The comparative diagnostic accuracy of conventional and liquid-based cytology in a colposcopic setting

BJOG: an International Journal of Obstetrics and Gynaecology November 2005, Vol. 112, pp. 1542 1546 DOI: 10.1111/j.1471-0528.2005.00699.x The comparative diagnostic accuracy of conventional and liquid-based cytology in a colposcopic setting T. Hussein, a M. Desai, a A. Tomlinson, b H.C. Kitchener b Objective This study was conducted to compare the performance of liquid-based cytology (LBC) and conventional cytology (CS) in the high prevalence setting of colposcopy clinic. Design A split sample of matched ThinPrep (TP) and conventional smear from 563 patients were evaluated blindly. The performance of both techniques was compared with the gold standard of biopsy results or normal colposcopy examination in 441 cases. Setting Colposcopy clinic of an inner city hospital for women and children. Sample Five hundred and sixty-three women referred to colposcopy clinic over 14-month period. Methods Cervical smears were taken from 563 women referred for colposcopy. Using the split-sample technique, the material was spread on a conventional (CS) slide and the remaining material rinsed in a PreservCyt solution. A T2000 processor was used to prepare LBC preparations. All women underwent colposcopy/biopsy according to local protocol. Four hundred and forty-one women met the diagnostic standard criteria of the study, which was either a normal colposcopy or histopathology result. Sensitivity, specificity and positive and negative predictive values were calculated for both methods of cytology preparations. Main outcome measures Matched TP and conventional smears, detection of abnormality, matched biopsies, sensitivity, specificity, and positive and negative predictive values. Results Inadequate rates for CS and LBC (TP) were 4.3% and 0.68%, respectively. In 73% of cases, the CS and the LBC preparations showed exact agreement, whereas 77% agreement was seen when comparison was made for amalgamated low grade and high grade abnormalities. Low grade cytological abnormalities accounted for 44% of LBC slides versus 37% in CS slides. High grade cytological abnormalities accounted for 22% of LBC versus 17% seen in CS cases (P < 0.001). LBC showed increased sensitivity in the detection of CIN2 or worse than CS (92% and 83%, respectively) and CS showed greater specificity than LBC (82% and 76%, respectively). Conclusions In high prevalence setting, LBC performed at least as well as CS. The inadequate rate was significantly lower with LBC. The numbers are too small, however, to make confident comments about increased sensitivity and negative predictive value with LBC. Larger studies are required to verify these findings. INTRODUCTION Liquid-based cytology (LBC) has been recommended by the National Institute of Clinical Excellence for implementation in the National Cervical Screening Programme 1 and has already been implemented in the Scottish Screening Programme. Its principal advantages are a major reduction in reported inadequate smears and increased laboratory throughput which improves cost efficacy. Increased sensitivity has been claimed for LBC in the detection of underlying cervical intraepithelial neoplasia, 2 5 but in the absence of randomised control trials, there remains a degree of uncertainty about this. In terms of cervical screening, information about LBC performance in primary screening is available from the NHS Pilot Studies where it was compared with conventional cytology (CS). 6 In a previously published systematic review, the sensitivity and specificity of the conventional Papanicolaou smear ranged from 30% to 87% and from 86% to 100%, respectively. We wished to compare the diagnostic accuracy of one form of LBC (ThinPrep [TP]) with CS in the colposcopy setting where women were referred with abnormal cytology. a Manchester Cytology Centre, Manchester Royal Infirmary, UK b Academic Unit, Obstetric and Gynaecology, St Mary s Hospital, Manchester, UK Correspondence: Dr T. Hussein, P.O. Box 541, Tartous, Syria. D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology METHODS Cervical samples were taken using a broom-like collection device from women referred to the colposcopy clinic. A conventional smear was prepared on a glass slide first. Following this, the residual material on the collection device www.blackwellpublishing.com/bjog

CONVENTIONAL AND LIQUID-BASED CYTOLOGY 1543 Table 1. TP versus conventional smear results. TP Conventional Total Negative BL** Mild Moderate Severe Sev./Inv.*?Glandular # Inadequate Negative 126 6 1 6 139 BL** 43 64 1 6 114 Mild 6 14 55 1 1 5 82 Moderate 3 9 12 47 1 72 Severe 4 23 27 Sev./Inv. * 2 2?Glandular # 2 2 Inadequate 2 1 3 Total 180 93 69 52 24 2 2 19 441 ** Borderline nuclear changes. was rinsed into a vial of PreservCyt solution (Cytyc) for the LBC slide preparation. All women underwent a colposcopic examination. If colposcopic appearances were normal, a biopsy was generally not performed. With colposcopic abnormalities, a punch, loop or cone biopsy was taken. In some women with low grade cytology, and low grade colposcopic changes, biopsies were not performed on the first visit. Biopsies were taken in all cases of suspected high grade and a proportion of low grade abnormalities. All specimens were reported at the Manchester Royal Infirmary, Pathology Department. Cytology slides were prepared at the Manchester Cytology Centre on a Cytyc ThinPrep 2000 processor for LBC according to the laboratory s standard staining policy of progressive Papanicolaou stain. The TP slides were reported blind to the histopathology and colposcopy data by T.H. who was supervised by M.D. for any query on the slides. Conventional smears were reported by screeners, checkers, MLSOs or medical staff and were also reviewed blindly by T.H. before comparing with the LBC slides. Discrepancies between both methods of cytology and/or the CS review were discussed with a consultant cytopathologist over a multi-head microscope, which allowed double-checking of the LBC slide and the conventional smear. Using McNemar s test, statistical analysis was assessed and means of sensitivity, specificity and positive or negative predictive values were calculated using either a negative colposcopy or a histopathology result as the gold standard. RESULTS One hundred and twenty-two cases were excluded due to lack of colposcopy report of normal examination or there was no satisfactory biopsy. This group comprised solely negative and low grade cytology. Four hundred and forty-one patients met the criteria for the diagnostic standard for the study. The mean age of the women was 32.5 years (SD F10). In 99 cases there was a negative colposcopy or negative biopsy and in 342 cases there was some form of histological proven abnormality ranging from HPV associated change up to invasive carcinoma. Nineteen women (4.3%) had inadequate smears with the conventional technique compared with only 3 (0.68%) for the LBC (P < 0.001). The comparison of CS and LBC is shown in Table 1. There was an exact agreement by Table 2. TP versus conventional smear results by low or high grade. TP Conventional Total Negative Low grade High grade Sev./Inv. *?Glandular # Inadequate Negative 126 7 6 139 Low grade 49 134 2 11 196 High grade 3 21 74 1 99 Sev./Inv. * 2 2?Glandular # 2 2 Inadequate 2 1 3 Total 180 162 76 2 2 19 441

1544 T. HUSSEIN ET AL. Table 3. Colposcopy versus cytology results. Colposcopy/ biopsy results TP vs Conventional Results Negative BL** Mild Moderate Severe Sev./Inv.*?Glandular # Inadequate Total TP CS TP CS TP CS TP CS TP CS TP CS TP CS TP CS Negative 99 119 54 38 28 19 3 1 1 1 2 9 187 HPV changes 32 45 42 30 15 12 3 1 2 2 1 5 95 CIN1 a 6 10 11 12 25 20 2 2 44 CIN2 1 5 7 13 11 15 60 44 14 13 3 93 CIN3 1 1 2 2 4 6 9 7 1 1 17 CGIN b 1 1 1 1 1 1 3 SQ.CA c 1 1 1 AD.CA d 1 1 1 Total 139 180 114 93 82 69 72 52 27 24 2 2 2 2 3 19 441 a Cervical intraepithelial neoplasia. b Cervical glandular intraepithelial neoplasia. c Squamous cell carcinoma. d Adenocarcinoma. ** Borderline nuclear changes. classification in 320 out of 441 cases (73%). Table 2 shows cytology classified as low grade (borderline and mild dyskaryosis) and high grade (moderate and severe dyskaryosis) with the correlation remaining almost identical. Out of all LBC smears, 44% showed low grade abnormalities and 22% high grade lesions compared with 37% low grade and 17% high grade lesions in the conventional smears (P < 0.001). The comparative cytology results with LBC and CS are shown against the colposcopy/biopsy results in Table 3. Negative cytology in the presence of CIN was more common with CS than LBC. LBC smears detected 86% of all CIN1 lesions found by biopsies compared with 73% by CS method, and 98% of all CIN2 and CIN3 were detected by LBC slides in comparison with 92% by CS. There were three cases of cervical glandular intraepithelial neoplasia, one of which was accurately predicted by either techniques. Both methods agreed in reporting the other cases as mild and severe dyskaryosis. LBC and CS Table 4. Comparison between TP and CS preparations for detection of CIN2 or worse by cytological categories of mild dyskaryosis or worse and moderate dyskaryosis or worse. Mild or TP Moderate or Mild or CS Moderate or Sensitivity 92 80 83 67 Specificity 76 97 82 98 PPV* 57 89 62 94 NPV** 96 93 93 89 * Positive predictive value. ** Negative predictive value. reports also agreed over a case of squamous cell carcinoma and another adenocarcinoma, however, they were reported as glandular atypia for the former and severe possible invasive, for the latter. Relative sensitivity, specificity, positive and negative predictive values are shown in Table 4. DISCUSSION The most valid comparison between LBC and CS is in the primary screening. Ideally, such studies should involve direct-to-vial collection as the split sample could result in non-comparable cellular representation. For example, split-sample studies found higher proportion of slides described as Satisfactory But Limited By: No Endocervical Component, in LBC smears, when compared with the CS method. 7,8 Direct-to-vial studies, however, had SBLB: No ECC rates similar to the current rate of conventional smears. 9 Nevertheless, it is important for laboratories moving to implement LBC to be confident that LBC performs at least as well as CS in terms of picking up underlying abnormalities. The inadequate rates in this study are well within normal ranges, suggesting that the split-sample cytology preparations were a valid basis for this study. Similar to previous studies, 10,11 LBC showed superiority in reducing smears described as suboptimal or inadequate and in increasing the detection of cervical abnormalities of all grades. 12 Using the Bethesda System terminology, 13 LBC increased ASCUS/AGUS rate, 10,14 which was not the case in all studies. 15 Applying the BSCC terminology, 16 however, borderline nuclear changes rate was almost equal or higher in CS smears than their TP counterparts. 11,17 Our

study showed an increase in borderline detection with LBC smears (26% of TP smears vs 21% in CS cases), with this occurring in cases reported negative or HPV changes in colposcopy/histology reports. We think that enhancing nuclear preservation with distinctive details in benign cells and the presence of dispersed small shrinking cells in the LBC smears might lead to this higher rate. One split-sample study on predominantly routine Papanicolaou tests showed that both techniques had similar detection rates for high grade squamous intraepithelial lesions, 14 although the LBC test had more low grade abnormalities. In this study, the clean background in LBC slides helped to pick up cells showing high grade abnormality more efficiently, as they stand out and become easily recognisable among other squamous cells. The LBC results experienced even better performance in detection of CIN1 and CIN2/3 than the direct-to-vial method. 18 Of course, it is difficult to exclude a bias towards a highly prevalence patient population in the colposcopy clinic. As has been described in other reports, we found a significant decrease of inadequate cytology in LBC. Concern has been expressed as to whether cytological material can be assumed to be adequate using liquid-based technology 19 when it is shown to be inadequate in CS. In this study, the adequacy by LBC of 19 split samples thought inadequate in CS was confirmed by the identification of 11 low grade and 1 high grade abnormalities confirmed by histology. The LBC technique detected low grade and high grade dyskaryosis in smears found to be inadequate in the conventional smears (Table 1). Although our reference standard of histological diagnosis or negative colposcopy is not error-free, 20,21 all abnormal cytology cases with normal biopsies or when they showed only HPV changes were considered as false positive. Both TP and CS slides agreed over five HSIL cases, which reported as either negative and having HPV changes in histology. Although the follow up smear showed negative cytology, the discrepancy between histology and cytology could be related to the procedure done on the cervix before the biopsy taking 22 and even the smear-taking procedure can remove a small amount of dysplastic epithelium. 23 Both cytology techniques missed two cases showing severe glandular atypia in biopsies, only the accompanying squamous lesions (mild and severe dyskaryosis) were identified. These lesions might be inaccessible by the sampling device 24 or overlooked by the presence of squamous abnormalities. 25 Two carcinoma cases (adenocarcinoma and squamous cell carcinoma) were not diagnosed accurately by both TP and CS methods. Similar to CS, the presence of microbiopsies in the smear can also be a pitfall in TP preparations. 26 As we found, some reports showed sensitivity improvement when the TP technique is used and better specificity in conventional smears. 27,28 On the other hand, some studies showed better sensitivity in the CS method 15,29 and only one study showed higher specificity in the TP technique. 15 It would appear from this study that in a high prevalence setting, LBC performs at least as well as CS. Numbers are too small to make confident comments about increased sensitivity and the predictive value that would be expected in the implementation of LBC. Larger studies should be performed to verify these findings. References CONVENTIONAL AND LIQUID-BASED CYTOLOGY 1545 1. National Institute for Clinical Excellence. Final Appraisal Determination Guidance on the use of liquid-based cytology for cervical screening. Review of existing guidance, number 5. Issue date: August 2003. 2. Hutchinson ML, Agarwal P, Denault T, Berger B, Cibas ES. A new look at cervical cytology: ThinPrep multi-centre trial results. Acta Cytol 1992;36:499 504. 3. Wilbur DC, Cibas ES, Merritt S, James LP, Berger BM, Bonfiglio TA. ThinPrep Processor, clinical trials demonstrate an increased detection rate of abnormal cervical cytologic specimens. Am J Clin Pathol 1994; 101:209 214. 4. Lee KR, Ashfaq R, Birdsong GG, Corkill ME, McIntosh KM, Inhorn SL. Comparison of conventional Papanicolaou smears and a fluidbased, thin layer system for cervical cancer screening. Obstet Gynaecol 1997;90:278 284. 5. Papillo JL, Zarka MA, St John TL. Evaluation of the ThinPrep Pap test in clinical practice: a seven- month, 16314-case experience in northern Vermont. Acta Cytol 1998;42:203 208. 6. http://www.cancerscreening.nhs.uk/cervical/lbc-pilot-evaluation.pdf. 7. Evans SK, Wilbur DC. Identification of endocervical cells and microorganisms on cervical thin-layer cytology specimens: comparison to paired conventional smears. Acta Cytol 1993;37:776. 8. Awen C, Hathway S, Eddy W, Voskuil R, Janes C. Efficacy of ThinPrep preparation of cervical smears: a 1,000-case, investigatorsponsored study. Diagn Cytopathol 1994;11:33 37. 9. Corkill M, Knapp D, Martin J, Hutchinson ML. Specimen adequacy of ThinPrep sample preparations in a direct-to-vial study. Acta Cytol 1997;41:39 44. 10. Hutchinson ML, Zahniser DJ, Sherman ME, et al. Utility of liquidbased cytology for cervical carcinoma screening. Cancer (Cancer Cytopathol) 1999;87:48 55. 11. Grace A, Mcbrearty P, Troost S, Thornhill M, Kay E, Leader M. Comparative study: conventional cervical and ThinPrep Pap test in a routine clinical setting. Cytopathology 2002;13:200 205. 12. Limaye A, Connor AJ, Huang X, Luff R. Comparative analysis of conventional Papanicolaou tests and a fluid-based thin-layer method. Arch Pathol Lab Med 2003;127:200 204. 13. Smith JHF. Bethesda 2001. Cytopathology 2002;13:4 10. 14. Biscotti CV, O Brien DL, Gero MA, Gramlich TL, Kennedy AW, Easley KA. Thin-layer Pap test vs conventional Pap smear. Analysis of 400 split samples. J Reprod Med 2002;47(1):9 13 (January). 15. Park IA, Lee SN, Chae SW, Park KH, Kim JW, Lee HP. Comparing the accuracy of ThinPrep Pap tests and conventional Papanicolaou smears on the basis of the histologic diagnosis: a clinical study of women with cervical abnormalities. Acta Cytol 2001;45:525 531. 16. Achievable Standards, Benchmarks for Reporting and Criteria for Evaluating Cervical Cytology. NHSCSP Publication No 1: May 2000. 17. Ring M, Bolger N, O Donnell M, et al. Evaluation of liquid-based cytology in cervical screening of high-risk populations: a split study of colposcopy and genitor-urinary medicine populations. Cytopathology 2002;13:152 159. 18. Pan Q, Belinson JL, Li L, et al. A thin-layer, liquid-based Pap test for mass screening in an area of China with a high incidence of cervical carcinoma. Acta Cytol 2003;47:45 50. 19. Herbert A, Johnson J. Personal view. Is it reality or an illusion that

1546 T. HUSSEIN ET AL. liquid based cytology is better than conventional cervical smears? Cytopathology 2001;12:383 389. 20. DiBonito L, Falconieri G, Tomasic G, Colautti I, Bonifacio D, Dudine S. Cervical cytopathology: an evaluation of its accuracy based on cytohistologic comparison. Cancer 1993;72:3002 3006. 21. Willocx F, de Somer ML, Van Roy J. Classification of cervical smears with discordance between the cytologic and/or histologic rating. Acta Cytol 1987;31:883 886. 22. Frost JK. Diagnostic accuracy of cervical smear. Obstet Gynecol Surv 1969;24:893 908. 23. Berner A, Hoeg K, Oppedal BR. Smear biopsies. A cause of negative follow-up biopsies in patients with premalignant conditions of the uterine cervix. Diagn Gynecol Obstet 1980;2:99 102. 24. Rylander E. Negative smears in women developing invasive cervical cancer. Acta Obstet Gynecol Scand 1977;56:115 118. 25. Jarmulowicz MR, Jenkins D, Bartons SE, Goodall AL, Hollingworth A, Singer A. Cytological status and lesion size: a further dimension in cervical intraepithelial neoplasia. Br J Obstet Gynaecol 1989;96: 1061 1066. 26. Herbert A, Johnson J, Patnick J. Report of a working party convened by the Royal College of Pathologists, British Society for Clinical Cytology and NHS Cervical Screening Programme. Achievable standards, benchmarks for reporting and criteria for evaluating cervical cytopathology. Cytopathology 1995;6(5):301 303 (October). 27. Sheets EE, Constantine NM, Dinisco S, Dean B, Cibas ES. Colposcopically directed biopsies provide a basis for comparing the accuracy of ThinPrep and Papanicolaou smears. J Gynecol Tech 1995; 1:27 34. 28. Ferenczy A, Robitaille J, Franco E, Arseneau J, Richart RM, Wright TC. Conventional cervical cytologic smears vs ThinPrep smears. A paired comparison study on cervical cytology. Acta Cytol 1996;40: 1136 1142. 29. Coste J, Cochand-Priollet B, de Cremoux P. Cross sectional study of conventional cervical smear, monolayer cytology, and human papillomavirus DNA testing for cervical cancer screening. BMJ 2003; 326:733 736. Accepted 18 April 2004