IJC International Journal of Cancer

Transcription

1 IJC International Journal of Cancer Unique recurrence patterns of cervical intraepithelial neoplasia after excision of the squamocolumnar junction Michael Herfs 1, Joan Somja 1, Brooke E. Howitt 2,3, Meggy Suarez-Carmona 1, Gaelle Kustermans 1, Pascale Hubert 1, Jean Doyen 4, Frederic Goffin 4, Frederic Kridelka 4, Christopher P. Crum 2,3* and Philippe Delvenne 1* 1 Department of Pathology, Laboratory of Experimental Pathology, GIGA-Cancer, University of Liege, Liege, Belgium 2 Department of Pathology, Brigham and Women s Hospital, Boston, MA 3 Division of Women s and Perinatal Pathology, Brigham and Women s Hospital, Boston, MA 4 Department of Gynecology, University of Liege, Liege, Belgium Recent studies have identified a putative cell of origin for cervical intraepithelial neoplasia (CIN) and cervical cancer at the squamocolumnar junction (SCJ) and suggest that these cells may not regenerate after excision (loop electrosurgical excision procedure). Our study addressed the impact of SCJ excision on the temporal dynamics, histologic and viral (human papillomavirus, HPV) characteristics of recurrent CIN. One hundred and thirty-one consecutive patients treated by excision and attending follow-up visits were enrolled. We compared recurrent and initial CIN with attention to excision margins, timing of recurrence, CIN grade, HPV types, p16 immunophenotype and SCJ immunophenotype. During the follow-up period (up to 4 years), 16 (12.2%) recurrences were identified. Four (25%) were identified at the first follow-up visit, closely resembled the initial CIN 2/3 in grade and HPV type and were typically SCJ marker positive [SCJ(1)], suggesting nonexcised (residual) disease. Twelve (75%) manifested after the first postoperative visit and all were in the ectocervix or in mature metaplastic epithelium. All of the 12 delayed recurrences were classified as CIN 1 and were SCJ (2). In total, 9 out of 11 SCJ (2) recurrences (82%) followed regressed spontaneously. Taken together, these results show that new lesions developing from any HPV infection are delayed and occur within the ectocervix or metaplastic epithelium. This markedly lower risk of CIN 2/3 after successful SCJ excision suggests that the removal of the SCJ could be a critical variable in reducing the risk of subsequent CIN 2/3 and cervical cancer. Key words: HPV, cervical intraepithelial neoplasia, excision, follow-up *C.P.C. and P.D. contributed equally as senior authors Grant sponsors: The Belgian Fund for Medical Scientific Research, The Centre Anti-Cancereux près l Universite de Liège, The Faculty of Medicine of the University of Liège; Grant sponsor: The National Cancer Institute; Grant number: 1R21CA ; Grant sponsor: The Fonds Leon Fredericq DOI: /ijc History: Received 18 Feb 2014; Accepted 9 May 2014; Online 16 May 2014 Correspondence to: Michael Herfs, Laboratory of Experimental Pathology, GIGA-Cancer, University of Liege, Liege 4000, Belgium, Tel.: , Fax: , M.Herfs@ulg.ac.be Human papillomavirus (HPV) has been linked to cervical cancer and its precursors from the early 1980s. 1,2 To date, more than 100 HPV genotypes have been characterized and vaccination programs of sexually naive individuals are anticipated to eventually achieve a significant reduction in cervical cancer incidence and mortality. 3,4 Despite this, there is a large population of women in both high- and low-resource settings, who have been exposed to HPV and will be at risk for cervical cancer irrespective of vaccination. Management of this subset of women remains within the realm of conventional screening and HPV testing with the removal of potentially dangerous precancers (cervical intraepithelial neoplasia or CIN 2/3) by cryoablation or excision (loop electrosurgical excision procedure [LEEP] or cone biopsy). Protocols using a pre-emptive see and treat strategy are being explored in low-resource settings and are designed to identify the candidate CIN by the application of acetic acid to the cervix followed by immediate ablation/excision of the acetowhite areas. Ablation/excision techniques are highly successful in both removing CIN 2/3 and preventing recurrence, achieving an approximately 90% cure rate with a very low risk of a cervical cancer outcome. 5,6 This high success rate has been generally credited to both the removal of the lesion and an acquired HPV type-specific immune response that protects the cervical mucosa from a recurrent infection. However, a third variable possibly influencing recurrence rates has been proposed with the discovery of a discrete population of cells at the cervical squamocolumnar junction (SCJ), a model that has its roots in a similar discovery of residual embryonic cells in the esophagogastric junction. 7 9 This cell type shares a common immunophenotype with cervical cancer and has been proposed as the cell of origin for this disease. In the cervix model, embryonic SCJ cells normally differentiate to form mature columnar and metaplastic squamous

