Gynecologic Cytopathology

Acta Cytologica 2012;56:506514 DOI: 10.1159/000338979

Received: December 9, 2011 Accepted after revision: April 16, 2012 Published online: September 27, 2012

Role of p16INK4a Cytology Testing as an Adjunct to Enhance the Diagnostic Specificity and Accuracy in Human Papillomavirus-Positive Women within an Organized Cervical Cancer Screening Program
Daniela Gustinucci a Basilio Passamonti a Elena Cesarini a Daniela Butera b Emiliano Antonio Palmieri b Simonetta Bulletti a Angela Carlani a Maria Staiano b Maria Rosaria DAmico a Valentina DAngelo a Eugenio Di Dato a Nadia Martinelli a Morena Malaspina a Nicoletta Spita a Beatrice Tintori a Franco Fulciniti b
a b

Azienda Sanitaria Regionale dellUmbria USL2, U.O.C. Diagnostica di Laboratorio, U.O. Citologia, Perugia, S.S.D. di Citopatologia, U.O.C. di Anatomia Patologica e Citopatologia, Istituto Nazionale Tumori Fondazione G. Pascale, Napoli, Italia

Key Words Cervical screening p16INK4a Human papillomavirus High-risk human papillomavirus genotypes Triage Atypical squamous cells of undetermined significance Low-grade squamous intraepithelial lesion High-grade squamous intraepithelial lesion Atypical squamous cells, cannot exclude high-grade intraepithelial squamous lesion

2012 S. Karger AG, Basel 00015547/12/05650506$38.00/0 Fax +41 61 306 12 34 E-Mail karger@karger.ch www.karger.com Accessible online at: www.karger.com/acy

Correspondence to: Dr. Franco Fulciniti S.S.D. di Citopatologia, U.O.C. di Anatomia Patologica Via M. Semmola IT80131 Naples (Italy) Tel. +39 08 1590 3849, E-Mail franco.fulciniti@gmail.com

Downloaded by: 187.151.28.231 - 11/21/2013 7:37:29 AM

Abstract Objective: We evaluated the performance of cytologic p16INK4a (p16) immunostaining within a cervical cancer screening program for the categories of atypical squamous cells of undetermined significance (ASC-US) and low-grade squamous intraepithelial lesion (LS after triage with highrisk human papillomavirus (HR-HPV) testing and atypical squamous cells, cannot exclude high-grade intraepithelial squamous lesion (ASC-H) and high-grade squamous intraepithelial lesion (HSIL). We also verified whether the routine introduction of p16 staining might enhance the specificity and positive predictive value (PPV) for cervical intraepithelial neoplasia grade 2 or higher (CIN2+) lesions predicted by a cytological screening test. Study Design: Performance of

the p16 cytology test was estimated in 578 cytological samples, of which 213 were HR-HPV+ ASC-US, 186 were HR-HPV+ LSIL, 74 were ASC-H, 56 were HSIL-CIN2 and 49 were HSILCIN3. All samples had histological follow-up. Results: In the ASC-US category, p16 sensitivity was 91% for CIN2+ and 100% for CIN3, while specificity was 64 and 58%, respectively, negative predictive value (NPV) was 96 and 100%, respectively, and PPV was 39%. In the LSIL category, sensitivity was 77 and 75%, respectively, for CIN2+ and CIN3, while specificity was 64 and 57%, NPV was 93 and 98% and PPV was 30%. Sensitivity for ASC-H and HSIL-CIN3 was 100% for CIN2+ and CIN3, while for HSIL-CIN2 it was 91 and 95%, respectively; NPV for ASC-H was 100%, and for HSIL-CIN2 it was 43 and 86%, respectively. Follow-up examinations of 8 cases diagnosed as p16+ ASC-H and HSIL-CIN3, but histologically negative or CIN1 on the first biopsy, showed 4 CIN2 and 4 CIN3 lesions. Conclusions: Sensitivity, specificity, PPV and NPV confirm the importance of the utilization of p16 in the categories ASC-US and LSIL after triage with an HR-HPV test. In the ASC-H and HSIL-CIN3 lesions, p16 was shown to be an excellent marker for picking up CIN2+ lesions, especially in cases with cytohistological discordance.
Copyright 2012 S. Karger AG, Basel

