2019 ASCCP Risk-Based Management Consensus Guidelines for Abnormal Cervical Cancer Screening Tests and Cancer Precursors

Table: of ContentsA. EXECUTIVE SUMMARY Updated US consensus guidelines for management of cervical screening abnormalities are needed to accommodate the 3 available cervical screening strategies: primary human papillomavirus (HPV) screening, cotesting with HPV testing and cervical cytology, and cervical cytology alone. New data indicate that a patient's risk of developing cervical precancer or cancer can be estimated using current screening test results and previous screening test and biopsy results, while considering personal factors such as age and immunosuppression. Routine screening applies only to asymptomatic individuals who do not require surveillance for prior abnormal screening results. The 2012 consensus guidelines were the first to be based on the principle of equal management for equal risk, specifically, the risk of a patient developing cervical cancer, estimated by the surrogate end point of the 5-year risk of cervical intraepithelial neoplasia (CIN) grade 3 (CIN 3) or more severe diagnoses (CIN 3+), regardless of which test combinations yielded this risk level. Introduction of risk-based guidelines in 2012 was a conceptual breakthrough, but the recommendations retained a continued reliance on complicated algorithms and insufficiently incorporated screening history. With a more nuanced understanding of how previous results affect risk, and more variables to consider, the 2019 guidelines further align management recommendations with current understanding of HPV natural history and cervical carcinogenesis. More frequent surveillance, colposcopy, and treatment are recommended for patients at progressively higher risk, whereas those at lower risk can defer colposcopy, undergo follow-up at longer surveillance intervals, and, when at sufficiently low risk, return to routine screening. Clearly defined risk thresholds to guide management are designed to continue functioning appropriately when population-level prevalence of CIN 3+ decreases because of HPV vaccination and also as new screening and triage tests are introduced. The revised guidelines provide a framework for incorporating new data and technologies as ongoing incremental recommendation revisions, minimizing the time needed to implement changes that are beneficial to patient care. B. INTRODUCTION This is the fourth American Society of Colposcopy and Cervical Pathology (ASCCP)-sponsored consensus guidelines for management of cervical cancer screening abnormalities, after the original consensus conferences in 20011 and subsequent updates in 20062 and 2012.3 An interim guidance publication providing management recommendations for primary HPV screening was released in 2015.4This document updates and replaces all previous guidance. The key difference between 2019 guidelines and previous versions is the change from primarily test results–based algorithms (e.g., “Colposcopy is recommended for patients with HPV-positive atypical squamous cells of undetermined significance [ASC-US], low-grade squamous intraepithelial lesion [LSIL],” etc.) to primarily “risk-based” guidelines (e.g., “Colposcopy is recommended for any combination of history and current test results yielding a 4.0% or greater probability of finding CIN 3+,” etc.). See Box 1 for essential changes. Tables of risk estimates for possible combinations of current screening test results and screening history (including unknown history) have been generated from a prospective longitudinal cohort of more than 1.5 million patients followed for more than a decade at Kaiser Permanente Northern California (KPNC). All KPNC estimates of risk underlying guideline decisions are detailed in the accompanying article by Egemen et al.5 The applicability of these risk estimates to other United States regions and populations has been confirmed in other data sets from screening programs and clinical trials.6 Many patients, especially those with minor abnormalities, can be managed by identifying their risk level using Tables 1A to 5B in Egemen et al5 and linking it to a recommended clinical action (return to routine screening, surveillance with repeat testing at 1- or 3-year intervals, colposcopy, or treatment). To facilitate use of these tables, the same information will be accessible via smartphone app (for purchase) and web (no cost) through http://www.asccp.org. Decision aids may facilitate use of the tables.7 Common abnormalities are managed using risk estimates outlined in Section E, and rare abnormalities are managed via the result-specific consensus recommendations outlined in Sections G-K.TABLE 1: Participating OrganizationsBox 1. Essential Changes From Prior Management Guidelines 1) Recommendations are based on risk, not results. Recommendations of colposcopy, treatment, or surveillance will be based on a patient's risk of CIN 3+ determined by a combination of current results and past history (including unknown history). The same current test results may yield different management recommendations depending on the history of recent past test results. 