Recurrent high-grade cervical lesion after primary conization is associated with persistent human papillomavirus infection in Norway

Gynecologic Oncology 133 (2014) 159–166

Contents lists available at ScienceDirect

Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno

Recurrent high-grade cervical lesion after primary conization is associated with persistent human papillomavirus infection in Norway☆,☆☆ O.K. Vintermyr a,b,⁎, O. Iversen c,d, S. Thoresen e, W. Quint f, A. Molijn f, S. de Souza g, D. Rosillon h, K. Holl h a

Department of Pathology, Haukeland University Hospital, Norway The Gade Laboratory for Pathology, Department of Clinical Medicine, University of Bergen, Bergen, Norway c Institute of Clinical Medicine, University of Bergen, Bergen, Norway d Women's Clinic, Haukeland University Hospital, Bergen, Norway e Abbvie AS, Norway f DDL Diagnostic Laboratory, Rijswijk, The Netherlands g 4Clinics, Paris, France h GlaxoSmithKline Vaccines, Wavre, Belgium b

H I G H L I G H T S • HPV-types in primary and recurrent cervical cone CIN2+ lesions were evaluated. • 94.8% of recurrent CIN2+ lesions had evidence of persistent HPV infection. • 70.9% of these persistent HPV lesions were associated with HPV types 16/18.

a r t i c l e

i n f o

Article history: Received 27 November 2013 Accepted 1 March 2014 Available online 11 March 2014 Keywords: Cervical intraepithelial neoplasia Conization High-risk HPV Human papillomavirus Recurrence

a b s t r a c t Objective. This retrospective registry-based study aimed to assess the human papillomavirus (HPV)-type distribution in primary and recurrent high-grade cervical intraepithelial neoplasia (CIN2+), and to discriminate pre-existing from newly-acquired infections. Methods. Cervical specimens from 58 women (median age (Q1–Q3): 37.6 (31.7–44.9)) who underwent primary (1998–2003) and repeat conizations were confirmed as CIN2+ during expert pathology review. HPV testing was performed using PCR MP-TS123 Luminex for 16 HPV types. Molecular HPV16 E6 and HPV18 LCR DNA sequencing was performed on specimens with persistent HPV16/18. Results. All 58 paired cones were HPV positive; 49 had CIN3+ in the primary cone. Forty-seven (95.9%) women with primary CIN3+ and recurrent CIN2+ had persistent high-risk (hr) HPV infection, of which 74.5% were HPV16/18. Two women had probable newly-acquired HPV16/52/56 and HPV39 infections. One woman with persistent HPV52 also had a probable new HPV16 E6 variant in the recurrent CIN2+. Median time delay (Q1–Q3) between conizations was 2.0 years (1.1–4.0), being shorter for women older than 40 years: 2.6 years (1.1–3.7) than for women younger than 40 years: 6.0 years (2.0–8.7). Primary conization histology revealed CIN3, cervical adenocarcinoma in situ and microinvasive carcinomas in 43 (87.8%), 5 (10.2%) and 1 (2.0%) women, respectively. Primary HPV16- and HPV18-infected CIN3+ had a shorter delay between conizations: 1.8 years (1.2–4.4) and 2.2 years (0.4–NE), respectively, compared to HPV33-: 3.8 years (3.3–7.8) or other HPV type-infected: 8.2 years (6.0–NE) CIN3+. Conclusions. Routine post-conization hr-HPV DNA testing together with cervical cytology may provide a better prediction for potential recurrent disease. Further, primary prevention through adolescent vaccination may prevent CIN2+ and its recurrence. © 2014 Elsevier Inc. All rights reserved.

☆ Previous congress activities, abstract/poster IPV 2012 San Juan, Puerto Rico. ☆☆ Clinical Trial Registration: NCT00924794 (www.clinicaltrials.gov). ⁎ Corresponding author at: Department of Pathology, Haukeland University Hospital, N-5021 Bergen, Norway. Fax: +47 55973158. E-mail address: [email protected] (O.K. Vintermyr).

http://dx.doi.org/10.1016/j.ygyno.2014.03.004 0090-8258/© 2014 Elsevier Inc. All rights reserved.

Introduction Invasive cervical cancer (ICC) is preceded by high-grade cervical intraepithelial neoplasia (CIN2+) lesions (i.e. CIN2/3 or adenocarcinoma in situ (AIS)) [1–3] which are effectively treated by conization or