Introduction

Role of p16 Stain in HR-HPV+ Women in a Cervical Cancer Screening Program

Acta Cytologica 2012;56:506514

507

Downloaded by: 187.151.28.231 - 11/21/2013 7:37:29 AM

Carcinoma of the uterine cervix ranks second in the world with respect to cancer-related incidence and mortality in women. The implementation of organized screening programs based on regular Pap cytology testing has been demonstrated to be a valuable tool for secondary prevention in many countries. However, one of the various limitations of Pap cytology-based testing is the presence of equivocal or mildly abnormal diagnostic categories, such as atypical squamous cells of undetermined significance (ASC-US) or low-grade squamous intraepithelial lesion (LSIL), according to the revised Bethesda classification [1]. In these cytologic categories, there is a substantial interobserver variability and a high degree of discordance with histological follow-up [2]. Routine referral of patients with equivocal cytology results to colposcopy may cause unnecessary anxiety and an excessive use of public funds which is not justified by the results. Epidemiological and molecular studies have shown that 99.7% of cervical carcinomas host genomic sequences of human papillomavirus (HPV) [3, 4], and it is now well known that persistent high-risk HPV (HRHPV) infections are the dominant causative etiological agents for the development of high-grade precancerous lesions that may progress to invasive carcinomas [5]. However, in reality, most HPV infections in young patients are transient [6, 7], and only HR-HPV-associated lesions that actively express the E6 and E7 oncoproteins in replication-competent epithelial cells may cause neoplastic transformation (transforming HPV infections) [6]. HR-HPV testing by various modalities has been validated in various studies, particularly for the triage of Pap cytology results categorized as ASC-US. It has been found to be helpful in the management of cases of equivocal cytology [810]. In fact, its usage has permitted a reduction in the number of patients referred to colposcopy, although there is substantial room for additional improvement. As a matter of fact, HPV testing has a high sensitivity for high-grade (grade 2 or higher) cervical intraepithelial neoplasia (CIN2+) when used for the triage of abnormal Pap cytology results, but generally a low specificity [11, 12], as it does not discriminate between transforming HPV infections and the large majority of transient infections. Various studies have analyzed the potential role of a biomarker called p16INK4a (p16), a cyclin-dependent kinase inhibitor, which is overexpressed in almost all of the high-grade squamous intraepithelial lesions (HSIL) of the cervix uteri and may be detected by immunocytochemi-

cal staining [1316]. The molecular mechanisms leading to the overexpression of p16 are well understood and were shown at the molecular level to be directly linked to the transforming activity of the viral E7 oncoprotein [14, 15, 17]. p16 may be expressed at detectable but typically low levels in individual epithelial cells of the lower genital tract, including mature and immature metaplastic cells, squamous atrophy, endocervical cells and endometrial cells [18], but it is usually strongly expressed in cells from high-grade precancerous and cancerous cervical lesions. The diagnostic usefulness of the immunohistochemical staining of cervical biopsies for p16 has been shown in many studies and is widely acknowledged [13, 15]. In cervical cytology samples, various studies [1820] have analyzed the potential utility of p16 staining for the triage of Pap cytology cases with equivocal or mildly abnormal results, as an alternative to the HPV test [12]. The aim of our study was to evaluate the usefulness of p16 immunocytochemical staining in various groups of Pap cytology cases as follows: (1) in ASC-US or LSIL in a cohort of HPV-positive patients to improve the specificity of the triage; (2) in the subgroup of Pap cytology cases classified as atypical squamous cells, cannot exclude high-grade intraepithelial squamous lesion (ASC-H) in order to better discriminate which lesions in this category will be histologically classified as CIN2 or higher (CIN2+) on cervical biopsies, and (3) in Pap cytology cases diagnosed as HSIL to discriminate which of those will be correctly confirmed as having underlying CIN2+ on cervical biopsy.

Material and Methods

Study Design Our Cytology Unit is the regional cervical screening center for the prevention of cervical cancer in the territory of the province of Perugia, Italy. The unit drains three different regional screening centers located in different local health authorities, with a total of 171,000 referred patients in the age range 2564 years. The total number of cervical smears taken (organized screening plus spontaneous screening) amounts to 50,000/year. The cytological reporting scheme is based on the Bethesda 2001 reporting system with its modifications. In the period considered for the purposes of the analyses reported in this paper (June 2007June 2009), 92,742 conventional Pap smears were considered, of which 60,975 (65.7%) were derived from organized screening programs. Among this latter group, 1,388 abnormal Pap smears (2.2% of all smears from our screening cohort) were identified, of which 600 (0.98%) were classified as ASC-US, 360 (0.59%) as LSIL, 156 (0.25%) as ASC-H, 214 (0.35%) as HSIL and 58 (0.09%) as atypical glandular cells. HR-HPV triage was started in our region on 1 June 2007 for smears classified as ASC-US, and from 1 January 2008 also for the

June 2007June 2009 PAP test screening 60,975

Abnormal PAP test 1,388 (2.2%)

ASC-US 600 (0.98%)

LSIL 360 (0.59%)

ASC-H 156 (0.25%)

HSIL 214 (0.35%)

AGC 58 (0.09%)

HR-HPV+ 275 (46%)

HR-HPV+ 268 (74%)

Collected during the colposcopy visit 87 (56%)

Collected during the colposcopy visit 127 (59%)

Collected during the colposcopy visit 0 (0%)

p16 stains performed 757

Samples with histology, follow-up and p16 staining evaluable 578 (76%)

Fig. 1. Study flow chart. AGC = Atypical

ASC-US 213 (37%)

LSIL 186 (32%)

ASC-H 74 (13%)

HSIL 105 (18%)

glandular cells.