2) Colposcopy can be deferred for certain patients. Repeat HPV testing or cotesting at 1 year is recommended for patients with minor screening abnormalities indicating HPV infection with low risk of underlying CIN 3+ (e.g., HPV-positive, low-grade cytologic abnormalities after a documented negative screening HPV test or cotest). 3) Guidance for expedited treatment is expanded (i.e., treatment without colposcopic biopsy). Expedited treatment was an option for patients with HSIL cytology in the 2012 guidelines; this guidance is now better defined. For non-pregnant patients 25 years or older, expedited treatment, defined as treatment without preceding colposcopic biopsy demonstrating CIN 2+, is preferred when the immediate risk of CIN 3+ is ≥60%, and is acceptable for those with risks between 25% and 60%. Expedited treatment is preferred for nonpregnant patients 25 years or older with high-grade squamous intraepithelial lesion (HSIL) cytology and concurrent positive testing for HPV genotype 16 (HPV 16) (i.e., HPV 16–positive HSIL cytology) and never or rarely screened patients with HPV-positive HSIL cytology regardless of HPV genotype. Shared decision-making should be used when considering expedited treatment, especially for patients with concerns about the potential impact of treatment on pregnancy outcomes. 4) Excisional treatment is preferred to ablative treatment for histologic HSIL (CIN 2 or CIN 3) in the United States. Excision is recommended for adenocarcinoma in situ (AIS). 5) Observation is preferred to treatment for CIN 1. 6) Histopathology reports based on Lower Anogenital Squamous Terminology (LAST)/World Health Organization (WHO) recommendations for reporting histologic HSIL should include CIN 2 or CIN 3 qualifiers, i.e., HSIL(CIN 2) and HSIL (CIN 3). 7) All positive primary HPV screening tests, regardless of genotype, should have additional reflex triage testing performed from the same laboratory specimen (e.g., reflex cytology). Additional testing from the same laboratory specimen is recommended because the findings may inform colposcopy practice. For example, those HPV-16 positive HSIL cytology qualify for expedited treatment. HPV 16 or 18 infections have the highest risk for CIN 3 and occult cancer, so additional evaluation (e.g., colposcopy with biopsy) is necessary even when cytology results are negative. If HPV 16 or 18 testing is positive, and additional laboratory testing of the same sample is not feasible, the patient should proceed directly to colposcopy. 8) Continued surveillance with HPV testing or cotesting at 3-year intervals for at least 25 years is recommended after treatment and initial post-treatment management of histologic HSIL, CIN 2, CIN 3, or AIS. Continued surveillance at 3-year intervals beyond 25 years is acceptable for as long as the patient's life expectancy and ability to be screened are not significantly compromised by serious health issues. The 2012 guidelines recommended return to 5-year screening intervals and did not specify when screening should cease. New evidence indicates that risk remains elevated for at least 25 years, with no evidence that treated patients ever return to risk levels compatible with 5-year intervals. 9) Surveillance with cytology alone is acceptable only if testing with HPV or cotesting is not feasible. Cytology is less sensitive than HPV testing for detection of precancer and is therefore recommended more often. Cytology is recommended at 6-month intervals when HPV testing or cotesting is recommended annually. Cytology is recommended annually when 3-year intervals are recommended for HPV or cotesting. 10) Human papilloma virus assays that are Food and Drug Administration (FDA)-approved for screening should be used for management according to their regulatory approval in the United States. (Note: all HPV testing in this document refers to testing for high-risk HPV types only). For all management indications, HPV mRNA and HPV DNA tests without FDA approval for primary screening alone should only be used as a cotest with cytology, unless sufficient, rigorous data are available to support use of these particular tests in management. The minimum amount of data required to generate a recommendation will include the patient's age and current test results, as we recognize that previous screening history is often not known. Increased precision of management guidance will be possible if information is available on test results within the past 5 years and previous precancer treatment within the past 25 years.3 Current results and past history are designed to generate the patient's risk estimate from data tables.5 Risk estimates are available for the following clinical situations: abnormal screening test results with unknown history, abnormal screening test results with medical record documentation of a preceding negative HPV test or cotest, surveillance of previous abnormal screening test results that did not require immediate colposcopic referral (e.g., follow-up after an HPV-positive cytology negative result), colposcopy/biopsy results, and follow-up surveillance tests after colposcopy or after treatment for, or resolution of, high-grade abnormalities (e.g., CIN 2+). The recognition that persistent HPV infection is necessary for developing precancer and cancer (defined as CIN 3+, which includes diagnoses of CIN 3, AIS, and cancer) underlies the 2019 guideline update. Prospective longitudinal data indicate that when a new abnormal screening test result follows a negative HPV test or cotest within the past 5 years, the estimated risk of CIN 3+ is reduced by approximately 50%.8 A negative cytology result within 3 years of a new abnormal screening test, however, does not confer a similar reduction in risk.9 The 2019 guidelines also recognize that a colposcopic examination performed according to accepted standards (e.g., using the KPNC colposcopy protocol or the ASCCP Colposcopy Standards10) confirming low-grade or normal histology reduces a patient's estimated risk of having precancer/cancer in the next 2 years.11 This allows patients with an HPV-positive ASC-US or LSIL result at their 1-year follow-up visit after a colposcopy confirming normal- or low-grade histology to return for repeat HPV or cotesting in 1 more year, rather than immediately return to colposcopy. Thus, incorporating a patient's history of previous HPV tests and colposcopy/biopsy results will permit detection and treatment of CIN 3+ while avoiding unnecessary interventions for patients with new HPV infections who are at lower risk.12 C. GUIDING PRINCIPLES Guidelines are based on several guiding principles. The first 4 guiding principles are new for 2019, whereas the others are from the 2012 guidelines. As the 2012 guidelines are familiar to providers, we changed management recommendations only when new evidence favored an altered management strategy. Note that management guidelines apply only to patients with current or previous abnormal screening test results; screening guidelines for individuals in the general population, that are not being followed for a screening abnormality, are addressed elsewhere.13,14 New 2019 Principles 1. HPV–based testing is the basis for risk estimation. The term HPV-based testing is used throughout this document and refers to use of either primary HPV testing alone or HPV testing in conjunction with cervical cytology (cotesting). Characteristics of HPV infections, including HPV type and the duration of infection, determine a patient's risk of CIN 3+.15–18 Although cytology has high specificity (apart from ASC-US) and can be helpful when estimating immediate risk, its lower sensitivity and lower negative predictive value compared with HPV testing reduces its utility for long-term risk prediction.9 The results of HPV tests alone or in conjunction with cytology are used to guide recommendations that allow lengthening of follow-up intervals and deferral of colposcopy for low-risk results. Of note, risk estimates underlying the 2019 management guidelines are based on HPV DNA testing. 2.Personalized risk-based management is possible with knowledge of current results and past history. A patient's risk of having or developing CIN 3+ is estimated based on current and previous results, as well as history of previous precancer treatment. Management recommendations use thresholds of risk.19 Recommendations of routine screening, 1-year or 3-year surveillance, colposcopy, or treatment correspond to a risk stratum, a range of risk for CIN 3+. The lower threshold of each risk stratum, called Clinical Action Threshold, defines the level at which the management recommendation changes. The Clinical Action Thresholds for each risk stratum were determined through the consensus process. Risks were estimated for all combination of current results and past history (including unknown history) for which adequate data were available at KPNC. Management can be determined via look-up tables,5 and use of the tables can be facilitated using decision aids. 3.Guidelines must allow updates to incorporate new test methods as they are validated, and to adjust for decreasing CIN3+ risks as more patients who received HPV vaccination reach screening age. The field of cervical cancer prevention is rapidly evolving, with new technologies being continually validated. Data on the validation of new technologies are being published frequently, and risk reduction from HPV vaccination is increasing as vaccine coverage increases and vaccinated individuals age into screening cohorts. Up to now, guideline revisions have required full consensus conferences, which are time-consuming, expensive, and not compatible with the rapid evolution of the field. The 2019 guidelines build a framework that allows incorporation of new technologies and modified strategies without requiring full consensus conferences, so that revisions may rapidly incorporate new findings and be quickly disseminated to optimize patient care. Clinical Action Thresholds for management created through the 2019 consensus process will remain in place, but as new tests become available and more long-term data accrue, the test combinations used to reach these thresholds will change. For example, at the 2019 consensus conference, HPV vaccination levels in the United States population currently 25 years or older were deemed too low to warrant incorporating HPV vaccination into the 2019 management recommendations. However, this is expected to change in the near future as more vaccinated patients, who have lower CIN 3+ risk, reach the age of 25 years and additional data accrue demonstrating the impact of vaccination on the CIN 3+ risk associated with abnormal test result combinations. The framework outlined here will allow guideline modification as robust data become available and are publicized. Because Clinical Action Thresholds remain constant, new data can be added while the Clinical Action Thresholds remain unchanged. This design is intentional to reduce clinician confusion associated with frequently changing guidelines. 4.Colposcopy practice must follow guidance detailed in the ASCCP Colposcopy Standards.10 Colposcopy with targeted biopsy remains the primary method of detecting precancers requiring treatment. Because patients are managed less aggressively after a colposcopic examination where CIN grade 2 or higher (CIN 2+) is not found, maximizing detection of CIN 2+ at each colposcopy visit is paramount. Evidence-based practice recommends that biopsies be taken of all discrete acetowhite areas, usually 2 to 4 biopsies at each colposcopic examination. For those at lowest risk, defined as less than HSIL cytology, no evidence of HPV 16/18 infection, and a completely normal colposcopic impression (i.e., no acetowhitening, metaplasia, or other visible abnormality, and a fully visualized squamocolumnar junction), untargeted (random) biopsies are not recommended and patients with a completely normal colposcopic impression can be observed without biopsy. To ensure that CIN 2+ is not missed, the ASCCP Colposcopy Standards emphasize the need for biopsies even when the colposcopic impression is normal but any degree of acetowhitening, metaplasia, or other abnormality is present. 2012 Principles Carried Forward 5.The primary goal of screening and management is cancer prevention through detection and treatment of cervical precancer. Numerous population-level studies indicate that incidence and mortality from cervical cancer decrease as detection and treatment of high-grade histologic cervical abnormalities (generally defined as CIN 2+) increases.20,21 Timely detection and treatment of the highest grade of precancers (CIN 3/AIS) have been the benchmark used for previous guidelines3 and remain the primary goal of the 2019 management guideline; a secondary goal (because of the relative rarity of this finding in the United States) is early diagnosis of cervical cancer to reduce related morbidity and mortality. A patient's risk of having or developing CIN 3+ is estimated based on current and previous results, as well as history of previous precancer treatment. Management recommendations are guided by risk thresholds.19 Recommendations of routine screening, 1- or 3-year surveillance, colposcopy, or treatment each correspond to a risk stratum. These risk strata (ranges of risk for CIN 3+) are defined by Clinical Action Thresholds that were determined through the consensus process (Section E). 6.Guidelines apply to all individuals with a cervix. Guidelines apply to women and transgender men with a cervix, including individuals who have undergone supracervical hysterectomy. Risk estimates were validated in individuals of diverse racial, ethnic, and socioeconomic backgrounds and shown to be comparable.6 Though not the primary focus of the 2019 guidelines, management recommendations are also provided for patients who have undergone hysterectomy with removal of the cervix and who have a previous diagnosis of histologic HSIL, CIN 2, CIN 2/3, CIN 3, and/or AIS, irrespective of whether the hysterectomy was performed for precancer treatment or another indication. 7.Equal management for equal risk. History and current test results are used to calculate a patient's current and future risk of CIN 3+. Similar risks are managed similarly, regardless of the combination of results/history used to estimate the risk. 8.Balancing benefits and harms. Providing the best care means balancing cancer prevention with overtesting and overtreatment. Preventing all cervical cancers is unfortunately not an achievable goal. Interventions to prevent cervical cancer can cause harm. The 2019 guidelines are designed to maximize cervical cancer prevention and minimize harms from overtesting and overtreating by managing patients according to their current and future risks of CIN 3+. High-risk patients require closer follow-up to maximize detection of CIN 3+, whereas low-risk patients require fewer tests and procedures. 9.Guidelines apply to asymptomatic patients that require management of abnormal cervical screening test results. Patients with symptoms such as abnormal uterine or vaginal bleeding or a visibly abnormal-appearing cervix require appropriate diagnostic testing as this may be a sign of cancer.22 This evaluation may include cervical cytology, colposcopy, diagnostic imaging, and cervical, endocervical, or endometrial biopsy. Guidelines cannot cover all clinical situations and clinical judgment is advised, especially in those circumstances which are not covered by the 2019 guidelines. 10.Guidelines are intended for use in the United States. Appropriate management may differ in countries with limited follow-up capabilities, less availability of colposcopy, limited pathology infrastructure, or different views of the trade-offs between cancer risk, cost, and overtesting/overtreatment. D. METHODS D.1 Process and Timeline The ASCCP and National Cancer Institute (NCI) established a Memorandum of Understanding in January 2017 to undertake the work of this guideline update. As with the previous 2001, 2006, and 2012 guidelines,1–3 NCI produced risk data and other scientific support for the consensus guideline process. The ASCCP sponsored the consensus effort to develop and ratify the guidelines. Stakeholder organizations representing best practice in the United States were identified and invited to participate. These included medical professional societies, patient advocacy groups, and federal agencies integral to cervical cancer screening and management of abnormal results (see Table 1). Participation of the stakeholder organizations included identifying organization representatives and, for nongovernment participants, sponsoring their travel to consensus conferences. Representatives from 19 organizations attended the initial meeting in February 2018. At that time, 7 working groups were convened. In previous consensus conferences, working groups considered specific test outcomes (e.g., ASC-US/HPV-positive) and special populations. In contrast, the 7 working groups for the 2019 guidelines were organized with the goal of establishing consensus Clinical Action Thresholds. The treatment group evaluated which risk levels of CIN 3+ warrant expedited treatment without confirmatory biopsy, as well as addressing treatment-related issues. The colposcopy group considered the threshold for colposcopy referral. The surveillance group created a hierarchy of retesting at shorter intervals than currently recommended for routine screening with either HPV primary testing or cotesting (5 years) and also examined when patients could return to routine screening. Patients undergoing surveillance include those with minimally abnormal screening results not requiring colposcopy (e.g., HPV-positive Negative for Intraepithelial Lesion or Malignancy [NILM]), after colposcopy with low-grade results, or after treatment for high-grade abnormality. The risk modification group evaluated factors that might change a patient's estimated risk or management, focusing on pregnancy and immunosuppression. The high value care group performed decision analyses related to proposed management strategies and will continue to assess value as the 2019 guidelines are implemented. The new technologies group evaluated laboratory terminology and emerging technologies specifically related to management. The communications group created and reviewed relevant content for public communication to both clinicians and the lay public about the guidelines and the development process. Working groups were composed of 2 to 8 members, including representatives of participating stakeholder organizations, content experts, and nonclinician representatives of patient advocacy organizations. Working groups met regularly from summer 2018 through fall 2019 to review data and develop guidelines for management. The consensus process was overseen by a 23-person steering committee convened by the ASCCP and was directed by a leadership team consisting of 1 NCI representative (M.S.) and 2 ASCCP representatives (R.G., R.P.). Because the guidelines represent a paradigm shift, the guidelines process included a deliberate and extensive process of stakeholder engagement. These included patient and provider surveys, a consensus meeting to review preliminary guidelines, and a 6-week open public comment period before the final consensus voting meeting in October 2019.23 D.2 Choice of CIN 3+ as Main Clinical End Point for Risk Estimates For the management guidelines, we chose CIN 3+ as the best surrogate for cancer risk. The definition of CIN 3+ as used in these guidelines includes CIN 3, AIS, and the rare cases of invasive cervical cancer that are found in screening programs. These management guidelines consider CIN 3+ risk at the time point relevant for the clinical action being considered—Clinical Action Thresholds for colposcopy and treatment consider immediate risks of CIN 3+, whereas longer-term surveillance recommendations use 5-year risks. CIN3+ was chosen as an endpoint instead of cancer because cancer is uncommon in the United States, and risk is profoundly decreased by precursor treatment. Cancers that are found in robust screening programs may represent cancers already prevalent at first screening, rare instances of aggressive or HPV-negative tumors not detectable by screening, or false negative results.24 CIN 3+ was chosen instead of CIN 2+ because it is a more pathologically reproducible diagnosis,25 the HPV type distribution in CIN 3+ lesions more closely approximates that of invasive cervical cancers than the larger range of types found in CIN 2,15–18,26 and CIN 2 has appreciable regression rates in the absence of treatment.27–29 The choice of CIN 3+ does have some limitations, as even among CIN 3/AIS lesions, risks of progression to cancer differ. Glandular lesions including AIS, lesions with HPV 16 and 18 infections, and those occurring in older patients have higher cancer risks than HPV-negative lesions and those occurring in younger patients.30 Different nomenclatures for cervical histopathology are in use in the United States. The LAST Project and the WHO recommend a 2-tiered terminology (histologic LSIL/HSIL) for reporting histopathology of HPV-associated squamous lesions, similar to the Bethesda system used for reporting cervical cytology.31,32 However, the CIN nomenclature is still commonly used, and data used to generate this set of guidelines relied on CIN nomenclature. Although no direct correlation is possible without use of the p16 biomarker, histologic HSIL is similar but not identical to CIN 2/3.33 D.3 Multiple Data Sets Used to Validate Risks Prior guidelines relied heavily on a large prospective data set including results of cytology, HPV testing, colposcopy, histology, and follow-up outcomes from KPNC, which adopted triennial cotesting as standard practice in 2003. The KPNC data continue to be the largest, most comprehensive data source in the United States for risk estimation of combinations of HPV DNA testing and cytology. For the 2019 guidelines, several additional databases were analyzed to ensure that results are applicable to patients of diverse racial, ethnic, and socioeconomic strata. Risk estimates were compared using screening and follow-up data from clinical trials (BD Onclarity registrational trials),34,35 a state registry (New Mexico HPV Pap Registry36,37), and the Centers for Disease Control and Prevention's (CDC's) National Breast and Cervical Cancer Early Detection Program, a national program that includes many low-income and minority patients.38 The populations vary in rates of abnormal screening results and the prevalence of CIN 3+. Nonetheless, the comparison showed that the risks of CIN 3+ for the specific combination of current results and screening history were similar in that they fell within the same risk bands for management. Cheung et al6 demonstrates the similarity of CIN 3+ risks associated with screening test result combinations among the different populations of screened patients from these data sets. In summary, different populations within the United States have higher or lower rates of CIN 3+ due to factors including access to screening and HPV infection prevalence. Nonetheless, patients with similar test results and screening history combinations have largely similar CIN 3+ risk, regardless of their geographic location, race, ethnicity, or socioeconomic status. D.4 Estimation of Risks Details of how risks of CIN 3+ were calculated for the many combinations of test results, including longitudinal series of tests over time, are described in the accompanying Methods article.6 In brief, for each combination of past and current test results, the risk of CIN 3+ was estimated using prevalence-incidence mixture models,39 which consist of joint estimation of prevalent CIN 3+ at the time of the current testing using a logistic regression model, and incident CIN 3+ at subsequent testing using a proportional hazards model. These joint models are designed to handle verification bias and interval censoring. Verification bias in this context means that histopathologic outcomes are only available for patients referred to colposcopy; thus, CIN3+ cases that occur in the setting of false negative screening or abnormal screening tests that were not referred for colposcopy will not be detected. Interval censoring in this context means that the CIN 3+ is diagnosed at colposcopy visits, but the actual time of onset of incident CIN 3+ cannot be determined as it is typically asymptomatic and occurs between testing visits. These flexible models are designed to provide risk estimates without forcing the data into a rigid distribution assumption (e.g., Weibull). D.5 Assigning Combinations of Test Results to Clinical Actions For each combination of current test results and screening histo