Year : 2009 | Volume
: 26 | Issue : 2 | Page : 69--73
An audit of cervicovaginal cytology in a teaching hospital: Are atypical glandular cells under-recognised on cytological screening?
Julian A Crasta, V Chaitra, CM Simi, Marjorie Correa
Department of Pathology, St. John's Medical College, Bangalore - 560 034, India
Julian A Crasta
Department of Pathology, St. John«SQ»s Medical College, Bangalore - 560 034, Karnataka
Background: Cervical cytology screening for carcinoma of the cervix in India is mainly opportunistic in nature and is practiced mainly in urban centres. The effectiveness of cervical cytology screening depends on various factors. The quality of cervicovaginal cytology service is assessed by various quality indices and by cyto-histology correlation, which is the most important quality assurance measure.
Aims: To describe the cervical cytology diagnoses, estimate the quality indices, and evaluate the discrepant cases on cytohistological correlation.
Settings and Design: Retrospective observational study from a tertiary care centre in South India.
Materials and Methods: Using a database search, all the cervicovaginal cytology reported during the period of 2002-2006 was retrieved and various diagnoses were described. The data was analysed to assess the quality indices. The cytohistologically discrepant cases were reviewed.
Results: A total of 10,787 cases were retrieved, of which 98.14% were labeled negative and 1.36% were unsatisfactory for evaluation. A few (0.81%) of the cases were labeled as squamous intraepithelial lesions and 0.38% as atypical squamous cells. The ASCUS: SIL ratio was 0.5. Cytohistological correlation revealed a total of ten cases with significant discrepancy. The majority of these were carcinomas that were misdiagnosed as atypical glandular cells. These cytology smears and the subsequent biopsies were reviewed to elucidate the reasons for the discrepancies.
Conclusions: The cervical cytology service at our centre is well within the accepted standards. An increased awareness of cytological features, especially of glandular lesions, a good clinician-laboratory communication and a regular cytohistological review would further improve the diagnostic standards.
|How to cite this article:|
Crasta JA, Chaitra V, Simi C M, Correa M. An audit of cervicovaginal cytology in a teaching hospital: Are atypical glandular cells under-recognised on cytological screening?.J Cytol 2009;26:69-73
|How to cite this URL:|
Crasta JA, Chaitra V, Simi C M, Correa M. An audit of cervicovaginal cytology in a teaching hospital: Are atypical glandular cells under-recognised on cytological screening?. J Cytol [serial online] 2009 [cited 2020 Jul 7 ];26:69-73
Available from: http://www.jcytol.org/text.asp?2009/26/2/69/55225
Carcinoma of the uterine cervix is the most common cancer among women in India and accounts for 33% of all female cancers in Bangalore, India according to hospital-based cancer registries.  The international medical community has recognized that cervical cancer is preventable and cervical cytology screening has been hailed as the most successful cancer screening method in medical history. Whereas cervical cytology is an organized cancer-screening programme in some countries, it is opportunistic and practised chiefly in urban centres in India. This has resulted in a gradual decrease in the incidence of cervical cancer in Indian urban centres. 
The effectiveness of cervical cytology screening depends on an intact chain of events, beginning with the patient coming in for screening and ending with the patient complying with appropriate follow-up or treatment in response to an optimally produced cytology report. The quality of cervical cytology screening can be assessed by quality indices such as relative percentages of the various diagnoses, including the unsatisfactory rate and the atypical squamous cells of undetermined significance (ASCUS) - squamous intraepithelial lesion (SIL) ratio.  The correlation of cytological diagnosis with available histological follow-up allows the evaluation of false positive and false negative cases.
Materials and Methods
The cytology division in the Department of Pathology of a 1200-bed tertiary care teaching hospital draws patients from Karnataka and the bordering states, particularly Tamil Nadu and Andhra Pradesh. Cervical screening is mainly opportunistic in nature: it is offered to patients who present to gynecology clinics with various gynecological or fertility-related problems, as well as to patients who attend annual health screenings. Therefore, we undertook an audit of our cervical cytology diagnoses for duration of five years in order to assess the quality of our cytology service and we also reviewed the cytohistological discrepancies.
Cervicovaginal specimens are prepared conventionally in our centre and consist of at least two smears from each case, containing both ectocervical scrapings and endocervical swab smears. Terminology used is as per recommendations by The Bethesda System 2001.  At least two pathologists read the smears. Using a database search, a total of 10,787cases, each having two smears was retrieved from 2002 to 2006. The numbers and percentages of unsatisfactory smears, atypical squamous cells of undetermined significance, atypical glandular cells (AGC), low- and high-grade squamous intraepithelial lesions (LSIL and HSIL), squamous cell carcinoma, and adenocarcinoma were compiled.
Follow-up histological findings extracted from a database (comprising of colposcopy-directed punch biopsies, cervical biopsies, cone biopsies, and hysterectomies) were recorded only for a year for the same cases, and cytohistological correlation was performed for all the cases with abnormal findings. The positive predictive value of LSIL and HSIL and the overall predictive value of a positive cytology diagnosis were calculated.
Cytohistological discrepancies of more than one grade (ASCUS, AGC, LSIL, HSIL) were retrieved for review. Both cytology and histology slides were reviewed by two pathologists (J.C. and M.C.) who were blinded for the original diagnoses, and an attempt was made at consensus diagnoses in case of divergent opinions.
A total of 10,787 cervical cytology cases were diagnosed from 2002 to 2006. Of these, 98.14% had negative diagnoses and both the smears were found to be inadequate for 147 cases (1.36%). Forty cases (0.38%) were diagnosed as ASCUS and 37 cases (0.35%) as AGC. There were 20 cases (0.19%) of LSIL and 66 cases of HSIL (0.62%). Nineteen cases of squamous cell carcinoma (0.18%) and 16 cases (0.15%) of adenocarcinoma were encountered. In most of the cases, both the smears showed abnormality. The ASCUS to SIL ratio was 0.5
Of the 198 cases with positive cytology, 136 had cytological or histological follow-up at our centre: 55% of ASCUS (23/40), 50% of AGC (19/37), 45% of LSIL (9/20), and 77.3% of HSIL (51/61) had subsequent histological follow-up. These cases were analysed as were the cases that were negative for dysplasia (false negatives) along with unsatisfactory cases that had a subsequent positive histology.
There were 147 unsatisfactory cases, of which 85 (57.8%) were followed up. Of these 85 cases, 80 (94.1%) were negative, one case had CIN I, two cases had CIN III, one case was squamous cell carcinoma, and one case was adenocarcinoma; these were detected histologically. The smears of these five cases were reviewed and no abnormality was detected. Sixty-two cases had no cytological or histological follow-up at our centre.
Cytohistological correlation of atypical cases
Of the 40 cases diagnosed as ASCUS, 23 (55%) were subsequently biopsied at our centre. Of these biopsied cases, 17 (77.3%) were negative, one case was CIN I, and one case was CIN III. There were two cases of squamous cell carcinoma and adenocarcinoma each. These cases on review were interpreted as malignant accounting for screening error at cytology. One case diagnosed as CIN III initially by both cytology and histology was interpreted as atrophy on review, using the same two modes. No case of atypical squamous cells-high grade squamous intraepithelial lesion cannot be excluded (ASC-H) was encountered.
Nineteen out of the 37 cases diagnosed as atypical glandular cells were subsequently followed up. Fifteen cases had normal histology and four cases gave abnormal results: one was a squamous cell carcinoma and three were adenocarcinoma, accounting for a significant cytological screening error.
Cytohistological correlation of squamous intraepithelial lesion
Of the 20 cases with a diagnosis of LSIL, only nine cases had follow-up histology. Three cases were CIN III, one case was CIN II, but five cases were negative according to histology. These biopsies were reviewed and no evidence of CIN- or human papilloma virus (HPV)-associated changes was detected.
Fifty-one (77.3%) of the 61 HSIL cases were subsequently biopsied at our centre. Of these, one was negative, one was CIN I, two were CIN II, 24 (45.1%) were CIN III, 20 (39.2%) were squamous cell carcinoma, and three were adenocarcinoma. The positive predictive value of smear cytology showing HSIL for histological lesions of CIN II or worse is 95.9%. The overall predictive value of a positive histological diagnosis is 93%.
Cytohistological correlation of carcinoma
Nineteen cases diagnosed as squamous cell carcinoma and 16 cases diagnosed as adenocarcinomas had subsequent biopsies / excisions that confirmed the diagnosis in all cases.
Negative for dysplasia or malignancy
Three cases that were reported as negative, later gave positive cervical biopsy results in the same year. The biopsy diagnosis was CIN I / koilocytosis in two cases and CIN II in one case. All the smears were reviewed and were confirmed to be negative for dysplasia or malignancy. Hence, the false negative diagnoses in these cases were possibly due to a sampling error in the initial cytology screen.
Review of significantly discrepant cases
A total of ten cases had a significant discrepancy between the cytological diagnoses and the subsequent histological diagnoses. These cases were subjected to a blind review by two pathologists (J.C. and M.C.) who were blinded to the diagnoses and a consensus diagnosis was arrived at, whenever there was a disagreement. The results of the review are outlined in [Table 1]. The reason for the discrepancy was categorized as sampling error in cytology or histology, and interpretive error [Table 1]. Three cases of AGC were upgraded as invasive carcinomas whereas two cases of ASCUS were upgraded as invasive carcinomas; one case was judged to be HSIL on review. The original cytological diagnosis was inaccurate in these cases. One case of ASCUS was downgraded to be negative, reflecting a sampling error in cytology. One case of atrophy was wrongly interpreted as ASCUS in cytology and CIN III in histology, which is a common diagnostic dilemma. Two cases of HSIL had a negative and lower grade CIN, reflecting the sampling error even in histology.
The relative percentages of various diagnoses are compared in [Table 2] with other studies and with benchmark data collected by the College of American Pathologists (CAP) Cytopathology Resource Committee.  Our data is comparable with the values of the National University of Singapore  but lower than CAP's. Our results are also considerably lower than another Indian study,  which was a multicentre study, and the population screened was mainly from the suburbs and rural areas all over India. Carcinoma of the cervix is a disease of the lower socioeconomic strata of society; hence, our study reflects the low risk for the urban population who are being screened at our centre and the opportunistic nature of cervical screening. However, it may also be due to the fact that some of these cases are part of routine annual health checkups due to which there is a higher rate of negative smears in the cohort.
The rate of unsatisfactory smears is high (CAP median: 0.5%)  but follow-up appears to be good, indicating good laboratory-clinician feedback. The specimen unsatisfactory rate is an important quality assurance indicator in cervical cytology as it identifies a group of women who are being inadequately screened. On follow-up, 5.9% of patients with unsatisfactory smears were found to have SIL or malignancy in our study, underscoring the importance of repeat cytology.
We have a very low rate of ASCUS and AGC with a low ASCUS / SIL ratio, indicating that this diagnosis is made very selectively in our centre. The clinical follow-up of ASCUS is variable; the options being repeat cytology, immediate colposcopy, or HPV DNA testing.  Most of our patients have colposcopically directed biopsy (55%) as the next procedure but only 22.7% had a positive yield on subsequent histology. The rates of SIL following ASCUS cytology are variable ( 70%), being attributed to individual variation in the diagnostic thresholds of SILs and ASCUS, as well as the type of follow-up.  But it is estimated that  and most of the lesions are known to regress spontaneously. As the percentage of SIL on biopsy follow-up is low in our study, at least one repeat smear before a biopsy, as recommended by CAP, would be appropriate follow-up in our financially restrained setup. However, clinicians' apprehension of the patients being lost for follow-up and the finding of 18.2% invasive malignancies in our series indicate that a diagnosis of ASCUS may often represent a more significant underlying lesion and may justify a colposcopy-guided biopsy at least in some cases. In these cases, HPV DNA testing can potentially triage ASCUS patients to either colposcopy with biopsy or continued cervical smear screening.  Although nearly equal numbers of cases were diagnosed as AGC, 21% were diagnosed as malignancies on biopsy follow-up, of which 15.8% were adenocarcinoma. AGC is a difficult diagnosis to make, especially in view of its rarity. It is a heterogeneous entity caused by a wide variety of squamous, endocervical, and endometrial lesions. This is reflected in our discrepant cases in which AGC were diagnosed as carcinomas, especially that of the endometrium, on review. An awareness of their cytological features including sub-classification and good clinician-laboratory communication may improve interpretive errors. As the natural history of AGC is not well defined and the sensitivity of repeated Pap tests for detection of high-grade uterine lesions is low, patients with AGCs should undergo colposcopy and endocervical sampling as an initial evaluation. 
The biopsy diagnoses of LSIL, HSIL, and invasive carcinomas showed good correlation with the cytological diagnoses. Histological sampling errors and interpretive errors probably contributed to the false positive diagnoses, as revealed by the review of discrepant cases. These factors are partly related to sampling technique as all biopsies were not done under colposcopic guidance and were not directly within our laboratory's control.  The false negative cases were confirmed on review to be false negative screening tests rather than false negative cytology. The false negatives could be either due to sampling error, as seen in previous studies,  or the natural history of the disease.
The overall figures of the cervical cytology service in our centre are well within the accepted standards. The figures reveal that the population screened is at low risk of developing cancer of the cervix. The relative low occurrence of SIL in follow-up biopsies of ASCUS reveal that these patients can be subjected to biopsy only if the lesion persists on repeat cytology after six months, or can have a HPV triage before a colposcopy-guided biopsy. However, in cases of AGC, a colposcopic examination and endocervical curettage is acceptable for initial evaluation as these are known to be associated with higher-grade lesions. Areas of continued improvement would be increased awareness of cytological features of various diagnoses, particularly atypical glandular cells, a good clinician-laboratory communication, and regular reviews of cytohistological discrepancies to analyse the causes of the discrepancies.
|1||Vallikad E. Cervical cancer: The Indian Perspective. Int J Obstet Gynecol 2006;95:S215-33.|
|2||Jones BA, Davey DD. Quality management in gynecologic cytology using interlaboratory comparison. Arch Pathol Lab Med 2000;124:672-81.|
|3||Solomon D, Davey D, Kurman R, Moriarty A, O'Connor D, Prey M, et al . The 2001 Bethesda system: Terminology for reporting results of cervical cytology. JAMA 2002;287:2114-9.|
|4||Davey DD, Nielsen ML, Naryshkin S, Robb JA, Cohen T, Kline TS. Atypical squamous cells of undetermined significance: Current laboratory practices of participants in the College of American Pathologists Interlaboratory Comparison Program in Cervicovaginal Cytology. Arch Pathol Lab Med 1996;120:440-4.|
|5||Thamboo TP, Salto-Tellez M, Tan KB, Nilsson B, Rajwanshi A. Cervical cytology: an audit in a Singapore teaching hospital. Singapore Med J 2003;44:256-60.|
|6||Sankaranarayanan R, Thara S, Sharma A, Roy C, Shastri S, Mahe C, et al . Accuracy of conventional cytology: Results from a multicentre screening study in India. J Med Screen 2004;11:77-84.|
|7||Wright TC, Cox JT, Massad SL, Twiggs LB, Wilkinson EJ. 2001 Consensus guidelines for the management of women with cervical cytological abnormalities. JAMA 2002;287:2120-9.|
|8||Raab SS, Bishop NS, Zaleski MS. Long term outcome and relative risk in women with typical, squamous cells of undetermined significance. Am J Clin Pathol 1999;112:57-62.|
|9||Schiffman M, Solomon D. Findings to date from the ASCUS-LSIL triage study (ALTS). Arch Path Lab Med 2003;127:946-9.|
|10||Chheing DC, Gallaspy S, Yang H, Roberson J, Eltoum I. Women with atypical glandular cells: A long term follow up study in a high- risk population. Am J Clin Pathol 2004; 122:575-9.|
|11||Tan KB, Chang SE, Soh VC, Thamboo TP, Nilsson B, Chan HL. Quality indices in a cervicovaginal cytology service: Before and after accreditation. Arch Pathol Lab Med 2004;128:303-7.|
|12||Sodhani P, Singh V, Das DK, Bhambhani S. Cytohistological correlation as a measure of quality assurance of a cytology laboratory. Cytopathology 1997;8:103-7.|