2 1044 Impact of SCJ excision on recurrent lesions What s new? A discrete population of cells at the squamo-columnar junction (SCJ) of the cervix may house the cells of origin the first to become genetically altered in cancer initiation for cervical intraepithelial neoplasia (CIN) and cervical cancer. In the present comparison of excision margins, timing of recurrence, CIN grade, HPV types, p16 immunophenotype, and SCJ immunophenotype for recurrent and initial CIN, recurrences were found to be invariably low grade and within the ectocervix or metaplastic epithelium. Hence, successful SCJ excision appears to reduce the risk of new CIN 2/3 lesions, suggesting that SCJ cryoablation could be effective in preventing cervical cancer. epithelium. 8 Based on the histochemical evidence, direct infection of the embryonic SCJ cells by high-risk HPV is required to produce CIN 2/3 and ultimately carcinoma. In contrast, infection of the mature metaplastic/ectocervical squamous epithelium more commonly leads to CIN 1 with a greater likelihood of regression. 10 The potential significance of this dichotomy lies in the observation that recurrence rates after attempted excision of the SCJ are low. The implication is that with excision of the SCJ cells the only remaining susceptible epithelium is the less cancer-prone metaplastic or ectocervical epithelium. 7 If excision of the SCJ eliminated the principal target population for HPV-induced CIN 2/3 and carcinomas, it would be expected to influence the location and histology of recurrent lesions that were discovered in follow-up after excision. Our study surveyed a consecutive series of women who had undergone cervical excision by LEEP for CIN 2/3, and determined the location, HPV type, biomarker expression pattern and SCJ immunophenotype (positive or negative) of recurrences. The goal was to better understand the nature of recurrences after treatment and the influence of SCJ excision on the above parameters, lesion grade and outcome. Material and Methods Case selection One hundred and ninety-four women (mean age, years [range, 22 67]) with histologically confirmed CIN were treated by excision in the Department of Obstetrics and Gynaecology (University Hospital of Liege, Belgium) between March 2006 and November All women were treated by LEEP. The indications to perform cervical excision were either histological diagnosis of CIN 2/3 (94.3%) or CIN 1 persistent for more than 2 years (5.7%). No adenocarcinoma or microinvasive squamous cell carcinoma were diagnosed. Pretreatment cytology, biopsy and carcinogenic HPV testing (hybrid capture 2 or HC2, Digene, Gaithersburg, MD) were available in all selected women. These samples were collected within 2 months before excision. All cases were examined independently by two pathologists using established criteria. Two additional independent pathologists reviewed the slides blindly with attention to lesion grading and distinguishing CIN 1 from CIN 2/3. Disagreements were resolved by a group review. Inclusion criteria were both CIN confirmed in the cone specimen and patient attending at least two follow-up visits. Sixty-three patients (32.5%) were lost during the follow-up and were excluded from the study. Therefore, 131 women were considered eligible and included in the database. The protocol was approved by the Ethics Committee of the University Hospital of Liege. After excision, the surgical margins (ectocervical and endocervical) were considered as positive if the preneoplastic lesion involved the resection margin. Postoperatively, vaginal intercourse was not advised for 4 weeks. Follow-up examination Postexcision examination was scheduled at 3, 6 and 12 months during the first year and every 6 months after this period for at least 2 years. Mean number of visits after treatment was 4 (range, 2 6) and the mean follow-up period was 89.5 weeks, with the range of weeks (Table 2). A Pap smear, high-risk HPV DNA test (AMPLICOR HPV test, Roche, Basel, Switzerland) and colposcopy were performed at each post-treatment visit. The cytobrush has been used for cervical cytology. The obtained smears were put onto a glass slide, stained by Papanicolaou method and were reviewed using the published criteria. When atypical cells (ASC-US or higher) were detected and/or carcinogenic HPV genotypes were observed (A 450 > 0.2) using the AMPLICOR HPV test (Roche, Basel, Switzerland), a colposcopy-directed punch biopsy was performed. These samples were embedded in paraffin and tissue sections were examined by experienced pathologists. HPV testing Prior treatment, residual material from all liquid-based cytology samples was tested for carcinogenic HPV using the HC2 assay according to the manufacturer s instructions (Digene, Gaithersburg, MD). This commercially available system contains specific probes for 13 high-risk HPV genotypes (HPV types, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59 and 68). The results were expressed as relative light units (RLUs) compared to the result for a high-risk positive control containing 1 pg/ml of HPV DNA (1 RLU). The RLU value (HPV load) of each individual sample was recorded. Postexcision, HPV DNA detection in cervical cytology samples was performed using the AMPLICOR HPV test according to the manufacturer s instructions (Roche, Basel, Switzerland). In this polymerase chain reaction (PCR)-based