LSIL cytologic category. The triage algorithm was implemented in agreement with European guidelines and the indications proposed by the Italian Group for Cervical Cancer Screening (Gruppo Italiano Screening del Cervicocarcinoma). It contemplates a second-level screening evaluation of HR-HPV-positive women with ASC-US or LSIL, while women with higher degrees of cytological abnormalities are sent directly to colposcopic evaluation. Case Selection and Sample Preparation In this paper, samples from 757 women aged 2564 years (mean age 38) with abnormal screening Pap cytology results were studied by immunocytochemical staining for p16 and HR-HPV testing. Both test methods were performed on a repeat cervical sample. Diagnostic categories were represented as follows: 275 ASC-US, 268 LSIL, 87 ASC-H and 127 HSIL. The 127 lesions diagnosed as HSIL were further subclassified for internal study purposes as HSIL favor CIN2 (n = 70) or HSIL favor severe dysplasia/CIN3 (n = 57; fig.1). The 275 smears classified as ASC-US and the 268 classified as LSIL included in this analysis represent the subgroup of cases that had tested positive for HR-HPV using the Digene HPV HC2 Hybrid Capture test (Qiagen) on a new sample taken using liquid-based cytology vials (Thin Prep Pap test, Hologic). The remaining cases (87 ASC-H, 70 HSIL favor CIN2 and 57 HSIL favor CIN3) were restudied with liquid-based cytology (Thin Prep Pap test) samples collected during the colposcopy visit performed as a second-level investigation. The cytological sample was taken by a modified Cervex-type

508

Acta Cytologica 2012;56:506514

Gustinucci et al.

Downloaded by: 187.151.28.231 - 11/21/2013 7:37:29 AM

spatula and endocervical brushing. The liquid-based cytology slides were prepared using a T2000 slide processor (Hologic). In all the repeat samples taken by the liquid-based technology, when atypical cells consistent with the original cytological diagnosis made on the traditional smear were lacking, p16 immunocytochemistry was performed on the original, conventional smears. Previous literature has, in fact, demonstrated that the same sensitivity and specificity between both conventional smears and thin-layer cytology preparations [2123]. The same conclusions were reached in a recent statement issued by the Italian Health Department concerning cervical screening (Art. 2 bis L 138/2004 and National Planning of Prevention 2005/2007). As stated above, in this study, all HR-HPV+ patients having an initial cytological diagnosis of ASC-US, LSIL, ASC-H or HSIL are routinely sent to second-level workup by colposcopy and biopsy if the latter is advised or deemed necessary. Results of these examinations are recorded in the files of our Regional Cytopathology Laboratory. A total of 179 out of 757 samples (24%) were excluded from this study due to missing biopsy follow-up or inadequate tissue samples, or because the p16 staining was not available. In total, 578 samples (76%) were included in this study, of which 213 (37%) were ASC-US, 186 (32%) were LSIL, 74 (13%) were ASC-H, 56 (10%) were HSIL-CIN2 and 49 (8%) were HSIL-CIN3, all with histological follow-up (fig.1). In 390 of these cases (67%), the results of the colposcopic examinations were known; while in the remaining 188 cases (33%), results were not immediately available, and they were therefore excluded from this study.

Color version available online

Fig. 2. Liquid-based cytology sample. This case had been cytologically diagnosed as LSIL. p16 staining shows intense nuclear positivity in slightly atypical superficial squamous cells. Immunocytochemistry. Original magnification !600. Papanicolaou counterstain.

Fig. 4. Liquid-based cytology sample. This case had been cyto-

logically diagnosed as HSIL favor CIN3. Notice intense nucleocytoplasmic p16 positivity in a sheet of atypical parabasal-like cells. Immunocytochemistry. Original magnification !600. Papanicolaou counterstain.

Color version available online

their coverslip by incubation in xylene and subsequently rehydrated through descending alcohol concentrations in triphosphate saline buffer (pH 7.4) before starting the immunostaining protocol. No additional, separate destaining procedure was performed since the slides are destained during the routinely used epitope retrieval step specified for the CINtec Cytology kit. For each staining run, a control slide containing both positive and negative control cells with respect to the immunoreactivity for p16 was used to qualify the staining procedure. The immunocytochemical staining was performed using an automated immunostainer (i6000 by Genomix). The slides were contrasted by the Papanicolaou stain [24] after the immunocytochemical staining. p16 Cytology Slide Interpretation In order to permit a faster and more certain interpretation of the staining, before p16 immunostaining, both traditional and thin-layer Pap smears underwent cytological evaluation and were marked on the back with a diamond-tip pen corresponding to the microscopic fields bearing morphologically significant cell changes. This technique reduces the risk of attributing a falsepositive score to metaplastic and cylindrical cells devoid of cytopathological changes. The immunocytochemical reaction was considered positive if at least one atypical stained cell was contained within the previously marked areas or if there were evident cytopathological atypias in conjunction with nuclear, cytoplasmic or nucleocytoplasmic p16 staining, even if outside the marked fields [25]. Papanicolaou restaining of immunocytochemically stained slides permits an easier global reevaluation of the slides compared to Harris Hematoxylin counterstain, because the overall features of the cytological slides are maintained in this way [24] (fig.24). All slides stained for p16 were evaluated by two laboratory operators blinded to the subsequent histology.