160

O.K. Vintermyr et al. / Gynecologic Oncology 133 (2014) 159–166

excision [4,5]. However, between 5 and 20% of women treated for CIN2+ with conization will develop recurrent disease within 3 years [6–12]. The main reasons for CIN2+ recurrence include residual disease with the incomplete removal of the primary lesion (i.e. treatment failure); persistent high-risk (hr)-human papillomavirus (HPV) infection [10,13]; or re-infection from a sexual partner or newly-acquired HPV infection [14]. There is evidence that the incomplete removal of the primary lesion and the on-going presence of persistent HPV infection leads to the need for repeat conization in 18% and 45% of cases, respectively [6,13,15,16]. However, there are limited data on the HPV type distribution in primary and recurrent CIN2+ and the time to recurrence of CIN2+ for individual HPV types. Disease-free margins in the cone specimens and a post-treatment negative hr-HPV test significantly decrease the risk of recurrence [10, 13,17,18]. A follow-up protocol by cytology and HPV testing to monitor for residual and recurrent disease is therefore recommended in some countries [12,19,20]. HPV vaccines have demonstrated a significant reduction in cervical HPV infections and precancerous lesions in clinical trials when administered to adolescents and young adults [21,22]. Previous vaccination with the quadrivalent HPV vaccine among women surgically-treated for HPV-related cervical disease significantly reduced the incidence of subsequent HPV-related disease, including high-grade disease [23]. In order to assess the role of type-specific HPV infection in recurrent CIN2+, we conducted a retrospective registry-based study in Norway to assess the distribution of HPV types in primary CIN3+ and recurrent CIN2+ post-cervical conization, discriminating persistent from newlyacquired HPV infections. Methodology Study design and study population The Cancer Registry of Norway (CRN) is one of the world's oldest national cancer registries, established in 1951. The activity of CRN is regulated by the Norwegian legislation, with compulsory registration of all cancer and pre-cancer cases. Notifications, by healthcare professional, are made using standardised reporting forms. The multiple reporting practice provides a very close to complete database. Each resident in Norway has a unique personal identification number (PIN) and overlapping reports on the same individual are identified by that number and duplicate cases are eliminated. Each year, approximately 140,000 notifications are received, of which 30,000 are newly diagnosed cancer cases [24]. CRN collects cervical precancer and cancer data from the Norwegian cervical cancer-screening programme offered to all women aged 25–69 years, as well as results of all women aged 18–24 years whom have had a conization/cervical biopsy. Recent evaluations indicate the CRN is close to complete when it comes to evaluating cervical cancer cases and their paraffin samples, partly through compulsory reporting practices and the use of multiple sources of information [25]. Out of 594 women registered in the CRN with at least one repeat conization between 1998 and 2003 (Table 1), 410 women, aged ≥18 years, with primary cone specimen treated for CIN2+ (including AIS) or microinvasive cervical carcinomas (MIC) were selected for this study (Fig. 1). Recurrent cones with CIN2+/MIC/ICC, obtained up until 2009, were identified for all women with primary cones. Of these, 56 women were excluded as they did not meet the inclusion criteria. The remaining 354 women, with available written informed consent and adequately preserved (formalin-fixed and paraffin-embedded (FFPE)) primary and repeat cone samples in designated local laboratories, were enrolled. To improve the integrity and consistency of histological diagnoses, conization pair samples from 175 women underwent centralised histological review by a central laboratory (DDL Diagnostic Laboratory, Rijswijk, The Netherlands) in the study timeframe; 110 (62.9%)

women had confirmed CIN2+/MIC diagnoses in the primary cone and confirmed CIN2+/MIC/ICC in the repeat cone. Samples were excluded from review if they held highly irregular tissue, were in poor condition or too thin, or on which histological examination of Haematoxylin and Eosin (H&E) sections before and after block(s) for HPV analysis did not confirm CIN2+. Among the 110 eligible women, two samples of women were selected for HPV DNA testing and a more detailed initial analysis of their pair cone specimens (Fig. 1). The first 29 women (Sample I) whose samples had been sent to and confirmed by the central laboratory with CIN2+ in both cones were enrolled into the study. The recruitment of Sample I was terminated at 29 subjects when it was determined that six women had a very narrow time interval (b 6 months) between conizations. It was therefore decided to include two additional selection criteria: 1) at least 6 months between primary and repeat conizations; and 2) at least one normal pap smear result registered between these two conizations. With the adoption of these additional selection criteria, only 29 out of the remaining 81 women were eligible for selection (Sample II). At least a 6-month delay was decided as the optimum time interval between the primary and repeat conizations so as to reduce the chances of a recurrent CIN2+ being due to incomplete removal of the CIN2+ during primary conization [26]. Cytology status after primary conization was checked for all women with repeat conizations. We have assumed that all surgical procedures and clinical follow-up surveillance have been performed according to standard clinical protocols [26]. No HPV results were registered along for these cytology results. The analysis was conducted on the subset of women with CIN3 + in their primary conisate. This study (CRT: NCT00924794) was approved by the Regional Committee for Ethics in Research, Norway (2009/1607-26) and conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Written informed consent was obtained from all participating women prior to any study procedure on clinical specimens. If the woman had deceased, the cone/biopsy samples were used according to prior approval from the Ethics Committee. Study procedures FFPE tissue samples were sectioned for histopathology review and HPV DNA testing by the central laboratory. All cone samples were cut and stained by H&E. P16 or other immunohistochemical stainings were used, when indicated, to investigate uncertain diagnosis of CIN. Centralised gynaecological histopathologists who examined the H&E sections were blinded for the initial diagnosis provided by the local laboratory. The specific HPV types were determined using polymerase chain reaction (PCR) MP-TS123 Luminex for 16 HPV types (HPV6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66 and 68) [27,28]. The central laboratory performed PCR testing for HPV DNA on two blocks with the most advanced CIN2 + diagnosis from the primary and repeat cones for each woman. A probable persistent HPV infection was defined as type-specific HPV (other than HPV16/18) infection present in both the primary and repeat cones. Persistent infection with HPV16 or HPV18 was confirmed by HPV16 E6 and HPV18 long control region (LCR) DNA sequencing, respectively, using the blocks from primary and repeat conizations from the same woman. Therefore, the complete HPV16 E6 region was amplified using 5 sets of HPV16 E6 specific primers. Furthermore variant regions of the HPV18 LCR were amplified using 5 sets of HPV18 LCR specific primers, the generated HPV18 LCR amplimers were reamplified with T7- and SP6-elongated primers. Amplimers were treated with ExoSAP-IT (Isogen Life Sciences, de Meern, The Netherlands) and analysed with the Big Dye Terminator v1.1 Cycle sequencing kit (Applied Biosystems, Foster City, CA, USA) using the same primers used in the PCR or T7 and SP6 primers, respectively. The ABI 3130 sequencer

