The New Bethesda System forReporting Results of Smears ofthe Uterine Cervix

Leopold G. Koss*

Over 40 years have elapsed since the large-scale introductionof the cervical-cancer smear as a cancer detection tool. Thesmear, which consists of a sample of cells from the epithelialsurfaces of the uterine cervix, is colloquially known as the "Paptest," after its discoverer, George N. Papanicolaou. The test isperformed annually on many millions of women. It has contrib-uted in a statistically significant way to the reduction of morbidityand mortality rates from invasive carcinoma of the uterus inappropriately screened populations {1-4). However, the test isprone to disturbing failures (5).

Theoretically, at least, the test is quite simple: the cell samples,placed on microscope slides and appropriately stained, are stud-ied under the microscope by trained cytotechnologists. Thepresence of precancerous lesions is heralded by the appearance ofnuclear and cytoplasmic abnormalities. Tissue biopsies, prefera-bly obtained with the colposcope, a magnifying instrument,provide further information on the scope and size of the lesion,which is subsequently removed or destroyed to prevent itsprogression to invasive cancer.

In practice, there are several problems with the Pap test. Theyare (a) the adequacy of the samples, (b) the difficulty of micro-scopic screening, based on recognizing few abnormal cellsamong many thousands of cells in each sample, and (c) thecompliance of patients with testing at appropriate intervals orwith follow-up procedures. These problems have been the subjectof considerable concern in the lay press and in scientific publica-tions (5). The failure to discover a precancerous lesion before itbecomes malignant has resulted in invasive cancer and death ofsome young patients.

Measures to improve the quality of screening have beenrecently proposed (Amendment to the Clinical Laboratory Im-provement Act: pending), but, without adequate funding, theimplementation and results of these measures are uncertain.

One of the key issues in providing adequate care to patients atrisk for cervical cancer is the manner in which the smears arereported to the primary care physicians and gynecologists. Papa-nicolaou, who was not a trained pathologist, recognized early inhis work that he was not qualified to render diagnostic verdictsbased on smears. He therefore devised a system of reportingbased on five classes. Class I smears were entirely benign, hencenormal. Class II smears disclosed minor cell abnormalities("atypia"), thought to be benign. Class III smears corresponded tocell abnormalities that were "suspicious," but not definitelycancerous. Class IV smears were "most likely malignant," andClass V smears were unequivocally malignant and diagnostic ofcancer.

This reporting system was widely adopted in the United Statesand abroad but was rarely used as originally intended. Instead, thesignificance of classes was often modified. In some laboratories,Class II smears were considered to be suspicious, Class III smearswere considered to indicate precancerous lesions, and Class IVsmears were considered to indicate invasive cancer. Other labo-ratories used only three classes. Still further variations in thetheme existed, including subdivision of the classes by letters,such as III A and IIIB.

The issue became even more complicated once tissue patholo-gists entered the act. It was recognized fairly early in themass-screening process that the rate of discovery of precancerouslesions of the cervical epithelium was much higher than the true orprojected incidence of invasive cancer. Hence, there were at leastsome precancerous lesions that did not have the potential forprogression to invasive cancer. Some observers thought that thecytologic and histologic appearance of these lesions might pro-vide the clue to their behavior. Thus, despite some dissentingvoices {6,7), the lesions were divided into two broad categories:dysplasia and carcinoma in situ. Dysplasia referred to lesionswith a lower degree of abnormality that were implicitly less likelyto progress to invasive cancer. Carcinoma in situ referred tolesions that were implicitly more likely to progress to invasivecancer. Dysplasia was further subdivided into three grades (mild,moderate, and severe) that were presumed to reflect the potentialfor "good" or "worse" behavior {8). It was also proposed that onecould accurately interpret the cervical smear in terms of theunderlying tissue pattern.

These additional classifications had a profound impact on thereporting system. Because of significant peer pressure, patholo-gists felt compel led to add a comment on the nature, and hence theprobable behavior, of the underlying lesion to every abnormalsmear report. This system of reporting was based on the premisethat the interpretation of smears and biopsies was reproducible.Unfortunately, in spite of some efforts {8), the precancerouslesions known as dysplasias have never been objectively defined.Numerous surveys among expert pathologists clearly showed thatthe diagnostic system was not reproducible and that one person'sdysplasia was another person's carcinoma and vice versa {9,10).Further, long-term follow-up studies documented that the behav-ior of precancerous lesions could not be predicted by morphology(6).

