher than that of VIA-negative women (hazard ratio 6.5; 95% CI, 1.6-27.1). 13 The risk was similar to that observed in VIA-positive women with colposcopically detected abnormalities who did not undergo biopsy or treatment, thus demonstrating the futility of colposcopy in this scenario.
Another major limitation of colposcopy as a triaging technique is its low specificity, which is approximately 50% for detecting highgrade cervical lesions, even in experienced hands. 14 The specificity can F I G U R E 1 Summary of management options for screen-positive women. Screening and treatment can be completed in a single visit if there is a point-of-care screening test and the lesion is suitable for ablative treatment with a simpler method such as cryotherapy or thermal coagulation. [Colour figure can be viewed at wileyonlinelibrary.com] be even lower when the specificity of the primary screening test is low, as is the case with the HPV test or VIA. 15 In a large community-based study in India, colposcopy was used to triage VIA and/or HPV positive women. 16 Colposcopy falsely suspected abnormalities in 68.8% of women with normal histopathology. A systematic, pooled analysis of the accuracy of colposcopy observed that for every 1000 screenpositive women referred for colposcopy, 464 would be falsely diagnosed to have CIN 2/CIN 3 and would be unnecessarily treated (in a "colposcopy-and-treat" scenario). 17 Based on such evidence, WHO has recommended direct referral of screen-positive women to treatment, bypassing colposcopy, a process now known as "screen and treat". 18

| TRIAGING OF HPV-POSITIVE WOMEN
As noted above, a major disadvantage of HPV testing is its low specificity-most infected women will clear the infection within 1-2 years and will not develop disease. 19 The pooled estimated specificity of HPV testing from 15 studies involving 45 783 participants was 88%, which implies that the test will be falsely positive in 12 out of every 100 normal women. 17 A risk stratification of HPV-positive women is needed to decide on further management, and several triaging strategies have been evaluated. Cytology is the most widely recommended test to triage HPV-positive women-where quality-assured cytology is available. 20 HPV-positive women with a cytology diagnosis of ASCUS or worse are referred for colposcopy, and the rest are advised to have repeat HPV testing after 1 year. Cytology performs better in a triaging scenario, since the prevalence of disease is high in the sample and cytologists have a limited number of specimens to evaluate. There is evidence that the competency of cytologists improves with prior knowledge of HPV status. 21 HPV-positive women may be further tested to know if they are positive for the most carcinogenic types, HPV 16 and 18, and such information can be used to triage these women to colposcopy. Due to such differential risks, genotyping for HPV 16/18 has been evaluated as a triaging test for women with nonspecific HPV-positive results. 22 The advantages are that the test is more reproducible than cytology, the original cervical sample can be used for triaging, and the test can In evaluation studies, the risk of high-grade CIN in HPV 16/18 positive women exceeded that in nontyped HPV-positive women with ASCUS cytology, signifying the need for colposcopy referrals of these T A B L E 1 Comparison of different management algorithms of screen-positive women by referral rates and their efficacies to detect or prevent CIN 3+ disease.

| Colposcopy-and-Treat Approach
In a colposcopy-and-treat approach, women reporting for colposcopy

| Screen-and-Treat Approach
Treatment of screen-positive women (screen-and-treat approach) without colposcopic or histopathologic verification is the most effective strategy to improve compliance, as this involves the least number of visits (Fig. 1). If screening and treatment are completed at the same sitting, this is known as the single-visit approach. The screen-and-treat strategy can be used in both VIA-and HPV-testing programs and usually involves treatment by an ablative technique.
For ablative treatment, the cervical squamocolumnar junction (SCJ) should be located on the ectocervix, the lesion should occupy less than 75% of the surface of the cervix, and there should be no suspicion of invasive cancer. 3 For VIA, these characteristics are assessed during the procedure itself, when 3%-5% acetic acid is applied to the cervix and lesions are revealed as tissue that appears white. The suitability for ablative treatment for the HPV-positive woman is assessed by similar application of 3%-5% acetic acid on the cervix and the same criteria as above are applied or there is no visible lesion. The women not eligible for ablative treatment are referred for excisional treatment.
The evidence for the strong recommendation of the screen-andtreat strategy by WHO was derived from a South African randomized controlled study, in which VIA-or HPV-positive women suitable for ablative treatment were treated with immediate cryotherapy in the study arm. 36,37 In the control arm the screen-positive women were not treated immediately but had colposcopic evaluation after 6 months.
Compared with the control arm, the treated HPV arm reported a 77% reduction of CIN 3+ lesions and the treated VIA arm a 38% reduction over 3 years of follow-up, suggesting high protection offered by such a simple algorithm. The benefits far outweigh the potential harms (discussed later) of overtreatment.
Cryotherapy of cervical premalignant lesions is highly effective, with reported cure rates of 90% for any CIN and 70% for CIN 3 disease even in a primary care setting. 38,39 Thermal coagulation (also known as cold coagulation) is as effective as cryotherapy to treat CIN and both techniques have excellent safety profiles even when performed by nonphysician providers. 40 The VIA screen-and-treat strategy has been used successfully to screen large numbers of women in the cervical cancer screening program in Zambia, with 56.4% of VIA-positive women being eligible for cryotherapy and 87% of the eligible women accepting same-day treatment. 41

| COST-EFFECTIVENESS OF DIFFERENT MANAGEMENT ALGORITHMS
Evidence from economic evaluation studies comparing methods for cervical cancer screening has been fairly consistent in showing that screening strategies that increase coverage and/or require fewer visits (thereby reducing loss to follow-up of screen-positive women) tend to be more cost-effective. One of the earliest cost-effectiveness studies was conducted in South Africa using mathematical modelling, and showed that a strategy of VIA or HPV testing followed immediately with cryotherapy was more cost-effective than strategies using conventional cytology. 44 A subsequent analysis by the same group using the same approach comparing screening strategies that could be performed at primary care health facilities in five LMICs largely confirmed these results. 45

| OUTSTANDING CHALLENGES
Even though the alternative screening and management algorithms discussed here have simplified the logistics of cervical cancer screening, implementation of the programs in LMICs is limited for several reasons. Among these are the needs to optimize fiscal and human resources, mobilize and educate communities, organize services that meet women's needs and preferences, and strengthen health information systems to track screen-positive women for follow-up. A truly point-of-care and affordable HPV test is still elusive. The higher specificity of HPV E6 oncoprotein detection observed in initial studies is encouraging and needs further evaluation. The real program effectiveness of the single-visit screenand-treat algorithm should be studied further in the countries that have implemented such a strategy. Where single-visit approaches are not feasible, strategies for improving follow-up (such as mobile phone reminders and outreach treatment services) should be evaluated.
An additional future consideration for screening and management is the advent of highly effective HPV vaccines, first against HPV 16 and 18, and now against additional high-risk types. Over a relatively short period, more than 80 countries or territories have introduced HPV vaccination into their national immunization programs, and 33 of these are LMICs, with many more implementing pilot projects. 51 The impact of large-scale HPV vaccination will be an eventual drastic reduction in the prevalence of disease, accompanied by a decline in both sensitivity and positive predictive value of cytology and VIA. 52 Furthermore, the early vaccines targeted just HPV 16 and 18, and the high-grade premalignant lesions caused by other oncogenic HPV types may have a more indolent natural history and may require a different management approach. These questions will pose new challenges for screening and management in LMICs and will need the attention of the research community.

AUTHOR CONTRIBUTIONS
PB and RS conceived the article, reviewed the evidence, and wrote the manuscript. FM, YC, RH, and FZ assisted in reviewing the evidence and writing the manuscript.