- Email: [email protected]
Cervicography combined with repeat Papanicolaou test as triage for low-grade cytologic abnormalities1
Cervicography Combined With Repeat Papanicolaou Test as Triage for Low-Grade Cytologic Abnormalities CAROLYN ESKRIDGE, MD, WALLACE P. BEGNEAUD, MD, AND CRAIG LANDWEHR, MD Objective: To determine the efficacy of combined repeat Papanicolaou test and cervicography for identification of high-grade squamous lesions among patients who previously have had atypical squamous cells of undetermined significance (ASCUS) or low-grade squamous intraepithelial lesion (LSIL) diagnosed by the Bethesda System. Methods: All patients who presented to a charity hospital gynecology clinic between July 1, 1994, and December 31, 1995, with a Papanicolaou test result of ASCUS or LSIL underwent repeat Papanicolaou test, cervicography, colposcopy with colposcopic-directed biopsy when appropriate, and an endocervical curettage. Pregnant patients and those who had undergone hysterectomy were excluded. The mean and median time from abnormal Papanicolaou test to clinical investigation was 61 and 58 days, respectively (range 6 –162 days). Results: One hundred eighty-seven of the patients had an initial ASCUS Papanicolaou test. Of these 187, 24 (13%) were found to have cervical intraepithelial neoplasia (CIN) grades II–III. The second Papanicolaou test would have resulted in the detection of 11 of 24 lesions (sensitivity 46%), whereas the cervigram would have detected 22 of 24 (sensitivity 92%). The combined Papanicolaou test and cervigram sensitivity for ASCUS was 91%. One hundred forty-one of the patients had an initial LSIL Papanicolaou test. Of these 141, 37 (26%) patients were found to have CIN II–III. The repeat Papanicolaou test would have detected 29 of 37 lesions (sensitivity 78%), whereas the cervigram would have detected 33 of 37 (sensitivity 89%). The combined Papanicolaou test and cervigram sensitivity for LSIL was 97%. Conclusion: Cervicography is a helpful adjunctive technique for detection of CIN II–III lesions in patients with previous ASCUS or LSIL Papanicolaou tests. (Obstet GyFrom the Department of Obstetrics and Gynecology, Ochsner Clinic, New Orleans, and the Departments of Obstetrics and Gynecology and Pathology, Leonard J. Chabert Medical Center, Houma, Louisiana. The authors thank the National Laboratories of Fenton, Missouri, for furnishing all cervicography equipment, materials, and cervigram interpretation free of charge, and to Michael J. Campion, MD, and Mitchell Greenburg, MD, Graduate Hospital, Philadelphia, Pennsylvania, for the interpretation of cervigrams.
VOL. 92, NO. 3, SEPTEMBER 1998
necol 1998;92:351–5. © 1998 by The American College of Obstetricians and Gynecologists.)
The Bethesda System for reporting cervical and vaginal cytologic diagnoses was introduced in 1988 and further revised in 1991 to facilitate communication of cytologic findings to the referring physician in unambiguous diagnostic terms that have clinical relevance and that reflect the current understanding of cervical and vaginal neoplasia.1,2 The Bethesda System has been adopted by more than 80% of cytology laboratories in the United States (Herbst AL, Jones H, Reid R. Interpreting the new Bethesda classification system. Contemp OB/GYN 1993;38:86 –107). However, the Bethesda System terminology was not pretested before its introduction, and the full implication of some categories is not known.3 Two such categories are atypical squamous cells of undetermined significance (ASCUS) and low-grade squamous intraepithelial lesion (LSIL). Clinicians have the greatest problem with these abnormalities because their behavior remains uncertain due to limited experience with the Bethesda System and because the largest volume of patients who require evaluation also have such abnormalities.4 In studies published before the introduction of the Bethesda System, 7–16% of women with atypical cells on Papanicolaou test were found to have underlying cervical intraepithelial neoplasia (CIN) grades II–III lesions.5– 8 In studies that used the Bethesda System, 4.8 –23% of patients with ASCUS Papanicolaou tests have such lesions.9 –12 In 1986, Campion et al13 reported that 26% of women with mild dysplasia on Papanicolaou test had moderate to severe dysplasia (CIN II–III) on colposcopic biopsy. Studies using the Bethesda System terminology have reported that approximately 22%
0029-7844/98/$19.00 PII S0029-7844(98)00199-9
351
of patients with LSIL Papanicolaou tests have CIN II–III.12,14 Because high-grade squamous lesions of the cervix progress to invasive cancer in a large percentage of cases, it is important to identify and treat patients with such lesions in a timely fashion.15–17 The annual cost of colposcopic examination and interventional therapy for low-grade cervical cytologic abnormalities (ASCUS and LSIL) has been estimated to be $6 billion in the United States.4 The third workshop of the National Cancer Institute on the Bethesda System convened in 1992 to address interim guidelines for the treatment of abnormal cervical cytologic findings and published its recommendations in 1994.4 Mentioned in this report was the possibility that human papillomavirus testing and cervicography may be helpful adjunctive techniques to prioritize patients with low-grade cytological abnormalities. However, the report stated that data from clinical trials of the efficacy of both methods are limited. The following study was conducted to determine the efficacy of using cervicography combined with repeat Papanicolaou test for identification of high-grade squamous cervical lesions among patients who present with ASCUS and LSIL Papanicolaou tests. The basic premise of the study is that either a suspicious lesion detected by cervigram or an epithelial cell abnormality (ASCUS, LSIL, and high-grade squamous intraepithelial lesion [HSIL]) on repeat Papanicolaou test would result in colposcopic examination and biopsy of any abnormal lesion found.
Materials and Methods The study was performed in the outpatient gynecology clinic at Leonard J. Chabert Medical Center in Houma, Louisiana. Chabert Medical Center is one of nine hospitals in the Louisiana Charity Hospital System that render care to indigent Louisiana residents. It is a major teaching hospital used by the Alton Ochsner Medical Foundation for graduate medical education. All patients who presented to the gynecology clinic between July 1, 1994 and December 31, 1995, with cervical Papanicolaou test reports of ASCUS or LSIL were included in the study. Pregnant patients and those who had undergone hysterectomy were excluded. The initial cytologic slides were not reviewed. Patients underwent investigation within 6 months of the date of the initial Papanicolaou test. The mean and median time from the Papanicolaou test to evaluation was 61 and 58 days, respectively (range 6 –162 days). Patients were not routinely checked for cervical or vaginal infection unless symptoms or signs were present to suggest cervicitis or vaginitis. If vaginitis was diagnosed, the patient was treated and returned at a
352 Eskridge et al
Cervicography/Repeat Papanicolaou Test
later date for follow-up examination. No cases of clinically diagnosed cervicitis occurred. Patients who had trichomoniasis on initial Papanicolaou test were treated and returned for follow up at a later date; those with Candida species or bacterial vaginosis on initial Papanicolaou test were not treated before examination. First, each patient had a repeat cervical cytologic smear performed using an Ayre spatula and cytobrush. All repeat cervical cytologic smears were processed and read by SmithKline Beecham Laboratories of New Orleans, Louisiana. All slides were screened by licensed cytotechnologists, and the results were reported using the Bethesda System terminology. The laboratory uses the following quality control measures: the slide of any patient with a history of an epithelial cell abnormality within the past 5 years is screened by two cytotechnologists, any specimen found to have an epithelial cell abnormality is rescreened by a cytopathologist, and 10% of Papanicolaou tests are randomly rescreened. For this study, abnormal Papanicolaou test results included ASCUS, LSIL, HSIL, and squamous cell carcinoma. Benign cellular and reactive changes were classified as normal. After the repeat Papanicolaou test, the cervix was cleansed with 5% acetic acid and a cerviscope was used to make two ectochrome exposures of the cervix as required by the National Testing Laboratories for cervigram interpretation.18 All cervigrams were processed by National Testing Laboratories and were viewed by one of two gynecologists experienced in cervigram interpretation. The categories of cervigram reporting are: normal, atypical, positive, and technically defective. The positive category is divided into four subcategories: probable normal variant, compatible with lowgrade lesion, compatible with high-grade lesion, and compatible with cancer. An atypical cervigram is one on which a lesion of doubtful significance is seen. National Testing Laboratories recommends a repeat Papanicolaou test and a repeat cervigram in 6 months for patients with atypical cervigrams to check for any change in the appearance of the lesion. For this study, atypical cervigrams were considered to be abnormal. The patient then underwent a colposcopic examination by gynecology resident physician; all colposcopies were viewed by one of two board-certified gynecology staff physicians with a combined colposcopic experience of 27 years. Colposcopic-directed biopsies were performed on patients with colposcopic abnormalities; no random biopsies were performed if no colposcopic abnormality was identified. Patients with unsatisfactory colposcopic findings who had an incompletely visualized lesion or an endocervical curettage showing dysplastic change or a repeat HSIL Papanicolaou test underwent either loop electrosurgical excision proce-
Obstetrics & Gynecology
Table 1. Colposcopic and Histologic Results Abnormality
Normal
CIN I
CIN II
CIN III
ASCUS (n 5 187) LSIL (n 5 141)
93 (50) 56 (40)
70 (37) 48 (34)
11 (6) 17 (12)
13 (7) 20 (14)
CIN 5 cervical intraepithelial neoplasia; ASCUS 5 atypical squamous cells of undetermined significance; LSIL 5 low-grade squamous intraepithelial lesion. Data are presented as n (%).
dure or cold knife conization of the cervix. The results of these procedures are included in the tabulation of our data. Patients with unsatisfactory colposcopic examination who had no colposcopically visible lesion with a normal endocervical curettage and a repeat Papanicolaou test result of normal, ASCUS, or LSIL were followed up with serial Papanicolaou tests. All biopsy and endocervical curettage specimens were processed in the pathology laboratory at Chabert Medical Center and were read by one of two boardcertified pathologists. No tissue specimens were reread in a blinded fashion by either the same or different pathologists. For purposes of reporting abnormal biopsy results, the categories cervical intraepithelial neoplasia (CIN) I, II, and III are used, where CIN I denotes mild dysplasia, CIN II denotes moderate dysplasia, and CIN III signifies severe dysplasia or carcinoma in situ. Because there is little distinction between koilocytosis and CIN I on the basis of clinical characteristics, histologic morphology, and human papillomavirus typing, condylomatous change or koilocytosis is categorized as CIN I.3 If there was a discrepancy between the colposcopic impression and histopathologic results, histopathologic results overruled the colposcopic impression.
Results There were 334 patients in the study; 191 (57%) were white, 137 (41%) were black, and six (2%) were Native American. The median age was 27 years (range 14 –73 years), and the median parity was 1 (range 0 –11). There were 189 patients with an initial Papanicolaou
Table 3. Sensitivity, Specificity, and Predictive Values for Repeat Papanicolaou Test, Cervicography, and Both Tests in the Atypical Squamous Cells of Undetermined Significance
Accuracy measure
Repeat Papanicolaou test (%)
Sensitivity Specificity Positive predictive value Negative predictive value
Repeat Papanicolaou plus Cervicography cervicography (%) (%)
46 64 16
91 30 30
91 45 19
56
98
97
test of ASCUS, of which two patients were not included in the study calculations because of a technically defective cervigram. In this category, 102 patients (54%) underwent colposcopic-directed biopsy. There were 145 patients with an initial Papanicolaou test of LSIL, of which two patients with technically defective cervigram and two patients with unsatisfactory Papanicolaou tests were not included in study calculations. In the LSIL group of 141 patients, 87 (62%) underwent colposcopicdirected biopsy. Table 1 depicts the results of colposcopic and histologic findings in the ASCUS and LSIL groups. It is noteworthy that 13% of the ASCUS group and 26% of the LSIL group were found to have CIN II–III. Table 2 correlates the colposcopic and histologic findings with the repeat Papanicolaou test and cervicography results for the ASCUS group. Table 3 depicts the sensitivity, specificity, positive predictive value, and negative predictive value for the repeat Papanicolaou test, for cervicography, and for combined repeat Papanicolaou test and cervicography in the ASCUS group. Table 4 correlates the colposcopic and histologic findings with the repeat Papanicolaou test and cervicography results for the LSIL group. Table 5 shows the sensitivity, specificity, positive predictive value, and negative predictive value for the repeat Papanicolaou
Table 2. Initial Atypical Squamous Cells of Undetermined Significance Papanicolaou Test (n 5 187*) Repeat Papanicolaou test
Repeat Papanicolaou plus cervicography
Cervicography
Colposcopichistologic findings
Normal
Abnormal
Normal
Abnormal
Normal
Abnormal
Normal (n 5 93) CIN I (n 5 70) CIN II–III (n 5 24)
70/93 (75%) 38/70 (54%) 13/24 (54%)
23/93 (25%) 32/70 (46%) 11/24 (46%)
74/93 (80%) 39/70 (56%) 2/24 (8%)
19/93 (20%) 31/70 (44%) 22/24 (92%)
55/93 (59%) 19/70 (27%) 2/24 (8%)
38/93 (41%) 51/70 (73%) 22/24 (92%)
CIN 5 cervical intraepithelial neoplasia. * Two patients with atypical squamous cells of undetermined significance were eliminated from the study because of a technically defective cervigram.
VOL. 92, NO. 3, SEPTEMBER 1998
Eskridge et al
Cervicography/Repeat Papanicolaou Test
353
Table 4. Initial Low-Grade Squamous Intraepithelial Lesion Papanicolaou Test (n 5 141*) Repeat Papanicolaou test
Cervicography
Repeat Papanicolaou plus cervicography
Colposcopichistologic findings
Normal
Abnormal
Normal
Abnormal
Normal
Abnormal
Normal (n 5 56) CIN I (n 5 48) CIN II–III (n 5 37)
29/56 (52%) 21/48 (44%) 8/37 (22%)
27/56 (48%) 27/48 (56%) 29/37 (78%)
42/56 (75%) 19/48 (40%) 4/37 (11%)
14/56 (25%) 29/48 (60%) 33/37 (89%)
21/56 (38%) 10/48 (21%) 1/37 (3%)
35/56 (62%) 38/48 (79%) 36/37 (97%)
CIN 5 cervical intraepithelial neoplasia. * Two patients were excluded because of a technically defective cervigram and two others because of an unsatisfactory Papanicolaou test.
test, for the cervicography, and for the combined Papanicolaou test and cervicography. There were 61 patients in the study with a high-grade squamous lesion of the cervix. The cervigram impression in these 61 cases was as follows: normal, 6; atypical, 8; compatible with low-grade lesion, 43; and compatible with high-grade lesion, 4. The colposcopic impression for these cases was: unsatisfactory, 3; CIN I lesion, 32; and CIN II–III, 26.
Discussion The findings of this study demonstrate that cervicography is a helpful adjunctive technique for detection of CIN II–III lesions in patients with previous ASCUS or LSIL Papanicolaou tests. Cervicography showed a superior sensitivity to repeat Papanicolaou test and detected more CIN II–III lesions in both the ASCUS and LSIL categories. Cervicography also showed a higher negative predictive value than did Papanicolaou test in both categories. It is both disturbing and noteworthy that in the ASCUS group the repeat Papanicolaou test would have led to detection of only 46% of CIN II–III lesions, whereas the cervigram would have detected 91% of such lesions. This finding leads us to question the recommendation of the 1992 National Cancer Institute workshop on the Bethesda System that repeat
Table 5. Sensitivity, Specificity, and Predictive Values for Repeat Papanicolaou Test, Cervicography, and Both Tests in the Low-Grade Squamous Intraepithelial Lesion Group
Accuracy measure Sensitivity Specificity Positive predictive value Negative predictive value
354 Eskridge et al
Repeat Papanicolaou test (%)
Repeat Papanicolaou plus Cervicography cervicography (%) (%)
78 48 34
89 59 44
97 32 36
83
90
97
Cervicography/Repeat Papanicolaou Test
Papanicolaou test alone is an acceptable follow-up strategy for ASCUS Papanicolaou tests. In fact, in our study, the repeat Papanicolaou test did not add to the sensitivity of the cervigram in detecting CIN II–III lesions in patients in the ASCUS group. However, in the LSIL group the repeat Papanicolaou test did increase the sensitivity and negative predictive value of the cervigram. Given a sensitivity of 91% for the ASCUS group and 97% for the LSIL group and a negative predictive value of 97% for both the ASCUS and LSIL groups, this method of triage could obviate the necessity for colposcopy in cases in which the repeat Papanicolaou test and cervigram are both normal. If we had followed this strategy in this group of patients, we would have reduced our colposcopic load by 36%, which would have resulted in cost savings. The cost of a cervigram is $25.00; in our locale the cost of a Papanicolaou test is $15.00, colposcopy fees average $100.00 (without biopsy), and the cost of an endocervical curettage is $75.00. The cost of performing a repeat Papanicolaou test, colposcopy, and an endocervical curettage on all 334 patients included in the study would be $63,435.00. If we had performed a repeat Papanicolaou test and cervicography on all 334 patients and had given only patients with an abnormal repeat Papanicolaou test and/or cervigram a colposcopy, we would have eliminated 120 colposcopies and endocervical curettages, and the total cost would be $51,225.00 (net savings of $12,210). If we had not performed colposcopy on the 120 patients with both a normal repeat Papanicolaou test and a normal cervigram, then three high-grade cervical lesions would have been missed. We cannot state that the money saved by this strategy justifies missing these cervical lesions. Our data also demonstrate that in the 61 CIN II–III lesions determined histopathologically, the cervigram classified the lesion as high grade in only four instances (8%). Most of these lesions were compatible with a low-grade or an atypical cervical lesion. Accordingly, the cervigram cannot be used to differentiate between low-grade and high-grade lesions. Its use is to deter-
Obstetrics & Gynecology
mine which patients have cervical lesions that need colposcopic evaluation. Our data indicate that this method of triage was not efficacious in detecting CIN I lesions, and it should not be used for this purpose. We do acknowledge that because the histologic slides were not reread in a blinded fashion, it is possible that some cases that were diagnosed as CIN I may have been metaplastic lesions that were interpreted incorrectly, with a resulting adverse effect on the sensitivity of the method. Grimes (Grimes D. Screening tests: What they are and what they aren’t. Contemporary OB/GYN 1982;20:69 – 78) suggested that the characteristics of an ideal screening test are as follows: the disease being screened for must be an important one which is prevalent and for which a recognizable, treatable latent stage exists; and the test must be safe, acceptable to patients, valid, and reasonably priced. Cervicography combined with repeat Papanicolaou test fulfills these requirements when used as a screening method to identify CIN II–III cervical lesions in patients with previous ASCUS or LSIL Papanicolaou tests. The apparent weakness of this triage method is the fact that 64% of the patients screened in this study would have still required colposcopic evaluation.
8.
9.
10.
11.
12.
13.
14.
15. 16. 17.
References 1. National Cancer Institute Workshop. The 1988 Bethesda system for reporting cervical/vaginal cytological diagnoses. JAMA 1989;262: 931– 4. 2. Broder S. From the National Institutes of Health. JAMA 1992;267: 1892. 3. American College of Obstetricians and Gynecologists. Cervical cytology: Evaluation and management of abnormalities. ACOG technical bulletin no. 183. Washington, DC: American College of Obstetricians and Gynecologists, 1993. 4. Kurman RJ, Henson DE, Herbst AL, Noller KL, Schiffman MH, for the 1992 National Cancer Institute Workshop. Interim guidelines for management of abnormal cervical cytology. JAMA 1994;271: 1866 –9. 5. Davis GL, Hernandez E, Davis JL, Miyazawa K. Atypical squamous cells in Papanicolaou smears. Obstet Gynecol 1987;69:43– 6. 6. Melamed MR, Flehinger BJ. Non-diagnostic squamous atypia in cervico-vaginal cytology as a risk factor for early neoplasia. Acta Cytol 1976;20:108 –10. 7. Paavonen J, Kiviat NB, Wolner-Hanssen P, Stevens CE, Vontver
VOL. 92, NO. 3, SEPTEMBER 1998
18.
LA, Brockway J, et al. Significance of mild cervical cytologic atypia in a sexually transmitted disease clinic population. Acta Cytol 1989;33:831– 8. Lindheim SR, Smith-Nguyen G. Aggressive evaluation for atypical squamous cells in Papanicolaou smears. J Reprod Med 1990;35: 971–3. Widra EA, Dookham D, Jordan A, McCue P, Bibbo M, Dunton CJ. Evaluation of the atypical cytologic smear. Validity of the 1991 Bethesda System. J Reprod Med 1994;39:682– 4. Delmore J, Williams CM, Nielsen ML, Kluzak T, Horbelt DV, Stembridge TW, et al. Atypical squamous cells of undetermined significance: Community incidence and management review. Kans Med 1995;96:133– 4. Pearlstone AC, Grigsby PW, Mutch DG. High rates of atypical cervical cytology: Occurrence and clinical significance. Obstet Gynecol 1992;80:191–5. Lonky NM, Navarre GL, Saunders S, Sadeghi M, Wolde-Tsadik G. Low-grade Papanicolaou smears and the Bethesda System: A prospective cytohistopathologic analysis. Obstet Gynecol 1995;85: 716 –20. Campion MJ, McCance DJ, Cuzick J, Singer A. Progressive potential of mild cervical atypia: Prospective cytological, colposcopic, and virological study. Lancet 1986;2:237– 40. Roland PY, Naumann RW, Alvarez RD, Kilgore LC, Partridge EE. A decision analysis of practice patterns used in evaluating and treating abnormal Pap smears. Gynecol Oncol 1995;59:75– 80. Richart RM, Barron BA. A follow-up study of patients with cervical dysplasia. Am J Obstet Gynecol 1969;105:386 –93. Kottmeier HL. The development and treatment of epitheliomas (Fr). Rev Fr Gynecol Obstet 1961;56:821– 6. Koss LG, Stewart FW, Foote FW, Jordan MJ, Bader GM, Day E. Some histological aspects of behavior of epidermoid carcinoma in situ and related lesions of the uterine cervix: A long-term prospective study. Cancer 1963;16:1160 –221. Stafl A. Cervicography: A new method for cervical cancer detection. Am J Obstet Gynecol 1981;139:815–25.
Address reprint requests to:
Wallace P. Begneaud, MD Chabert Medical Center 1978 Industrial Boulevard Houma, LA 70363
Received November 18, 1997. Received in revised form March 16, 1998. Accepted March 26, 1998. Copyright © 1998 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
Eskridge et al
Cervicography/Repeat Papanicolaou Test
355
VOL. 92, NO. 3, SEPTEMBER 1998
necol 1998;92:351–5. © 1998 by The American College of Obstetricians and Gynecologists.)
The Bethesda System for reporting cervical and vaginal cytologic diagnoses was introduced in 1988 and further revised in 1991 to facilitate communication of cytologic findings to the referring physician in unambiguous diagnostic terms that have clinical relevance and that reflect the current understanding of cervical and vaginal neoplasia.1,2 The Bethesda System has been adopted by more than 80% of cytology laboratories in the United States (Herbst AL, Jones H, Reid R. Interpreting the new Bethesda classification system. Contemp OB/GYN 1993;38:86 –107). However, the Bethesda System terminology was not pretested before its introduction, and the full implication of some categories is not known.3 Two such categories are atypical squamous cells of undetermined significance (ASCUS) and low-grade squamous intraepithelial lesion (LSIL). Clinicians have the greatest problem with these abnormalities because their behavior remains uncertain due to limited experience with the Bethesda System and because the largest volume of patients who require evaluation also have such abnormalities.4 In studies published before the introduction of the Bethesda System, 7–16% of women with atypical cells on Papanicolaou test were found to have underlying cervical intraepithelial neoplasia (CIN) grades II–III lesions.5– 8 In studies that used the Bethesda System, 4.8 –23% of patients with ASCUS Papanicolaou tests have such lesions.9 –12 In 1986, Campion et al13 reported that 26% of women with mild dysplasia on Papanicolaou test had moderate to severe dysplasia (CIN II–III) on colposcopic biopsy. Studies using the Bethesda System terminology have reported that approximately 22%
0029-7844/98/$19.00 PII S0029-7844(98)00199-9
351
of patients with LSIL Papanicolaou tests have CIN II–III.12,14 Because high-grade squamous lesions of the cervix progress to invasive cancer in a large percentage of cases, it is important to identify and treat patients with such lesions in a timely fashion.15–17 The annual cost of colposcopic examination and interventional therapy for low-grade cervical cytologic abnormalities (ASCUS and LSIL) has been estimated to be $6 billion in the United States.4 The third workshop of the National Cancer Institute on the Bethesda System convened in 1992 to address interim guidelines for the treatment of abnormal cervical cytologic findings and published its recommendations in 1994.4 Mentioned in this report was the possibility that human papillomavirus testing and cervicography may be helpful adjunctive techniques to prioritize patients with low-grade cytological abnormalities. However, the report stated that data from clinical trials of the efficacy of both methods are limited. The following study was conducted to determine the efficacy of using cervicography combined with repeat Papanicolaou test for identification of high-grade squamous cervical lesions among patients who present with ASCUS and LSIL Papanicolaou tests. The basic premise of the study is that either a suspicious lesion detected by cervigram or an epithelial cell abnormality (ASCUS, LSIL, and high-grade squamous intraepithelial lesion [HSIL]) on repeat Papanicolaou test would result in colposcopic examination and biopsy of any abnormal lesion found.
Materials and Methods The study was performed in the outpatient gynecology clinic at Leonard J. Chabert Medical Center in Houma, Louisiana. Chabert Medical Center is one of nine hospitals in the Louisiana Charity Hospital System that render care to indigent Louisiana residents. It is a major teaching hospital used by the Alton Ochsner Medical Foundation for graduate medical education. All patients who presented to the gynecology clinic between July 1, 1994 and December 31, 1995, with cervical Papanicolaou test reports of ASCUS or LSIL were included in the study. Pregnant patients and those who had undergone hysterectomy were excluded. The initial cytologic slides were not reviewed. Patients underwent investigation within 6 months of the date of the initial Papanicolaou test. The mean and median time from the Papanicolaou test to evaluation was 61 and 58 days, respectively (range 6 –162 days). Patients were not routinely checked for cervical or vaginal infection unless symptoms or signs were present to suggest cervicitis or vaginitis. If vaginitis was diagnosed, the patient was treated and returned at a
352 Eskridge et al
Cervicography/Repeat Papanicolaou Test
later date for follow-up examination. No cases of clinically diagnosed cervicitis occurred. Patients who had trichomoniasis on initial Papanicolaou test were treated and returned for follow up at a later date; those with Candida species or bacterial vaginosis on initial Papanicolaou test were not treated before examination. First, each patient had a repeat cervical cytologic smear performed using an Ayre spatula and cytobrush. All repeat cervical cytologic smears were processed and read by SmithKline Beecham Laboratories of New Orleans, Louisiana. All slides were screened by licensed cytotechnologists, and the results were reported using the Bethesda System terminology. The laboratory uses the following quality control measures: the slide of any patient with a history of an epithelial cell abnormality within the past 5 years is screened by two cytotechnologists, any specimen found to have an epithelial cell abnormality is rescreened by a cytopathologist, and 10% of Papanicolaou tests are randomly rescreened. For this study, abnormal Papanicolaou test results included ASCUS, LSIL, HSIL, and squamous cell carcinoma. Benign cellular and reactive changes were classified as normal. After the repeat Papanicolaou test, the cervix was cleansed with 5% acetic acid and a cerviscope was used to make two ectochrome exposures of the cervix as required by the National Testing Laboratories for cervigram interpretation.18 All cervigrams were processed by National Testing Laboratories and were viewed by one of two gynecologists experienced in cervigram interpretation. The categories of cervigram reporting are: normal, atypical, positive, and technically defective. The positive category is divided into four subcategories: probable normal variant, compatible with lowgrade lesion, compatible with high-grade lesion, and compatible with cancer. An atypical cervigram is one on which a lesion of doubtful significance is seen. National Testing Laboratories recommends a repeat Papanicolaou test and a repeat cervigram in 6 months for patients with atypical cervigrams to check for any change in the appearance of the lesion. For this study, atypical cervigrams were considered to be abnormal. The patient then underwent a colposcopic examination by gynecology resident physician; all colposcopies were viewed by one of two board-certified gynecology staff physicians with a combined colposcopic experience of 27 years. Colposcopic-directed biopsies were performed on patients with colposcopic abnormalities; no random biopsies were performed if no colposcopic abnormality was identified. Patients with unsatisfactory colposcopic findings who had an incompletely visualized lesion or an endocervical curettage showing dysplastic change or a repeat HSIL Papanicolaou test underwent either loop electrosurgical excision proce-
Obstetrics & Gynecology
Table 1. Colposcopic and Histologic Results Abnormality
Normal
CIN I
CIN II
CIN III
ASCUS (n 5 187) LSIL (n 5 141)
93 (50) 56 (40)
70 (37) 48 (34)
11 (6) 17 (12)
13 (7) 20 (14)
CIN 5 cervical intraepithelial neoplasia; ASCUS 5 atypical squamous cells of undetermined significance; LSIL 5 low-grade squamous intraepithelial lesion. Data are presented as n (%).
dure or cold knife conization of the cervix. The results of these procedures are included in the tabulation of our data. Patients with unsatisfactory colposcopic examination who had no colposcopically visible lesion with a normal endocervical curettage and a repeat Papanicolaou test result of normal, ASCUS, or LSIL were followed up with serial Papanicolaou tests. All biopsy and endocervical curettage specimens were processed in the pathology laboratory at Chabert Medical Center and were read by one of two boardcertified pathologists. No tissue specimens were reread in a blinded fashion by either the same or different pathologists. For purposes of reporting abnormal biopsy results, the categories cervical intraepithelial neoplasia (CIN) I, II, and III are used, where CIN I denotes mild dysplasia, CIN II denotes moderate dysplasia, and CIN III signifies severe dysplasia or carcinoma in situ. Because there is little distinction between koilocytosis and CIN I on the basis of clinical characteristics, histologic morphology, and human papillomavirus typing, condylomatous change or koilocytosis is categorized as CIN I.3 If there was a discrepancy between the colposcopic impression and histopathologic results, histopathologic results overruled the colposcopic impression.
Results There were 334 patients in the study; 191 (57%) were white, 137 (41%) were black, and six (2%) were Native American. The median age was 27 years (range 14 –73 years), and the median parity was 1 (range 0 –11). There were 189 patients with an initial Papanicolaou
Table 3. Sensitivity, Specificity, and Predictive Values for Repeat Papanicolaou Test, Cervicography, and Both Tests in the Atypical Squamous Cells of Undetermined Significance
Accuracy measure
Repeat Papanicolaou test (%)
Sensitivity Specificity Positive predictive value Negative predictive value
Repeat Papanicolaou plus Cervicography cervicography (%) (%)
46 64 16
91 30 30
91 45 19
56
98
97
test of ASCUS, of which two patients were not included in the study calculations because of a technically defective cervigram. In this category, 102 patients (54%) underwent colposcopic-directed biopsy. There were 145 patients with an initial Papanicolaou test of LSIL, of which two patients with technically defective cervigram and two patients with unsatisfactory Papanicolaou tests were not included in study calculations. In the LSIL group of 141 patients, 87 (62%) underwent colposcopicdirected biopsy. Table 1 depicts the results of colposcopic and histologic findings in the ASCUS and LSIL groups. It is noteworthy that 13% of the ASCUS group and 26% of the LSIL group were found to have CIN II–III. Table 2 correlates the colposcopic and histologic findings with the repeat Papanicolaou test and cervicography results for the ASCUS group. Table 3 depicts the sensitivity, specificity, positive predictive value, and negative predictive value for the repeat Papanicolaou test, for cervicography, and for combined repeat Papanicolaou test and cervicography in the ASCUS group. Table 4 correlates the colposcopic and histologic findings with the repeat Papanicolaou test and cervicography results for the LSIL group. Table 5 shows the sensitivity, specificity, positive predictive value, and negative predictive value for the repeat Papanicolaou
Table 2. Initial Atypical Squamous Cells of Undetermined Significance Papanicolaou Test (n 5 187*) Repeat Papanicolaou test
Repeat Papanicolaou plus cervicography
Cervicography
Colposcopichistologic findings
Normal
Abnormal
Normal
Abnormal
Normal
Abnormal
Normal (n 5 93) CIN I (n 5 70) CIN II–III (n 5 24)
70/93 (75%) 38/70 (54%) 13/24 (54%)
23/93 (25%) 32/70 (46%) 11/24 (46%)
74/93 (80%) 39/70 (56%) 2/24 (8%)
19/93 (20%) 31/70 (44%) 22/24 (92%)
55/93 (59%) 19/70 (27%) 2/24 (8%)
38/93 (41%) 51/70 (73%) 22/24 (92%)
CIN 5 cervical intraepithelial neoplasia. * Two patients with atypical squamous cells of undetermined significance were eliminated from the study because of a technically defective cervigram.
VOL. 92, NO. 3, SEPTEMBER 1998
Eskridge et al
Cervicography/Repeat Papanicolaou Test
353
Table 4. Initial Low-Grade Squamous Intraepithelial Lesion Papanicolaou Test (n 5 141*) Repeat Papanicolaou test
Cervicography
Repeat Papanicolaou plus cervicography
Colposcopichistologic findings
Normal
Abnormal
Normal
Abnormal
Normal
Abnormal
Normal (n 5 56) CIN I (n 5 48) CIN II–III (n 5 37)
29/56 (52%) 21/48 (44%) 8/37 (22%)
27/56 (48%) 27/48 (56%) 29/37 (78%)
42/56 (75%) 19/48 (40%) 4/37 (11%)
14/56 (25%) 29/48 (60%) 33/37 (89%)
21/56 (38%) 10/48 (21%) 1/37 (3%)
35/56 (62%) 38/48 (79%) 36/37 (97%)
CIN 5 cervical intraepithelial neoplasia. * Two patients were excluded because of a technically defective cervigram and two others because of an unsatisfactory Papanicolaou test.
test, for the cervicography, and for the combined Papanicolaou test and cervicography. There were 61 patients in the study with a high-grade squamous lesion of the cervix. The cervigram impression in these 61 cases was as follows: normal, 6; atypical, 8; compatible with low-grade lesion, 43; and compatible with high-grade lesion, 4. The colposcopic impression for these cases was: unsatisfactory, 3; CIN I lesion, 32; and CIN II–III, 26.
Discussion The findings of this study demonstrate that cervicography is a helpful adjunctive technique for detection of CIN II–III lesions in patients with previous ASCUS or LSIL Papanicolaou tests. Cervicography showed a superior sensitivity to repeat Papanicolaou test and detected more CIN II–III lesions in both the ASCUS and LSIL categories. Cervicography also showed a higher negative predictive value than did Papanicolaou test in both categories. It is both disturbing and noteworthy that in the ASCUS group the repeat Papanicolaou test would have led to detection of only 46% of CIN II–III lesions, whereas the cervigram would have detected 91% of such lesions. This finding leads us to question the recommendation of the 1992 National Cancer Institute workshop on the Bethesda System that repeat
Table 5. Sensitivity, Specificity, and Predictive Values for Repeat Papanicolaou Test, Cervicography, and Both Tests in the Low-Grade Squamous Intraepithelial Lesion Group
Accuracy measure Sensitivity Specificity Positive predictive value Negative predictive value
354 Eskridge et al
Repeat Papanicolaou test (%)
Repeat Papanicolaou plus Cervicography cervicography (%) (%)
78 48 34
89 59 44
97 32 36
83
90
97
Cervicography/Repeat Papanicolaou Test
Papanicolaou test alone is an acceptable follow-up strategy for ASCUS Papanicolaou tests. In fact, in our study, the repeat Papanicolaou test did not add to the sensitivity of the cervigram in detecting CIN II–III lesions in patients in the ASCUS group. However, in the LSIL group the repeat Papanicolaou test did increase the sensitivity and negative predictive value of the cervigram. Given a sensitivity of 91% for the ASCUS group and 97% for the LSIL group and a negative predictive value of 97% for both the ASCUS and LSIL groups, this method of triage could obviate the necessity for colposcopy in cases in which the repeat Papanicolaou test and cervigram are both normal. If we had followed this strategy in this group of patients, we would have reduced our colposcopic load by 36%, which would have resulted in cost savings. The cost of a cervigram is $25.00; in our locale the cost of a Papanicolaou test is $15.00, colposcopy fees average $100.00 (without biopsy), and the cost of an endocervical curettage is $75.00. The cost of performing a repeat Papanicolaou test, colposcopy, and an endocervical curettage on all 334 patients included in the study would be $63,435.00. If we had performed a repeat Papanicolaou test and cervicography on all 334 patients and had given only patients with an abnormal repeat Papanicolaou test and/or cervigram a colposcopy, we would have eliminated 120 colposcopies and endocervical curettages, and the total cost would be $51,225.00 (net savings of $12,210). If we had not performed colposcopy on the 120 patients with both a normal repeat Papanicolaou test and a normal cervigram, then three high-grade cervical lesions would have been missed. We cannot state that the money saved by this strategy justifies missing these cervical lesions. Our data also demonstrate that in the 61 CIN II–III lesions determined histopathologically, the cervigram classified the lesion as high grade in only four instances (8%). Most of these lesions were compatible with a low-grade or an atypical cervical lesion. Accordingly, the cervigram cannot be used to differentiate between low-grade and high-grade lesions. Its use is to deter-
Obstetrics & Gynecology
mine which patients have cervical lesions that need colposcopic evaluation. Our data indicate that this method of triage was not efficacious in detecting CIN I lesions, and it should not be used for this purpose. We do acknowledge that because the histologic slides were not reread in a blinded fashion, it is possible that some cases that were diagnosed as CIN I may have been metaplastic lesions that were interpreted incorrectly, with a resulting adverse effect on the sensitivity of the method. Grimes (Grimes D. Screening tests: What they are and what they aren’t. Contemporary OB/GYN 1982;20:69 – 78) suggested that the characteristics of an ideal screening test are as follows: the disease being screened for must be an important one which is prevalent and for which a recognizable, treatable latent stage exists; and the test must be safe, acceptable to patients, valid, and reasonably priced. Cervicography combined with repeat Papanicolaou test fulfills these requirements when used as a screening method to identify CIN II–III cervical lesions in patients with previous ASCUS or LSIL Papanicolaou tests. The apparent weakness of this triage method is the fact that 64% of the patients screened in this study would have still required colposcopic evaluation.
8.
9.
10.
11.
12.
13.
14.
15. 16. 17.
References 1. National Cancer Institute Workshop. The 1988 Bethesda system for reporting cervical/vaginal cytological diagnoses. JAMA 1989;262: 931– 4. 2. Broder S. From the National Institutes of Health. JAMA 1992;267: 1892. 3. American College of Obstetricians and Gynecologists. Cervical cytology: Evaluation and management of abnormalities. ACOG technical bulletin no. 183. Washington, DC: American College of Obstetricians and Gynecologists, 1993. 4. Kurman RJ, Henson DE, Herbst AL, Noller KL, Schiffman MH, for the 1992 National Cancer Institute Workshop. Interim guidelines for management of abnormal cervical cytology. JAMA 1994;271: 1866 –9. 5. Davis GL, Hernandez E, Davis JL, Miyazawa K. Atypical squamous cells in Papanicolaou smears. Obstet Gynecol 1987;69:43– 6. 6. Melamed MR, Flehinger BJ. Non-diagnostic squamous atypia in cervico-vaginal cytology as a risk factor for early neoplasia. Acta Cytol 1976;20:108 –10. 7. Paavonen J, Kiviat NB, Wolner-Hanssen P, Stevens CE, Vontver
VOL. 92, NO. 3, SEPTEMBER 1998
18.
LA, Brockway J, et al. Significance of mild cervical cytologic atypia in a sexually transmitted disease clinic population. Acta Cytol 1989;33:831– 8. Lindheim SR, Smith-Nguyen G. Aggressive evaluation for atypical squamous cells in Papanicolaou smears. J Reprod Med 1990;35: 971–3. Widra EA, Dookham D, Jordan A, McCue P, Bibbo M, Dunton CJ. Evaluation of the atypical cytologic smear. Validity of the 1991 Bethesda System. J Reprod Med 1994;39:682– 4. Delmore J, Williams CM, Nielsen ML, Kluzak T, Horbelt DV, Stembridge TW, et al. Atypical squamous cells of undetermined significance: Community incidence and management review. Kans Med 1995;96:133– 4. Pearlstone AC, Grigsby PW, Mutch DG. High rates of atypical cervical cytology: Occurrence and clinical significance. Obstet Gynecol 1992;80:191–5. Lonky NM, Navarre GL, Saunders S, Sadeghi M, Wolde-Tsadik G. Low-grade Papanicolaou smears and the Bethesda System: A prospective cytohistopathologic analysis. Obstet Gynecol 1995;85: 716 –20. Campion MJ, McCance DJ, Cuzick J, Singer A. Progressive potential of mild cervical atypia: Prospective cytological, colposcopic, and virological study. Lancet 1986;2:237– 40. Roland PY, Naumann RW, Alvarez RD, Kilgore LC, Partridge EE. A decision analysis of practice patterns used in evaluating and treating abnormal Pap smears. Gynecol Oncol 1995;59:75– 80. Richart RM, Barron BA. A follow-up study of patients with cervical dysplasia. Am J Obstet Gynecol 1969;105:386 –93. Kottmeier HL. The development and treatment of epitheliomas (Fr). Rev Fr Gynecol Obstet 1961;56:821– 6. Koss LG, Stewart FW, Foote FW, Jordan MJ, Bader GM, Day E. Some histological aspects of behavior of epidermoid carcinoma in situ and related lesions of the uterine cervix: A long-term prospective study. Cancer 1963;16:1160 –221. Stafl A. Cervicography: A new method for cervical cancer detection. Am J Obstet Gynecol 1981;139:815–25.
Address reprint requests to:
Wallace P. Begneaud, MD Chabert Medical Center 1978 Industrial Boulevard Houma, LA 70363
Received November 18, 1997. Received in revised form March 16, 1998. Accepted March 26, 1998. Copyright © 1998 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
Eskridge et al
Cervicography/Repeat Papanicolaou Test
355