18 genotyping in high-risk HPV positive self-sampled specimens to identify women with high grade CIN or cervical cancer

Gynecologic Oncology 135 (2014) 58–63

Contents lists available at ScienceDirect

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

Methylation marker analysis and HPV16/18 genotyping in high-risk HPV positive self-sampled specimens to identify women with high grade CIN or cervical cancer Viola M.J. Verhoef a,1, Daniëlle A.M. Heideman a, Folkert J. van Kemenade a,2, Lawrence Rozendaal a, Remko P. Bosgraaf b, Albertus T. Hesselink a, Ruud L.M. Bekkers b, Leon F.A.G. Massuger b, Renske D.M. Steenbergen a, Peter J.F. Snijders a, Johannes Berkhof c, Chris J.L.M. Meijer a,⁎ a b c

Department of Pathology, VU University Medical Center, De Boelelaan 1117, 1081 HV Amsterdam, The Netherlands Department of Obstetrics and Gynaecology, Radboud University Medical Center, PO Box 9101, 6500 HB Nijmegen, The Netherlands Department of Epidemiology and Biostatistics, VU University Medical Center, De Boelelaan 1117, 1081 HV Amsterdam, The Netherlands

H I G H L I G H T S • HPV positive women with high methylation levels of MAL/miR-124-2 have a substantial risk of cervical (pre-)cancer. • Specificity of MAL/miR-124-2 methylation testing can be enhanced by increasing the threshold and adding HPV16/18 genotyping to the panel. • This study supports HPV self-sampling for cervical screening with triage by a combination of molecular tests.

a r t i c l e

i n f o

Article history: Received 18 May 2014 Accepted 3 August 2014 Available online 8 August 2014 Keywords: Human papillomavirus Self-sampling DNA methylation HPV genotyping Cervical intraepithelial neoplasia Cervical screening

a b s t r a c t Objectives. Methylation marker analysis using bi-marker panel MAL/miR-124-2 is a promising triage test for identifying cervical (pre)cancer in high-risk human papillomavirus (hrHPV) positive women. Bi-marker panel MAL/miR-124-2 can be applied directly on self-sampled cervico-vaginal material and its sensitivity is noninferior to that of cytology, yet at the cost of more colposcopy referrals. Our objective was to increase specificity of MAL/miR-124-2 methylation analysis by varying the assay thresholds and adding HPV16/18 genotyping. Methods. 1019 hrHPV-positive women were selected from a randomized controlled self-sampling trial (PROHTECT-3; 33–63 years, n = 46,001) and nine triage strategies with methylation testing of MAL/miR-124-2 and HPV16/18 genotyping were evaluated. The methylation assay threshold was set at four different predefined levels which correspond with clinical specificities for end-point cervical intra-epithelial grade 3 or worse (CIN3+) of 50%, 60%, 70%, and 80%. Results. The CIN3+ sensitivity of methylation analysis decreased (73.5 to 44.9%) while specificity increased (47.2 to 83.4%) when increasing the assay threshold. CIN3+ sensitivity and specificity of HPV16/18 genotyping were 68.0% and 65.6%, respectively. Combined methylation analysis at threshold-80 and HPV16/18 genotyping yielded similar CIN3+ sensitivity as that of methylation only at threshold-50 (77.6%) with an increased specificity (54.8%). Conclusions. Combined triage by MAL/miR-124-2 methylation analysis with threshold-80 and HPV16/18 genotyping reaches high CIN3 + sensitivity with increased specificity to identify women with cervical (pre)cancer among HPV self-sample positive women. The combined strategy is attractive as it is fully molecular and identifies women at the highest risk of cervical (pre)cancer because of strongly elevated methylation levels and/or HPV16/18 positivity. © 2014 Published by Elsevier Inc.

⁎ Corresponding author at: VU University Medical Centre, Department of Pathology, PO Box 7057, 1007 MB Amsterdam, The Netherlands. E-mail address: [email protected] (C.J.L.M. Meijer). 1 Present address: Department of Obstetrics and Gynaecology, Onze Lieve Vrouwe Gasthuis, Amsterdam, The Netherlands. 2 Present address: Department of Pathology, Erasmus MC University Medical Centre, PO Box 2040, 3000 CA Rotterdam, The Netherlands.

http://dx.doi.org/10.1016/j.ygyno.2014.08.003 0090-8258/© 2014 Published by Elsevier Inc.

