Cervicography: An Intermediate Triage Test for the Evaluation o f Cervical Atypia Daron G. Ferris, MD; Peter Payne, MD; Lawrence E. Frisch, MD Augusta and Athens, Georgia, and Areata, California Background. A substantial number of abnormal Papani­ pia. The histologic specimens confirmed evidence of colaou (Pap) smears are reported as demonstrating “cy­ cervical dysplasia for 166 women. Of these women, tologic atypia.” This finding may actually represent pre- cervicography detected 74.7% of those who had mild malignant cervical disease. Some of these patients may dysplasia, 87.5% of those who had moderate dysplasia, not be able to afford definitive colposcopic examina­ 75% of those who had severe dysplasia, and the one tions, and simply repeating cytologic testing may result patient who had cervical cancer. Most (93%) of the in missed treatable disease. The purpose of this study women with dysplasia that was undetected by cervicog­ was to evaluate the use of cervicography as an interme­ raphy had mild dysplasia. diate triage test for women with atypical cervical cytol­ ogy- Conclusions. Cervicography may be an effective inter­ Methods. Women with a recent smear demonstrating mediate triage test for the evaluation of young women otologic atypia were evaluated using colposcopy, bi­ with Pap smears demonstrating cytologic atypia. opsy, and cervicography. Key words. Cervix dysplasia; cervix neoplasms; cytolog- Results. Colposcopically directed cervical biopsies were ical techniques; cytology; cervicitis; photography. obtained from 224 of 685 women with cytologic aty­ ( / Fam Proa 1993; 37:463-468) Of the abnormal Papanicolaou (Pap) smears collected cytology smear.3 Benign cellular changes of infection, from young women, 56% are reported as demonstrating inflammation, repair, and atypia, once known collectively cytologic atypia, or, more specifically, atypical squamous as a class II Pap smear, are now reported as separate cells of undetermined significance (ASCUS).1 The reason entities. for this high rate of reported atypia is multifactorial. Approximately 15% to 25% of women with a Pap smear report of cytologic atypia actually have significant Closer scrutiny of smears by cytotechnologists, prompted premalignant cervical disease.4 Therefore, although there by the Clinical Laboratory Improvement Amendment of is no consensus, many physicians suggest that women 1988,2 may be one explanation. The new Bethesda Sys­ with cytologic atypia should have a colposcopic exami­ tem of cervical cytology classification has also modified nation. Such an aggressive, unselective threshold for col­ the morphologic categorizations of interpretation of the poscopy evaluation of minor abnormalities strains clinical capabilities and resources. Frequently, delays in colpos­ copy scheduling exacerbate patient anxiety over the ab­ Submitted, July 7, 1993. normal Pap smear report. These emotional concerns arc This paper was presented in part at the national meeting o f the American College of typically compounded by the expense of expert colpo­ Health Associations, San Francisco, California, May 1992. scopic examination. From the M edical Effectiveness Education and Research Program, Department o f A less aggressive management option is to obtain Family Medicine, Medical College o f Georgia, Augusta; the Student Health Service, subsequent Pap smears at more frequent intervals. This University of Georgia, Athens, Georgia; and the Student Health Service, Humboldt State University, Areata, California. Requests for reprints should be addressed to Daron option, however, involves serial office visits for the pa­ G. Ferris, M D, Department ofFam ilv Medicine, Medical C ollege o f Georgia, Augusta, tient. Furthermore, this means continuing to monitor the G4 30912-3500. © 1993 Appleton & Lange ISSN 0094-3509 463 The fournal of Family Practice, Vol. 37, No. 5, 1993 Cervicography Ferris, Payne, and Frisch abnormality by means of a screening test that is relatively cally uninterpretable. Cervigram and histologic evalua­ insensitive compared with colposcopy. A more moderate tors were mutually blinded to the results. management option or intermediate triage test to evalu­ Colposcopy was performed, and colposcopically di- ate women with atypical cervical cytology, one that was rected biopsies were obtained by traditional methods. 12 technically uncomplicated, quick, inexpensive, and sensi­ Only subjects with colposcopically significant cervical tive, could be useful. changes of the abnormal transformation zone had a tissue Cervicography (National Testing Laboratories, Fen­ biopsy. Histologic specimens were evaluated at each clin­ ton, Mo) is a rapid and relatively inexpensive test used to ical site reference laboratory by certified pathologists. detect cervical abnormalities.5-6 Cervicography is most commonly used as a Pap smear adjunct to increase the Statistical Methods sensitivity of detecting premalignant and malignant cer­ vical disease.7-8 Cervicography is based on colposcopic For data analyses, atypical cervigrams were considered as principles, and includes a photographic examination of positive tests, since frequently the findings are indicative the cervix after the application of acetic acid. The “cervi- of minimally significant premalignant change. “High- gram,” or photographic slide, is taken by suitably trained risk” patients were defined as those having a current or clinicians and interpreted by expert evaluators. A written prior human papillomavirus infection of the lower geni­ evaluation, along with the cervigram, is returned to the tal tract; having a history of an abnormal Pap smear; or clinician and included in the medical record. Cervicogra­ being older than 50 years of age. phy has been shown to be more sensitive but less specific The proportion of patients in each dysplasia cate­ than the Pap smear in detecting premalignant cervical gory who were correctly classified as atypical or positive disease.8-11 by cervigram was calculated as a measure of the effective­ The purpose of this investigation was to evaluate the ness of cervicography as a triage test. Ninety-five percent use of cervicography as an intermediate triage test for confidence intervals (Cl) for the proportions were calcu­ women with atypical cervical cytology smear reports. lated, based on the F-distribution. The Mantel-Haenszei chi-square statistic was used to test the statistical signif­ icance of the association of cervigram category and his­ tology category, incorporating the ordinal nature of the Methods classifications. Kendall’s tau statistic was calculated as a measure of the strength of association. Subjects were recruited from the Family Medicine Center and Student Health Service, Medical College of Georgia, Augusta, Georgia; the Student Health Service, Univer­ Results sity of Georgia, Athens, Georgia; and the Student Health Center, Humboldt State University, Areata, California. A total of 1449 subjects were screened by a Pap smear The inclusion criteria were that the woman be at least 18 and cervicography. Pap smears detected cytologic atypia years of age with a recent Pap smear indicating cytologic in 685 women. The majority of these women received an atypia.3 The exclusion criteria were clinically apparent examination by colposcopy; however, some subjects re­ cervicitis, menses, surgical absence of the cervix uteri, and fused colposcopy, opting for management by repeat Pap cytologic evidence of a low-grade or high-grade squa­ smears. Occasionally women received colposcopy exam­ mous intraepithelial lesion (mild to severe dysplasia). inations by another clinician whose findings were un­ Cervigrams were taken by manufacturer-trained cli­ available. Evaluation and treatment for some women nicians according to the specified technique. The cervix were also managed by the clinicians, based on the ap­ was gently swabbed twice with a solution of 5% acetic pearance of the cervigram. acid. It was clearly visualized, focused through the cervi- Colposcopically directed cervical biopsies were ob­ scope (camera), and two cervigrams were taken. The film tained from 224 women. Cervical dysplasia was histolog­ was returned to the manufacturer for processing and then ically confirmed in 166 of the 685 (24% ) women with interpreted by certified evaluators as follows: “negative” atypical cytology. Cervigrams were technically adequate if normal; “atypical” if there was evidence of an ace- for 220 o f 224 (98% ) subjects. Obscuring cervical mucus towhite lesion outside the transformation zone, or inside was noted on the remaining four cervigrams. the transformation zone but of doubtful significance, or Demographic data, based on those women who had of atypical immature squamous metaplasia; “positive” if a cervical biopsy, revealed a mean subject age of 23.8 there was evidence of minor or major grade lesion or years. No subjects were pregnant, and 86% were nullip- cancer; and “technically defective” if the film was techni­ arous. A history of previous human papillomavirus in- 46 4 The Journal of Family Practice, Vol. 37, No. 5, 1993 Cervicography Ferris, Payne, and Frisch Cer\'icography Use for Triage of AtypicalPapanicolaou Smear Reports* (N = 224) Cervicography Histology Negative! Atypical! Positive§ Totals Negative 2 7 3 12 Atypia 11 13 21 45 Mild dysplasia 37 48 61 146 Moderate dvsplasia 2 5 9 16 Severe dysplasia 1 0 3 4 Cancer 0 0 1 1 Totals 53 73 98 224 *Mantel-Haenszel chi-square - 1.425; P - .233 for association o f cerrigram and histology. Kendall’s tau = .06 for measure o f association.