O.K. Vintermyr et al. / Gynecologic Oncology 133 (2014) 159–166

161

Table 1 Total number and estimated incidence of primary and repeat conizations performed in Norway, 1999–2003. Source: National Cancer Registry, downloaded Sept 2013. Calendar year

No. of conizations

Total no. of 25–69 yo females, (Norwegian population)

Incidence (n/1000 per year)

No. of 2nd conizations

No. of 3rd conizations

No. of 4th conizations

Total no. of reconizations

% reconization

1999 2000 2001 2002 2003 Total

2528 2450 2334 2685 2739 12,736

1,248,423 1,258,154 1,267,140 1,277,714 1,288,672 6,340,103

2.02 1.95 1.84 2.1 2.13 2.01

87 110 114 105 142 558

4 7 6 5 8 30

0 1 1 2 2 6

91 118 121 112 152 594

3.60% 4.82% 5.18% 4.17% 5.55% 4.66%

(Applied Biosystems, Foster City, CA, USA) was used for the analysis of the sequencing reactions. For each patient, obtained sequences were aligned and investigated for mutation/variation using SeqScape version 2.6.0. Persistent infection HPV16 or HPV18 denotes the same HPV type (including no nucleotide difference for E6 in HPV16 or for LCR in HPV18), in the primary and recurrent CIN2+ lesions. A probable new infection was defined as a HPV type found in the repeat cone that was not identified in the primary cone. For ‘probable’

new infection with HPV16, additional cutting or PCR testing on all blocks from the primary or repeat cone confirmed the HPV type. Statistical methods Subject characteristics at primary and repeat conizations are summarised by descriptive statistics. Time between primary CIN3 + and recurrent CIN2 + associated with a persistent HPV infection was

Fig. 1. Disposition of enrolled women. Key: NDS — non-diagnostic sample (e.g. quality of sample not adequate for pathology diagnosis, inappropriate sample); NEG — negative result; CIN 1/2/3 — cervical intraepithelial neoplasia (grade 1/2/3); ADC — adenocarcinoma of the cervix; HPV — human papillomavirus.

162

O.K. Vintermyr et al. / Gynecologic Oncology 133 (2014) 159–166

further analysed using the Kaplan–Meier curves, as well univariate and multivariable and Cox regression models. In these analyses, the statistical unit was HPV infection, and variance estimates adjusted for the correlation within subjects were obtained using the robust estimation method [29,30]. Crude hazard ratios as well as adjusted hazard ratios for age at primary conization and HPV infection were computed with their 95% confidence intervals (95% CIs). All statistical calculations were performed using SAS 9.2.

respectively) compared to the median ages of HPV31/33 infected women (45.5 or 49.3) (Table 3). The median time between primary CIN3+ and recurrent CIN2+ with persistent HPV infection was shorter for HPV16 (1.8 years) and HPV18 (2.2 years) than for the other HPV types (Table 4; Fig. 2). The age-adjusted risk of persistent HPV infection among recurrent CIN2+ was 6.3 (95% CIs: 1.7–22.6) times higher for HPV18 and 3.8 (1.9–7.7) times higher for HPV16, when compared to the other HPV types (Table 4).

Results

Discussion

Between 1999 and 2003, there were 12,736 cervical conizations performed in women aged 25–69 years in Norway with an overall incidence of conization of approximately 2.01/1000 person-years; 594 (4.7%) of these were repeat conizations (Table 1). These estimations were consistent across different years. Among the 354 women enrolled with a CRN diagnosis of CIN2+ in both primary and repeat cones, cervical intraepithelial neoplasia grade 3 (CIN3) diagnoses (78.8%) were the most frequent (Table 2). The median age of women (Q1–Q3) at the time of primary conization was 37.8 (31.6–45.1) years. Post-primary conization cytology was either negative or missing for 172 (48.6%) and 182 (51.4%) women, respectively. The median time delay (Q1–Q3) between primary and repeat conizations was 1.1 (0.5–3.5) years, with 71.5% of women having a repeat conization within 3 years of the primary conization (Table 2). Fifty-eight paired CIN2+ cone samples were tested for HPV infection and all (100%) were confirmed hr-HPV positive (Table 2). CIN3 was the most common diagnosis (74.1%) in the primary cone. The median age at the time of the primary conization for CIN2 + was 37.6 (31.7–44.9) years (Table 2). The median time interval between primary and recurrent cones was 2.2 (1.2–4.8) years (Sample I: 1.3 years; Sample II: 3.7 years). Post-primary conization cytology was available; 44 (75.9%) women had at least one normal cytology result post-primary conization (Sample I: 15 (51.7%); Sample II: 29 (100%)). Of these 58 women, 49 women had CIN3+ in the primary cone. The median age for these 49 women at the time of the primary conization for CIN3 + was 37.5 (31.7–43.7) years (Table 2). Post-primary conization cytology was available; 36 (73.5%) women had at least one normal cytology result post-primary CIN3+ conization. The prevalence of the most common HPV types in primary CIN3+ and recurrent CIN2+ was similar for HPV16 (67.4% and 65.3%), whereas the prevalence of HPV33 (18.4% and 14.3%), HPV18 (18.4% and 14.3%), and HPV31 (16.3% and 6.1%) tended to be lower in the recurrent CIN2+ (Table 3). In the analysis of the CIN3+ endpoint, HPV18 is less frequent in the recurrent cones compared to the frequency of HPV18 in the CIN2+ endpoint (Suppl. Table 3A). The prevalence of multiple HPV infection tended to be also lower in recurrent CIN2+ (Table 2). Two women with primary CIN3+ had probable newly-acquired HPV16/52/56 and HPV39 infections, respectively, in their recurrent CIN2 +. The time between the primary CIN3 + and the recurrent CIN2 + was: 0.8 years (HPV39) and 5.4 years (HPV16/ 52/56). Cytology was normal for each of these cases post-initial conization. Forty-seven (95.9%) of the 49 women with recurrent CIN2 + had persistent HPV infection, of which 74.5% were HPV16/18 (Table 3). One woman with persistent HPV52 also had a probable new HPV16 E6 variant, in her recurrent CIN2+. The median time (Q1–Q3) between the primary CIN3+ and recurrent CIN2+ with persistent HPV was 2.6 (1.1–3.7) years in women older than 40 years and 6.0 (2.0–8.7) years in women younger than 40 years (p-value b 0.0001) (Table 4). After adjustment for HPV-type, we determined that women older than 40 years harboured a 2.7 times (95% CIs: 1.3–5.3) relative increased risk of persistent infection in their recurrent CIN2+ as compared to women younger than 40 years (Table 4). Women with primary CIN3 + infected with HPV16/18/other HPV were diagnosed at a younger median age (38.1, 37.0 or 30.0 years,