3 Herfs et al Table 1. Risk of residual/recurrent/new lesion after electrosurgical cone resection of the cervix according to several demographical features Characteristics No recurrence (n 5 115) (87.8%) Recurrence (n 5 16) (12.2%) p-value Age (23 67) (years) <35 55 (47.8%) 9 (56.3%) >35 60 (52.2%) 7 (43.7%) Smoking status 41 (35.7%) 5 (31.3%) Hormonal contraception 76 (66.1%) 10 (62.5%) Menopausal status 15 (13%) 3 (18.8%) HIV infection 0 (0%) 0 (0%) HPV test prior excision (HC2) Negative 5 (4.3%) 1 (6.2%) Positive 110 (95.7%) 15 (93.8%) HPV load (prior excision) Negative 5 (4.3%) 1 (6.2%) <1 RLU 0 (0%) 0 (0%) 1to<2 RLU 1 (0.9%) 0 (0%) 2to<4 RLU 3 (2.6%) 0 (0%) >4 RLU 106 (92.2%) 15 (93.8%) Histopathological diagnosis of the cone 1 CIN 1 3 (2.6%) 0 (0%) CIN 2/3 112 (97.4%) 16 (100%) Microinvasive carcinoma 0 (0%) 0 (0%) Surgical margins Negative 86 (74.8%) 11 (68.8%) Positive (ectocervical) 16 (13.9%) 2 (12.5%) Positive (endocervical) 8 (7%) 2 (12.5%) Double positive 5 (4.3%) 1 (6.2%) Size excised (mean diameter of the base [ecto] 3 mean diameter of the apex [endo] 3 the height) 2.1 (ecto) (endo) (h) cm 1.9 (ecto) (endo) (h) cm p16ink4 Staining (cone) Negative or patchy 0 (0%) 0 (0%) Diffuse basal 8 (7%) 2 (12.5%) Diffuse full 107 (93%) 14 (87.5%) HPV test at first follow-up visit (AMPLICOR) Negative 63 (54.8%) 4 (25%) Positive 52 (45.2%) 12 (75%) Optical density (A450) (at first follow-up visit) Negative (<0.2) 63 (54.8%) 4 (25%) 0.2 to <1 11 (9.6%) 1 (6.3%) 1to<2 17 (14.8%) 1 (6.3%) 2to<4 24 (20.8%) 9 (56.1%) >4 0 (0%) 1 (6.3%) Visits post conization 4 (2 6) 4 (3 6) Follow-up (weeks) 90.5 (21 184) 74.9 (35 123)

4 1046 Impact of SCJ excision on recurrent lesions Table 2. Viral characteristics of post-treatment CIN Viral characteristics CIN1 (n 5 12) (75%) CIN2/3 (n 5 4) (25%) p-value Single infections 8 (66.7%) 2 (50%) Multiple infections 4 (33.3%) 2 (50%) HPV genotypes High-risk types HPV 16 1 (8.3%) 4 (100%) HPV 18 1 (8.3%) 1 (25%) 0.45 Others 6 (50%) 0 (0%) Possible high-risk types HPV 26 0 (0%) 0 (0%) HPV 53 1 (8.3%) 0 (0%) 1 HPV 66 2 (16.7%) 1 (25%) 1 Low-risk types HPV 6 1 (8.3%) 0 (0%) 1 HPV 11 1 (8.3%) 0 (0%) 1 Others 3 (25%) 2 (50%) test, the primers are designed to amplify HPV DNA from the same 13 high-risk HPV types detected by the HC2 assay. After PCR, hybridization and optical density measurement (A 450 ), the samples were considered HPV positive when the absorbance reading was higher than 0.2. The amplification of a region within the beta-globin gene was used as control. The sensitivity and specificity of AMPLICOR are similar to that of HC2 test. 11 HPV genotyping The Linear Array HPV genotyping assay (Roche, Basel, Switzerland) was used to simultaneously detect and genotype 37 different high- and low-risk HPV types in residual/recurrent/ new CIN and corresponding pretreatment samples. To confirm both the data obtained using this latter assay and to detect potential additional HPV types (HPV 43 and 44), the PapilloCheck HPV-screening test (Greiner bio-one, Wemmel, Belgium) was also used. A region within the beta-globin (Linear Array) and ADAT1 (PapilloCheck) genes was simultaneously amplified in the same reaction as a control to assess cell adequacy, DNA extraction and quality. After PCR amplification, hybridization and detection were performed according to the manufacturer s instructions. Immunostaining Immunohistochemical analysis was performed as described previously. 8,12 For the primary reaction, the following antibodies were used: anti-p16 ink4 (Santa Cruz Biotechnology, Santa Cruz, CA), anti-cytokeratin 7 (Thermo Scientific, Waltham, MA), anti-agr2 (Proteintech, Chicago, IL), anti-gda (Sigma-Aldrich, Saint-Louis, MO) and anti-mmp7 (R&D systems, Minneapolis, MN). The four latter are SCJ-specific proteins. Immunoperoxidase staining was performed using the Envision kit (Dako, Glostrup, Denmark) or the Vectastain ABC kit (Vector Laboratories, Burlingame, CA) according to the supplier s recommendations. Positive cells were visualized using a 3,3 0 -diaminobenzidine substrate and the sections were counterstained with Mayer s hematoxylin. A control IgG was used as negative control (Santa Cruz Biotechnology, Santa Cruz, CA). Immunostaining assessment Two independent pathologists evaluated the p16 ink4 -immunolabeled tissues by using a semi-quantitative score of the intensity and extent of the staining. Thus, scoring of p16 ink4 included both nuclear and cytoplasmic staining and was graded as 0 (negative), 1 (rare dispersed positive cells), 2 (continuous over one-third of the epithelial ) and 3 (strong and diffuse staining, uniform from basal layer to epithelial surface). Regarding SCJ-specific markers, AGR2, GDA and MMP7 were classified as positive when diffuse cytoplasmic immunoreactivity of the entire preneoplastic epithelium was observed. Cytokeratin 7 was either expressed by the suprabasal and/or apical cell layers (positive staining) or were not expressed (negative). These scoring systems were described previously. 10 Criteria for residual, recurrent, new (pre)neoplastic lesions Women who had two or more consecutive negative postoperative Pap smears and/or biopsies were considered as being free of residual lesions irrespective of the results of the HPV test (AMPLICOR). Residual/recurrent/new CIN of any grade (1, 2 or 3) was diagnosed on the basis of a colposcopydirected biopsy at the follow-up evaluation. Women, in whom CIN (HPV genotype(s) and expression status for p16 ink4 and SCJ markers similar to the initial lesion) was found at the first visit after excision, were defined as having residual disease. If, after one or more negative postoperative Pap smear, another CIN was diagnosed, we assumed that it was a recurrent lesion when both HPV tests were positive at the first and repeated follow-up visit(s) and similar HPV genotypes were detected in pre- and postexcision lesions. New lesions on the cervix were characterized by a negative HPV test result at the first follow-up visit. Both different HPV genotypes and different expression status for p16 ink4 and SCJ markers compared to the excised specimen also further confirmed this diagnosis. However, the negativity of the HPV test was the main feature to define a newly posttreatment acquired lesion. In case of positive histology, the decision whether to perform a second excision was based on the severity (CIN 2/3) of the preneoplastic lesion. Statistical analysis Multivariate analysis was performed to identify any factor that could affect the risk of residual/recurrent/new preneoplastic lesion after electrosurgical cone resection of the uterine cervix. The viral characteristics of post-treatment CIN

5 Herfs et al Table 3. Characteristics of cervical cone biopsies and associated residual/recurrent/new lesions Follow-up Patient Number Indication to excision Cytology/ biopsy margins Excision specimen First visit Residual/recurrent/new disease HPV SCJ genotype(s) p16 ink4 biomarkers Weeks HPV test Weeks Cytology/ SCJ biopsy Genotype(s) p16 ink4 biomarkers Treatment/ repeated visits Group 1 1 HSIL/CIN 3 Positive 16, 18, 54, 66, 70 Diffuse full Positive 18 Positive 18 ASC-US/CIN 3 16, 18, 66, 70 Diffuse full Positive Reconization 2 HSIL/CIN 3 Positive 16 Diffuse full Positive 12 Positive 12 LSIL/CIN 3 16 Diffuse full Positive Reconization 3 HSIL/CIN 3 Positive 16 Diffuse basal Negative 14 Positive 14 LSIL/CIN 2 16 Diffuse basal Negative Reconization 4 HSIL/CIN 3 Positive 16, 62 Diffuse full Positive 13 Positive 13 HSIL/CIN 2 16, 62 Diffuse full Positive Reconization Group 2 5 HSIL/CIN 3 Negative 51 Diffuse full Positive 14 Positive 101 Normal/CIN 1 51, 82 Diffuse basal Negative Persistence (Week 123) 6 HSIL/CIN 3 Positive 58 Diffuse full Positive 12 Positive 23 LSIL/CIN 1 58 Patchy Negative Regression (Week 36) 7 HSIL/CIN 2 Negative 16 Diffuse full Positive 7 Positive 7 LSIL/CIN 1 16 Diffuse basal Negative Regression (Week 40) 8 HSIL/CIN 3 Negative 31, 70 Diffuse full Positive 12 Positive 26 LSIL/CIN 1 31 Diffuse basal Negative Persistence (Week 95) 9 HSIL/CIN 3 Negative 18 Diffuse full Positive 25 Positive 35 LSIL/CIN 1 18, 59 Diffuse full Negative Regression (Week 80) 10 HSIL/CIN 2 Negative 33, 53, 61 Diffuse full Positive 13 Positive 13 Normal/CIN 1 53, 61 Patchy Negative Regression (Week 32) 11 HSIL/CIN 2 Negative 6, 16, 58 Diffuse full Positive 11 Positive 28 Normal/CIN 1 6 Patchy Negative Regression (Week 104) 12 LSIL/CIN 3 Negative 6, 54, 58, 66, 84 Diffuse full Positive 12 Positive 26 LSIL/CIN 1 58, 66, 84 Diffuse basal Negative Regression (Week 53) Group 3 13 HSIL/CIN 2 Negative 39 Diffuse full Positive 13 Negative 96 LSIL/CIN 1 11 Negative Negative Unknown 14 HSIL/CIN 2 Negative 16 Diffuse full Positive 13 Negative 28 LSIL/CIN 1 84 Patchy Negative Regression (Week 52) 15 HSIL/CIN 2 Negative 31 Diffuse basal Positive 28 Negative 48 Normal/CIN1 58 Diffuse basal Negative Regression (Week 64) 16 ASC-US/CIN 3 Negative 16 Diffuse full Positive 14 Negative 29 LSIL/CIN 1 66 Patchy Negative Regression (Week 68)

6 1048 Impact of SCJ excision on recurrent lesions were also compared. Statistical analysis (v 2 statistical analysis or Fisher s exact test) was performed with the Instat 3 software (Graph-Pad Software, San Diego, CA) and the differences were considered statistically significant when p-values were <0.05. Results Evaluable patient population During the study period, 194 patients with histologically confirmed CIN underwent an excision procedure. From this group, three (1.5%) women underwent radical hysterectomy for non-hpv-related uterine disease weeks after index excision and were, therefore, excluded from the study. Sixty (30.9%) patients were lost because they underwent follow-up examinations in different hospitals and no data were available. The remaining 131 women were enrolled in our study and evaluated for statistical analysis. Clinical characteristics of post-treatment CIN After a median follow-up of 89.5 weeks (range, weeks), postexcision cervical lesions were identified in 16 patients (12.2%) on the basis of a colposcopy-directed biopsy. HPV16 was the most common HPV type found in posttreatment CIN. The characteristics of the two groups (no recurrence vs. recurrence) are summarized in Table 1. No statistical difference was observed between the two groups regarding age (p ), smoking status (p ) and hormonal contraception (p ). In total, 34 (25.9%) cone specimens displayed positive surgical margins. The ectocervical margin was involved in 18 (52.9%), the endocervical margin in 10 (29.4%) and both margins in 6 cases (17.7%). There was no significant difference in risk of residual/recurrent disease based on the margin status (p ). In contrast, a positive HPV test (AMPLICOR) at the first follow-up visit was significantly associated with a CIN outcome (p ). HPV status, histology and immunohistochemistry of posttreatment CIN A positive HPV test and an abnormal cytology (ASC-US or higher) at first visit (3 months) after excision was observed in 64 (48.8%) and 14 (10.7%) women, respectively. Six months after treatment, 32 out of 131 (24.4%) still showed persisting HPV infection with high-risk HPV. Overall, 78 (59.5%) and 26 (19.8%) patients had at least one positive HPV test and/or abnormal cytology during the follow-up period, respectively. As mentioned above, biopsy-proven CIN was found in 16 women post-treatment (12.2%); four (25%) with CIN 2/3 and 12 (75%) with CIN 1 (Tables 2 and 3). No squamous cell carcinoma or adenocarcinoma was diagnosed during the follow-up period. All CINs were HPV DNA positive. Postexcision HPV (AMPLICOR and genotyping) tests and p16 ink4 immunoreactivity (patchy, basal or diffuse full ) were positive in 16 (100%) and 15 (93.8%) postexcision CINs, respectively (Table 3). SCJ biomarker expression was also evaluated and we observed that 15 (93.8%) cone specimens and 3 (18.8%) post-treatment high-grade lesions stained positive for SCJ-specific proteins (keratin 7, AGR2, MMP7 and GDA). The immunostaining results are shown in Figures 1 and 2. Table 3 summarizes the postexcision CINs in detail. Using the parameters of SCJ status (SCJ(1) or SCJ (2)) and HPV type, three categories of postexcision CIN were identified. All but one initial lesion was SCJ(1). Group 1 consisted of four CINs that were identical in SCJ marker status and high-risk HPV type to the initial lesions. All were CIN2 or CIN3, infected with HPV 16 and included the one SCJ(2) CIN. Group 2 consisted of eight CINs that shared the same highrisk HPV type as the initial but were all SCJ(2). All eight were classified as CIN 1 in contrast to the initial diagnosis of CIN2/3. Group 3 consisted of four CINs that were SCJ(2) and contained HPV types different from the initial lesion. Compared to both Groups 2 and 3, HPV 16 infection was significantly more observed in Group 1 (p ) (Table 2). The differences in frequency of postexcision CIN 1 between Group 1 and Group 2 or Groups 2 and 3 were significant at p and p < As summarized and shown in Table 3 and Figure 1, cases in Group 1, which were the only postexcision CIN 2/3s, were found at the first follow-up visit between 12 and 18 weeks postexcision, and HPV genotype(s), p16 expression pattern and SCJ biomarker status matched the initial CINs. In addition, the excision specimen margins in all four were positive (for CIN) in contrast to one of the 12 in Groups 2 and 3 (p ). The CINs in Group 1 were thus considered to most likely signify residual disease and these patients underwent a second excision procedure. CINs in Groups 2 and 3 were mostly not appreciated on the initial postexcision visit and were discovered later in the follow-up period, with a median time between excision and diagnosis of CIN of 38.3 weeks (range, weeks). Given the contrast in histology and SCJ biomarker status between these CIN and the initial lesions combined with the prior negative excision margins and later discovery, they were presumed to have developed de novo after excision, either from latent infection of one of the initial HPV(s) (Group 2) or from a new HPV infection (Group 3). Representative examples are shown in Figure 2. Out of the 12 recurrent/new CIN 1 diagnosed in our study, follow-up information was available for 11 (91.7%) cases (Table 3). None of the recurrent/new lesions was followed by a CIN 2/3 diagnosis (on cytology or histology) with an overall spontaneous regression rate of 81.8% for more than a 1.5- year window. Discussion In a population of women with CIN 2/3, success in preventing cancer on follow-up depends on completely removing the abnormality and ensuring that new CIN do not develop in the follow-up period that might pose a threat by progressing to malignancy. Both goals are achieved through regular

7 Herfs et al Figure 1. Residual high-grade CIN (right panel) and associated cone biopsy (left panel) stained for p16 ink4 and the SCJ-specific markers (krt 7, AGR2, MMP7 and GDA). The similar HPV genotype and the uniform staining displayed by these lesions should be noted. follow-up to exclude a persisting or new CIN. 13,14 Discriminating between residual and new disease in the cervix can be problematic. In our study, multiple parameters were evaluated to tease out residual from recurrent or new disease, including margin status at excision, differences in lesion grade between initial and postexcision CIN, timing of the recurrence, HPV types detected, location of the postexcision CIN and SCJ marker status. What emerged was evidence that some postexcision CINs are identical in all respects to the initial lesion and likely represent residual (unexcised) disease (Group 1). All these latter lesions diagnosed in our study were infected with HPV16. Interestingly, a previous study showed that women treated for HPV16-associated CIN 3 had an increased risk of post-treatment (residual/recurrent) CIN3 and should, therefore, be monitored more closely. 15 Our data as well as the observation that SCJ cells appear to be preferentially infected with HPV 16, suggesting that the HPV genotype should be included in post-treatment monitoring strategies. Besides these unexcised lesions, many CIN developing postexcision, including those that share the same HPV type, exhibit characteristics that strongly suggest a new site of origin (Groups 2 and 3). In this analysis (Table 3) consistent features distinguishing the latter two groups from the first were the absence of SCJ marker expression and a low-grade histopathology. Recent studies from our laboratory and others surmised that cells sharing the SCJ immunophenotype were responsible for more than 90% of CIN 2/3. 7,10,16 Lesions lacking SCJ markers were usually CIN 1 and the absence of SCJ marker expression suggested that they arose either in the mature metaplastic epithelium of the cervix or in the ectocervix. In a small retrospective study of 54 cases of SCJ(2) CIN 1, none developed evidence of CIN 2/3, either on cytology or in follow-up biopsy. The conclusion was that SCJ negative CIN are (i) more often CIN 1 and (ii) less likely than SCJ(1) CIN1 to be associated with a CIN 2/3 outcome on followup. 10 The implication was that the SCJ cell was more vulnerable to developing CIN 2/3 after carcinogenic HPV infection

8 1050 Impact of SCJ excision on recurrent lesions Figure 2. Representative examples of recurrent/new low-grade lesions analyzed in our study. These postexcision HPV-related lesions were exclusively located in the ectocervical squamous epithelium. The absence of staining for SCJ-specific biomarkers should be noted. in contrast to epithelium with clear-cut squamous differentiation, such as metaplastic or ectocervical epithelium. This disparity in risk of CIN 2/3 between the SCJ and the mature metaplastic epithelium is mirrored in the significant differences in cancer risk between the cervix and the lower genital tract vaginal or vulvar mucosa, both of which are SCJ marker negative. 7 In a previous study, we found that after excisional procedures (LEEP or cone biopsy) the new squamocolumnar interface was usually devoid of SCJ cells. 7 We thus concluded that the SCJ cells were not being spontaneously regenerated after the removal. Thus, we hypothesized that postexcision, new lesions developing on the cervix would preferentially be found on the ectocervix or mature metaplastic epithelium. In addition, given the higher predilection of the metaplastic/ ectocervical epithelium for CINs of lower histologic grade, we hypothesized that many postexcision lesions would be CIN In our study, it would not be unexpected that new lesions associated with newly acquired HPV types the four cases in Group 3 would be of low grade. The unexpected finding was that half of the postexcision CIN (Group 2) occurring after negative excision margins exhibited the same carcinogenic HPVs as the initial CIN 2/3 but were lower grade (CIN 1). Most regressed on follow-up without further treatment. Based on the cone margins, timing of lesion development, histology and SCJ marker status, all of the CINs in Groups 2 and 3 fulfilled the requirements for new CIN. Although it could be argued that defining a recurrence as new, based on the absence of SCJ markers, is self-fulfilling, none of the 12 cases physically involved the SCJ or endocervix. Although most CIN 2/3 are diagnosed within 2 years after treatment, 17,18 we cannot ascertain that the recurrent/new lesions are always going to be CIN 1. Long-term follow-up studies indicated that recurrent CIN 2/3 are sometimes detected many years after ablation in women with HPV

9 Herfs et al persistency. 19 Such lesions could emerge from metaplastic or ectocervical epithelium or signify small residual lesions that went undetected at the initial follow-up visit (after 3 months); hence, a negative initial follow-up cytology would not ensure that the discovered lesion is SCJ(2). It is interesting, however, that the risk of recurrence diminished with the certainty that a persistent lesion was not present, dropping to zero risk of new CIN 3 for those with both negative cytology and negative HPV testing. 19 Another limitation of our study is the low number of high-grade lesions diagnosed during the follow-up period. Post-treatment CIN 2/3 rate was approximately 3% in our study which is lower than that seen in the study by Kocken et al. 19 and could influence interpretation of the success of this procedure in preventing new CIN 2/3 lesions. In addition, the young age of our population might enrich for success. Indeed, Strander et al. 20 recently showed that women aged >60 years, previously treated for CIN 3, are at increased risk of developing cervical/vaginal (pre)cancer. Finally, our cohort did not include patients with immunodeficiency. Although an increased frequency of post-treatment CIN was reported in immunodeficient patients, the nature of recurrent/new lesions in HIV-positive women remains unknown. Conclusions Our study offers a compelling argument that the removal of the SCJ could play an important role in determining the nature of postexcision CIN. It also begs the question of whether pre-emptive ablation or excision of the SCJ would reduce the risk of subsequent high-grade CIN and cancer. Addressing this question is not simple, given that as many as 67% of CIN 2 might resolve spontaneously in young women, a fact that has prompted more conservative management of CIN2 in this population. 24,25 Still, the frequency of new highgrade CIN postexcision is surprisingly low, given the HPV References exposure rate in the population and the multiplicity of HPVs that can be recovered from the genital tract of a single individual (Tables 2 and 3). It is noteworthy that one study reported a significantly lower rate of cervical HPV infection in women treated with prophylactic cryoablation. 26 Anecdotal reports have also suggested a cancer protective effect from electrocautery of the cervical os post partum. 27 Randomized controlled trials would be needed to determine the accessibility of the SCJ to less traumatic prophylactic ablation techniques (such as cryotherapy), estimate the degree to which prophylactic SCJ ablation protects an individual from developing CIN 2/3 and test its role in protecting the more cancer-prone women (e.g., HIV-infected/multipartner women). However, given the number of women in need of protection who may not be served by prophylactic vaccines, this concept merits further study. Acknowledgements Our study was supported by the Belgian Fund for Medical Scientific Research, by the Centre Anti-Cancereux près l Universite de Liège, by the Faculty of Medicine of the University of Liège, by a grant from the National Cancer Institute (1R21CA to C.P.C.) and by the Fonds Leon Fredericq. M.H. and M.S.C. are Postdoctoral Researcher and Research fellow of the Belgian National Fund for Scientific Research, respectively. The authors also thank Dr. Wa Xian and Dr. Frank McKeon from the Jackson Laboratory for Genomic Medicine in Farmington Connecticut (USA) for helpful discussions. Author Contributions M.H., C.P.C. and P.D. designed the study. M.H. and J.S. performed the experiments. J.D., F.G. and F.K. collected the tissue samples. B.H., C.P.C. and P.D. reviewed all cases. M.H., J.S., M.S.-C., G.K., C.P.C. and P.D. interpreted the data. M.H. generated the figures. M.H. and C.P.C. wrote the manuscript. All authors had final approval of the submitted manuscript. 1. Crum CP, Ikenberg H, Richart RM, et al. Human papillomavirus type 16 and early cervical neoplasia. N Engl J Med 1984;310: Durst M, Gissmann L, Ikenberg H, et al. A papillomavirus DNA from a cervical carcinoma and its prevalence in cancer biopsy samples from different geographic regions. Proc Natl Acad Sci USA 1983;80: Lehtinen M, Paavonen J, Wheeler CM, et al. Overall efficacy of HPV-16/18 AS04-adjuvanted vaccine against grade 3 or greater cervical intraepithelial neoplasia: 4-year end-of-study analysis of the randomised, double-blind PATRICIA trial. Lancet Oncol 2012;13: Paavonen J, Naud P, Salmeron J, et al. Efficacy of human papillomavirus (HPV)216/18 AS04- adjuvanted vaccine against cervical infection and precancer caused by oncogenic HPV types (PAT- RICIA): final analysis of a double-blind, randomised study in young women. Lancet 2009;374: Milojkovic M. Residual and recurrent lesions after conization for cervical intraepithelial neoplasia grade 3. Int J Gynaecol Obstet 2002;76: Skjeldestad FE, Hagen B, Lie AK, et al. Residual and recurrent disease after laser conization for cervical intraepithelial neoplasia. Obstet Gynecol 1997;90: Herfs M, Yamamoto Y, Laury A, et al. A discrete population of squamocolumnar junction cells implicated in the pathogenesis of cervical cancer. Proc Natl Acad Sci USA 2012;109: Herfs M, Vargas SO, Yamamoto Y, et al. A novel blueprint for top down differentiation defines the cervical squamocolumnar junction during development, reproductive life, and neoplasia. J Pathol 2013;229: Wang X, Ouyang H, Yamamoto Y, et al. Residual embryonic cells as precursors of a Barrett s-like metaplasia. Cell 2011;145: Herfs M, Parra-Herran C, Howitt BE, et al. Cervical squamocolumnar junction-specific markers define distinct, clinically relevant subsets of lowgrade squamous intraepithelial lesions. Am J Surg Pathol 2013;37: Mo LZ, Monnier-Benoit S, Kantelip B, et al. Comparison of AMPLICOR and Hybrid Capture II assays for high risk HPV detection in normal and abnormal liquid-based cytology: use of INNO-LiPA Genotyping assay to screen the discordant results. J Clin Virol 2008;41: Herfs M, Hubert P, Poirrier AL, et al. Proinflammatory cytokines induce bronchial hyperplasia and squamous metaplasia in smokers: implications for chronic obstructive pulmonary disease therapy. Am J Respir Cell Mol Biol 2012;47: Wright TC Jr, Massad LS, Dunton CJ, et al Consensus guidelines for the management of women with cervical intraepithelial neoplasia or adenocarcinoma in situ. Am J Obstet Gynecol 2007;197: Wright TC Jr, Massad LS, Dunton CJ, et al Consensus guidelines for the management of women with abnormal cervical cancer screening tests. Am J Obstet Gynecol 2007;197: G ok M, Coupe VM, Berkhof J, et al. HPV16 and increased risk of recurrence after treatment for CIN. Gynecol Oncol 2007;104: Van der Marel J, Van Baars R, Alonso I, et al. Oncogenic human papillomavirus-infected immature metaplastic cells and cervical neoplasia. Am J Surg Pathol 2014;38:470 9.

10 1052 Impact of SCJ excision on recurrent lesions 17. Prato B, Ghelardi A, Gadducci A, et al. Correlation of recurrence rates and times with posttreatment human papillomavirus status in patients treated with loop electrosurgical excision procedure conization for cervical squamous intraepithelial lesions. Int J Gynecol Cancer 2008;18: Bais AG, Eijkemans MJ, Rebolj M, et al. Posttreatment CIN: randomized clinical trial using hrhpv testing for prediction of residual/recurrent disease. Int J Cancer 2009;124: Kocken M, Helmerhorst TJ, Berkhof J, et al. Risk of recurrent high-grade cervical intraepithelial neoplasia after successful treatment: a long-term multi-cohort study. Lancet Oncol 2011;12: Strander B, H allgren J, Sparen P. Effect of ageing on cervical or vaginal cancer in Swedish women previously treated for cervical intraepithelial neoplasia grade 3: population based cohort study of long term incidence and mortality. Br Med J 2014;348:f Petry KU, Scheffel D, Bode U, et al. Cellular immunodeficiency enhances the progression of human papillomavirus-associated cervical lesions. Int J Cancer 1994;57: Lima MI, Tafuri A, Araujo AC, et al. Cervical Intraepithelial neoplasia recurrence after conization in HIV-positive and HIV-negative women. Int J Gynaecol Obstet 2009;104: Lodi CT, Michelin MA, Lima MI, et al. Factors associated with recurrence of cervical intraepithelial neoplasia after conization in HIV-infected and noninfected women. Arch Gynecol Obstet 2011;284: McAllum B, Sykes PH, Sadler L, et al. Is the treatment of CIN 2 always necessary in women under 25 years old? Am J Obstet Gynecol 2011; 205: Moscicki AB, Ma Y, Wibbelsman C, et al. Rate of and risks for regression of cervical intraepithelial neoplasia 2 in adolescents and young women. Obstet Gynecol 2010;116: Taylor S, Wang C, Wright TC, et al. Reduced acquisition and reactivation of human papillomavirus infections among older women treated with cryotherapy: results from a randomized trial in South Africa. Biomed Chromatogr Med 2010;8: Younge PA. Cancer of the uterine cervix; a preventable disease. Obstet Gynecol 1957;10: