Effect of margin status on cervical intraepithelial neoplasia recurrence following LLETZ in women over 50 years

DOI: 10.1111/j.1471-0528.2008.01853.x www.blackwellpublishing.com/bjog Gynaecological oncology Effect of margin status on cervical intraepithelial neoplasia recurrence following LLETZ in women over 50 years R Manchanda, P Baldwin, R Crawford, SL Vowler, R Moseley, J Latimer, K Welton, M Shafi Department of Gynaecology, Division of Gynaecological Oncology, Addenbrooke s Hospital, Cambridge, UK Correspondence: Dr R Manchanda, Department of Gynaecological Oncology, EGA Institute for Women s Health, University College London, First Floor, Maple House, 149 Tottenham Court Road, London W1T 7NF, UK. Email r.manchanda@ucl.ac.uk Accepted 11 June 2008. Objective To establish the effect of margin status on recurrence following large loop excision of the transformation zone (LLETZ) in women over 50 years. Study design Prospectively collected data of women over 50 years, who underwent LLETZ for suspected cervical intraepithelial neoplasia between 1998 and 2003, were analysed. Women were followed up for up to over 6 years. Setting District colposcopy service based at a gynae-oncology cancer centre. Main outcome measures The main outcome measure included histologically detected recurrence. Any abnormal cytology on follow up was also documented. Methods Prospectively collected data were analysed from the colposcopy database. Recurrence was analysed using Kaplan Meir plots and Cox regression. Fisher s exact test was used to determine the association between margins and grade. The Kruskal Wallis and Mann Whitney U tests were used to compare age and duration of follow up between groups. Results A total of 118 women underwent LLETZ and 92 were included in the final analysis. Margins were designated as clear (n = 62), involved (n = 22) or uncertain (n = 8). Histological recurrence occurred in 12 while abnormal cytology was demonstrated in 17 women. One woman with involved margins developed cervical cancer. Individuals with clear margins were less likely to have recurrence than those with involved margins (Hazard Ratio (HR) 0.18, 95% CI: 0.06 0.59). Involved margins were more common with high-grade than low-grade lesions (P = 0.002). Conclusion The data show an association between disease recurrence and the finding of involved margins in this cohort. Keywords CIN, LLETZ, margin status, over 50 years, recurrence. Please cite this paper as: Manchanda R, Baldwin P, Crawford R, Vowler SL, Moseley R, Latimer J, Welton K, Shafi M. Effect of margin status on cervical intraepithelial neoplasia recurrence following LLETZ in women over 50 years. BJOG 2008;115:1238 1242. Introduction Large loop excision of the transformation zone (LLETZ), also known as loop electrosurgical excision procedure in North America, is currently the most widely used method for treating cervical intraepithelial neoplasia (CIN) in the UK. This has been shown to reduce the risk of invasive cervical cancer by up to 95%. 1 Only a few studies have documented the pattern of long-term recurrence following LLETZ. 2 4 Limited data suggest that age more than 50 years 3,5 and the presence of disease at resected margins 5 7 are independent risk factors for recurrence. Revised British National guidelines (2004) recommend that women over the age of 50 years, with high-grade disease at the resected margin should have a repeat LLETZ to try to obtain clear margins. 8 Women with both these risk factors may be at particular risk of developing recurrent CIN and could benefit from active intervention in the form of repeat treatment rather than merely undergoing early follow up. Evidence is, however, lacking in this regard. This study aims to establish the effect of margin status on the risk of recurrence following LLETZ in women over the age of 50 years. Methods An analysis of prospectively collected data was performed using the colposcopy database at Addenbrooke s Gynaecological Oncology Centre, Cambridge, UK. The centre provides a district colposcopy service. The data of all women over the age of 50, who underwent LLETZ for suspected CIN between 1238 ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

Recurrence after LLETZ for CIN in women over 50 1998 and 2003, were examined. All women were treated by accredited colposcopists at the cancer centre. LLETZ was performed routinely using a standard technique that has already been described elsewhere. 9 A see and treat policy was typically used for high-grade cytology and defer and treat for low-grade cytology. Following histological assessment, specimens were classified into high-grade lesions (HGL; CIN2, CIN3) and low-grade lesions (LGL; less than CIN2). Women were followed up, up to a period of over 6 years (median 4.3, range: 0.5 6.8 years) in a manner consistent with existing national guidelines. Colposcopy and smear were performed at first follow up at 6 months in the treatment centre. Thereafter, follow-up smears were performed in primary care at 12 months post-treatment and subsequently at yearly intervals. Those women with abnormal smears were referred back for colposcopic assessment (at the treatment centre) with biopsy as necessary. Statistical analysis was carried out in SPSS V13.0 (SPSS Inc., Chicago, IL, USA), using Kaplan Meier plots and Cox regression. The Fisher s exact test was used to look at the association between margins and grade. Age and duration of follow up were not sufficiently normally distributed, and hence were compared between the different margin groups using nonparametric Kruskal Wallis test. As age showed a significant difference, post hoc pairwise tests were carried out using the Mann Whitney U test. The confidence interval on the difference of two proportions (converted to percentages) was used to compare margin status and grade between the different groups, in Stat-Xact V4.0 (Cytel Corp., Cambridge, MA, USA). Results From the database, 118 women over the age of 50 years who underwent LLETZ between 1998 and 2003 were identified. Of these, 92 were included in the final analysis. For the purpose of analysis, histology has been considered as the gold standard for diagnosis. Hence, those nine women who initially had inadequate (n = 3) or abnormal smears (three with borderline nuclear abnormalities and two with mild dyskaryosis and one with severe dyskaryosis) but no histologically proven CIN on either cervical biopsy or loop excision were excluded from the analysis. Other reasons for exclusion included the finding of cervical cancer (n = 6) and endometrial cancer (n = 1). There was absence of any follow-up data in seven women. Following histopathology review, three women underwent a repeat LLETZ within 6 8 weeks of the first procedure and were excluded from the analysis. Histological margins were found to be clear (complete excision) in 62 and involved (incomplete excision) in 22 histological specimens. In eight women, fragmentation or the poor orientation of the histological specimen prevented accurate assessment of the margin status. This group were deemed to have uncertain margins (Table 1). Table 1. Margin status, grade and recurrence Margins High grade, Low grade, Total Histological recurrence, Cytological recurrence, Involved 16 (72.7) 6 (27.3) 22 7 (31.8) 7 (31.8) Uncertain 4 (50.0) 4 (50.0) 8 0 (0.0) 1 (12.5) Clear 19 (30.6) 43 (69.4) 62 5 (8.1) 9 (14.5) Total 39 (42.4) 53 (57.6) 92 12 17 Kaplan Meier plots were used to analyse recurrence. Overall mean time to recurrence was 73.2 months (95% CI: 68.6 77.9) (Figure 1). When assessed by margin status, mean time to recurrence was 47.4 months (95% CI: 38.3 56.5) in those with involved margins and 76.8 months (95% CI: 72.4 81.2) in those with clear margins (Figure 2). Cox regression analysis demonstrates the significant effect of margin status on recurrence course (Table 2). Women with clear margins were less likely to have a recurrence than either those with involved margins (HR 0.18; 95% CI: 0.06 0.59; P = 0.004) or the study group as a whole (P = 0.02). If the specimens were classified by histology into HGL (CIN2, CIN3) and LGL, a mean time to recurrence of 70.7 months (95% CI: 62.8 79.0) was found in the HGL group and 71.6 months (95% CI: 66.8 76.5) in the LGL group. It was not possible to compare the survival courses of these two groups as the proportional hazards assumption is not met. The relationship between margin status, grade of lesion and recurrence is given in Table 1. HGL are more likely to be associated with involved margins and LGL with clear margins Figure 1. Kaplan Meier plot for recurrence for whole cohort. Mean time to recurrence is 73.2 months (95% CI: 68.6 77.9) in the whole cohort. ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology 1239

Manchanda et al. Figure 2. Kaplan Meier plot for recurrence according to margin status. Margins involved mean time to recurrence 47.4 months (95% CI: 38.3 56.5). Clear margins mean time to recurrence 76.8 months (95% CI: 72.4 81.2). (Fisher s exact test, FE(x) = 11.9; P = 0.002; Table 1). The percentage of HGL in the involved margin group was higher than the clear margin group by 42.1% (95% CI: 17.7 66.6%; P = 0.003). However, there was no evidence of an increase (22.7%, 95% CI: 18.0 to 61.7%) in HGL in the involved margins group compared with the uncertain margins group (P = 0.28). Those with involved margins can be further subclassified into the type of margin affected, that is, ectocervical, endocervical and deep stromal or lateral cervical (Table 3). However, the number of recurrences was insufficient to allow further investigation of these variables. Histologically proven recurrence occurred in 12 women and abnormal smears or cytological recurrence was found in 17 women (Table 1). One woman with involved margins developed cervical cancer. Her initial histology at LLETZ showed CIN2. Her follow up was uneventful until 56 months when she was found to have severe dyskaryosis. Subsequent cervical biopsy at colposcopy revealed CIN3. Moderately differentiated squamous cell carcinoma was diagnosed in Table 2. Results of Cox regression analysis Contrast df Significance HR (95% CI) Overall 2 0.02 Uncertain margins 1 0.99 versus margins involved Clear margins versus margins involved 1 0.004 0.18 (0.06 0.59) Cox regression analysis shows that there is a significant effect of margin involvement on the survival course. The hazard of recurrence was 0.18 (95% CI: 0.06 0.59) in those with clear margins compared with those with margins involved (P 5 0.004). the apical segment of the LLETZ specimen. Two women with involved margins who were detected to have histological recurrence also had a concurrent negative smear. Five women with involved margins and four with clear margins who had cytological recurrence at follow up showed no evidence of concurrent histological recurrence. The age distribution of those with involved, clear and uncertain margins ranged from 52 to 74 years, 55 66 years and 55 57 years, respectively. There was no evidence of a difference in the age distribution between the two main groups (involved and clear margins) using a Mann Whitney U test (U = 659, P = 0.82). However, overall, there was a significant difference in distribution of age between the three groups (Kruskal Wallis test, x 2 = 6.81, df =2,P = 0.033). This was due to the significant difference in age between clear margins and uncertain margins (U = 105.5, P = 0.006), and also between uncertain margins and involved margins (U = 44, P = 0.037). Using the Kruskal Wallis test, there was no evidence of a difference in follow up between the three groups (x 2 kw = 3.70, df =2,P = 0.16). Discussion Age over 50 years 5 or even 40 years 3 together with presence of disease at the resected margin 3,5 7 have been documented as independent risk factors for recurrence following LLETZ. Individuals with both these factors would constitute a group at significantly higher risk for recurrence. Our study clearly demonstrates the association between disease recurrence and margin status in women over 50 years by comparing those with involved and clear margins. To the best of our knowledge, this is the first study that looks at recurrence in this fashion, specifically in this composite high-risk group. Previous studies have looked at risks for recurrence (involved margins) in all ages 5 7 or compared outcomes in different age groups. 3,5 Although conservative treatment for CIN reduces the risk of cervical cancer by 95%, the risk is still around five times greater than the general population. 1 The majority of recurrences (80 87%) are reported to occur in the first 18 24 months after LLETZ. 4,5 It has been suggested that earlier recurrence is more common with HGL. 4 It was not possible to perform a Cox regression analysis comparing recurrence course in HGL and LGL with our data as the proportional hazards assumption is not met. Six of the 12 recurrences occurred after 2 years in our study, with two of these occurring after 4.5 years. The recurrences were equally associated with HGL and LGL (six each). Two of these six HGL and four of the six LGL recurred after 2 years. Such late recurrences could reflect viral latency, re-infection or the natural history of persistent residual disease. Typically, recurrences within 2 years have been classified as persistent disease and those beyond this termed as true recurrences. 2,4,10 This classification 1240 ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology

Recurrence after LLETZ for CIN in women over 50 Table 3. Type of margin involved, grade and recurrence Involved margin n HGL, LGL, Histological recurrence, Cytological recurrence, Ectocervical only 8 5 (62.5) 3 (37.5) 3 (37.5) 3 (37.5) Endocervical only 11 9 (81.8) 2 (18.2) 4 (36.4) 3 (27.3) Lateral cervical 2 2 (100) 0 0 1 (50) or deep stromal Endocervical and deep stromal 1 0 1 (100) 0 0 is somewhat arbitrary and we prefer not to distinguish between the two, as it is not possible to be sure which represents true recurrence or residual disease. Hence, we have categorised all as recurrent CIN. The finding of late recurrence in this study is in agreement with previous publications 1,4 and emphasises the need for long-term follow up following LLETZ for CIN. Current British guidelines recommend that women should have annual follow up for at least 10 years after the treatment of CIN2 or worse before returning to the routine screening interval, whereas women treated for CIN1 may be returned to routine recall after 2 years of negative post-treatment cytology. Of the four late recurrences associated with LGL, three were from the clear margin group and one from the involved margin group. Our findings suggest that even women with LGL in this age group need to be followed up for a longer duration than the currently recommended 2 years. Although pathological evidence of incomplete excision of margins does not imply residual disease, 11 the risk of recurrence is clearly increased in these women. Almost one-third of women with involved margins had histologically proven recurrence in our series. Those with clear margins were less likely to develop recurrent CIN (HR 0.18 [95% CI: 0.06 0.59]; P = 0.004; Figure 2, Table 2). The overall rate of recurrence was higher than has been previous documented. 11 This may be explained by the fact that our study group consists exclusively of women over 50 years age and hence, is inherently at a higher risk. Lesion size or the number of involved sectors of atypical transformation have been described as risk factors for recurrent CIN after treatment. 7 These data were not specifically recorded on our database and hence, could not be assessed. However, we did find that HGL were associated with a higher recurrence rate and it is these lesions which are likely to be larger in size 12,13 or more extensive. 14 HGL were more commonly associated with involved margins and LGL with clear margins (Table 1). Kruskal Wallis and Mann Whitney tests show no evidence of a difference in the age distribution and duration of follow up between the involved and clear margin groups. This further validates the inference of a clear association between risk of disease recurrence and margin status in these women. Among those with involved margins, the ectocervical margin was involved in 8, endocervical in 11, lateral cervical or deep stromal in 2 and both endocervical and deep stromal in 1 woman. It has been suggested that the risk of recurrence is mainly due to CIN at the endocervical margin. 5 However, in our series three recurrences were associated with ectocervical and four with endocervical involvement (Table 3). Unfortunately, the low number of recurrences prevents statistical analysis of the contribution of site of the involved margin towards the risk of recurrence. It has been proposed that cytology alone may be adequate to detect residual CIN at follow up. 11 However, if histological assessment is used as the gold standard for disease recurrence, the use of cervical cytology for post-treatment surveillance has a high false-positive rate (negative concurrent histology) of up to 43%. 15 Consistent with this, nine women with cytological abnormality at follow up showed no evidence of histological recurrence in our series. These nine women included five with involved margins. The use of colposcopy in the post-treatment setting could mitigate against such a potential for over diagnosis. However, post-treatment inflammatory changes may lead to the acetowhite of regenerating and undifferentiated epithelium being confused with that of CIN, and repeat LLETZ based on colposcopy alone is therefore associated with the finding of non-dysplastic histology in a greater proportion of women than at primary treatment. 2,3 Delaying initial follow up until 6 months may reduce the risk of confusing acetowhite of metaplastic epithelium with CIN. Colposcopic follow up has also been recommended following LLETZ where the excision has been ruled incomplete on the basis of histologically involved margins. 6,11 In the current series, two women with involved margins at primary treatment had histological recurrence detected at colposcopy following normal cervical smears. False-negative cytology may be common in this setting occurring in up to 47% of women. 3,11 Post-treatment colposcopic review offers the potential benefit of early diagnosis of treatment failures associated with false-negative smears. 3 The relatively poor sensitivity and specificity of cervical cytology must be balanced against the cost of routine colposcopy post-lletz and the risk of over treatment due to over interpretation of benign ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology 1241

Manchanda et al. cervical change. To date, it is our practise to perform a colposcopic examination and cervical cytology at first follow up in women over 50 years with involved margins. However, further prospective studies are needed to establish the benefit of colposcopy over cytology alone. Current UK national guidelines recommend repeat excision for the presence of CIN3 at the endocervical margin. 8 In this study, two of the seven (28%) recurrences in the involved margin group occurred in LGL in our series. Furthermore, three of the seven (42%) recurrences in this group were in women with ectocervical margin involvement only. The data suggest that it is worth considering performing a repeat LLETZ in all such women over the age of 50, irrespective of the type of margin involved. A second LLETZ may, however, increase the risk of cervical stenosis, thus making follow up more difficult. Ideally, an appropriately powered large randomised controlled trial would be needed to establish the efficacy of such a strategy. Conclusion There is a strong association between involved margins and recurrence of CIN in women over the age of 50 years. This finding suggests consideration should be given to repeat LLETZ in this high-risk group of women. It was not possible to establish the contribution of location of the involved margin to the recurrence risk and larger studies will be needed to resolve this issue. Disclosure of interest This article is original; does not infringe upon any copyright or other proprietary right of any third party; is not under consideration by another publication and has not been previously published. The authors confirm that the final manuscript has been read and that each author s contribution has been approved by the appropriate author. The authors declare no conflict of interest. Contribution to authorship R.M. was involved in data collection, analysis, drafting and writing of the paper. P.B., J.L., K.W., R.C. and M.S. contributed to writing of the manuscript and were responsible for the clinical care of the women. Robin Mosley reviewed the histological specimens and the final manuscript. S.L.V. performed the statistical analysis and has contributed to writing the statistical sections of the manuscript. The final draft was prepared by R.M., P.B., R.C., M.S. and approved by the others. Ethics approval The project was referred to the Chair of the Research Ethics committee (NHNNIN Joint REC reference number 07L 245). Under the Research Governance Framework, the project was deemed to be a clinical audit, and permission for data analysis and submission for publication was given. j References 1 Soutter WP, Lopes A, Fletcher A, Monaghan JM, Duncan ID, Paraskevaidis E, et al. Invasive cervical cancer after conservative therapy for cervical intraepithelial neoplasia. Lancet 1997;349:978 80. 2 Bigrigg A, Haffenden DK, Sheehan AL, Codling BW, Read MD. Efficacy and safety of large-loop excision of the transformation zone. Lancet 1994;343:32 4. 3 Flannelly G, Langhan H, Jandial L, Mann E, Campbell M, Kitchener H. A study of treatment failures following large loop excision of the transformation zone for the treatment of cervical intraepithelial neoplasia. Br J Obstet Gynaecol 1997;104:718 22. 4 van Hamont D, van Ham MAPC, Struick-van der Zanden PHTH, Keijser KGG, Bulten J, Melchers WJG, et al. Long-term follow-up after large-loop excision of the transformation zone: evaluation of 22 years treatment of high-grade cervical intraepithelial neoplasia. Int J Gynecol Cancer 2006;16:615 619. 5 Flannelly G, Bolger B, Fawzi H, Lopes A, Monaghan JM. Follow up after LLETZ: could schedules be modified according to risk of recurrence. BJOG 2001;108:1025 30. 6 Dobbs SP, Asmussen T, Nunns D, Hollingworth J, Brown LJR, Ireland D. Does histological incomplete excision of cervical intraepithelial neoplasia following large loop excision of transformation zone increase recurrence rates? A six year cytological follow up. BJOG 2000;107:1298 301. 7 Shafi MI, Dunn JA, Buxton EJ, Finn CB, Jordan JA, Luesley DM. Abnormal cervical cytology following large loop excision of the transformation zone: a case controlled study. Br J Obstet Gynaecol 1993;100:145 8. 8 Colposcopy and Programme Management. Guidelines for the NHS Cervical Screening Programme. NHSCSP Publication No. 20. 2004. Sheffield, UK: NHS Cancer Screening Programmes. 9 Murdoch JB, Grimshaw RN, Monaghan JM. Loop diathermy excision of the abnormal cervical transformation zone. Int J Gynecol Cancer 1991; 1:105 12. 10 Paraskevaidis E, Jandial L, Mann EM, Fisher PM, Kitchener HC. Pattern of treatment failure following laser for cervical intraepithelial neoplasia: implications for follow-up protocol. Obstet Gynecol 1991;78:80 3. 11 Murdoch JB, Morgan PR, Lopes A, Monaghan JM. Histological incomplete excision of CIN after large loop excision of the transformation zone (LLETZ) merits careful follow up, not treatment. Br J Obstet Gynaecol 1992;99:990 3. 12 Shafi MI, Finn CB, Luesley DM, Jordan JA, Dunn J. Lesion size and histology of atypical transformation zone. Br J Obstet Gynaecol 1991; 98:490 2. 13 Jarmulowicz MR, Jenkins D, Barton SE Goodall AL, Hollongworth A, Singer A. Cytological status and lesion size: a further dimension in cervical intraepithelial neoplasia. Br J Obstet Gynaecol 1989;96:1061 6. 14 Anderson MC, Hartley RB. Cervical crypt involvement by intraepithelial neoplasia. Obstet Gynecol 1980;55:546 50. 15 Buxton EJ, Luesley DM, Wade-Evans T, Jordan JA. Residual disease after cone biopsy: completeness of excision and follow-up cytology as predictive factors. Obstet Gynecol 1987;70:529 32. 1242 ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology