Atypical squamous cells of undetermined significance : do the guidelines actually help? - ASCUS - Editorial

Journal of Family Practice, April, 1997 by W. Douglas Everett, Marcia J. Chesebro

Family physicians are deluged with recommendations from various professional groups to provide a host of preventive services, some of which may conflict. The important question that we should ask of these recommendations is whether they are evidence-based, and if so, what references were used? In 1994, a National Cancer Institute (NCI) workshop issued interim guidelines for the management of abnormal cervical cytology and published them in the Journal of the American Medical Association. A review of these guidelines appeared later in the American Family Physician. (1,2) We examined the evidence as it applies to the interim guidelines for evaluating the minimally abnormal Papanicolaou (Pap) smear and allow family physicians to judge for themselves whether the guidelines agree with the evidence.

The US Preventive Services Task Force (3) has examined the evidence for and against several clinical preventive services, including in some cases era analysis of the costs involved. Frame (4) has also published a well-respected analysis of clinical preventive services. Both reports agree that routine Pap smears are effective in decreasing mortality from invasive cervical cancer. The area of controversy, however, lies in the appropriate follow-up of abnormal cytologic findings.

In general, there is little controversy over the management of the findings are clearly dysplastic, namely, high-grade squamous intraepithelial lesions (moderate and severe dysplasia). With lesser degrees of cytologic abnormality, however, the management becomes less clear. To complicate this issue, clinicians have perceived a burgeoning volume of minimally abnormal smears, those with atypical squamous cells of unknown significance (ASCUS). The NCI concluded that "in most populations, a diagnosis of ASCUS may be expected in no more than 5% of Pap test findings. A greater frequency may represent overuse of that diagnosis." (1) No references for the finding were cited. The Centers for Disease Control and Prevention, in a comprehensive report of the true incidence of ASCUS smears, found a rate of 5.4% and did not list ASCUS to be an abnormal result. (5) In some communities, up to 30% of the Pap smears ate interpreted as unqualified ASCUS. (6)

THE RELATIONSHIP BETWEEN ASCUS AND CERVICAL CANCER

Sherman and Kelly (7) reviewed the Pap smears that preceded the diagnosis of carcinoma in situ (CIS) in 18 women and in two women with invasive cervical cancer (ICC). Of the average number of 9.5 preceding smears per patient collected over a median of 93.5 months, 30.1% were classified as ASCUS. The patients were from the outpatient population of Johns Hopkins Hospital between 1958 and 1989 (perhaps a population that would be prone to have a previous ASCUS smear). The study did not consider a control group of patients without CIS or ICC. It is thus impossible to draw conclusions about the relative risk of ASCUS for CIS or ICC; patients without CIS or ICC might have a similar rate of preceding ASCUS cytologic test results. A mathematical model illustrates the problem: If we assume a rate of ASCUS cytologic findings in the normal population to be 5.4%, and the average patient has had 9.5 cytologic examinations, then 40.1% of the patients would have at least one cytologic finding of ASCUS. The equation for this is:

cumulative probability = 1 [-e.sup.-(Ir)(t)] where Ir is the incidence rate (0.054) and t is the number of times (9.5) the cytologic test is performed. (8) The implication that an ASCUS cytologic finding might be a risk factor for ICC is, thus, unproved and there may likely be no association.

IS ASCUS A REPRODUCIBLE CYTOLOGIC FINDING?

Wright presented unpublished data at a recent American Society for Colposcopy and Cervical Pathology meeting (Orlando, Fla, March 1996) showing little or no agreement among six internationally acclaimed experts for identifying ASCUS cytologic findings. An earlier report on observer variation on histopathologic diagnosis and grading of cervical intraepithelial neoplasia (CIN) showed that the most important source of disagreement lay in the distinction of reactive squamous proliferation (equivalent to the current ASCUS) from CIN grade 1. (9) The Kappa test ratio for normal cytologic findings was 0.3798, for reactive findings 0.430, for any dysplasia 0.660, and for invasive carcinoma 0.832. The Kappa test ratio is calculated using the following formula:

(overall percent agreement--total agreement

expected by chance)/

(maximum possible agreement--total agreement

expected by chance)

Kappa test ratios below 0.40 are interpreted as minimal or poor agreement, 0.40 to 0.60 as fair agreement, 0.60 to 0.80 as good agreement, and greater than 0.80 as excellent agreement. (10) With such poor interreader reliability for reading cytologic findings without dysplasia, how can we effectively react to a report of ASCUS?

THE HPV EFFECT AND LGSIL

The Bethesda System added human papillomavirus (HPV) effect to the diagnosis of mild dysplasia to define low-grade squamous intraepithelial lesion (LGSIL). The arguments for this pairing were that (1) it is difficult to distinguish HPV effect from mild dysplasia, and (2) HPV is associated with invasive cervical cancer. Though a logical change, it had a substantial clinical effect. The number of LGSIL cytologic findings increased, leading to more referrals for colposcopy and biopsy. However, HPV is a very common vaginal inhabitant in sexually active women. Koutsky found a 16% HPV infection rate among college women at entrance into college, increasing to 60% after 3 years in college. In those with 5 to 23 sexual partners, the HPV infection rate was 85%. Within 24 months, 30% had a squamous intraepithehal lesion on cytology, which commonly reverted to normal under observation. (Unpublished data. American Society for Colposcopy and Cervical Pathology meeting. Orlando, Fla, March 19, 1996).