Gynaecological surgery. Introduction. MJ Camargo, a FB Russomano, a MA Trist~ao, a G Huf, b W Prendiville c

Transcription

1 DOI: / Gynaecological surgery Large loop versus straight-wire excision of the transformation zone for treatment of cervical intraepithelial neoplasia: a randomised controlled trial of electrosurgical techniques MJ Camargo, a FB Russomano, a MA Trist~ao, a G Huf, b W Prendiville c a Fernandes Figueira Institute, Oswaldo Cruz Foundation, Rio de Janeiro, Brazil b National Institute of Quality Control in Health, Oswaldo Cruz Foundation, Rio de Janeiro, Brazil c Walter Prendiville, International Agency for Research on Cancer, World Health Organization, Lyon, France Correspondence: MJ Camargo, Fernandes Figueira Institute, Oswaldo Cruz Foundation, Rio de Janeiro , Brazil. mariaj. camargo@hotmail.com Accepted 13 October Published Online 17 December Objective To compare two electrosurgical techniques, straight-wire excision of transformation zone (SWETZ) with large loop excision of transformation zone, as a cone procedure (LLETZ-cone), for the treatment of cervical intraepithelial neoplasia (CIN), when disease is present at the cervical canal. Design Randomised controlled trial. Setting Two public hospitals, one in Rio de Janeiro, Brazil and one in Dublin, Ireland. Population One hundred and three women with indication to treat CIN located at cervical canal. Methods Women were randomised to receive LLETZ-cone or SWETZ. Outcomes Main outcome was the incidence of complete excision of disease at endocervical margin of the surgical specimen. Secondary outcomes were complete excision at ectocervical and stromal margins, time to complete the procedure, specimen fragmentation, blood loss and death after 1 year. Results Fifty-two women were allocated to LLETZ-cone and 51 to SWETZ. Ten women were lost for main outcome because of damaged specimens. Forty-two women in the LLETZ-cone group had free endocervical margin versus 43 women in the SWETZ group (relative risk 1.04, 95% confidence interval [95% CI] ; P = 0.64). For secondary outcomes related to margins, we observed a relative risk of 1.15 (95% CI ; P = 0.15) for ectocervical free margin. For free stromal margin, the relative risk was 1.07 (95% CI ; P = 0.47). No death was observed. Conclusions This study was inconclusive; SWETZ and LLETZcone were equally effective to treat endocervical disease, with no difference in protecting against margin involvement. Higher, but not severe, blood loss and longer surgical time were observed in the SWETZ group. Keywords Cervical intraepithelial neoplasia, cervix dysplasia, cone biopsy, electrosurgery, treatment margins. Please cite this paper as: Camargo MJ, Russomano FB, Trist~ao MA, Huf G, Prendiville W. Large loop versus straight-wire excision of the transformation zone for treatment of cervical intraepithelial neoplasia: a randomised controlled trial of electrosurgical techniques. BJOG 2015;122: Introduction With new cases every year, cervical cancer was the fourth most common cancer affecting women worldwide in There were estimated deaths from cervical cancer in 2012, accounting for 7.5% of all female cancer Trial registration: ClinicalTrials.gov. NCT (Diathermy cone biopsy: a study of two techniques) deaths. Almost 87% of those deaths occur in underdeveloped regions. 1 In Brazil, new cases of cervical cancers are expected in Organised screening programmes followed by treatment of screen-positive women are associated with a dramatic reduction in deaths. Large loop excision of transformation zone (LLETZ) is a commonly used treatment for cervical intraepithelial neoplasia (CIN) and its popularity is the result of a number of advantages that include ease of technique, outpatient set- 552 ª 2014 Royal College of Obstetricians and Gynaecologists

2 LLETZ-cone versus SWETZ for the treatment of CIN ting, histological analysis of the specimen and low cost. 3 However, for endocervical pre-invasive disease LLETZ is not always adequately performed. There is relatively little information on this form of treatment of CIN, particularly on women with a type 3 transformation zone or glandular disease. The excision, in different degrees of depth at the cervix canal, may not completely remove the disease. Endocervical CIN demands a different approach, a cone biopsy classified as type three excision. 4 The usual method uses LLETZ with the objective of achieving 20-mm depth at the endocervical canal. Another technique, straight-wire excision of transformation zone (SWETZ), is a method of electrosurgically excising the endocervical transformation zone or glandular disease, using a 1-cm straight-wire electrode. The activated wire is used in the same way as a cold knife, fashioning a cone biopsy. 5 It has also been described as needle excision of the transformation zone (NETZ). 