The proportion of women with repeat conization (4.7%) estimated for Norway is in the range of those reported in other studies (3%–20%) [6,10,12]. There was no major clinical difference observed between the CIN2+ and the CIN3+ group of women in this study cohort from Norway. The majority of recurrent CIN2+ in this study was associated with at least one persistent type-specific hr-HPV infection. Acquisition of a new HPV infection following primary conization and resulting in recurrent CIN2+ was very uncommon. A similar result was observed previously in which the majority of recurrent CIN2+ were due to the same type-persistent hr-HPV infection [15,16]. In another smaller study prospectively following women treated for CIN2+, 46% of the persistently HPV-positive women redeveloped CIN, with this recurrence occurring from month 4 to month 10 post-conization [9]. It has previously been suggested that the presence of HPV16, 18, 33 and 45, as well as multiple hr-HPV types in the primary conisate is associated with a higher risk of recurrent CIN [31]. We found that over 75% of the recurrent CIN2+ was associated with HPV16/18. The prevalence of HPV16 was the same in both the primary and recurrent CIN2+ cone diagnoses while the prevalence of HPV18, 31 or 33 was lower in the recurrent CIN2+ cone diagnoses. When HPV31 or 33 were present along with HPV16 or 18 in primary CIN2 + cones, these hr-HPV types were less likely to be present in the recurrent CIN3+ cones. In general, the prevalence of multiple infections was lower in the recurrent CIN2 + cones than in the primary CIN3 + cones, supporting the concept that only one hr-HPV type drives the development of CIN2+ [32]. Furthermore, the shorter time between primary CIN3+ and recurrent CIN2+ lesions driven by HPV16/18 as compared to the other HPV types is compatible with the notion that these hr-HPV types are clinically more aggressive, especially that of HPV18. Also, for hr-HPV16/18-infected lesions, the median time-interval for recurrent CIN2+ lesions to develop was considerably shorter in women older than 40 years than in women younger than 40 years. In our study, the women were followed for ten years after primary conization. Still, 60% of the recurrent CIN2+ lesions developed within three years of primary conization (median: 2.0 years, range: 0.2– 8.7 years). As many previously reported prospective data in this field are hampered by a rather short follow-up period of less than three years [33], our data, and that reported by Melnikow and colleagues, give further credit to the notion that the risk of developing CIN2+ becomes progressively less likely after five years from the primary conization [8]. The fact that a higher proportion of recurrent disease occurred in women older than 40 years may reflect the impact of an ageing immune system [19,34], a positive selection over time towards the persistence of a virus with higher oncogenic risk [11], or a tendency of older women to have a squamous columnar junction deeper inside their cervical canal compared to younger women, thereby hampering the removal of the initial CIN2 +. Our findings were consistent with observations in another study showing that women older than 35 years have a significantly higher risk of persistent HPV infection following conization [35]. There were only 2 women with probable newly acquired hr-HPVrelated recurrent CIN2 +. The time lag between primary and repeat conizations in these women ranged from less than 1 year (1 case) to more than 5 years (1 case). A timeframe of less than 1 year for

O.K. Vintermyr et al. / Gynecologic Oncology 133 (2014) 159–166

163

Table 2 Subject characteristics.

Key: CIN 2/3 — cervical intraepithelial neoplasia (grade 2/3); ADC — adenocarcinoma of the cervix; MIC — microinvasive carcinoma; AIS — adenocarcinoma in situ of the cervix; SCC — squamous cell carcinoma; HPV — human papillomavirus.

developing a de novo CIN2+ lesion is very short, giving some support to speculate that this lesion could have evolved from the pre-existing CIN2+ even in the absence of a negative hr-HPV result that would otherwise having affirmed that possibility. For the other new CIN2+ lesion

the de novo hr-HPV infection is likely to have occurred in the absence of a recurrent hr-HPV type in the second cone. The bulk of our results also shows that the persistent HPV16 and HPV18 infections in the recurrent cone are of the same strain genetically

164

O.K. Vintermyr et al. / Gynecologic Oncology 133 (2014) 159–166

Table 3 HPV type distribution in primary CIN3+ and recurrent CIN2+, median age at primary conization, overall and by HPV type (HPV+ cohort).

