Colposcopy was introduced by Hinselmann1,2 in 1927 as a method of visualizing the cervical epithelium under low-power magnification to identify a specific morphological change that he called a “cancer nodule,” which might indicate the presence of a cervical cancer before it was clinically apparent.
Despite some enthusiasm in Europe for this technique during the 1930s, it was not until the early 1970s that colposcopy started to be used extensively in North America.3–7 Before this time, the treatment of women with abnormal Pap tests involved proceeding directly to a cone biopsy.8,9 However, that approach often led to high morbidity10,11 and to a significant number of unnecessary procedures. In contrast, colposcopy not only permitted accurate identification of lesions but also allowed for the use of ablative procedures, which resulted in fewer unnecessary morbid surgical procedures.
In the province of British Columbia, a provincial colposcopy service was established in 197312 and integrated into a similar province-wide cervical cancer–screening program that had been operative since 1949.13–16 Thus, in this province, colposcopy has served for roughly 30 years as a valuable diagnostic and evaluative intermediary between an abnormal Pap test and treatment. However, despite its inherent value, it has long been recognized that accurate colposcopic diagnosis also requires considerable skill and, hence, is a technique for which an appropriate initial training period is essential.
The accuracy of colposcopy as a diagnostic method in assessing cervical lesions has been questioned and has led some to advocate the liberal use of diagnostic excisional procedures for the initial investigation of women with abnormal Pap tests.17–19 This form of “see-and-treat” colposcopy, in which an abnormal Pap test is followed by colposcopy and immediate excision, potentially brings the clinician back full circle, negating the need for colposcopic proficiency, and results in a high rate of surgical procedures, often with minimal or no pathology, particularly if done for low-grade cytology. However, the actual abandonment of colposcopy would only be warranted if it can be shown that a high degree of colposcopic expertise cannot be maintained on a community-wide basis.
Quality control activities are integral to a cytology laboratory’s activities and are essential if accreditation status is to be maintained. However, similar requirements and standards for colposcopic practice have been elusive and difficult to determine.
The present study undertook to examine the quality of colposcopic practice in the province of British Columbia during the 2001 calendar year. The aim was to determine whether an analysis of the colposcopy and related cytology and histology data would reveal a sufficiently high standard of colposcopy practice being maintained on a province-wide basis.
MATERIALS AND METHODS
Currently, 37 gynecologists provide colposcopic services through 24 community-based hospital clinics. Referral to the various colposcopy clinics for diagnostic assessment and evaluation typically takes place when an abnormal Pap test is detected. It has been the policy of British Columbia Cancer Agency to recommend that all women with a Pap test showing moderate dyskaryosis or greater, ie, high-grade squamous intraepithelial lesions (HSIL), be referred for immediate colposcopic assessment. Colposcopy is additionally recommended if mild atypia or low-grade squamous intraepithelial lesions (LSIL) changes persist for 2 years. Recently, an increasing number of women are also being referred for colposcopy because of Pap test reports called “unsatisfactory.”
Although the cytology screening program makes recommendations according to the above guidelines, the final decision as to when colposcopy should be carried out and for what reason is left to the patient and her personal physician. Once a colposcopic evaluation has been completed, including any biopsy results, a general recommendation for further treatment or follow-up is made to the referring physician who then discusses this with the patient. If the report includes a suggestion for further therapy that could be provided through an ambulatory clinic, it is the responsibility of the primary care provider to arrange for the patient to return for treatment.
This organizational model was created to encourage orderly continuity of care and to minimize disruptions in the referral and practice patterns among local gynecologists, not all of whom may be directly involved in providing colposcopic services.
Gynecologists who are approved to provide diagnostic colposcopy services are required to undergo a standardized training period consisting of an intense 4-week clinical colposcopy preceptorship with cytology and histopathology modules. This standardized training period is provided to all participants regardless of previous training and experience. At the completion of this preceptorship, all prospective colposcopists are expected to successfully complete a multiple-choice examination, as well as maintain agreed-upon practice proficiency levels.
Each colposcopist’s correlation scores among cytology, colposcopy, and histology are evaluated annually and used to measure performance compared with that of the mean value for all peers. An individual’s performance results, together with the mean score for the entire group of colposcopists, is then made available for one’s personal evaluation. Confidentiality and anonymity of results for each colposcopist are maintained and are known only to the program director and the individual colposcopist. Gynecologists whose performance falls consistently below that of their peers are encouraged to return for further updates and training. To date, several practitioners have requested additional preceptorship training of their own volition, and it has not been necessary to request that individuals return for additional training.
All gynecologists performing colposcopy in regional British Columbia clinics use a standardized report form and terminology for colposcopic examinations. A copy of the results of each colposcopic examination is sent to the cervical screening program data management section at the British Columbia Cancer Agency, where it is incorporated into the patient’s cervical screening record. This information is then readily available for future reference to the colposcopist if subsequent examinations are undertaken at any approved clinic. This again insures a more complete continuity of care.
The central colposcopy program also sponsors an annual workshop in which the accumulated data of the previous year are reviewed and presented to the group with a variety of colposcopically related topics. This usually includes case presentations of unusual or difficult management problems, which are submitted by the various members. This activity helps maintain skills, competence, and interest in the program. In addition, any topics that colposcopists wish to have reviewed are also presented.
In the British Columbia screening program, Pap tests are processed and interpreted through one central cytology laboratory. This approach facilitates data recording and analysis.
In the calendar year 2001, a total of 579,430 cervical/endocervical smears from 534,461 women were submitted to the cytology screening program for interpretation. The numbers and percentages of patients with tests showing either mild atypia (LSIL) or moderate atypia or higher (HSIL) are presented in Table 1 in 10-year age groups. The estimated female population distribution, according to the latest official census figures, is also presented in Table 1 for comparison.
For the present analysis, data from every colposcopic examination and biopsy, which was forwarded to the data management section from January 1 to December 31, 2001, and incorporated into the patient’s cervical cancer screening record, were examined. Of the total 13,308 colposcopic assessments performed, 12,527 (94.1%) were for cervical examinations. These records were classified as either new patients (7,720), defined as a first colposcopy clinic visit, or a repeat visit (4,807) of previously treated individuals. The latter category was excluded from further analysis in this study to avoid duplication of records and its potential bias.
To provide an optimum cohort for determining the diagnostic proficiency of the colposcopists in the program, patients with unsatisfactory colposcopy were also excluded from the study (1,768 women). Colposcopy was called unsatisfactory if all margins of the transformation zone or the limits of any lesion visualized were not clear. The remaining 5,952 patients constituted the study population. The rate of unsatisfactory colposcopy was most commonly related to the patient’s age or whether there had been some prior cervical injury or surgery.
Three items—the referral cytology, colposcopic impression, and any related directed biopsy—were paired with one another and these 3 sets of paired comparisons—cytology versus colposcopic impression, cytology versus histology, and colposcopic impression versus histology—were then used as a guide to practitioner performance. Satisfactory agreement or correlation was defined as that point at which the cytology, histology, and/or colposcopic impression were within 1 degree of each other.
The histology of directed biopsies or any excisional procedures were read locally but were available for review if needed. Sensitivity was defined as the ability of colposcopy to correctly identify the presence of disease in the biopsy and was calculated by dividing the number of test-positive subjects by the number of diseased subjects. Specificity was defined as the ability of colposcopy to identify correctly the absence of disease in the biopsy. Specificity was calculated by dividing the number of test-negative subjects by the number of nondiseased subjects.
All patient records, information, and results in this study were used in accordance with the Institutional Research Guidelines for the British Columbia Cancer Agency and the University of British Columbia. Approval for this study was obtained from the University of British Columbia–British Columbia Cancer Agency Research Ethics Board.
Figure 1 presents the nomenclature for cytology smears as used in British Columbia, which is similar to that of the British Society of Clinical Cytology Classification (1986).20 The corresponding Bethesda 2001 cytology categories are also listed for comparison.
The most common cause of a new-patient referral for colposcopy was abnormal cytology (76.7%), with a further 14.9% seen because of a clinical abnormality of the cervix, and the remaining 8.4% for follow-up of treatments received in other provinces.
Recommendations for colposcopy are issued by the cytopathologist according to agreed-upon criteria established in joint consultation with the colposcopists. Figure 2 shows the level of compliance with the recommendation for colposcopy by patient’s age group and category of cytological atypia. Overall, the rate of compliance was high, exceeding 80% for women aged more than 30 years who had moderate atypia or higher. There was a general trend for increasing compliance with increasing age and a slight trend for better compliance when the atypias were moderate or higher. The poorest level of compliance occurred in women aged less than 20 years with mild atypia. It should be noted that the rate of compliance was determined by both the patient and the primary care provider.
The correlation between the most recent cytology findings leading to colposcopy and the results of the corresponding colposcopic impression is shown in Table 2. Cytology was not available for those patients who had their referral Pap test or treatment out of province (843 individuals). Agreement was defined as satisfactory when the colposcopic impression and the cytological diagnosis were within 1 degree of each other. This correlation occurred in 89.8% of patients. This suggests how prior knowledge of the presenting cytology may influence the colposcopist’s impression of any lesion seen. Nevertheless, in 460 (14.4%) of the 3,205 patients, the colposcopists did not see lesions or patterns to account for the high-grade cytological changes that were documented. Forty-four women had colposcopic findings that suggested more advanced pathology than what would be expected according to the cytological atypia seen. Overall, of the patients with high-grade Pap tests, a corresponding lesion was expected by the colposcopists in 87.5% of cases.
Table 3 presents the comparison between the referring cytology and histology for all new patients seen for the year 2001. A biopsy specimen was not obtained in 945 patients. A satisfactory correlation was determined to exist when the presenting cytology and the histology of the colposcopically obtained biopsies were essentially within one degree of each other. Satisfactory agreement was observed in 79.4% of all patients, but this percentage rises to 82.2% if the 1,137 women with incompletely visualized transformation zones or lesion boundaries (“unsatisfactory colposcopy”) are removed from the analysis. Cases falling outside of the highlighted bars in Table 3 were those in which the correlation was unsatisfactory. The 91 (1.8%) cases below the line appear to be instances of cytology underestimating disease, but these are essentially Pap test sampling errors because review of their cytology remained unchanged. Although interpretative errors can and do occur, review of our data show this to be an uncommon occurrence in this series. Four of these patients however had colposcopically overt cancers, which normally would not have been expected by their cytology grade.
Agreement was directly related to the degree of cytological atypia, with those patients having the most abnormal tests having the better correlation scores. A total of 943 (18.8%) patients appears as cytological “overcalls” or overestimates of disease. Many of these patients have gone on to further investigations, including conizations or loop excisional procedures. Analysis of these data are as yet incomplete but would be expected to improve the correlation rate. In using a cytological classification system that has no atypical squamous cells of undetermined significance (ASC-US) category, the cytologists are forced to call Pap tests either “within normal limits” or indicative of an intraepithelial lesion (SIL), which is responsible for some of the discordant cases.
Of the 97.4% of patients seen because of positive cytology, 64.7% were women with high-grade Pap tests. The sensitivity for a positive cytology in this study was 98.6%, with the predictive value for SIL of 63%.
Table 4 presents the comparison between the colposcopic impression of any lesions seen and the histology of the colposcopically directed biopsies. Missing information, either the colposcopic impression or histology, occurred in 684 (11.4%) cases. Good predictive scores were present in approximately 87% of patients. In 259 (5%) patients, the colposcopist underestimated the severity of the pathology found.
This would indicate that, although the correlation was not as expected, the colposcopist was still able to identify the area with the most significant pathology. Conversely, 438 (8.3%) women had less severe pathology on their directed biopsy than what was anticipated. Many of these patients, particularly if cytology confirmed high-grade Pap tests, would go on to further investigation, with the expected result of better overall correlation.
The data in the lower portion of Table 4, which uses a combined Bethesda/cervical intraepithelial neoplasia (CIN) terminology, shows the actual correlation between disease predicted and the actual disease found.
The sensitivity of colposcopy in this series was calculated as 90.3%, with a predictive value of a positive colposcopy of 75.9%. The specificity for no disease was lower (57.3%).
A review of the correlation indices by all colposcopists is presented in Table 5. Satisfactory performance was considered to have been achieved if one’s score in all of the comparison tables was within 10% of the mean for the entire group. Three individual’s scores for presenting cytology versus directed biopsies fell below this standard. In evaluating predictive or colposcopic diagnostic ability, 5 colposcopists had scores below the levels of performance as defined by the group.
Colposcopy, when first reintroduced 30 years ago, had as one of its main objectives the prevention of unnecessary cone biopsies. Conization was the usual method of evaluating women found to have Pap test abnormalities. Cone biopsies were often accompanied by significant degrees of morbidity and often found to have little or no pathology on these specimens. As cervical screening programs grew in their coverage and scope, concern was expressed that many of these procedures were being done on young, reproductive aged–group women with potentially adverse effects.
Colposcopy was shown in numerous studies3–6,21 to be a safe and effective procedure for assessing such women to rule out the presence of invasive disease so that less aggressive methods of therapy might be used in those with identified, high-grade premalignant lesions. Cryosurgery22–25 and carbon dioxide26–29 laser vaporization and, more recently, loop electrosurgical excision procedures (LEEP)17–19,30 have emerged as effective treatment methods for cervical dysplasia and related conditions in selected cases. These procedures can be applied in an ambulatory office setting and represent a significant savings to the patient and health care system over traditional methods of therapy. If ablative methods of treatment for CIN are to be used, it is crucial that the colposcopic assessment carried out before therapy be diagnostically accurate if treatment errors are to be avoided.
Management of patients with high-grade cytology is well understood, whereas women with lesser cytological atypias have not been handled in a uniform fashion. Data from the ASC-US LSIL Triage Studies Trial31,32 have provided good direction in this regard and were largely influential in shaping the recently published American Society for Colposcopy and Cervical Pathology (ASCCP)/National Cancer Institute (NCI) 2001 Consensus Guidelines33 for the management of CIN.
The results of this study show that the colposcopists rely on and are influenced by the patient’s presenting cytology in predicting the nature of any lesions seen.
In analyzing the cytology-histology correlations, it is obvious that low-grade cytology change, in particular, can seriously underestimate high-grade lesions including invasive cancer. These findings underscore the need for adequate cytology sampling if these types of errors are to be kept to a minimum. Prior studies34–42 have reported rates of disagreement between cytology and histology of 11% to 47%. In this study, the discordance between cytology and histology was noted to be 20.5%, not unlike the 16.5% reported by Jones and Novis37 in an American College of Pathology Q Probe study. Most of the cytology-histology correlation studies use histology as the arbiter with which the cytology results are compared or evaluated. Our results and those of several other studies suggest that it is the cytology results that are most indicative of the patient’s condition, rather than the biopsy. When discordant pairs are subjected to review, the cytological diagnosis rarely changes. Our policy has been to recommend further tissue sampling, particularly if high-grade cytology is confirmed. These additional tissue samples most often will correspond to the referral cytology and show a lesion that was missed with the initial small colposcopic biopsy. Our internal laboratory reviews confirm that biopsy sampling errors are most often the cause of poor correlation.
Similarly, the 13.3% discordance between colposcopic impression and directed biopsy results are predominantly biopsy sampling errors, particularly in those that were overcalled. Massad, Collins, and Cejtin43 concluded that colposcopy tends to overestimate the grade of disease but found agreement within 1 degree in 71.9% of cases. Correlation between impression and histology in other reports has noted agreement in 75% to 95% of cases.
The sensitivity of both cytology and colposcopy in this study were satisfactory and similar to other reports.36,42,43 Specificity, as has been shown, usually bears an inverse relationship to sensitivity and was moderate in this study.
The results of this analysis confirm the value of colposcopy as an integrated adjunct to cervical cancer screening programs. The program, as established in British Columbia, offers a uniform approach to all women, regardless of where they reside in the province. It also permits a careful analysis of virtually all of the colposcopic practices in the province in any given year. This, we believe, has resulted in a persistently high level of colposcopic practice. These results also indicate what might realistically be expected from colposcopy used in a traditional manner to assess large groups of women with varying degrees of cytologic abnormalities. The fact that 9 of 20 invasive cancers identified were in women in whom cytology suggested much less serious disease confirms the skills of the colposcopists in the program. Nonetheless, 7 of the 18 invasive cancers found on directed biopsy were not associated with the colposcopic features usually associated with invasive cancer. It would seem reasonable that not all superficially invasive lesions will have specific telltale features that enable their identification by even experienced colposcopists. This has, of course, been one of the reasons why liberal use of diagnostic excisional procedures has become so popular in colposcopic practice today.
In conclusion, the results of the current report indicate that the level of colposcopy as practiced in the community is acceptable and confirms the value of an integrated cytology colposcopic service to aid in the monitoring of the performance and practice of the colposcopy and cervical cancer screening program in British Columbia.
1.Hinselmann H. Zur Kenntnis der pracancerosen veranderungen der portio. Zentralbl Gynaecol 1927;51:901.
2.Hinselmann H. Einfuhrung in die kolposkopic. Hamburg: Hartung; 1933.
3.Stafl A, Mattingly R. Colposcopic diagnosis of cervical neoplasia. Obstet Gynecol 1973;41:168–76.
4.Donohue L, Meriwether W. Colposcopy as a diagnosis tool in the investigation of cervical neoplasias. Am J Obstet Gynecol 1972;113:107–10.
5.Townsend D, Ostergard D, Mishell D, Hirose F. Abnormal Papanicolaou smears: evaluation by colposcopy, biopsies, and endocervical curettage. Am J Obstet Gynecol 1970;108:429–34.
6.Benedet JL, Boyes DA, Nicols TM, Millner A. Colposcopic evaluation of patients with abnormal cervical cytology. Br J Obstet Gynaecol 1976;83:177–82.
7.Ortiz R, Newton M, Langlois P. Colposcopic biopsy in the diagnosis of carcinoma of the cervix. Obstet Gynecol 1969;34:303–6.
8.Ahlgren M, Ingemarsson I, Lindberg L, Nordqvist M, Nordqvist S. Conization as treatment of carcinoma in situ of the uterine cervix. Obstet Gynecol 1975;46:135–40.
9.Burghardt E, Holtzer E. Treatment of carcinoma in situ: evaluation of 1609 cases. Obstet Gynecol 1980;55:539–45.
10.Claman A, Lee N. Factors that relate to complications of cone. Am J Obstet Gynecol 1974;120:124–8.
11.Bjerre B, Eliasson G, Linell F, Soderberg H, Sjoberg N. Conization as only treatment of carcinoma in situ of the uterine cervix. Am J Obstet Gynecol 1976;125:143–52.
12.Benedet JL, Anderson GH. Cervical intraepithelial neoplasia in British Columbia: a comprehensive program for detection, diagnosis and treatment. Gynecol Oncol 1981;12:S280–91.
13.Fidler HK, Boyes DA, Worth AJ. Cervical Cancer detection in British Columbia: a progress report. J Obstet Gynaecol Br Commonw 1968;75:392–404.
14.Fidler HK, Boyes DA, Lock D. Intraepithelial carcinoma of the cervix: 214 cases with emphasis on investigation by cytology and cone biopsy. Can Med Assoc J 1957;77:79–85.
15.Boyes DA, Worth AJ, Anderson GH. Experience with cervical screening in British Columbia. Gynecol Oncol 1981;12:S143–55.
16.Anderson G, Boyes DA, Benedet JL, Le Riche J, Matisic J, Suen K. Organization and results of the cervical cytology screening programme in British Columbia, 1955–85. Br Med J 1988;296:75–8.
17.Bigrigg MA, Colding BW, Pearson P, Read MD, Swingler GR. Colposcopic diagnosis and treatment of cervical dysplasia at a single clinic visit. Lancet 1990;336:229–31.
18.Keijser K, Kenemans P, van der Zanden H, Schijf C, Vooijs P, Rolland R. Diathermy loops excision in the management of cervical intraepithelial neoplasia: diagnosis and treatment in one procedure. Am J Obstet Gynecol 1992;166:1281–7.
19.Holschneider C, Ghosh K, Montz FJ. See-and-treat in the management of high-grade squamous intraepithelial lesions of the cervix: a resource utilization analysis. Obstet Gynecol 1999;94:377–85.
20.Evans D, Halson EA, Brown CL, Boddington MM, Hughes HE, Mackenzie EF, et al. Terminology in gynecological cytopathology: report of the working party of the British Society for Clinical Cytology. J Clin Pathol 1986;39:933–44.
21.Benedet JL, Anderson GH. The accuracy of colposcopy in the diagnosis of preclinical invasive squamous carcinoma of the cervix. Colpo Gynecol Laser Surg 1987;3:129–34.
22.Townsend D, Ostergard D. Cryocauterization for pre-invasive cervical neoplasia. J Reprod Med 1971;6:171–6.
23.Lickrish G, Fortier M. Conservative management of intraepithelial cervical neoplasia. Can Med Assoc J 1977;116:641–3.
24.Charles E, Savage E. Cryosurgery treatment of cervical intraepithelial neoplasia. Obstet Gynecol Surv 1980;35:539–48.
25.Benedet JL, Miller DM, Nickerson KG, Anderson GH. The results of cryosurgical treatment of cervical intraepithelial neoplasia at one, five, and ten years. Am J Obstet Gynecol 1987;157:268–73.
26.Benedet JL, Nickerson KG, White GW. Laser therapy for cervical intraepithelial neoplasia. Obstet Gynecol 1981;58:188–91.
27.Baggish MS, Dorsey JH, Adelson M. A ten year experience treating cervical intraepithelial neoplasia with the CO2
laser. Am J Obstet Gynecol 1989;161:60–8.
28.Wright VC. Laser surgery for cervical intraepithelial neoplasia. Acta Obstet Gynecol Scand Suppl 1984;125:17–23.
29.Stewart JW. Treatment of cervical intraepithelial neoplasia with the CO2
laser: laser versus cryotherapy. A review of effectiveness and cost. Obstet Gynecol Surv 1984;39:469–73.
30.Krebs H, Pastore L, Helmkamp F. Loop electrosurgical excision procedures for cervical dysplasia: experience in a community hospital. Am J Obstet Gynecol 1993;169:289–95.
31.Solomon D, Schiffman R, Tarone R. ASCUS LSIL triage studies (ALTS) conclusions reaffirmed: response to a November 2001 commentary. Obstet Gynecol 2002;99:671–4.
32.Tarone R; ALTS Study group. Solomon D, Schiffman R, Comparison of three management strategies for patients with atypical squamous cells of undetermined significance: baseline results from a randomized trial. J Natl Cancer Inst 2001;93:293–9.
33.Wright TC, Cox JT, Massad LS, Twiggs LB, Wilkinson EJ. 2001 Consensus Guidelines for the management of women with cervical cytological abnormalities. JAMA 2002;287:2120–9.
34.Joste NE, Crum CP, Cibas ES. Cytologic/histologic correlation for quality control in cervicovaginal cytology: experience with 1,582 paired cases. Am J Clin Pathol 1995;103:32–4.
35.Tritz DM, Weeks JA, Spires SE, Sattich M, Banks H, Cibull M, et al. Etiologies for non-correlating cervical cytologies and biopsies. Am J Clin Pathol 1995;103:594–7.
36.Massad LS, Collins YC, Meyer PM. Biopsy correlates of abnormal cervical cytology classified using the Bethesda system. Gynecol Oncol 2001;82:516–22.
37.Jones BA, Novis DA. Cervical biopsy-cytology correlation. A College of American Pathologists Q-Probes study of 22,439 correlations in 348 laboratories. Arch Pathol Lab Med 1996;120:523–31.
38.Ibrahim SN, Krigman HR, Coogan AC, Wax TD, Dodd LG, Bentley RC, et al. Prospective correlation of cervicovaginal cytologic and histologic specimens. Am J Clin Pathol 1994;319–24.
39.Dodd LG, Sniege N, Villarreal Y, Fanning CV, Staerkel GA, Caraway NP, et al. Quality-assurance study of simultaneously sampled, non-correlating cervical cytology and biopsies. Diag Cytopathol 1993;9:138–44.
40.Contreras-Melendez L, Herbert A, Millward GH, Moore IE, Masson GM, Camillieri AP, et al. Assessment of the accuracy of cytology in women referred for colposcopy and biopsy: the results of a 1-year audit. Cytopathology 1992;3:267–74.
41.Kierkegaard O, Byrjalsen C, Fransden K, Frydenberg M. Diagnostic accuracy of cytology and colposcopy in cervical squamous intraepithelial lesions. Acta Obstet Gynecol Scand 1994;73:648–51.
42.Mitchell MF. The accuracy of colposcopy. Clin Consult Obstet Gynecol 1994;6:70–3.
43.Massad LS, Collins YC, Cejtin HE. The accuracy of colposcopic impression among women undergoing directed biopsy. J Low Gen Tr Dis 1999;3:57–8.