Introduction In developed countries, cytology-based cervical screening programs have reduced the incidence of and mortality from cervical cancer [1,2]. However, participation rates in the current screening programs are still suboptimal [3]. Offering cervico-vaginal self-sampling for human

V.M.J. Verhoef et al. / Gynecologic Oncology 135 (2014) 58–63

papillomavirus (HPV) testing (HPV self-sampling) could lead to an improved protection as it may re-attract non-attendees into cervical screening [4,5]. Moreover, HPV self-sampling enables cervical screening in low- and middle resource countries [6,7]. A recent meta-analysis showed that the accuracy of HPV self-sampling for detecting cervical precancerous disease was higher than that of cytology [8]. However, HPV testing also involves the detection of transient infections; thus, further triage testing in high-risk HPV (hrHPV)-positive women is needed to identify women with the highest risk for cervical intraepithelial neoplasia (CIN) grade 3 or worse (CIN3+) [9,10]. Presently, cytology is the most frequently used triage test in hrHPVpositive women [11,12]. Since cytology is not feasible on self-sampled material [13,14], triage by cytology would require an extra visit to a physician for a cervical smear in women with a hrHPV-positive selfsample. As such, non-morphology based triage tests (i.e., HPV16/ HPV18 genotyping and molecular biomarkers) that can be directly applied to self-sampled specimens are receiving growing interest. HPV16/18 has shown to be a feasible triage test on cervical smears as well as on self-sampled material [12,15–17] but does not detect all cervical cancers since only ~ 70% is HPV16/18 positive [18,19]. DNA methylation analysis of promoter regions of tumor suppressor genes by quantitative methylation specific PCR (qMSP) has shown to be a valuable, objective triage tool for hrHPV positive women [20–23]. DNA promoter methylation is an important event in cervical carcinogenesis and DNA methylation levels increase by duration of the HPV infection and degree of CIN lesions, and are very high in cervical cancer [24]. The bi-marker panel CADM1/MAL was shown to be equally discriminatory for CIN3 + as cytology when applied to cervical smears of hrHPV positive women [20]. Methylation marker analysis has also shown to be feasible as a direct triage tool on self-sampled material taken by a lavage device in women who were tested hrHPV-positive [17,25]. A recent prospective randomized controlled trial in which self-sampling lavage devices were offered to non-attendees of regular cervical screening revealed that methylation analysis of genes MAL and miR-124-2 genes with a preset assay threshold giving rise to a CIN3+ specificity of 50% was non-inferior to cytology testing in terms of detecting CIN grade 2 or worse (CIN2 +) [26]. This method obviates the need for a visit to a physician and reduces time to CIN2 + diagnosis, but at the cost of more colposcopy referrals than by cytology triage. By increasing the MAL/miR-124-2 methylation assay thresholds, the number of colposcopy referrals will decrease, but the sensitivity for detection of CIN2+ will decrease as well. The drop in sensitivity may be remedied by using a combination of different molecular markers. Here, we evaluated molecular strategies using MAL/miR-124-2 methylation analysis at an increased assay threshold combined with HPV16/18 genotyping. To this end, we conducted a post hoc analysis using self-collected lavage specimens of 1019 hrHPV positive women belonging to either the molecular or cytology triage group of the PRotection by Offering HPV TEsting on self-sampled Cervico-vaginal specimens Trial-3 (PROHTECT3; N = 46,001; 33–63 years). Material and methods Study design PROHTECT-3 study The PROHTECT-3 study (Trial register: NTR6026) is a randomized controlled trial designed to compare direct triage by methylation marker testing on self-sampled material of hrHPV positive women to cytology on a subsequent smear taken at the physician's office. The primary end-point of the study was CIN2 +. The study was ethically approved by the Ministry of Public Health (No. 2010/WBO04) and participants gave written informed consent. The design and results of this study were described previously [26]. In short, 46,001 women from the regions Noord-Holland, Utrecht, Flevoland and Gelderland in the Netherlands who did not respond to an invitation for a cervical cytology smear in 2007 were offered a self-sampling lavage device