Fig. 3. Liquid-based cytology sample. This case had been cyto-

logically diagnosed as ASCH-H. Notice intense nucleocytoplasmic positivity for p16 in an atypical parabasal-like epithelial cell. Immunocytochemistry. Original magnification !600. Papanicolaou counterstain.

Role of p16 Stain in HR-HPV+ Women in a Cervical Cancer Screening Program

Acta Cytologica 2012;56:506514

509

Downloaded by: 187.151.28.231 - 11/21/2013 7:37:29 AM

p16 Immunocytochemistry p16 immunocytochemical staining of cervical cytology samples was performed using the CINtec Cytology Kit (MTM Laboratories, Milan, Italy), which is based on the validated antibody clone E6H4, according to the manufacturers instructions. We used routinely stained Pap smears which had been deprived of

Color version available online

Statistical Analysis Estimates of the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of p16 cytology testing for histologically confirmed CIN2+/CIN3+ were calculated from cross-tabulation of binary test results and are presented with their exact 95% confidence intervals. For study purposes, all histological follow-up results (including surgical specimens from ablative treatment) obtained within 1 year from the date of referral to colposcopy were used to establish the clinical endpoint. The most severe diagnosis for each woman was used.

Table 1. p16 positivity rates and distribution of p16 immunoreac-

tivity in different diagnostic categories Cytologic diagnosis ASC-US LSIL ASC-H HSIL-CIN2 HSIL-CIN3 Total p16+ n 100 80 67 49 49 345 % 47 43 90 87 100 60 p16 n 113 106 7 7 0 233 % 53 57 10 13 0 40 213 186 74 56 49 578 Total

Results

510

Acta Cytologica 2012;56:506514

Gustinucci et al.

Downloaded by: 187.151.28.231 - 11/21/2013 7:37:29 AM

The results are summarized in figure 1. With regard to the ASC-US and LSIL diagnostic categories, p16 immunostaining was performed only on HR-HPV-positive samples, which represented 49% of all ASC-US cases (30 50% in the New Technologies in Cervical Cancer study) and 75% of all LSIL cases (5080% in the New Technologies in Cervical Cancer study). p16 immunostaining was performed in all patients with ASC-H or HSIL cytologic results, for which referral to colposcopy is routinely adopted. Table 1 shows the distribution of p16 immunocytochemistry positivity rates for the various diagnostic categories. Test positivity rates increase with the severity of the cytologic lesion, with values lower than 50% in the group of HR-HPV-positive ASC-US and LSIL cases and higher than 87% in categories comprising high-grade lesions. The correlation between the individual Pap cytology categories, their p16 immunoreactivity status and histologic diagnoses established on cervical biopsies is shown in table2. The PPVs for the detection of underlying CIN2+ obtained for ASC-US and LSIL Pap cytology results after HR-HPV triage as well as for ASC-H, HSIL-CIN2 and HSIL-CIN3 results were compared to those obtained after subsequent p16 immunostaining for the respective diagnostic cytology classes (table3). An increase in the PPV of 19% was observed in the ASC-US/HR-HPV+ category, and increases of 13, 6 and 3% were seen in LSIL/HRHPV+, ASC-H and HSIL-CIN2, respectively. For the group of HSIL-CIN3 cytology results, the PPV for underlying CIN2+ was not further improved. Table 4 summarizes the sensitivity, specificity and NPV estimates for underlying CIN2+ or CIN3+ disease in the various diagnostic categories tested for p16. In the ASC-US/HR-HPV+ category, the sensitivity of p16 cytology was 91% for CIN2+ and 100% for CIN3, with a re-

spective specificity of 64 and 58%. A slightly lower sensitivity was recorded in the LSIL/HR-HPV+ category, i.e. 77 and 75% for CIN2+ and CIN3+, respectively, with a corresponding specificity of 64 and 57%. The most interesting value in the ASC-US/HR-HPV+ category is represented by the NPV, which was found to be 96 and 100% for CIN2+ and CIN3+, respectively, and 93 and 98% for CIN2+ and CIN3+, respectively, in the LSIL/HR-HPV+ category. In ASC-H, the sensitivity and NPV of p16 cytology were 100% for both CIN2+ and CIN3+. Specificity was found to be 26% (CIN2+ threshold) and 15% (CIN3+), and the PPV of positive p16 cytology results for underlying CIN2+ was found to be 6% higher. In the HSIL-CIN2 category, the sensitivity of p16 for CIN2+ and CIN3+ was 91 and 95%, respectively, and the NPV of a negative p16 test result for CIN2+ and CIN3+ was 43 and 86%, respectively. As expected, this category showed a rather low specificity value (23 and 17% for CIN2+ and CIN3+, respectively), due to the small number of p16-negative smears, similar to the ASC-H category. All the HSIL-CIN3 cytology cases were p16 positive, with a sensitivity of 100% for underlying high-grade disease. Table 5 shows the distribution of the colposcopic scores in relation to the cytologic diagnostic categories and to subsequent histopathological diagnoses for the 390 patients with Pap cytology smears for which colposcopic follow-up results were available. Of the HRHPV-positive ASC-US and LSIL cases that showed CIN2+ on biopsy, 77% had a G1 colposcopic score, correlating with atypias of minor significance, while for HSIL cases, only 37% were evaluated as G1, while the remaining 63% were classified as G2.

Table 2. Distribution of histologic diagnoses with respect to the different cytologic categories and p16 immunoreactivity

Cytologic diagnosis

Histologic diagnosis negative p16+ p16 25 (71) 17 (59) 1 (17) 2 (33) 0 45 (57) CIN1 p16+ 51 (38) 44 (35) 15 (71) 6 (86) 2 (100) 118 (41) p16 84 (62) 82 (65) 6 (29) 1 (14) 0 173 (59) CIN2 p16+ 22 (85) 18 (78) 14 (100) 20 (87) 4 (100) 78 (87) p16 4 (15) 5 (22) 0 3 (13) 0 12 (13) CIN3 p16+ 17 (100) 6 (75) 33 (100) 19 (95) 40 (100) 115 (97) p16 0 2 (25) 0 1 (5) 0 3 (3)

Total

ASC-US LSIL ASC-H HSIL-CIN2 HSIL-CIN3 Total

10 (29) 12 (41) 5 (83) 4 (67) 3 (100) 34 (43)

213 186 74 56 49 578

Numbers in parentheses represent percentages and refer to the presence or absence of p16 staining within diagnostic classes or histologic diagnoses.

Table 3. Comparison between the PPVs vs. CIN2+ of HR-HPV+ ASC-US and LSIL, ASC-H, HSIL-CIN2 and HSIL-CIN3 results without and with p16 testing in the same diagnostic categories

PPV 95% CI CIN2+ HR-HPV+ ASC-US HR-HPV+ LSIL ASC-H HSIL-CIN2 HSIL-CIN3 All diagnostic categories 20% 17% 64% 77% 90% 36% 0.260.15 0.220.11 0.740.53 0.880.66 0.980.81 0.400.32

95% CI PPV CIN2+ (p16+) 39% 30% 70% 80% 90% 56% 0.490.29 0.400.20 0.810.59 0.910.68 0.980.81 0.610.51

CI = Confidence interval.

Discussion

Role of p16 Stain in HR-HPV+ Women in a Cervical Cancer Screening Program

Acta Cytologica 2012;56:506514

511

Downloaded by: 187.151.28.231 - 11/21/2013 7:37:29 AM

The aim of this paper was to evaluate the impact of p16 immunostaining in gynecological cytologic samples that had been classified as ASC-US or LSIL using the Bethesda classification and subsequently tested positive for HRHPV, or which had been classified as ASC-H or HSIL. Our intention for these cases was to verify whether the introduction of p16 staining could cause a possible reduction of public costs in the national health service, while helping to identify patients really needing further diagnostic and therapeutic workup in the various diagnostic categories and at the same time avoiding unjustified anxiety associated with an abnormal test result.

It is well known that HR-HPV testing, though a highly sensitive diagnostic test and hence potentially useful in the triage of low-grade squamous atypias like ASC-US and LSIL, has a low specificity in the identification of HSIL [12]. We also know that p16 expression in atypical cells is related to a disturbance in cell cycle regulation induced by the oncogenic E7 HR-HPV viral protein, favoring a higher probability of neoplastic transformation. On the other hand, as also shown in table5, colposcopic evaluation has a limited value in predicting which lesions will fall in the CIN2+ histologic category, especially when a previous diagnosis of low-grade cytologic atypia (ASCUS or LSIL) has been formulated. In fact, in these groups of cytologic diagnoses, CIN2+ lesions were found in 77% of the cases in colposcopic categories scored as G1, while only 23% of the cases were scored as G2 during colposcopy. These histological lesions are probably too deeply located in the transformation zone and in the endocervical canal and so they might be scarcely evident on Pap smears and even on more extensive colposcopic inspections of the cervix. This latter technique demonstrated a higher sensitivity in the cytologic diagnostic category of HSIL, in which 63% of CIN2+ lesions were scored as G2. By considering all cytologic diagnostic categories, 97% of histologically diagnosed CIN3+ lesions were positively stained by p16. Only 3 cases with biopsy-confirmed CIN3+ were not immunoreactive for p16. A possible explanation for the lack of p16 reactivity in CIN 3 lesions may be the presence of a mutation in the p16 gene (deletion or promoter hypermethylation), this latter event being observed especially in young women (27 to 30 years old) who are heavy smokers [26].

Table 4. Sensitivity, specificity and NPV of p16 testing for underlying CIN2+ or CIN3+ disease

Diagnostic category HR-HPV+ ASC-US HR-HPV+ LSIL ASC-H HSIL-CIN2 HSIL-CIN3 CIN2+ CIN3 CIN2+ CIN3 CIN2+ CIN3 CIN2+ CIN3 CIN2+ CIN3

Sensitivity % 91 100 77 75 100 100 91 95 100 100

95% CI 0.990.82 1.001.00 0.920.63 1.050.45 1.001.00 1.001.00 0.990.82 1.050.85 1.001.00 1.001.00

Specificity % 64 58 64 57 26 17 23 17 0 0

95% CI 0.710.57 0.650.51 0.710.56 0.650.50 0.420.09 0.290.06 0.460.00 0.290.04 0.000.00 0.000.00

NPV % 96 100 93 98 100 100 43 86 0 0

95% CI 1.000.93 1.001.00 0.980.89 1.010.95 1.001.00 1.001.00 0.800.06 1.120.60 0.000.00 0.000.00

Table 5. Distribution of colposcopic grades with respect to the diagnostic categories and subsequent histologic diagnoses

Cytologic diagnosis

Histologic diagnosis negative G0 G1 G2 CIN1 G0 G1 G2 CIN2 G0 1 (3) 0 0 1 (1) G1 G2 CIN3 G0 0 0 0 0 G1 G2

Total

ASC-US/LSIL ASCH HSIL Total

16 (39) 25 (61) 0 0 1 (33) 2 (67) 2 (29) 2 (29) 3 (42) 18 (35) 28 (55) 5 (10)

8 (5) 146 (92) 5 (3) 2 (17) 9 (75) 1 (8) 0 6 (86) 1 (14) 10 (6) 161 (90) 7 (4)

30 (81) 6 (16) 8 (62) 5 (38) 13 (54) 11 (46) 51 (69) 22 (30)

13 (93) 1 (7) 17 (65) 9 (35) 13 (28) 34 (72) 43 (49) 44 (51)

251 54 85 390

Numbers in parentheses refer to the percent distribution of colposcopic grades with respect to the cyto-histologic diagnostic categories.

512

Acta Cytologica 2012;56:506514

Gustinucci et al.

Downloaded by: 187.151.28.231 - 11/21/2013 7:37:29 AM

With regard to CIN2 lesions, 87% of these were p16 positive. p16 negativity in the remaining 12 cases with a histological diagnosis of CIN2 may be explained by the significant possibility of a spontaneous regression of the respective lesion and also by the diagnostic challenge associated with the differential diagnosis of CIN2 versus CIN1, especially when evaluating histologically bizarre lesions. European 2008 guidelines quote a table by Ostor [27], who already in 1993 had reported a regression rate of 40% for histologic CIN2 lesions; subsequently, Castle et al. [28] reconfirmed these data, in particular for lesions not caused by HPV 16, and most recently, Milicic-Juhas et al. [29] demonstrated a regression rate of up to 51% for CIN2. By evaluating the specific impact within the single diagnostic categories in HR-HPV-positive women, it can be noted that the PPV of p16-positive ASC-US cases was improved to 39%, compared to a PPV of 20% for HR-HPVpositive ASC-US results only. The reported sensitivity for the p16 cytology test of 91% for CIN2+ and 100% for

CIN3+ in our series is largely comparable to that reported in the literature for the HPV test [12]. Specificity values of 64% (CIN2+) and 58% (CIN3+) almost doubled for positive p16 stains. Probably the most significant data are represented by the high NPV of 96% for CIN2+ and 100% for CIN3+ lesions associated with a negative p16 staining result. This means that p16-negative ASC-US lesions have an almost null probability of harboring a high-grade lesion. Table2 shows that only 47% of HR-HPV-positive ASC-US lesions were p16 positive; hence, the utilization of this biomarker to discriminate patients to be sent to colposcopy might further reduce their number by 50% compared to the already lower number of referrals due to HR-HPV testing (average positivity in our center had been 49%). The analysis of the LSIL category leads to similar considerations. In fact, as for the ASC-US category, the PPV of p16-positive, HR-HPV-positive LSIL cytology results was 30%, a value almost doubled compared to that obtained for the HR-HPV test alone (17%). Sensitivity was