Overall HPV HPV16 HPV18 HPV31 HPV33 Other HPV typesb

Primary CIN3+ (N = 49)

Recurrent CIN2+ (N = 49)

HPV prevalence n (%)

HPV prevalence n (%)

Persistent HPV typea n (%)

49 (100) 33 (67.4) 9 (18.4) 8 (16.3) 9 (18.4) 15 (30.6)

49 (100) 32 (65.3) 7 (14.3) 3 (6.1) 7 (14.3) 11 (22.5)

47 (95.9) 30 (63.8) 5 (10.6) 2 (4.2) 6 (12.8) 7 (14.9)

Median age (Q1–Q3) (years) at primary conization of women with persistent HPV infectionb 37.7 (32.3–44.9) 38.1 (33.8–43.7) 37.0 (31.6–37.7) 45.5 (45.1–45.9) 49.3 (34.4–57.7) 30.0 (28.7–33.2)

Key: N = number of women in the HPV+ cohort; n = number of women in given category. a Persistent infection means CIN3+ cases for which at least one specific HPV type is present in both cones. b Among the other HPV types there were 5 (10.2%) cases of HPV52 in primary CIN3+ and 6 (12.2%) cases of HPV52 in recurrent CIN2+.

that caused the primary CIN2+ lesion. Nucleotide variations in the E6 of the HPV16 genome and in the LRC of the HPV18 genome were examined using DNA sequencing. Variations in the nucleotides were identified in five cone pairs of HPV16 strains; no differences were detected in the HPV18 paired genomes (data not shown). There was a complete nucleotide substitution identified in only one HPV16 cone pair and this difference in the DNA sequence was considered consistent with infection by a new HPV16 strain variant (174C in the primary cone and 174T in the repeat cone). We cannot however exclude the possibility that the 174C/T variant was missed in the primary or recurrent cone as all data are based on population sequencing which allows minor variants to be detected only when present at N 10%. The four other HPV16 cone pairs demonstrated a minor nucleotide variation between the strains in the primary cone compared to the strains in the recurrent cones, which involved the shift in prevalence between present variants. These minor DNA changes are interpreted as being mostly consistent with the presence of a sub-clone already being present in the primary CIN2+ lesion and that the variant HPV16 was dominant in the repeat cone. Strengths of this study include the population-based source of data, unique data linkage of cancer registry with hospital databases to retrieve the histological specimens from women with CIN2 +, standardised central histopathological review, and the use of a highly-sensitive MPTS123 Luminex PCR method. Further, the E6-based multiplex typespecific HPV assay has proven its clinical distinction in terms of sensitivity, reproducibility, coverage of many hr-HPV types and with respect to detection of hr-HPV of multiple HPV infected high-grade lesions [27,28]. Molecular sequencing was also performed to discriminate between persistent primary from newly acquired de novo infections employing E6/E7 for HPV16 and LCR for HPV 18 as PCR target for better identification of HPV16 and HPV18 variants, respectively [28]. All these above factors helped to improve and secure the clinicopathological classification of the CIN2+ lesions and to obtain HPV typing with high sensitivity and specificity. Subject characteristics (frequency of histological

diagnoses, age, etc.) of the 58 women included in our study were comparable to those women who were initially selected from CRN between 1998 and 2003, and in being so, are most likely representative of the entire selected population of women with recurrent CIN2 +. Finally, FFPE tissue specimens are a useful source of DNA to identify HPV types in archival material [36]. Of note, the selection criteria for women in Sample II were more stringent than for women in Sample I. This was done to reduce the chance of selecting cases with residual disease due to poor initial clearance. This was ascertained by selecting women with at least one normal cytology between primary and repeat conizations for Sample II. Thus, the median times between conizations were somewhat different between Sample I and Sample II. Despite the additional selection criteria for Sample II, the women in both samples were found to be statistically similar with respect to all other baseline clinicopathological variables (i.e. age, year of collection and histological diagnoses). Therefore all data from the women in the two samples were pooled in this study. This study was not designed to determine whether the recurrence of CIN2+ was caused by incomplete removal of infected tissue (i.e. residual disease), due to a more aggressive persistent HPV infection (i.e. new lesion), or due to an effect of auto-infection (field-effect) or re-infection with the same HPV type from the same sexual partner. Other studies have shown that women testing hr-HPV negative immediately postconization have a high probability of having no recurrent or residual disease [18,37]. Lack of data on margin clearance could hamper our conclusions. That being said, the presence of recurrence or disease persistence is not directly related to the cone margin status [6,38], with the only factor that was associated with CIN2 + recurrence in women with negative margins post-conization being the presence of the same hr-HPV and mainly HPV16 in both primary and repeat cones if controlled for viral load, age, cytology, punch biopsy histology and conization histology [39]. Of note, we have clearly shown that normal cytology between conizations was not indicative of HPV clearance. This is further

Table 4 Median time (Q1 to Q3) to and the risk of detection of persistent HPV infection in recurrent CIN2+ after conization of primary CIN3+ (HPV+ cohort). Exposure variable at primary conization.