59

[14,27]. In total, 12,819 women participated by returning self-sampled material to the laboratory for hrHPV testing (hrHPV GP5 +/6 +-PCR; EIA HPV GP HR kit; Diassay, Voorburg, The Netherlands). All women who tested hrHPV negative were advised to participate in the next screening round. All hrHPV-positive women (n = 1038) were randomized (1:1 ratio) to either cytology triage (i.e., invited for a physiciantaken cervical smear for cytology) or molecular triage (i.e., methylation analysis of bi-marker panel MAL/miR-124-2 [17]), performed directly on the original submitted self-sample. Cytology reports were classified as either normal or abnormal; results classified as atypical squamous cells of undetermined significance (ASC-US) or more were considered abnormal [28,29]. On the left-over self-sampled material of all hrHPV positive women, methylation analysis of bi-marker panel MAL/miR-124-2 by quantitative methylation specific PCR (qMSP) was performed [17]. The methylation test results were used to triage women in the molecular triage group but were blinded in the cytology triage group. All women with a positive triage test (i.e. abnormal cytology in cytology triage group or positive methylation test in molecular triage group) were referred for colposcopy-directed biopsy. Women with a negative triage test (normal cytology or negative methylation test) were advised to undergo an exit-test after six months (consisting of both HPV testing and cytology), and referred for colposcopy-directed biopsy in case of abnormal cytology and/or HPV positive test. Histology was performed in different pathology laboratories in The Netherlands, according to standard guidelines, and data was tracked through the nationwide network and registry of pathology in the Netherlands (PALGA, Bunnik, The Netherlands) [30]. The study endpoint was achieved if either a histological outcome was reported or if the exit-test was double negative (i.e., normal cytology and HPV negative). Women without a study-endpoint at January 2013, which is 12 to 24 months after entering the study, were labeled ‘no CIN2 + detected’. HPV genotyping and methylation marker analysis hrHPV genotyping was performed on all HPV positive samples using Luminex suspension array technology as described previously [31]. Bisulfite treatment was performed on left-over isolated DNA by using the EZ DNA Methylation Kit (Zymo Research, Orange, CA) as earlier described [17,23]. To detect promoter hypermethylation of the MALm1 region and the miR124-2 region, we used primers and probes for qMSP on the ABI 7500 real-time PCR system (Applied Biosystems, USA) as was previously described [20,32]. As a reference, a PCR for the bisulfite converted housekeeping gene beta-actin was performed. Cycle thresholds (Ct) values were measured at fixed fluorescence threshold (i.e., 0.01) [20]. Ct ratios between the methylation level of the target genes relative to the reference gene beta-actin were calculated as described previously to quantify the level of methylation [17,20]. Samples were scored invalid if tested both negative for the target gene and negative for beta-actin. In the prospective trial [26], specimens were recorded as positive for methylation when either or both markers had a qMSP outcome above predefined qMSP thresholds. These thresholds were set as Ct ratios that gave rise to CIN3 + specificity value of 50% (referred to as threshold-50) in a previous training set analysis [17]. For this study, we considered three additional predefined thresholds that have previously been set as Ct ratios corresponding to CIN3 + specificity values of 60%, 70%, and 80% (further referred as threshold-60, threshold-70, and threshold-80, respectively) in hrHPV-positive self-collected lavage specimens [17]. Statistical analysis The primary outcome was CIN3+ detection and the secondary outcome was CIN2+ detection. Nine strategies were obtained by defining

60

V.M.J. Verhoef et al. / Gynecologic Oncology 135 (2014) 58–63

four different thresholds for scoring methylation positivity, HPV16/18 genotyping, and combinations of the two instruments. For these strategies, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for CIN3 + and CIN2 + were calculated. Statistical uncertainty was expressed by 95% Wald confidence intervals. Sensitivity and specificity of triage strategies were compared with the McNemar test. Furthermore, the colposcopy referral rate was calculated as the proportion of hrHPV positive women who had a positive triage test result. The Pearson chi-square test was used to assess whether the performance of the methylation test was similar in the molecular triage group and in the cytology triage group. The analyses were carried out in SPSS version 20 and STATA version 11.

Table 1 Age strata for hrHPV positivity rate, detected CIN2+ and CIN3+. Age group

30–34 35–39 40–44 45–49 50–54 55–59 60–64

years years years years years years years

hrHPV positive

CIN2+

n

%

n

%

n

CIN3+ %

254 254 187 126 101 63 34 1019

25 25 18 12 10 6 3 100

67 64 42 28 15 9 0 225

30 28 19 12 7 4 0 100

46 45 27 18 9 2 0 147

31 31 18 12 6 1 0 100

Results Characteristics of the PROHTECT-3 study participants From November 2010 to December 2011, 46,001 women who did not respond to an invitation for regular screening were invited to submit a self-sample for HPV testing and 12,819 women participated (Fig. 1). Of the 1024 HPV self-sample positive women eligible for randomization, 509 were randomized to the cytology triage and 515 to molecular triage group. After excluding five women with invalid

qMSP results, the cytology and molecular triage group comprised 505 and 514 women, respectively. In total, 772 women (76.8%) reached a study-endpoint including thirteen carcinomas, 134 CIN3, 78 CIN2 and 547 ≤ CIN1. This latter group consisted of 231 negative histology cases (no CIN), 151 CIN1, and 165 double negative exit-tests (i.e., normal cytology and HPV negative). The remaining 247 women without a study-endpoint were considered as ‘no CIN2 detected’ in the analyses. The hrHPV positivity rate and number of CIN3 + and CIN2 + per age strata are shown in Table 1.

46,001 women were offered self-sampling for hrHPV testing 33,182 excluded no participation

12,819 women participated by returning cervicovaginal self-sampled material 11,781 excluded 11,755 HPV negative 26 invalid result

1,038 women were tested hrHPV positive on selfsampled material and were randomly assigned over cytology- and methylation group (1:1) 14 excluded 4 hysterectomy 7 ≥BMD in last 2 years 1 CIN2 in last 2 years 2 deceased

1,024 eligible participants 5 excluded Invalid methylation marker test result

1,019 participants with valid methylation test result and HPV16/18 genotyping test result 505 in cytology triage group 514 in molecular triage group

772 with study endpoint 165 normal smear/HPV negative 231 No CIN 151 CIN 1 78 CIN 2 134 CIN 3 13 SCC 247 without study endpoint

Fig. 1. Flowchart of eligible women of PROHTECT-3 study.

V.M.J. Verhoef et al. / Gynecologic Oncology 135 (2014) 58–63

Strategies in women tested HPV positive on self-sampled material The performance of the bi-marker methylation test was similar in the cytology and methylation triage groups, for end-points CIN3 + (p N 0.05) and CIN2 + (p N 0.05) (Supplementary Table). Therefore, we decided to pool the methylation marker test results in the two study groups. The positivity rate of the bi-marker methylation tests with four separate assay thresholds (i.e., threshold-50, -60, -70 and -80, as described in the Methods section), HPV16/18 genotyping, and combinations thereof in self-samples of hrHPV-positive women, overall and stratified per study endpoint, is shown in Table 2. The performance of the triage strategies with regard to CIN3+ and CIN2+ is listed in Table 3. The strategy with methylation marker testing at threshold-50 showed a sensitivity for CIN3+ of 73.5% and a specificity of 47.2%. Methylation marker testing with threshold-60 resulted in a somewhat lower sensitivity (68.0%) for CIN3 + in combination with a significantly higher specificity (55.3%). For methylation testing with threshold-70 and threshold-80, sensitivities were 55.1% and 44.9% and specificities were 67.5% and 83.4%, respectively. The sensitivity, specificity, and PPV strongly varied with the methylation threshold whereas the NPV was rather flat over the range of threshold values. The strategy with threshold-50 resulted in a referral rate of 55.7% in the group of hrHPV-positive women. The referral rates at elevated thresholds were 48.1% at threshold-60, 35.7% at threshold-70, and 20.7% at threshold80. For comparison, the figures of cytology triage as performed on a subsequent physician-taken smear (cytology triage group) [26] are also depicted in Table 3. Cytology triage yielded similar sensitivity, a higher specificity, and lower colposcopy referral rate compared with methylation analysis at threshold-80 combined with HPV16/18 genotyping. For HPV16/18 genotyping (without methylation testing), CIN3 + sensitivity was 68.0%, the specificity was 65.6%, the PPV was 25.0%, the NPV was 92.4%, and the referral rate of hrHPV positives was 39.3%. When combining methylation marker testing at threshold-80 with HPV16/18 genotyping, the sensitivity for CIN3+ was similar to that of the strategy with sole methylation marker testing at threshold-50 (77.6%, 95%CI 70.8–84.3 versus 73.5%, 95%CI 66.3–80.6, respectively, p-value 0.42). However, CIN3 + specificity value of the combined threshold-80 and HPV16/18 genotyping strategy was significantly higher than that observed at threshold-50 without additional HPV16/ 18 genotyping (54.8%, 95%CI 51.5–58.1 versus 47.2%, 95%CI 43.9–50.6, respectively, p-value 0.001). Lowering the threshold in the combined testing setting (methylation testing with HPV16/18 genotyping) further increased the sensitivity but at the cost of a substantial decrease in specificity. This is displayed in Fig. 2 where receiver operating characteristic (ROC) curves are presented for methylation testing with and without

61

HPV16/18 genotyping. The PPV and NPV of the combination strategies are presented in Table 3. Again, the NPV was rather insensitive to the methylation threshold but the NPV of combination testing was markedly higher than the threshold of single methylation testing (without HPV16/18 genotyping). The referral rate of the combined strategy with threshold-80 and HPV16/18 genotyping (49.9%) was significantly lower than that of sole methylation testing with threshold-50 (55.7%) (p-value 0.03). Discussion We evaluated qMSP results of the bi-marker MAL/miR-124-2 methylation test with four different assay thresholds with and without HPV16/18 genotyping in women with HPV positive self-collected cervico-vaginal specimens. By combining methylation marker testing and HPV16/18 genotyping, the methylation assay threshold can be increased from threshold-50 to threshold-80 which yields a significant increase in specificity without loss of sensitivity. The CIN3+ sensitivity of sole methylation marker testing at threshold-50 found in this study (73.5%) was similar to that of cytology testing in our previous study (77.9%) [26]. Of note, cytology in our previous study was performed only in the cytology triage group, whereas here methylation marker testing was done in women of both the molecular and cytology triage groups. The observed CIN3+ NPV with methylation marker testing at threshold-50 and threshold-60, and the combination of methylation marker threshold-80 with HPV16/18 genotyping are within the same range as that of women with a normal cytology triage test in previous studies [11,33]. Bierkens et al. showed that promoter methylation levels of CADM1 and MAL increased with increasing severity and longer duration of cervical lesions and were very high in cervical cancers [24]. Therefore, bi-marker methylation testing seems promising in detecting more advanced high-grade precursor lesions and cervical cancer. However, in this study, one woman with cervical cancer (a micro-invasive carcinoma) was tested methylation test negative on self-sampled cervicovaginal material but methylation positive on the subsequent cervical smear; this suggests a sampling error. Two cervical cancers were missed by using sole HPV16/18 genotyping, because these cancers were typed as HPV59. With combined methylation marker testing and HPV16/18 genotyping, all cervical cancers in this study were detected, even when methylation threshold for positivity was increased to threshold-80. This study showed that the performances of bi-marker MAL/ miR-124-2 methylation testing in the cytology triage group and the molecular triage group were similar for detection of both CIN2+ and CIN3 + indicating a high reproducibility of the previously reported methylation test results [26]. The advantage of qMSP is that thresholds

Table 2 Positive methylation marker test results, positive HPV16/18 genotyping results, and combinations thereof in hrHPV positive women and with regard to study-endpoint. Number and proportion of positive triage tests

Threshold-50 Threshold-60 Threshold-70 Threshold-80 HPV16/18 Threshold-50/HPV16/18 Threshold-60/HPV16/18 Threshold-70/HPV16/18 Threshold-80/HPV16/18

In all HPV positive women n = 1019

Per endpoint No CIN2+ detected n = 794

CIN2 n = 78

CIN3 n = 134

Cancer n = 13

568 490 364 211 400 742 692 611 508

405 340 253 135 267 549 506 443 357

55 50 30 10 33 63 59 48 37

96 88 69 55 89 117 114 107 101

12a 12a 12a 11a 11b 13 13 13 13

55.7% 48.1% 35.7% 20.7% 39.3% 72.8% 67.9% 60.0% 49.9%

51.0% 42.8% 31.9% 17.0% 35.6% 73.3% 67.6% 59.1% 47.7%

70.5% 64.1% 38.5% 12.8% 42.3% 80.8% 75.6% 61.5% 47.4%

71.6% 65.7% 51.5% 41.0% 66.4% 87.3% 85.1% 79.9% 75.4%

92.3% 92.3% 92.3% 84.6% 84.6% 100% 100% 100% 100%

Threshold = thresholds corresponding with predefined specificity for CIN3+ in an earlier study. CIN = cervical intraepithelial neoplasia (grade 2 or grade 3). a One micro-invasive carcinoma tested methylation marker negative on self-sampled material; a subsequent cervical smear after 5 weeks revealed a positive methylation marker test result. b Two cervical carcinomas were classified with genotyping as type HPV59.

62

V.M.J. Verhoef et al. / Gynecologic Oncology 135 (2014) 58–63

Table 3 Performance and referral rates of triage strategies in hrHPV positive women (of whom 505 were allocated to the cytology triage group and 514 to the molecular triage group). CIN3+

Threshold-50 Threshold-60 Threshold-70 Threshold-80 HPV16/18 genotyping Threshold-50/HPV16/18 Threshold-60/HPV16/18 Threshold-70/HPV16/18 Threshold-80/HPV16/18 Cytologya

CIN2+

Sensitivity

Specificity

PPV

NPV

Sensitivity

Specificity

PPV

NPV

Referrals

% (95% CI)

% (95% CI)

% (95% CI)

% (95% CI)

% (95% CI)

% (95% CI)

% (95% CI)

% (95% CI)

%

73.5 (66.3–80.6) 68.0 (60.5–75.6) 55.1 (47.1–63.1) 44.9 (36.9–52.9) 68.0 (60.5–75.6) 88.4 (83.3–93.6) 86.4 (80.9–91.9) 81.6 (75.4–87.9) 77.6 (70.8–84.3) 77.9 (68.1–87.8)

47.2 (43.9–50.6) 55.3 (52.0–58.6) 67.5 (64.4–70.7) 83.4 (80.9–85.8) 65.6 (62.4–68.7) 29.8 (26.8–32.9) 35.2 (32.0–38.4) 43.7 (40.4–47.0) 54.8 (51.5–58.1) 77.6 (73.6–81.5)

19.0 (15.8–22.2) 20.4 (16.8–24.0) 22.3 (18.0–26.5) 31.3 (25.0–37.5) 25.0 (20.8–29.2) 17.5 (14.8–20.3) 18.4 (15.5–21.2) 19.6 (16.5–22.8) 22.4 (18.8–26.1) 35.6 (27.9–43.3)

91.4 (88.8–93.9) 91.1 (88.7–93.5) 89.8 (87.6–92.2) 90.0 (87.9–92.0) 92.4 (90.3–94.5) 93.9 (91.0–96.7) 93.9 (91.3–96.5) 93.4 (91.0–95.8) 93.5 (91.4–95.7) 95.7 (93.5–97.8)

72.4 (66.6–78.3) 66.7 (60.5–72.8) 49.3 (42.8–55.9) 33.8 (27.6–40.0) 59.1 (52.7–65.5) 85.8 (81.2–90.3) 82.7 (77.7–87.6) 74.7 (69.0–80.3) 67.1 (61.0–73.2) 74.3 (65.7–82.8)

49.0 (45.5–52.5) 57.2 (53.7–60.6) 68.1 (64.9–71.4) 83.0 (80.4–85.6) 66.4 (63.1–69.7) 30.9 (27.6–34.1) 36.3 (32.9–39.6) 44.2 (40.8–47.7) 55.0 (51.6–58.5) 81.3 (77.4–85.1)

28.7 (25.0–32.4) 30.6 (26.5–34.7) 30.5 (25.8–35.2) 36.0 (29.5–42.5) 33.3 (28.6–37.9) 26.0 (22.9–29.2) 26.9 (23.6–30.2) 27.5 (24.0–31.0) 29.7 (25.7–33.7) 50.3 (42.3–58.4)

86.3 (83.1–89.4) 85.8 (82.8–88.8) 82.6 (79.7–85.5) 81.6 (78.9–84.2) 85.1 (82.3–87.9) 88.4 (84.7–92.2) 88.1 (84.6–91.6) 86.0 (82.7–89.4) 85.5 (82.5–88.6) 92.5 (89.7–95.3)

55.7 48.1 35.7 20.7 39.3 72.8 67.9 60.0 49.9 30.0

Threshold = thresholds corresponding with predefined specificity for CIN3+ in an earlier study. CIN = cervical intraepithelial neoplasia (grade 2 or grade 3 or higher). CI = confidence interval. a Cytology was performed on an additional physician-taken scrape only in 505 women of the cytology triage group.

for positivity could easily be varied in search for the optimal balance between the safety of the methylation triage test (NPV) and the colposcopy burden (PPV and referral rate) for patients and gynecologists. Since the methylation marker test with predefined specificity threshold-50 resulted in a relative high sensitivity in combination

Fig. 2. Comparison of different triage strategies in women with HPV positive self-sampled material. Receiver operating characteristic (ROC) curve of the MAL/miR-124-2 methylation test, HPV16/18 genotyping and combination of MAL/miR-124-2 methylation and HPV16/18 genotyping are shown for CIN2+ and CIN3+.

with an acceptable NPV, this could be the strategy of first choice. The combination of methylation testing at threshold-80 with HPV16/18 genotyping showed a similar performance in terms of clinical sensitivity with a significantly increased specificity compared to the threshold-50 strategy. This combination is a feasible strategy by referring those women with the most advanced CIN2+ lesions and cancer, associated with the highest methylation marker levels, and/or the highest risk for cervical cancer based on HPV16/18 positivity [19]. In this study, the referral rate of the combined strategy with threshold-80 and HPV16/18 genotyping (49.9%) was statistically significantly lower than that of sole methylation testing using threshold-50 (55.7%), yet higher than that of cytology triage on a subsequent physician-taken smear (30.0%). However, cytology is a subjective test with a performance that is largely variable between countries [34] and even within a country [11,35,36]. Moreover, cytology triage requires an additional visit at baseline to the physician which results in a lower triage completion rate. The advantages of our study are its large size and the setting within non-attendees of the nation-wide screening program, making the study results applicable to non-attendees of organized cervical screening in other countries. Since this study was performed in non-attendees of screening with a higher CIN2+ risk than responders of screening [37], the performance of these screening strategies should preferably be studied in a responder population first, before it could be implemented in the setting of a regular screening program. A limitation was that our results could possibly be biased because no study-endpoint was available in a substantial number of HPV positive women due to loss to follow-up (i.e., 23.2% of all women did not reach a study-endpoint). However, we think it is unlikely that this aspect will markedly influence the outcome. After all, of all women without a study-endpoint, 175 (71%) had normal cytology as their last documented report. It should be noted that the qMSP assay used to detect MAL/ miR-124-2 methylation is a laboratory-developed test [17], validated prospectively in a randomized controlled trial [26], but not yet validated by other laboratories or certified as IVD assay. The randomized controlled trial contains a methylation group and a cytology group, but methylation was tested in a blinded fashion in the cytology group. When the methylation status of samples in the cytology group was

V.M.J. Verhoef et al. / Gynecologic Oncology 135 (2014) 58–63

unblinded, a similar performance of these markers was observed in terms of CIN2+ and CIN3+ detection in comparison to the methylation arm. Together with the fact that previously these markers revealed a high reproducibility and similar clinical performance [17], this supports robustness of the assay. In conclusion, this pooled post-hoc analysis on both triage groups of PROHTECT-3 confirms our previous finding on the molecular triage group [26] that bi-marker MAL/miR124-2 methylation testing at threshold-50 can be used as direct triage on hrHPV-positive selfsampled material in terms of sensitivity, specificity and NPV for CIN3+. This circumvents the requirement for a cytology triage-visit to the physician, but at the cost of a higher referral rate. In addition, the current study shows that bi-marker methylation testing with a higher threshold for positivity (threshold-80) combined with HPV16/18 genotyping gives a lower colposcopy referral rate, while retaining a sufficiently high sensitivity for detecting clinically meaningful cervical disease. Thus, the bi-marker MAL/miR-124-2 methylation test (threshold-80) with HPV16/18 genotyping or sole bi-marker MAL/miR-124-2 methylation testing may allow full molecular screening directly on self-sampled material. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2014.08.003.

[15]

[16]

[17]

[18] [19]

[20]

[21]

[22]

[23]

Conflict of interest statement CJLMM, PJFS, RDMS, and DAMH are shareholders of Self-screen BV, a spin-off company of VU University Medical Center. RLMB is a member of the advisory board of Takeda International and consultant for Roche Diagnostics. All other authors declare that they have no conflicts of interest.

[24]

[25]

References [26] [1] Arbyn M, Raifu AO, Weiderpass E, Bray F, Anttila A. Trends of cervical cancer mortality in the member states of the European Union. Eur J Cancer 2009;45:2640–8. [2] Peto J, Gilham C, Fletcher O, Matthews FE. The cervical cancer epidemic that screening has prevented in the UK. Lancet 2004;364:249–56. [3] Bos AB, Rebolj M, Habbema JD, van Ballegooijen M. Nonattendance is still the main limitation for the effectiveness of screening for cervical cancer in the Netherlands. Int J Cancer 2006;119:2372–5. [4] Snijders PJ, Verhoef VM, Arbyn M, Ogilvie G, Minozzi S, Banzi R, et al. High-risk HPV testing on self-sampled versus clinician-collected specimens: a review on the clinical accuracy and impact on population attendance in cervical cancer screening. Int J Cancer 2013;132:2223–36. [5] Zhao FH, Lewkowitz AK, Chen F, Lin MJ, Hu SY, Zhang X, et al. Pooled analysis of a self-sampling HPV DNA Test as a cervical cancer primary screening method. J Natl Cancer Inst 2012;104:178–88. [6] Lazcano-Ponce E, Lorincz AT, Cruz-Valdez A, Salmeron J, Uribe P, VelascoMondragon E, et al. Self-collection of vaginal specimens for human papillomavirus testing in cervical cancer prevention (MARCH): a community-based randomised controlled trial. Lancet 2011;378:1868–73. [7] Qiao YL, Sellors JW, Eder PS, Bao YP, Lim JM, Zhao FH, et al. A new HPV-DNA test for cervical-cancer screening in developing regions: a cross-sectional study of clinical accuracy in rural China. Lancet Oncol 2008;9:929–36. [8] Arbyn M, Verdoodt F, Snijders PJ, Verhoef VM, Suonio E, Dillner L, et al. Accuracy of human papillomavirus testing on self-collected versus clinician-collected samples: a meta-analysis. Lancet Oncol 2014;15:172–83. [9] Cuzick J, Clavel C, Petry KU, Meijer CJ, Hoyer H, Ratnam S, et al. Overview of the European and North American studies on HPV testing in primary cervical cancer screening. Int J Cancer 2006;119:1095–101. [10] Ronco G, Giorgi-Rossi P, Carozzi F, Confortini M, Dalla PP, Del MA, et al. Results at recruitment from a randomized controlled trial comparing human papillomavirus testing alone with conventional cytology as the primary cervical cancer screening test. J Natl Cancer Inst 2008;100:492–501. [11] Rijkaart DC, Berkhof J, van Kemenade FJ, Coupe VM, Hesselink AT, Rozendaal L, et al. Evaluation of 14 triage strategies for HPV DNA-positive women in population-based cervical screening. Int J Cancer 2012;130:602–10. [12] Castle PE, Stoler MH, Wright Jr TC, Sharma A, Wright TL, Behrens CM. Performance of carcinogenic human papillomavirus (HPV) testing and HPV16 or HPV18 genotyping for cervical cancer screening of women aged 25 years and older: a subanalysis of the ATHENA study. Lancet Oncol 2011;12:880–90. [13] Garcia F, Barker B, Santos C, Brown EM, Nuno T, Giuliano A, et al. Cross-sectional study of patient- and physician-collected cervical cytology and human papillomavirus. Obstet Gynecol 2003;102:266–72. [14] Brink AA, Meijer CJ, Wiegerinck MA, Nieboer TE, Kruitwagen RF, van Kemenade F, et al. High concordance of results of testing for human papillomavirus in cervicovaginal samples collected by two methods, with comparison of a novel

[27]

[28]

[29]

[30]

[31] [32]

[33]

[34]

[35]

[36]

[37]

63

self-sampling device to a conventional endocervical brush. J Clin Microbiol 2006; 44:2518–23. Khan MJ, Castle PE, Lorincz AT, Wacholder S, Sherman M, Scott DR, et al. The elevated 10-year risk of cervical precancer and cancer in women with human papillomavirus (HPV) type 16 or 18 and the possible utility of type-specific HPV testing in clinical practice. J Natl Cancer Inst 2005;97:1072–9. Wright Jr TC, Stoler MH, Behrens CM, Apple R, Derion T, Wright TL. The ATHENA human papillomavirus study: design, methods, and baseline results. Am J Obstet Gynecol 2012;206:46. Hesselink AT, Heideman DA, Steenbergen RD, Gok M, van Kemenade FJ, Wilting SM, et al. Methylation marker analysis of self-sampled cervico-vaginal lavage specimens to triage high-risk HPV-positive women for colposcopy. Int J Cancer 2014;135:880–6. Bosch FX, Lorincz A, Munoz N, Meijer CJ, Shah KV. The causal relation between human papillomavirus and cervical cancer. J Clin Pathol 2002;55:244–65. Munoz N, Bosch FX, de Sanjose S, Herrero R, Castellsague X, Shah KV, et al. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N Engl J Med 2003;348:518–27. Hesselink AT, Heideman DA, Steenbergen RD, Coupe VM, Overmeer RM, Rijkaart D, et al. Combined promoter methylation analysis of CADM1 and MAL: an objective triage tool for high-risk human papillomavirus DNA-positive women. Clin Cancer Res 2011;17:2459–65. Wentzensen N, Sherman ME, Schiffman M, Wang SS. Utility of methylation markers in cervical cancer early detection: appraisal of the state-of-the-science. Gynecol Oncol 2009;112:293–9. Eijsink JJ, Lendvai A, Deregowski V, Klip HG, Verpooten G, Dehaspe L, et al. A fourgene methylation marker panel as triage test in high-risk human papillomavirus positive patients. Int J Cancer 2012;130:1861–9. Overmeer RM, Louwers JA, Meijer CJ, van Kemenade FJ, Hesselink AT, Daalmeijer NF, et al. Combined CADM1 and MAL promoter methylation analysis to detect (pre-) malignant cervical lesions in high-risk HPV-positive women. Int J Cancer 2011; 129:2218–25. Bierkens M, Hesselink AT, Meijer CJ, Heideman DA, Wisman GB, Van Der Zee AG, et al. CADM1 and MAL promoter methylation levels in hrHPV-positive cervical scrapes increase proportional to degree and duration of underlying cervical disease. Int J Cancer 2013;133:1293–9. Eijsink JJ, Yang N, Lendvai A, Klip HG, Volders HH, Buikema HJ, et al. Detection of cervical neoplasia by DNA methylation analysis in cervico-vaginal lavages, a feasibility study. Gynecol Oncol 2011;120:280–3. Verhoef VM, Bosgraaf RP, van Kemenade FJ, Rozendaal L, Heideman DA, Hesselink AT, et al. Triage by methylation-marker testing versus cytology in women who test HPV-positive on self-collected cervicovaginal specimens (PROHTECT-3): a randomised controlled non-inferiority trial. Lancet Oncol 2014;15:315–22. Verhoef VM, Dijkstra MG, Bosgraaf RP, Hesselink AT, Melchers WJ, Bekkers RL, et al. A second generation cervico-vaginal lavage device shows similar performance as its preceding version with respect to DNA yield and HPV DNA results. BMC Womens Health 2013;13:21. Bulk S, van Kemenade FJ, Rozendaal L, Meijer CJ. The Dutch CISOE-A framework for cytology reporting increases efficacy of screening upon standardisation since 1996. J Clin Pathol 2004;57:388–93. Rijkaart DC, Berkhof J, Rozendaal L, van Kemenade FJ, Bulkmans NW, Heideman DA, et al. Human papillomavirus testing for the detection of high-grade cervical intraepithelial neoplasia and cancer: final results of the POBASCAM randomised controlled trial. Lancet Oncol 2012;13:78–88. Casparie M, Tiebosch AT, Burger G, Blauwgeers H, van de Pol A, van Krieken JH, et al. Pathology databanking and biobanking in The Netherlands, a central role for PALGA, the nationwide histopathology and cytopathology data network and archive. Cell Oncol 2007;29:19–24. Schmitt M, Bravo IG, Snijders PJ, Gissmann L, Pawlita M, Waterboer T. Bead-based multiplex genotyping of human papillomaviruses. J Clin Microbiol 2006;44:504–12. Wilting SM, van Boerdonk RA, Henken FE, Meijer CJ, Diosdado B, Meijer GA, et al. Methylation-mediated silencing and tumour suppressive function of hsa-miR-124 in cervical cancer. Mol Cancer 2010;9:167. Dijkstra M, van Niekerk D, Rijkaart D, van Kemenade FJ, Heideman DA, Snijders P, et al. Primary hrHPV DNA testing in Cervical Cancer screening: how to manage screen positive women? A POBASCAM Trial sub study. Cancer Epidemiol Biomarkers Prev 2013;23:55–63. Cuzick J, Clavel C, Petry KU, Meijer CJ, Hoyer H, Ratnam S, et al. Overview of the European and North American studies on HPV testing in primary cervical cancer screening. Int J Cancer 2006;119:1095–101. Rijkaart DC, Berkhof J, Rozendaal L, van Kemenade FJ, Bulkmans NW, Heideman DA, et al. Human papillomavirus testing for the detection of high-grade cervical intraepithelial neoplasia and cancer: final results of the POBASCAM randomised controlled trial. Lancet Oncol 2012;13:78–88. Bosgraaf RP, Verhoef VM, Massuger LF, Siebers AG, Bulten J, de Kuyper-de Ridder GM, et al. Comparative performance of novel self-sampling methods in detecting high-risk human papillomavirus in 30,130 women not attending cervical screening. Int J Cancer 2014. http://dx.doi.org/10.1002/ijc.29026 [Epub ahead of print]. Gok M, Heideman DA, van Kemenade FJ, de Vries AL, Berkhof J, Rozendaal L, et al. Offering self-sampling for human papillomavirus testing to non-attendees of the cervical screening programme: characteristics of the responders. Eur J Cancer 2011;48:1799–808.