The clinicians at the receiving end of the cytologic reports wereoften unaware of the substance and nature of the controversy andacted (or failed to act) according to their interpretation of theclasses and verbal comments. There is evidence that women withClass III smears or "dysplasia" were not always treated andsubsequently developed invasive cancer {11). Because of thediagnostic chaos, it became impossible to compare the resultsfrom laboratories and individual pathologists.

Received May 1, 1990; accepted May 2, 1990.*Correspondence to: Leopold G. Koss. M.D., Department of Pathology,

Montefiore Medical Center and Albert Einstein College of Medicine. H I E .210th St.. Bronx, NY 10467.

988 Journal of the National Cancer Institute

Richart's introduction of the concept of "cervical intraepithe-lial neoplasia" (CIN) (12), encompassing all of the precancerouslesions of the epithelium of the uterine cervix, constituted a majorbreakthrough. Many gynecologists recognized that reports oncervical smears were not always accurate and that many womenwith "atypia" or "mild dysplasia" have precancerous lesionsrequiring colposcopy and treatment. This knowledge, however,did not necessarily drift down to the primary care physicians oralleviate the need for a uniform reporting system of cervicalsmears.

A group of experts, representing a broad spectrum of interestedprofessional organizations, met in Bethesda, Md, under theauspices of the National Cancer Institute on December 12 and 13,1988. They met in an attempt to redefine the reporting system of

the cervical smears and to address several other pertinent issues(13). The resulting document, now known as The BethesdaSystem (table 1), was adopted by consensus and, hopefully, willintroduce order and uniformity into the diagnostic chaos thatprevailed until now.

The key points of the Bethesda System are as follows:I. The cytology report is a medical consultation. This state-

ment separates the cervical smear from machine-generated clini-cal laboratory tests and proposes a different role for the smear inthe cancer prevention system. It holds clinician responsible forobtaining an adequate sample and for providing an adequatemedical history. It obligates the cytopathologist to assess theadequacy of the sample, that is, whether it is representative of thestatus of the cervical epithelium. The statement also obligates the

Table 1. The 1988 Bethesda System for reporting cervical/vaginal cytological diagnoses.

Statement on Specimen Adequacy

Satisfactory for interpretationLess than optimalUnsatisfactory

Explanation for less than optimal/unsatisfactory specimen:

—Scant cellularity—Poor fixation or preservation—Presence of foreign material (e.g., lubricant)—Partially or completely obscuring inflammation—Partially or completely obscuring blood—Excessive cytolysis or autolysis—No endocervical component in a premenopausal woman who has a

cervix—Not representative of the anatomic site—Other

General Categorization

Within normal limitsOther:

See descriptive diagnosesFurther action recommended

Descriptive Diagnoses

INFECTIONFungal

Fungal organisms morphologically consistent with Candida speciesOther

BacterialMicroorganisms morphologically consistent with Gardnerella speciesMicroorganisms morphologically consistent with Actinomyces speciesCellular changes suggestive of Chlamydia species infection, subject

to confirmatory studiesOther

ProtozoanTrichomonas vaginalisOther

ViralCellular changes associated with cytomegalovirusCellular changes associated with herpesvirus simplexOther(Note: for human papillomavirus IHPVJ, refer to "Epithelial Cell

Abnormalities, Squamous Cell")Other

REACTIVE AND REPARATIVE CHANGES

InflammationAssociated cellular changesFollicular cervicitis

Miscellaneous (as related to patient history)Effects of therapy

Ionizing radiationChemotherapyEffects of mechanical devices (e.g., intrauterine contraceptive device)Effects of nonsteroidal estrogen exposure (e.g., diethylstilbestrol)Other

EPITHELIAL CELL ABNORMALITIES

Squamous Cell• Atypical squamous cells of undetermined significance

(recommended follow-up and/or type of further investigation:specify)

• Squamous intraepithelial lesion (SIL) (comment on presence ofcellular changes associated with HPV if applicable)

Low-grade squamous intraepithelial lesion, encompassing:Cellular changes associated with HPVMild (slight) dysplasia/cervical intraepithelial neoplasia

grade 1 (CIN I)High-grade squamous intraepithelial lesion, encompassing:

Moderate dysplasia/CIN IISevere dysplasia/CIN IIICarcinoma in situ/CIN III

• Squamous cell carcinoma

Glandular Cell

• Presence of endometrial cells in one of the followingcircumstances:Out of phase in a menstruating womanIn a postmenopausal womanNo menstrual history available

• Atypical glandular cells of undetermined significance(recommended follow-up and/or type of further investigation:specify)EndometrialEndocervicalNot otherwise specified

• AdenocarcinomaSpecify probable site of origin: endocervical, endometrial,extrauterineNot otherwise specified

• Other epithelial malignant neoplasm: specify

NONEPITHELIAL MALIGNANT NEOPLASM: SPECIFY

HORMONAL EVALUATION (APPUES TO VAGINAL SMEARS ONLY)

• Hormonal pattern compatible with age and history• Hormonal pattern incompatible with age and history: specify• Hormonal evaluation not possible

Cervical specimenInflammationInsufficient patient history

OTHER

Vol. 82, No. 12, June 20, 1990 EDITORIAL 989

pathologist to provide guidance to the clinician in further evalu-ation of the patient, if required.

II. The Papanicolaou reporting system of smear classes wasnot considered adequate. The lack of uniformity and uncertaintyof the message conveyed by classes was judged to be ill-suited tothe practice of preventive medicine.

III. A new reporting system was proposed. The proposedreporting system is the heart of the Bethesda document because itrequires that each report should determine several aspects of thecervical sample. (The Bethesda System proposal does not pre-clude the use of other nomenclature as an addendum to theprincipal reporting system.) Requirements for the report are asfollows:

A. Adequacy of smears. The adequacy of each sample must beassessed. If the sample is thought to be inadequate, anexplanation of the reasons for this judgment must beprovided. This provision has significant fiscal implica-tions: for a smear to be judged as inadequate, it must beprocessed and screened. The costs of this procedure mustbe borne by the patients or third-party payers, and obtain-ing another smear clearly implies additional costs. Thus, itis in the interest of the clinicians and the patients to provideadequate smears on the first attempt. On the other hand,rejecting a smear as inadequate requires good judgment bythe pathologist, and this decision must be based on solidgrounds. Because the makeup of the smear may varyaccording to the clinical situation, sampling instruments,age, and menstrual status of the patient, these factors mustbe taken into account. Needless to say, reporting aninadequate smear as negative in a patient who subsequentlydevelops invasive cervical cancer may have major legalconsequences.

B. Primary assessment. The reporting of the sample falls intotwo categories: within normal limits and "other," whichcalls for descriptive diagnosis.

C. Descriptive diagnosis of benign abnormalities. The de-scriptive diagnosis of benign abnormalities includes abroad variety of infections and infestations and reactivechanges.

D. Descriptive diagnosis of precancerous lesions. In referenceto the precancerous intraepithelial lesions, the significantthrust of the Bethesda System is the classification of theselesions into two categories: low-grade and high-grade. Thelow-grade lesions include all neoplastic changes, previ-ously classified as "mild dysplasia," CIN I, or lesions withmorphologic changes suggestive of human papillomavirusinfections (e.g., flat condylomas). The high-grade lesionscomprise all other precancerous events or lesions previ-ously classified as moderate or marked dysplasia andcarcinoma in situ (CIN II and III). Thus, the reporting ofprecancerous lesions is much simplified.

E. Descriptive diagnosis of cancer. Additional reporting cate-gories have been provided to encompass other abnormali-ties, including various types of invasive cancer that can berecognized in the cervical samples.

The goal of the Bethesda System is to introduce a uniformreporting system for cervical smears. This is an important firststep in introducing a national system of quality control. The

mandatory reporting of inadequate samples may have a majorimpact on the quality of the smears. Patients' compliance withtesting at appropriate intervals and/or with follow-up procedures,on the other hand, will not be directly affected, short of a majormedia campaign.

Is the Bethesda System a final document that will once and forall eliminate all problems with providing, screening, and report-ing cervical smears? Will it introduce uniform clinical handling ofthese lesions? And finally, will the system eliminate invasivecervical cancer, which is a preventable disease? This writer hassome doubts about it. The reporting system as such is only onepart of a complex cancer detection process (5). The reproducibil-ity of the diagnostic system, even with the simplified classifica-tion of low-grade and high-grade precancerous lesions must stillbe tested. While this subdivision is possible in the morphologicsense, it still does not imply that the handling of the patientsshould be different.

Although there is fairly good evidence that many of thelow-grade lesions will disappear spontaneously, many exceptionsto this rule exist. Further, no statistical analysis performed to dateindicates how many high-grade lesions are represented in thesmears classified as low-grade lesions. Anecdotal experiencesuggests that this occurrence is not uncommon. Thus, colposcopyof all the lesions, whether low- or high-grade, is still the prudentway to proceed.

Undoubtedly, with the passage of time, perhaps in the nextcentury, molecular biologic probes will become available. Theseprobes will allow an accurate prediction of the behavior ofprecancerous lesions, a task that unfortunately is not possibletoday by morphologic examination of the cytologic samples ortissue samples or by typing of human papillomaviruses (14).

However, the Bethesda System is a good beginning. As thewriters of the document stated, like any other human endeavor,this document can be amended in the future. At this time, it isworthy of a major trial on a national scale, provided that theresults are collated and evaluated. To my knowledge, no efforthas been made so far to create an agency that would assess theresults.

References(/) BOYES DA: The value of a Pap smear program and suggestions for its

implementation. Cancer 48:613-621, 1981(2) MILLER AB, LINDSAY J, HILL GB: Mortality from cancer of the uterus in

Canada and its relationship to screening for cancer of the cervix. Int J Cancer17:602-612, 1986

(3) GEIRSSON G: Organization of screening in technically advanced countries:Iceland. In Screening for Cancer of the Uterine Cervix (Hakama M, MillerAB, Day NE, eds). Lyons: 1ARC, 1986, pp 239-250

(4) HAKAMA M: Trends in the incidence of cervical cancer in the Nordiccountries. In Trends in Cancer Incidence, Causes and Practical Implications(Magnus K, ed). Washington, DC: Hemisphere 1988, pp 279-292

(5) Koss LG: The Papanicolaou test for cervical cancer detection: A triumphand a tragedy. JAMA 261:737-743, 1989

(6) Koss LG, STEWART FW, FOOTE FW JR, ET AL: Some histologic aspects ofbehavior of epidermoid carcinoma in situ and related lesions of the uterinecervix: A long-term prospective study. Cancer 12:1171-1193, 1963

(7) Koss LG: Dysplasia: A real concept or a misnomer? Obstet Gynecol51:374-379, 1978

(8) PATTEN SF JR: Diagnostic Cytology of the Uterine Cervix. 2nd ed. Basel: SKarger, 1978

(9) COCKER J, Fox H, LANCLEY FA: Consistency in the histological diagnosis ofepithelial abnormalities of the uterine cervix. J Clin PathoF 21:67-70, 1968

990 Journal of the National Cancer Institute

(10) SEYBOLT JF, JOHNSON WD: Cervical cytodiagnostic problems: A survey.Am J Obstet Gynecol 109:1089-1103, 1971

(//) RYLANDER E: Negative smears in women developing invasive cervicalcancer. Acta Obstet Gynecol Scand 56:115-118, 1977

(12) RICHART RM: Cervical intraepithelial neoplasia: A review. Pathol Annu3:301-328, 1973

(13) NATIONAL CANCER INSTITUTE WORKSHOP: The 1988 Bethesda system for

reporting vaginal/cervical cytological diagnoses. JAMA 262:567-576,1989

(14) LORINCZ AT, LANCASTER WD, KURMAN RJ, ET AL: Characterization ofhuman papillomaviruses in cervical neoplasias and their detection in routineclinical screening. In Viral Etiology of Cervical Cancer: Banbury Report 21(Peto R, zur Hausen H, eds). Cold Spring Harbor, NY: Cold Spring HarborLaboratory, 1986, pp 225-237

Cancer Is ComplexIs Simple

For the latest cancer informationfrom the National CancerInstitute's Cancer InformationService, write:

The National Cancer InstituteBuilding 31, Room 10A24Bethesda, Maryland 20892-3100U.S.A.

Or Call:1-800-4-CANCER*(1-800-422-6237)

Vol. 82, No. 12, June 20, 1990 EDITORIAL 991