Role of p16 Stain in HR-HPV+ Women in a Cervical Cancer Screening Program

Acta Cytologica 2012;56:506514

513

Downloaded by: 187.151.28.231 - 11/21/2013 7:37:29 AM

77% for CIN2+ and 75% for CIN3+, lower than that recorded for HR-HPV-positive ASC-US, but the NPV for CIN2+ and CIN3+ was 93 and 98%, respectively. Specificity data overlap perfectly: 64% for CIN2+ and 57% for CIN3+. Considering that the HR-HPV test reduces on average only 25% of referrals to colposcopy, the routine introduction of p16 testing would permit reduction in the number of diagnostic workups by a further 57%, as shown in table2. The ASC-H category was introduced in this study since it is an expression of a generic cytological suspicion versus a high-grade lesion. We wanted to speculate whether p16 staining could function as a marker to distinguish histologically negative ASC-H and ASC-H cases with CIN1 histology from ASC-H cases with histologically confirmed CIN2+, with a resulting increase in the PPV, which in our laboratory varies between 50 and 60% in this group of lesions. The analysis of results has shown the extreme usefulness of p16 staining to identify highgrade lesions, with a sensitivity of 100%; none of the 7 p16-negative samples shown to be CIN2+ on biopsy (NPV 100%), nor indeed any of these samples, became positive during follow-up over the subsequent screening interval (between 12 and 18 months). The PPV did not show a substantial increase (6%), and specificity was not shown to be significantly high (26% for CIN2+ and 16% for CIN3+). In fact, 20 of the p16-positive Pap cytology tests classified as ASC-H had a negative or CIN1 biopsy result, which might of course also be related to the known nonoptimal sensitivity of colposcopic procedures for the detection of existing high-grade disease. In the ASC-H group, the results of the next follow-up after the first diagnostic procedure (biopsy or cone biopsy) were also tabulated; in 5 cases we registered a recurrence after conization, a CIN3 in 1 case and a CIN2 in 4 cases. So, if p16 staining in this group of lesions does not seem to be useful to discriminate patients to send to colposcopic examination (190% of ASC-H were p16+), it might be useful in intralaboratory controversial cases or in those with negative findings on biopsy (by representing an indication for close follow-up). Moreover, it could represent an instrument (as in our laboratories) to refine the cytologic diagnostic categorization and could help to alert the colposcopist to a closer scrutiny of some patients. The utilization of p16 staining on HSIL lesions needs a foreword. In our laboratory, we adopted the Bethesda classification system and hence also the usage of the HSIL category, comprising under the same heading CIN2 and CIN3/in situ carcinoma. While adopting this terminology, we kept the habit of indicating in our reports whether we favor CIN2 or CIN3 in each case. The PPVs as cal-

culated as part of our yearly quality control program are very satisfying for the HSIL-CIN3 group (about 89%) but less satisfying for HSIL-CIN2 (about 65%), because a relevant percentage of these latter have a CIN1 biopsy result. Thus, p16 testing in HSIL-CIN2 might have been useful in this group of lesions, as it is also in histology, to refine the cytological diagnosis with a subsequent increase in the PPV. The comparison with the performance of the p16 test in CIN3 could have served as a referral point. The analysis of these results generates some observations. The PPV for HSIL-CIN2 in our study population (77%) was notably higher than the values derived from our periodic quality controls (about 65%). The contribution of the added 3% increment in the PPV due to p16 staining in this group is more difficult to evaluate. There were 49 p16positive samples in total, of which 39 (80%) showed CIN2+ on biopsy, while 10 cases (20%) showed CIN1 or were negative for dysplasia. Of these latter 10 cases, 9 had follow-up in a period between 6 and 24 months with subsequent cytologies or histologies confirming the initial diagnosis. So, in none of these cases was a progression observed in the grade of the initial lesion. Moreover, notwithstanding the high sensitivity of 91 and 95% for CIN2+ and CIN3+, respectively, the NPV was 43% (among the 7 p16-negative smears, 3 resulted in CIN2 and 1 in CIN3 on histology). The obtained results do not seem to encourage p16 testing in the HSIL-CIN2 category, since p16 positivity does not seem to add discriminating power between negative or CIN1 lesions on biopsy versus CIN2+ lesions. A different pattern was found in the 49 HSIL-CIN3 cases. All of these cases tested p16 positive, with a PPV of 90% and sensitivity of 100% for both CIN2+ and CIN3+. The most interesting data emerged from the follow-up of the 5 p16-positive HSIL-CIN3 cytology results with either negative or CIN1 biopsy results. Of these, 4 were subsequently biopsied and/or treated, with 3 cone resections and 1 biopsy performed within 1215 months with a CIN3 result. p16 staining is a wonderful marker for the presence of high-grade lesions. Its usefulness in screening still needs to be determined, but our analyses already indicate that it might represent an important diagnostic tool in cases with discordant cytohistologic correlation, for colposcopists for a closer scrutiny and in follow-up of p16-positive patients.
Acknowledgements
The authors would like to thank Dr. Ornella Sacco for data management and Drs. Alessandra Trocino and Rosario Romanelli for their excellent bibliographical services.

References
1 Solomon D, Davey D, Kurman R, Moriarty A, OConnor D, Prey M, Raab S, Sherman M, Wilbur D, Wright T Jr, Young N; Forum Group Members; Bethesda 2001 Workshop: The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA 2002;287:21142119. 2 Stoler MH, Schiffman M; Atypical Squamous Cells of Undetermined SignificanceLow-grade Squamous Intraepithelial Lesion Triage Study (ALTS) Group: Interobserver reproducibility of cervical cytologic and histologic interpretations: realistic estimates from the ASCUS-LSIL Triage Study. JAMA 2001;285:15001505. 3 Parkin DM, Bray F, Ferlay J, Pisani P: Estimating the world cancer burden: Globocan 2000. Int J Cancer 2001;94:153156. 4 Walboomers JM, Jacobs MV, Manos MM, et al: Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999; 189:1219. 5 Bosch FX, Manos MM, Munoz N, Sherman M, Jansen AM, Peto J, Schiffman MH, Moreno V, Kurman R, Shah KV: Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International Biological Study on Cervical Cancer (IBSCC) Study Group. J Natl Cancer Inst 1995; 87: 796802. 6 Ho GY, Bierman R, Beardsley L, Chang CJ, Burk RD: Natural history of cervicovaginal papillomavirus infection in young women. N Eng J Med 1998;338:423428. 7 Schiffman M, Castle PE, Jeronimo J, et al: Human papillomavirus and cervical cancer. Lancet 2007;370:890907. 8 Solomon D, Schiffman M, Tarone R; ALTS Study Group: Comparison of three management strategies for patients with atypical squamous cells of undetermined significance: baseline results from a randomized trial. J Natl Cancer Inst 2001;93:293299. 9 Buntinx F, Van Ranst M, Paraskevaidis E, Martin-Hirsch J, Dillner J: Virologic versus cytologic triage of women with equivocal Pap smears: a meta-analysis of the accuracy to detect high-grade intraepithelial neoplasia. J Natl Cancer Inst 2004;96:280293. 10 Solomon D, Schiffman M: Have we resolved how to triage equivocal cervical cytology? J Natl Cancer Inst 2004;96:250251. 11 Szarewski A, Ambroisine L, Cadman L, Austin J, et al: Comparison of predictor for highgrade cervical intraepithelial neoplasia in women with abnormal smears. Cancer Epidemiol Biomarkers Prev 2008;17:30333042. 12 Denton KJ, Bergeron C, Klement P, Trunk MJ, et al: The sensitivity and specificity of p16INK4a cytology vs. HPV testing for detecting high-grade cervical disease in the triage of ASC-US and LSIL pap cytology results. Am J Clin Pathol 2010; 134:1221. 13 Cuschieri K, Wentzensen N: Human papillomavirus and p16 detection as biomarkers for the improved diagnosis of cervical neoplasia. Cancer Epidemiol Biomarkers Prev 2008;17:25362545. 14 Wentzensen N, von Knebel Doeberitz M: Biomarkers in cervical cancer screening. Dis Markers 2007;23:315330. 15 Tsoumpou I, Arbyn M, Kyrgiou M, et al: p16INK4a immunostaining in cytological and histological specimens from the uterine cervix: a systematic review and meta-analysis. Cancer Treat Rev 2009;35:210220. 16 Klaes R, Friedrich T, Spitkovsky D, et al: Overexpression of p16INK4a as a specific marker for dysplastic and neoplastic epithelial cells of the cervix uteri. Int J Cancer 2001; 92:276284. 17 Bibbo M, Klump WJ, De Cecco J, Kovatich AJ: Procedure for immunocytochemical detection of p16INK4a antigen in thin-layer, liquid-based specimens. Acta Cytol 2002; 46: 2529. 18 Bibbo M, De Cecco J, Kovatich AJ: p16INK4a as an adjunct in liquid-based cytology. Anal Quant Cytol Histol 2003;25:811. 19 Klaes R, Friedrich T, Spitkovsky D, et al: Over-expression of p16INK4a as a specific marker for dysplastic and neoplastic epithelial cells of the cervix uteri. Int J Cancer 2001; 92:276284. 20 Agoff N, Lin P, Morihara J, Mao C, Kiviat N, Koutsky L: p16INK4a expression correlates with degree of cervical neoplasia: a comparison with Ki-67 expression and detection of high-risk HPV types. Mod Pathol 2003; 16: 665673. 21 Arbyn M, Bergeron C, Klinkhamer P, Martin-Hirsch P, Siebers AG, Bulten J: Liquid compared with conventional cytology: a systematic review and meta-analysis. Obstet Gynecol 2008;111:167177. 22 Obwegeser J, Brack S: Does liquid-based technology really improve detection of cervical neoplasia? A prospective, randomized trial comparing the ThinPrep Pap Test with the conventional Pap Test, including followup of HSIL cases. Acta Cytol 2001; 45: 709 714. 23 Davey E, Barrat A, Irwig L: Effect of study design and quality on unsatisfactory rates, cytology classification, and accuracy in liquid-based versus conventional cervical cytology: a systematic review. Lancet 2006;367: 122132. 24 Negri G, Moretto G, Menia E, et al: p16INK4a immunocytochemistry in liquid-based cervico-vaginal specimens with modified Papanicolaou counterstaining. J Clin Pathol 2006;59:827830. 25 Trunk MJ, Dallenbach-Hellweg G, Ridder R, et al: Morphologic characteristics of p16INK4a positive cells in cervical cytology samples. Acta Cytol 2004;48:771782. 26 Lea JS, Coleman R, Kurien A, Schorge J, et al: Aberrant p16 methylation is a biomarker for tobacco exposure in cervical squamous cell carcinogenesis. Am J Obstet Gynecol 2004; 190:674679. 27 Ostor AG: Natural history of cervical intraepithelial neoplasia: a critical review. Int J Gynecol Pathol 1993; 12:186192. 28 Castle PE, Schiffman M, Wheeler CM, Solomon D: Evidence for frequent regression of cervical intraepithelial neoplasia-grade 2. Obstet Gynecol 2009;113:1825. 29 Milicic-Juhas V, Pajtler M: Is the HSIL subclassification cytologically real and clinically justified? Coll Antropol 2010;34:395400.

514

Acta Cytologica 2012;56:506514

Gustinucci et al.

Downloaded by: 187.151.28.231 - 11/21/2013 7:37:29 AM