No. of HPV infections (N = 82)

No. of HPV infections with an event (n = 50)

Median time, years (interquartiles)a

Univariate analysis (N = 82) Hazard ratio (95% CIs)

Multivariable analysis (N = 82) Hazard ratio (95% CIs)

Age ≤40 years N40 years p-Value

56 26

30 20

6.0 (2.0–8.7) 2.6 (1.1–3.7) b0.0001

1 3.2 (1.7–6.2) 0.0004

1 2.7 (1.3–5.3) 0.0059

HPV types HPV other HPV16 HPV18 HPV31 HPV33 p-Value

23 33 9 8 9

7 30 5 2 6

8.2 (6.0–NE) 1.8 (1.2–4.4) 2.2 (0.4–NE) NE (2.6–NE) 3.8 (3.3–7.8) 0.0001

1 5.1 (2.6–9.8) 5.2 (1.5–17.9) 1.0 (0.2–5.3) 2.3 (1.0–5.7) b0.0001

1 3.8 (1.9–7.7) 6.3 (1.7–22.6) 0.8 (0.2–3.6) 1.9 (0.8–4.4) 0.0002

Key: Hazard ratio (95% CIs) = hazard ratio and 95% confidence intervals; event = persistent HPV type in recurrent CIN3+; statistical unit = HPV infection; NE = not estimated. p-Value is the Log-rank p-value. a Kaplan–Meier analysis.

O.K. Vintermyr et al. / Gynecologic Oncology 133 (2014) 159–166

165

Fig. 2. Kaplan–Meier analysis of HPV types in primary cone. Probability for detecting persistent HPV infection in recurrent cone (CIN3+ subjects).

consistent with other studies having demonstrated high false-negative rate of pap smears post-cervical treatment for CIN2+ [12,37,40]. Further, knowledge of true ‘HPV negative’ or ‘HPV positive’ resection margins is not routinely tested for in the clinical setting and would be hard to assess based on morphology alone. The selection criteria for a minimum duration for recurrent CIN2 + disease including a requirement for at least one registered normal Pap smear result between primary and recurrent conizations for Sample II were adopted so as to identify the HPV types in women with no apparent clinical relapse from their primary cervical lesion. In the initial planning of the study, a relatively large cohort of 354 patients was enrolled. Within the study timeframe, samples from 179 women were sent to the central laboratory for centralised histopathological review; 110 women were determined to have CIN2 +. From this group of women a small subset was selected with the purpose of determining the most optimal testing procedures to identify newlyacquired HPV infection from persistent HPV as the cause of the recurrent CIN2 +. An initial sample of 29 women was selected based on study inclusion criteria as described above. A second sample of 29 women was selected based on a more than 6 month delay between primary and repeat conizations and with reported at least one normal cytology between conizations as described above. In undertaking the analysis on this subset of samples, it became evident that the recurrent CIN2+ infection was mainly caused by the same hr-HPV type so no additional women were included in the study. In conclusion, the majority (95.9%) of recurrent CIN2+ in this study of primary CIN3+ was associated with persistent hr-HPV infection, of which 74.5% were HPV16/18. More than 60% of the recurrent CIN2 + developed within three years of primary conization. A shorter median time-interval between conizations was observed for lesions infected with HPV16 and 18; in contrast HPV31/33 tended to be less prevalent in recurrent CIN2 + with a longer time interval to repeat conization. Women older than 40 years developed recurrent CIN2 + within shorter time interval than younger women. Normal cytology between conizations was not indicative of HPV clearance. Routine

post-conization hr-HPV testing together with cervical cytology may provide a better prediction for potential of recurrent disease [19]. Primary prevention through adolescent vaccination may prevent disease and, in doing so, have the potential to prevent its recurrence. Conflict of interest statement D Rosillon, S Thoresen and K Holl are employees of GlaxoSmithKline group of companies and hold stock options. O Iversen reports payment made by GlaxoSmithKline SA for travel support to meetings for the study or other purposes and communication to ethics committee and patients. O Vintermy reports payment made by GlaxoSmithKline SA for congress registration and to his institution for fees for participating in review activities such as data monitoring boards, statistical analysis and endpoint committees and provision of writing assistance, medicines, equipment, or administrative support. W. Quint reports no conflict of interest. A Molijn reports payment made by GlaxoSmithKline SA to his institution for travel support to meetings for the study or other purposes and fees for participation in review activities such as data monitoring boards, statistical analysis and endpoint committees. S Collas de Souza reports payment made by GlaxoSmithKline SA to her institution for consultancy, consulting fees and honorarium, fees for participating in review activities such as data monitoring boards, statistical analysis and endpoint committees, manuscript preparation, writing and reviewing.

Funding source GlaxoSmithKline Biologicals SA was the funding source and was involved in all stages of the study conduct and analysis. GSK Biologicals SA also funded all costs associated with the development and the publishing of the present manuscript. All authors had full access to the data and agreed with the submission of the publication. Acknowledgements The authors would like to thank Brigitte Colau, Alice Raillard, MarieCécile Bozonnat and Aurélie Le Plain (4Clinics France). They also would like to thank Roeland Van Kerckhoven (Keyrus Biopharma on behalf of GlaxoSmithKline Vaccines), Jérémie Dedessus le Moutier (Business and Decision Life Sciences on behalf of GlaxoSmithKline Vaccines) for manuscript coordination and Dr Catherine Streeton (Streeton Associates) for writing support.

166

O.K. Vintermyr et al. / Gynecologic Oncology 133 (2014) 159–166

Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2014.03.004.

[20] [21]

References [22] [1] McCredie MRE, Sharples K, Paul C, Baranyai J, Medley G, Jones RW, et al. Natural history of cervical neoplasia and risk of invasive cancer in women with cervical intraepithelial neoplasia 3: a retrospective cohort study. Lancet Oncol 2008; 9:425–34. http://dx.doi.org/10.1016/S1470-2045(08)70103-7. [2] Ostör AG. Natural history of cervical intraepithelial neoplasia: a critical review. Int J Gynecol Pathol 1993;12:186–92. [3] Schiffman M, Castle PE, Jeronimo J, Rodriguez AC, Wacholder S. Human papillomavirus and cervical cancer. Lancet 2007;370:890–7. [4] Elfgren K, Bistoletti P, Dillner L, Walboomers JM, Meijer CJ, Dillner J. Conization for cervical intraepithelial neoplasia is followed by disappearance of human papillomavirus deoxyribonucleic acid and a decline in serum and cervical mucus antibodies against human papillomavirus antigens. Am J Obstet Gynecol 1996;174:937–42. [5] Paraskevaidis E, Koliopoulos G, Malamou-Mitsi V, Zikopoulos K, Paschopoulos M, Pappa L, et al. Large loop excision of the transformation zone for treating cervical intraepithelial neoplasia: a 12-year experience. Anticancer Res 2001;21:3097–9. [6] 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;8:985–93. http://dx.doi.org/10.1016/S1470-2045(07)70283-8. [7] Katki HA, Schiffman M, Castle PE, Fetterman B, Poitras NE, Lorey T, et al. Five-year risk of recurrence after treatment of CIN2, CIN3, or AIS: performance of HPV and pap contesting in posttreatment management. J Lower Gen Tract Dis 2013;17: S78–84. http://dx.doi.org/10.1097/LGT.0b013e3182854269. [8] Melnikow J, McGahan C, Sawaya GF, Ehlen T, Coldman A. Cervical intraepithelial neoplasia outcomes after treatment: long-term follow-up from the British Columbia Cohort Study. J Natl Cancer Inst 2009;101:721–8. http://dx.doi.org/10.1093/jnci/ djp089. [9] Nagai Y, Maehama T, Asato T, Kanazawa K. Persistence of human papillomavirus infection after therapeutic conization for CIN 3: is it an alarm for disease recurrence? Gynecol Oncol 2000;79:294–9. http://dx.doi.org/10.1006/gyno.2000.5952. [10] Torne´ A, Fuste´ P, Rodrı'guez-Carunchio L, Alonso I, del Pino M, Nonell R, et al. Intraoperative post-conisation human papillomavirus testing for early detection of treatment failure in patients with cervical intraepithelial neoplasia: a pilot study. BJOG 2013;120:392–9. [11] Verguts J, Bronselaer B, Donders G, Arbyn M, Van Eldere J, Drijkoningen M, et al. High-risk HPV presence in cervical specimens after a large loop excision of the cervical transformation zone: significance of newly detected hr-HPV genotypes. J Med Virol 2007;79:314–9. http://dx.doi.org/10.1002/jmv.20800. [12] Verguts J, et al. Prediction of recurrence after treatment for high-grade cervical intraepithelial neoplasia: the role of human papillomavirus testing and age at conisation. BJOG 2006;113:1303–7. http://dx.doi.org/10.1111/j.1471-0528.2006. 01063.x. [13] Ryu A, Nam K, Kwak J, Kim J, Jeon S. Early human papillomavirus testing predicts residual/recurrent disease after LEEP. J Gynecol Oncol 2012;23:217–25. http:// dx.doi.org/10.3802/jgo.2012.23.4.217. [14] Elfgren K, Jacobs M, Walboomers JM, Meijer CJ, Dillner J. Rate of human papillomavirus clearance after treatment of cervical intraepithelial neoplasia. Obstet Gynecol 2002;100:965–71. http://dx.doi.org/10.1016/S0029-7844(02)02280-9. [15] Kreimer AR, Schiffman M, Herrero R, Hildesheim A, González P, Burk RD, et al. Longterm risk of recurrent cervical human papillomavirus infection and precancer and cancer following excisional treatment. Int J Cancer 2012;131:211–8. http:// dx.doi.org/10.1002/ijc.26349. [16] Livasy CA, Moore DT, Van Le L. The clinical significance of a negative loop electrosurgical cone biopsy for high-grade dysplasia. Obstet Gynecol 2004;104:250–4. http:// dx.doi.org/10.1097/01.AOG.0000132803.88049.84. [17] Prato B, Ghelardi A, Gadducci I, Marchetti C, Di Cristofano C, Di Coscio G, et al. Correlation of recurrence rates and times with post-treatment human papillomavirus status in patients treated with loop electrosurgical excision procedure conization for cervical squamous intraepithelial lesions. Int J Gynecol Cancer 2008;18:90–4. http://dx.doi.org/10.1111/j.1525-1438.2007.00965.x. [18] Rodriguez-Manfredi A, Alonso I, del Pino M, Fusté P, Torné A, Ordi J. Predictors of absence of cervical intraepithelial neoplasia in the conization specimen. Gynecol Oncol 2013;128:271–6. http://dx.doi.org/10.1016/j.ygyno.2012.10.020. [19] Kocken M, Uijterwaal MH, de Vries ALM, Berkhof J, Ket JCF, Helmerhorst TJM, et al. High-risk human papillomavirus testing versus cytology in predicting posttreatment disease in women treated for high-grade cervical disease: a systematic re-

[23]

[24]

[25]

[26]

[27]

[28]