6 A systematic review evaluating surgery for CIN included many types of interventions, and concluded that there was no obviously superior surgical technique for treating CIN. 7 Two randomised trials were described in this review comparing NETZ with LLETZ. Panoskaltsis et al. 8 indicated better performance of NETZ in terms of specimen fragmentation, free margins and risk of residual disease. We did not find any randomised trial comparing these techniques to treat women with endocervical disease. Incomplete excision exposes women to a substantial risk of high-grade post-treatment disease, especially if deep margins are involved The objective of this study was to compare these techniques in terms of achieving free margins of disease for the treatment of women with a type 3 transformation zone or glandular disease. Papanicolaou smear presenting high-grade squamous intraepithelial lesion (HSIL) in an incomplete visible transformation zone, suspicion of microinvasive or invasive carcinoma and suspicion of glandular disease. Ectocervical premalignant condition has not been considered an indication for cone biopsy. Patients were excluded from the trial if they presented only ectocervical disease (type one or two transformation zone), coagulation disorders, cervicitis, pregnancy or refusal to participate in the study. Interventions The standard procedure, LLETZ-cone, was performed with a large loop electrode of 2-cm depth. The activated loop was applied to the cervix outside the lateral margin of the transformation zone and brought slowly to the contralateral margin with the ambition of removing a 2-cm length of endocervical epithelium, using a blend two coagulation setting of W (Figure 1) SWETZ uses a 1-cm straight, 0.2-cm wire electrode, to fashion a 2-cm excision, using a blend two coagulation setting of 40 W (Figure 2). After excision, the biopsy specimen was opened and pinned to a cord or polystyrene pad before immersion in formalin, thereby facilitating serial section histology. Three surgeons, two in Rio de Janeiro and one in Dublin, participated in the study. All of them were colposcopists and trained in both techniques. Some medical residents had also operated women, under supervision. The procedures were under- Methods Design and setting This was an open randomised trial with two participating public hospitals: Instituto Fernandes Figueira at Oswaldo Cruz Foundation (IFF/Fiocruz) in Rio de Janeiro, Brazil and Coombe Women & Infants University Hospital, in Dublin, Ireland. Both hospitals provide medical assistance for women requiring maternity, neonatal and gynaecology care and are leading institutions in healthcare research, education and training. Participants At IFF/Fiocruz, since 1999 every woman with premalignant cervical disease and indication for a cone biopsy has been submitted to electrosurgical conisation. From this cohort of women, people were recruited to take part in the trial. Patients were eligible if they presented endocervical disease (type 3 transformation zone or suspicion of glandular disease). Common indications for a cone biopsy included: Figure 1. LLETZ-cone technique. 3 ª 2014 Royal College of Obstetricians and Gynaecologists 553

3 Camargo et al. concealment. Once selected to trial entry, the name of the woman was written on the envelope and it was opened just before the procedure. This constituted trial entry. If for any reason the procedure was not performed, the woman remained in the study and the envelope was returned to the trial coordinator. Sample size According to the analysis of the cohort of women usually submitted to cone biopsy in our colposcopy clinic, LLETZcone is associated with approximately 80% rate of complete endocervical margin excision. We thought that an increase of 15% of complete excision might influence clinicians to consider changing techniques, if this difference was achieved by SWETZ. One hundred and forty-six women would be required to have an 80% chance of detecting an increase in the rate of complete endocervical margin excision, from 80% in the LLETZ-cone group to 95% in the SWETZ group, as significant at the 5% level. 12 Figure 2. SWETZ technique. 3 taken in the operating room and in most cases local anaesthesia plus sedation was used. Outcomes Our primary outcome was the proportion of women achieving complete excision of dysplasia at the endocervical margin, as recognised histologically in each group. The pathologist was blinded to the used technique. Secondary outcomes were perioperative and histological evaluations. Operative outcomes were: time to complete the procedure; fragmentation of the surgical specimen, assessed by the number of excised pieces; and blood loss. Procedure time was recorded from insertion until removal of the vaginal speculum. Blood loss was categorised according to surgeon s visual observations during the procedure as: <20 ml, between 20 and 100 ml and >100 ml, and was recorded just after the end of surgery. Mortality in 1 year was also assessed. Histological secondary outcomes were ectocervical and stromal margins free of disease. Randomisation The random allocation generation was generated centrally in the RCSI Department of Obstetrics and Gynaecology at the Coombe Women s Hospital using computer-generated random numbers. Randomisation was stratified for each centre. Consecutively numbered treatment options were placed in opaque, sealed envelopes to ensure allocation Data analyses We assessed the success of randomisation by comparing sociodemographic and clinical characteristics between the two treatment groups and calculated relative risks (with 95% confidence intervals [95% CI]) for primary and secondary outcomes using an intention-to-treat analysis. For continuous variables we used means, standard deviations and Student s t-test. Data were analysed using SPSS 17.0 (SPSS Inc., Chicago, IL, USA). Deviations from protocol The trial stopped early, before the estimated number of women was achieved. The reason was that the participation of some voluntary pathologists could not be maintained until the end of the study. Results Between November 1999 and December 2004, 127 women were assessed for eligibility and 103 were randomised, the majority in Rio de Janeiro, seven participants in Dublin. Five women refused to participate in the study and 19 did not meet the inclusion criteria (Figure 3). Fifty-two women were allocated to LLETZ-cone and fifty-one to SWETZ. For the primary outcome, data were available for 93 women (90.3%). Ten women had unknown margin status because of fragmentation of the specimen or thermal damage. They were analysed according to an intention-to-treat approach. There were no losses for perioperative outcomes, which include procedure time, blood loss and number of pieces. Clinical and sociodemographic characteristics of participants are described in Table 1. Patients in the two groups 554 ª 2014 Royal College of Obstetricians and Gynaecologists

4 LLETZ-cone versus SWETZ for the treatment of CIN Table 1. Characteristics of women by intervention group LLETZ-cone n = 52 SWETZ n = 51 Age (years), mean (SD) 45.6 (11.5) 43.7 (9.9) Parity, mean (SD) 3.3 (11.5) 3.1 (2.5) Papanicolaou smear CIGN 1 (1.9) 4 (7.8) LSIL 3 (5.8) 2 (3.9) HSIL 37 (71.2) 44 (86.3) Invasion 11 (21.2) 1 (2.0) Colposcopic diagnosis HSIL (major changes) 37 (71.2) 28 (54.9) LSIL (minor changes) 5 (9.6) 9 (17.6) Invasion 1 (1.9) Inconclusive/Unsatisfactory 9 (17.3) 14 (27.5) Cone indication AIS 1 (1.9) 2 (3.9) Recurrence 1 (1.9) Cancer suspicion 5 (9.6) 1 (2) TZ incompletely seen 43 (82.7) 47 (92.2) Others 2 (3.8) 1 (2) Length of the excision attempted (cm), mean (SD) 2.84 (0.79) 2.67 (0.69) Figure 3. Flow diagram of the study. Values are numbers (percentage) unless stated otherwise. AIS, adenocarcinoma in situ; CIGN, cervical intraepithelial glandular neoplasia; HSIL, high-grade squamous intraepithelial lesion; LSIL, low-grade squamous intraepithelial lesion; TZ, transformation zone. had similar baseline characteristics, except for cytological diagnosis. A Papanicolaou smear suggesting invasion was present in 11 of 52 women in the LLETZ-cone group compared with one woman of 51 in the SWETZ group. Most women enrolled in this study were in their fifth decade, with a medium parity of three and with high-grade disease; the transformation zone being incompletely seen was the main indication for surgery. For the primary outcome (Table 2), 80.8% of women in the LLETZ-cone group were reported as having an endocervical margin that was free of disease compared with 84.3% of women in the SWETZ group. The relative risk for complete excision was 1.04 (95% CI ; P = 0.64). For secondary outcomes related to margins, we observed a relative risk of 1.15 (95% CI ; P = 0.15) for ectocervical margin free of disease. For stromal margin, the relative risk of a free margin was 1.07 (95% CI ; P = 0.47). No statistical significance was observed for incomplete excision at any margin for women with endocervical disease. When perioperative outcomes were analysed, the surgical time was significantly longer for women submitted to SWETZ, with a mean of 23.2 minutes (SD 8.1), compared with LLETZ-cone (mean 13.5 minutes; SD 5.3), difference of means 9.7 ( ; P = 0.000). The fragmentation of the specimen during the excision was more frequent in the LLETZ-cone group than in the SWETZ group, with 86% of women having the excision performed in one fragment by LLETZ-cone compared with 96% by SWETZ. The relative risk was 1.11 (95% CI ; P = 0.09). The blood loss was small for both groups, only one woman lost >100 ml in the LLETZ-cone group; but surgical bleeding was clinically and statistically greater in the SWETZ group. Five women (5.7%) in the LLETZ-cone group lost between 20 and 100 ml of blood and 14 women (27.4%) submitted to SWETZ lost the same amount of blood. Nobody had died by 1 year after the procedure. The definitive histological diagnosis is presented in Table 3, and there were no important differences between groups. Overall, eight women had no disease, 15 had lowgrade disease (CIN 1, HPV), 72 had high-grade disease (CIN 2 3, AIS), two had microinvasive disease and four had invasive cancer. Discussion Main findings Both surgical procedures were effective in achieving clear histological margins, with complete excision ranging from ª 2014 Royal College of Obstetricians and Gynaecologists 555

5 Camargo et al. Table 2. Histological and operatives outcomes Patients outcomes Free endocervical margin (primary outcome) Free ectocervical margin Free stromal margin Procedure time in minutes (SD) Specimen not fragmented Blood loss <20 ml LLETZ-cone n = 52 SWETZ n = 51 Relative risk (CI) P-value 42 (80.8) 43 (84.3) 1.04 ( ) (75.0) 44 (86.3) 1.15 ( ) (78.8) 43 (84.3) 1.07 ( ) (5.3) 23.2 (8.1) 9.7 ( )* (86) 49 (96) 1.11 ( ) (88.5) 36 (71) 0.80 ( ) 0.03 Values are numbers (percentages) unless stated otherwise. *Difference of means. Table 3. Definitive histological diagnosis LLETZ-cone n = 52 SWETZ n = 51 Adenocarcinoma 1 (2) AIS 1 (1.9) 2 (3.9) CIN (not specified) 1 (3.9) HSIL (CIN2, CIN3) 35 (67.3) 33 (64.7) LSIL (HPV, CIN1) 6 (11.5) 9 (17.6) Squamous carcinoma 2 (3.8) 1 (2) Microinvasive carcinoma 2 (3.8) Negative 5 (9.6) 3 (5.9) Not stated 1 (1.9) 1 (2) Values are numbers (percentages). AIS, adenocarcinoma in situ; CIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; HSIL, high-grade squamous intraepithelial lesion; LSIL, low-grade squamous intraepithelial lesion. 75% to 86%, with no statistical difference between the two techniques. There was no statistically significant difference concerning surgical fragmentation, an important issue when it comes to the diagnosis of microinvasive disease. Statistical significance was obtained for procedure time and blood loss. SWETZ demanded longer surgical time (23.2 minutes versus 13.5 minutes), which reflects the complexity of the procedure with the need to fashion the cone using the straight wire. This could also suggest that not all the colposcopists involved were familiar with the technique. Both procedures took longer than reported in the literature. 6,8 Women in the SWETZ group were significantly at greater risk of bleeding, even so, only four women in both groups combined had >100 ml blood loss, all solved during surgery with hemostatic sutures. Strengths and limitations This study was planned to be conducted with minimal funding, little or no interference in the routine of the service and voluntary collaboration. Unfortunately, an important part of these volunteers had to leave the hospital when approximately 70% of women had been included. Stopping the trial before achieving the necessary number of participants was a drawback, probably influencing the statistical significance for most outcomes. Nevertheless, to our knowledge, this was the first study to evaluate electrosurgical excision procedures for women with transformation zone type 3 or glandular disease. Consequently, the main strength of our study is to select only those women with endocervical disease and to choose the appropriate devices for an adequate depth of excision (a loop with 2-cm depth and a 1-cm straight wire used to achieve 2-cm depth). Losses for the primary outcome were minimal, about 10%, because of limitation of electrosurgical techniques, such as thermal damage and fragmentation. For perioperative outcomes no loss of follow-up was observed. The concealment of allocation and the blinded rating of the main outcome are other strengths of this study. Interpretations Incomplete excision of CIN increases the risk of post-treatment high-grade disease six-fold; 9 therefore, every effort to avoid incomplete excision should be welcome. There were some reasons for the inconclusive results, such as stopping the study earlier or anticipating a bigger difference between the two groups than what was actually found. We based our assumptions on literature reports, finding 12 27% differences favouring needle excision for margin protection. 8,13 LLETZ is usually performed in a flexible way, allowing different depths of loop, but NETZ is normally performed in a more standardised way; 6,8,13 to establish a standardiasation of these devices could be a way to enhance the treatment of women with pre-invasive endocervical disease. Most of the randomised studies evaluating electrosurgical procedures for CIN treatment do not distinguish women with ectocervical and endocervical dysplasia. In the trial by Panoskaltsis et al., 404 women were included with only 29 women treated for endocervical disease and it was not possible to access the number of women under the same conditions in the studies by Sadek and by Basu et al. 6,13 The major implication for this lack of selection criteria in many studies is that it enhances 556 ª 2014 Royal College of Obstetricians and Gynaecologists

