The Bethesda System for Reporting Cervical Cytology: A Historical Perspective

The Bethesda System (TBS) was the direct result of a series of articles published in 1987 in The Wall Street Journal by investigative reporter Walt Bogdanich, about lax practices and inaccuracies in medical laboratories. In spite of the well-accepted contribution of the Papanicolaou (Pap) test in reducing the incidence of cervical cancer, much of this breaking news focused on so called “Pap mills” in which cervical cytology specimens were being processed under less than optimal conditions. Public outcry and legislative action ensued. In response to the public's concerns about the quality of laboratory testing, particularly in physician's office laboratories and concerning Pap test interpretation, the US Congress passed the Clinical Laboratory Improvement Amendments (CLIA) in 1988. The CLIA detailed a number of quality assurance practices that laboratories must comply with, based on the type and complexity of testing they perform.

Prior to TBS, US laboratories used a variety of nonreproducible terms for reporting cervical cytology including Pap test class numbers (which were completely nondescriptive) amongst other terminologies, e.g., 3-tier dysplasia (mild/moderate/severe, or carcinoma in situ) and cervical intraepithelial neoplasia (CIN). Much was made of the clinically and biologically irrelevant morphologic distinctions between human papillomavirus (HPV) cytopathic effects and low-grade dysplasia as well as the nonreproducible distinction between the grades of precancer, and these differences were often used as a clinical threshold for significant interventions. To address these issues, the National Cancer Institute (NCI) convened a workshop, under the leadership of Drs. Diane Solomon and Robert Kurman, with the aim of developing a system for reporting Pap tests that would communicate the cytology findings to clinicians in a clear and relevant fashion. A group of individuals, with expertise in gynecologic cytopathology, histopathology, and patient management, met in Bethesda, MD, USA, in December 1988, and formulated the first version of TBS [1]. This lexicon for practice was founded on 3 basic principles that continue to hold true today - 3 decades and 3 updates later! Terminology must be (1) clinically relevant, (2) reasonably reproducible and flexible, and (3) reflect the most current understanding of cervical neoplasia.

TBS 1988 created a standardized framework for cervical cytology reports that included an evaluation of specimen adequacy, an optional general categorization, and a descriptive diagnosis. The inclusion of recommendations for adequacy limitations or elaborations on morphologic findings to guide management, as appropriate, was suggested. The introduction of adequacy has been considered by many to be the most important quality assurance component of TBS. The general categorization, although optional, clearly separated “normal” from “abnormal” reports and is aimed at assisting clinical providers in triaging reports for subsequent patient management. The descriptive diagnosis, on the other hand, is designed to be the essential interpretive component of the report that details the morphologic findings, i.e., infection, reactive/reparative changes, and/or epithelial cell abnormalities, both squamous and glandular. For squamous lesions, TBS introduced a new 2-tier terminology for reporting HPV-associated squamous intraepithelial lesions (SIL): low-grade (LSIL) and high-grade (HSIL). It also allowed for a period of transition during which previously used grades of CIN, degrees of dysplasia, and the presence of HPV cytopathic changes could be appended. The new 2-tier terminology was expected to improve interobserver reproducibility over that seen with 3-tier systems. A major component of TBS 1988 was a clarification of the misunderstood term “atypia.” As an equivocal finding, atypia had been overused and applied in a highly inconsistent manner, leading to considerable confusion about its appropriate management. In TBS, atypia was specifically defined to include only cases in which the cytologic findings were of “undetermined significance”; the term was not to be used for cases in which inflammatory, reactive, preneoplastic, or neoplastic changes were clearly present. This definition was also applied to cases of glandular epithelial abnormality. This gave rise to the now-familiar terms “atypical squamous cells of undetermined significance” (ASCUS) and “atypical glandular cells of undetermined significance” (AGUS).

Within a few years of its introduction, TBS gained widespread acceptance in clinical practice in the USA and, thereafter, internationally. Encouraged by the early benefits of this new standardized terminology, the NCI organized a follow-up workshop in Bethesda, MD, in April 1991 to critically evaluate TBS and consider areas for improvement. The 1991 Workshop was attended by 182 invited experts in cytopathology and gynecology, and approximately 65 presentations were made by various laboratories to share experiences regarding the use of TBS in clinical practice. The discussions resulted in the refinement of the report format and terminology, and the development of precise criteria for the evaluation of adequacy and specific diagnostic terms. A discussion about the management of abnormalities detected by cervical cytology acknowledged the need to have clinical trial data to help in development, so interim management guidelines were published in 1991 which would be finalized after further data was gathered. The updates were presented to the cytopathology community in the form of the first Bethesda Atlas “blue book” in 1994 [2].

Of all the changes introduced by TBS, the concept of ASCUS proved to be by far the most controversial for clinicians and pathologists alike. It highlighted the intrinsic limitations of morphologic interpretation, and led to significant variability in management and outcomes which caused frustration amongst clinical providers. With increasing implementation of TBS terminology in the early 1990s, it became clear that the vast majority of abnormal cervical cytology results fell into the equivocal categories of ASCUS or LSIL. In light of the increasing clarity of HPV biology, the large volume of women with ASCUS/LSIL cytology, which had the potential to overburden the available colposcopy resources, and the recent availability of HPV testing for potential triage, the NCI conceived the ASCUS-LSIL Triage Study (ALTS) to determine how best to manage women with these equivocal results. The ALTS was designed as a randomized multicenter trial to compare the 3 management strategies available: immediate colposcopy, HPV triage, and conservative follow-up. Women were recruited between 1996 and 1998, with the aim of establishing how best to avoid overreferral and overtreatment for the majority of women with clinically insignificant changes, while providing a safety net for the small percentage of women with underdiagnosed high-grade lesions [3].

The ALTS was a pivotal trial that not only achieved its main goal, but greatly informed our understanding of cervical cancer biology, which, in turn, significantly impacted clinical practice patterns. The study results established molecular testing for high-risk HPV as the most cost-effective approach to clarify equivocal cytologic findings. Follow-up data showed that the risk of CIN2+ for high-risk HPV-positive ASCUS is the same as for LSIL and should therefore be managed as such, while HPV-negative ASCUS is relatively safe and does not require immediate colposcopy. The implementation of reflex HPV testing for ASCUS decreased colposcopy referrals by up to 50%. From the ALTS, we were also able to demonstrate that in screening and diagnostic workup for cervical cancer, there is no gold standard, and, indeed, cytology, colposcopy, and histology are all subject to variable performance. Interobserver “non-reproducibility” of pathologic interpretations was shown to be comparable, for both histopathology and cytology and, at best, average [4]. Similarly, intake colposcopy in the ALTS had a sensitivity of 53% for subsequent CIN2+ and an overall sensitivity of 70% for CIN3+. Thus colposcopy, like cytology and histology, was found to be subjective and detected only about 2/3rd of CIN3+ [5].

From the onset, TBS was expected to continue to evolve with time in response to advances in the understanding of cervical neoplasia and the changing needs of the cytopathology and clinical communities. In 2001, the third TBS workshop was held to consider updates based on practice changes, and advances in science and technology. In order to broaden participation both within the USA and internationally, draft recommendations were posted on an Internet bulletin board, and over 2,000 comments were considered prior to the multidisciplinary workshop. TBS 2001 was attended by over 400 people from 24 countries and 44 professional societies. A number of significant updates occurred in 2001 that can be summarized as follows:

1. General. (a) The terms “interpretation” or “result” were recommended in lieu of “diagnosis” in the header of the cervical cytology report, since it is a screening test, which serves as a medical consultation by providing an interpretation that contributes to a diagnosis when integrated with patient history, clinical findings, and the results of other laboratory tests such as cervical biopsy. (b) Acknowledging that TBS was developed primarily for cervical cytology, it was decided that specimens from other sites in the lower anogenital tract, such as the vagina and anus, could also be reported using this terminology.

2. Terminology. (a) Squamous lesions: consideration was given to the possibility of eliminating the ASCUS interpretive category, but this was not done due the unacceptable loss of sensitivity and positive predictive value for HSIL that it would portend. Instead, it was modified to ASC, with a narrower definition and a simple dichotomous system of qualifiers, i.e., ASC-US and ASC-H, which also mirrored HPV biology. (b) Glandular lesions: while TBS emphasizes that cervical cytology is not a screening test for adenocarcinoma, improvements in sampling devices and experience with cytomorphologic recognition of glandular abnormalities led to the addition of “endocervical adenocarcinoma in situ” (AIS) as a distinct interpretative category, and the “atypical glandular cells” (AGC) category was refined to specify the glandular cell type and better reflect concern for neoplasia so as to provide more guidance for management.

The NCI and the American Society of Cytopathology collaborated, under the leadership of Drs. Diane Solomon and Ritu Nayar, to facilitate the dissemination of the TBS 2001 update. Three educational methods were utilized:

1. The 2nd edition of the Bethesda print atlas addressed the reporting changes, and was significantly expanded with a new chapter-based format and the addition of background, detailed interpretive criteria, ample illustrations, explanatory notes, management guidelines, and sample reports. New technologies such as liquid-based cytology, automation, computer-assisted imaging, and HPV testing were also included [6].

2. A companion Bethesda Web atlas was established in conjunction with the print atlas. It had more images than the print atlas, and the ability to perform searches by morphology, preparation type, etc., as well as a self-test [7].

3. The Web-based Bethesda Interobserver Reproducibility Study (BIRST), using a subset of images from the 2nd edition of the Bethesda atlas prior to its publication, was undertaken to gain additional insight into consensus among cytologists with varied training and experience, and to identify specific cytomorphologic features and reporting categories that represented sources of poor agreement. The results confirmed that the a priori difficulty of the image was a more important determinant of agreement than the characteristics of the participant [8].

Immediately following the 2001 Bethesda Workshop, the American Society for Colposcopy and Cervical Pathology (ASCCP) held a multidisciplinary consensus conference to develop management strategies for cytologic abnormalities conforming to TBS categories [9]. The results of ALTS and other clinical research formed the basis for guideline development. This was a historic event where, for the first time, reporting terminology correlated with both biology and clinical management. These evidence-based management guidelines were updated in 2006 and 2012, based on additional data from subsequent trials and clinical experience in the USA, and they are now widely accepted as the standard of care [10,11].

Between 2001 and 2014, with the approval and implementation of prophylactic HPV vaccination and primary HPV screening as an alternative option for the secondary prevention of cervical cancer, the role and performance characteristics of cervical cytology were gradually but surely changing in various parts of the world. In vaccinated populations, the removal of the HPV16 and HPV18 genotypes will lead to a subsequent decreased prevalence of high-grade lesions. In the primary HPV screening algorithm with reflex cytology only for women who are HPV-positive but do not have HPV16/18, the Pap test will take on more of a “diagnostic” role. Thus, it will be imperative to maintain the sensitivity of cytology for it to be utilized in this application. In addition, another decade of experience with liquid-based cytology, and other technologic advances such as immunocytochemistry and molecular process was gained, leading to the need for another TBS update.

The 2014 TBS revision, which occurred under the leadership of Drs. Ritu Nayar and David C. Wilbur, anticipated minimal changes in terminology, and it was largely undertaken to provide cytologists and clinicians with an extensive educational resource to include basic concepts, more morphologic clues, pitfalls, and histologic correlations, and also to update management guidelines. An additional non-morphologic chapter was also added, detailing the current conceptual approach to overall risk assessment-based management. The 3rd edition of the Bethesda atlas was significantly larger and published in print and electronic formats in 2015 [12,13]. It was also accompanied by a companion Bethesda 2014 website which contains images from the BIRST-2 project conducted in conjunction with this update [14,15]. It was gratifying to see that the prior educational efforts and experience resulted in a marked improvement in almost all interpretative categories when compared to BIRST-1, although the performance of glandular lesions is still less reproducible than that of squamous lesions and the image difficulty remains the main determinant of agreement.

One of the most significant contributions of the process is that TBS has served as a model for the development of other standardized reporting systems in cytopathology and histopathology. The Bethesda System for Thyroid Cytopathology, the Paris System for Urinary Cytopathology, the Milan System for Salivary Gland Cytology, and the Papanicolaou Society of Cytopathology Guidelines for Pancreaticobiliary Cytology have all followed educational efforts similar to TBS [16,17,18,19]. Extension of the 2-tiered terminology of LSIL and HSIL is now recognized by the World Health Organization, and by the ASCCP, and College of American Pathologists for reporting histopathology of specimens from the lower anogenital tract [20,21].

The success of TBS rests on sound principles that were key to adopting and implementing it as an international standardized reporting terminology for cervical cytology. TBS (a) is based on evidence, inclusion and consensus; (b) correlates with biology and management; (c) is concise and practical; (d) is flexible enough to be adopted by different geographic/laboratory settings; (e) allows for change with the increase of knowledge and experience; and (f) incorporates multiple educational endeavors (atlases, websites, and BIRST). Editors and authors of the cervical cytology Bethesda atlases have always been committed to making the educational materials affordable, and therefore widely accessible, by foregoing all honoraria or royalties.

In summary, TBS serves as a landmark and exemplary contribution to standardized reporting. The following excerpts are an appropriate end to this discourse:

It is paradoxical that instead of TBS responding to new developments and changes, it has actually led the way in many areas. TBS played a vital role in initiating research in the biology of cervical cancer, in exploring new approaches and strategies in patient management and in incorporating new technologies into cervical cancer screening. (Foreword by Robert Kurman in the 2nd edition of The Cervical Cytology Bethesda Atlas, 2004 [6].)

I never imagined that a small meeting on the campus of the National Institutes of Health in Bethesda, Maryland, one snowy weekend in December 1988 would begin a process that has changed the practice of cervical cytology - in both the laboratory and the clinician's office - around the world. (Preface by Diane Solomon to the 3rd edition of The Cervical Cytology Bethesda Atlas, 2015 [13].)

The authors have no conflicts of interest to declare.