[29] [30] [31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

view and meta-analysis. Gynecol Oncol 2012;125:500–7. http://dx.doi.org/10.1016/ j.ygyno.2012.01.015. NHMRC. http://www.cancerscreening.gov.au/internet/screening/publishing.nsf/ Content/cv-hpv-dna; 2009. [Accessed 2 Sept 2013]. Munoz N, Kjaer SK, Sigurdsson K, Iversen O-E, Hernandez-Avila M, Wheeler CM, et al. Impact of human papillomavirus (HPV)-6/11/16/18 vaccine on all HPVassociated genital diseases in young women. J Natl Cancer Inst 2010;102:325–39. http://dx.doi.org/10.1093/jnci/djp534. Lehtinen M, Paavonen J, Wheeler CM, Jaisamrarn U, Garland SM, Castellsagué X, 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;13:89–99. http:// dx.doi.org/10.1016/S1470-2045(11)70286-8. Joura EA1, Garland SM, Paavonen J, Ferris DG, Perez G, Ault KA, et al. Effect of the human papillomavirus (HPV) quadrivalent vaccine in a subgroup of women with cervical and vulvar disease: retrospective pooled analysis of trial data. BMJ 2012;344:7851. http://dx.doi.org/10.1136/bmj.e1401. Cancer Registry of Norway (CRN). About the cancer registry. Last updated 16 Apr 2009 http://www.kreftregisteret.no/en/General/About-the-Cancer-Registry/Aboutthe-organization/. [25 May 2013, date last accessed]. Bilet EF, Langseth H, Thoresen SO, Bray F. Completeness of invasive cervical cancer at the Cancer Registry of Norway. Acta Oncol 2009;48:1070–86. http://dx.doi.org/ 10.1080/02841860902862529. Norsk gynekologisk forening. Den Norske Legeforening. Premaligne lidelser cervix. “Vedlegg 3”. located at http://legeforeningen.no/PageFiles/2422/Premaligne% 20lidelser%20i%20cervix%20.pdf; 2009. [Accessed on 24 April 2013]. Morris BJ. Cervical human papillomavirus screening by PCR: advantages of targeting the E6/E7 region. Clin Chem Lab Med 2005;43:1171–7. http://dx.doi.org/10.1515/ CCLM.2005.203. van Alewijk D, Kleter B, Vent M, Delroisse J-M, de Koning M, van Doorn L-J, et al. A human papilloma virus testing algorithm comprising a combination of the L1 broad-spectrum SPF10 PCR assay and a novel E6 high-risk multiplex typespecific genotyping PCR assay. J Clin Microbiol 2013;51:1171. http://dx.doi.org/ 10.1128/JCM.02831-12. Lin DY, Wei LJ. The robust inference for the Cox proportional hazards model. J Am Stat Assoc 1989;84:1074–8. Zeger SL, Liang K-Y. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121–30. Wu D, Zheng Y, Chen W, Guo C, Yu J, Chen G, et al. Prediction of residual/recurrent disease by HPV genotype after loop excision procedure for high-grade cervical intraepithelial neoplasia with negative margins. ANZJOG 2011;51:114–8. http:// dx.doi.org/10.1111/j.1479-828X.2010.01280.x. Quint W, Jenkins D, Molijn A, Struijk L, van de Sandt M, Doorbar J, et al. One virus, one lesion-individual components of CIN lesions contain a specific HPV type. J Pathol 2012;227:62–71. http://dx.doi.org/10.1002/path.3970. Chan BK, Melnikow J, Slee CA, Arellanes R, Sawaya GF. Posttreatment human papillomavirus testing for recurrent cervical intraepithelial neoplasia: a systematic review. Am J Obstet Gynecol 2009;200:422. Brockmeyer AD, Wright JD, Gao F, Powell MA. Persistent and recurrent cervical dysplasia after loop electrosurgical excision procedure. Am J Obstet Gynecol 2005;192:1379–81. http://dx.doi.org/10.1016/j.ajog.2004.12.044. Sarian LOZ, Derchain SFM, da Rocha Pitta D, Morais SS, Rabelo-Santos SH. Factors associated with HPV persistence after treatment for high-grade cervical intraepithelial neoplasia with large loop excision of the transformation zone (LLETZ). J Clin Virol 2004;31:270–4. http://dx.doi.org/10.1016/j.jcv.2004.05.012. Bertelsen BI, Kugarajh K, Skar R, Laerum OD. HPV subtypes in cervical cancer biopsies between 1930 and 2004: detection using general primer pair PCR and sequencing. Virchows Arch 2006;449:141–7. http://dx.doi.org/10.1007/s00428006-0232-3. Jain S, Tseng C-J, Horng S-G, Soong Y-K, Pao CC. Negative predictive value of human papillomavirus test following conisation of the cervix uteri. Gynecol Oncol 2001;82:177–80. http://dx.doi.org/10.1006/gyno.2001.6241. Simoes RB, Campaner AB. Post-cervical conization outcomes in patients with highgrade intraepithelial lesions. APMIS 2014. http://dx.doi.org/10.1111/apm.12064 [in press]. Nam K, Chung S, Kim J, Jeon S, Bae D. Factors associated with HPV persistence after conization in patients with negative margins. J Gynecol Oncol 2009;20:91–5. http:// dx.doi.org/10.3802/jgo.2009.20.2.91. Paraskevaidis E, Koliopoulos G, Alamanos Y, Malamou-Mitsi V, Lolis ED, Kitchener HC. Human papillomavirus testing and the outcome of treatment for cervical intraepithelial neoplasia. Obstet Gynecol 2001;98:833–6. http://dx.doi.org/10. 1016/S0029-7844(01)01535-6.