6 LLETZ-cone versus SWETZ for the treatment of CIN the possibility of undertreating those women with dysplastic epithelium that is not accessible by colposcopy. The differences in endocervical margins are small (4%) and might be related to different degrees of surgeon s experience with the new technique. One of the main criticisms of eletrosurgical excisional procedures is the impact on women s reproductive life, but if a straight selective criterion is applied, and the endocervical disease is focused, only the oldest women will be submitted to more extensive excision such as cone biopsy: in our study, the average age was 44.6 years. We believe that the lack of a good screening programme in Brazil could raise the prevalence of pre-invasive disease in older women, compared with developed countries with better screening programmes the low number of women recruited in Dublin could reflect this reality. We understand that the presence of disease during the follow up provides better evaluation of therapeutic success. In this study we could not access this information and future studies should focus on cytological follow-up results after LLETZ/SWETZ. Because there are no long-term followup data is it not possible to demonstrate if the two techniques have different cure rates. Conclusion SWETZ and LLETZ-cone were both effective in achieving clear histological margins with no statistical difference between the two techniques. There was, however, higher but not severe blood loss and longer performance time for SWETZ compared with LLETZ-cone. Further studies focused on the treatment of women with CIN located at the cervical canal including follow-up results are needed to improve therapeutic success. Disclosure of interest The authors declared no conflicts of interest. Contribution to authorship WP conceived and designed the experiments. MJC was the principle investigator, searched for publications, wrote the original protocol with WP, and performed the experiments in cooperation with FR, WP and MAT. FR designed the database. GH and FR performed the analysis. MJC and GH wrote the paper. Details of ethics approval The study protocol was approved by Ethics Committees in Brazil (CONEP: 552/99) and Ireland (Coombe Women s Hospital MT/ff 17 October 1997). Funding The authors received no funding for this study. All funding was intramural. Acknowledgements We would like to thank Dr Aldo F F Reis, Dr Maria Jose Maisonette, Dr Fernanda Borzino and Dr Nathalie Raibolt for their help with this study. & References 1 Cervical cancer. Estimated incidence, mortality and prevalence worldwide in [ cervix-new.asp]. Accessed 21 April Cancer Incidence in Brazil. Rio de Janeiro: INCA, [ Accessed 21 April Prendiville W. LLETZ. Theoretical rationale, practical aspects, clinical experience, optimizing the technique. In: Prendiville W (ed) Colposcopy: Management Options, London: Elsevier limited; p IFCPC Nomenclature. [ nomenclature]. Accessed 15 June Prendiville W. Large loop excision of the transformation zone. In: Prendiville W (ed) Large Loop Excision of the Transformation Zone: A Practical Guide to LLETZ; London: Chapman and Hall; p Sadek AL. Needle excision of the transformation zone: a new method of treatment of cervical intraepithelial neoplasia. Am J Obstet Gynecol 2000;182: Martin-Hirsch PP, Paraskevaidis E, Bryant A, Dickinson HO, Keep SL. Surgery for cervical intraepithelial neoplasia. Cochrane Database Syst Rev 2010:CD Panoskaltsis T, Ind TE, Perryman K, Dina R, Abrahams Y, Soutter WP. Needle versus loop diathermy excision of the transformation zone for the treatment of cervical intraepithelial neoplasia: a randomized controlled trial. BJOG 2004;111: Ghaem-Maghami S, Sagi S, Majeed G, Soutter WP. Incomplete excision of cervical intraepithelial neoplasia and risk of treatment failure: a meta-analysis. Lancet Oncol 2007;8: Gardeil F, Barry-Walsh C, Prendiville W, Clinch J, Turner MJ. Persistent intraepithelial neoplasia after excision for cervical intraepithelial neoplasia grade III. Obstet Gynecol 1997;89: Oliveira CA, Russomano FB, Gomes Junior SCS, Corr^ea FM. Risk of persistent high-grade squamous intraepithelial lesion after electrosurgical excisional treatment with positive margins: a metaanalysis. Sao Paulo Med J 2012;130: Herbert R. Confidence interval calculator. [ direct.org/applied_statistics/binomial_distribution/ref/cicalculator.xls]. Accessed 15 December Basu PS, D Arcy T, McIndoe A, Soutter WP. Is needle diathermy excision of the transformation zone a better treatment for cervical intraepithelial neoplasia than large loop excision? Lancet 1999;353: ª 2014 Royal College of Obstetricians and Gynaecologists 557