The p16/ki-67 assay is a safe, effective and rapid approach to triage women with mild cervical lesions

Objective The aim of this study was to evaluate the diagnostic accuracy and efficiency of p16/ki-67 dual stain in the identification of CIN2+ lesions, in Greek women with ASCUS or LSIL cytology. Methods A total of 200 women, 20 to 60 years old, were enrolled in the study. All samples were cytologically evaluated and performed for p16/ki-67 and high-risk HPV (HR-HPV) test. All patients were referred to colposcopy for biopsy and histological evaluation. Three cervical cancer (CC) screening strategies were designed and the total direct medical costs of the procedures during our clinical trial were evaluated, from a healthcare perspective. Results HPV 16 as expected was the most common HR-HPV type followed by HPV 31 and HPV 51. The risk for CIN2+ was significantly higher in HPV 16/18 positive cases. p16/ki-67 demonstrated a high sensitivity for CIN2+ identification in both ASCUS and LSIL groups (90.4% and 95%, respectively). HR-HPV test with sensitivity 52.3% and 65.5%, as well as colposcopy with sensitivity 14.3% and 36% respectively in ASCUS and LSIL group, showed inferior results compared to p16/ki-67. The specificity of p16/ki-67 for ASCUS and LSIL was 97.2% and 95.2% respectively, inferior only to colposcopy: 100% and 100%, lacking however statistical significance. HR-HPV test instead, presented the lowest specificity: 76.4% and 71.4% respectively in comparison to the other two methods. From a healthcare perspective, the costs and benefits of the tests implementation for the annual screening and triaging, in three CC screening strategies, were also calculated and discussed. Conclusions The results of the study indicate that p16/ki-67 is a safe and rapid assay that could be used to detect CIN2+ among women with mild cervical lesions, presenting both high sensitivity and specificity and could minimize the psychological and economic burden of HPV screening.

Introduction subpopulation of a larger study. These women attended the clinic from March 2018 through April 2019 for their routine cytology-based cervical cancer screening and after being thoroughly informed participated in the study. Liquid-based cytology (LBC) samples were collected and after HPV testing and p16/ki-67 dual stain were performed, women were referred to colposcopy for evaluation and cervical biopsies. Three cases were excluded as invalid cytology and one as invalid p16/ki-67 due to low cell count.
The exclusion criteria of the current study were HR-HPV negativity or biopsy result of CIN0, treatment for CIN, prior hysterectomy and pregnancy. Written informed consent, medical history and demographic information were obtained from all the 196 enrolled subjects (Table 1). Medical history involved previous cytology tests, HPV vaccination and other vaginal infections. Demographic variables included: age, sexual and reproductive behavior and smoking status. The study was approved by the Ethics Committee of "Alexandra" Hospital.

Cytology-Procedure
Liquid-based cytology samples were stored in PreservCyt solution (Thinprep, Hologic, U.S.A) vials according to the manufacturer's instructions. A first slide was prepared using Thinprep T2000 processor system (Hologic, U.S.A) for cytological evaluation in accordance to the 2014 Bethesda system [17]. For this study we enrolled cases which presented ASCUS and LSIL HPV testing Cobas 1 HPV test. Qualitative in vitro HPV genotyping analysis was performed using Cobas 1 HPV test (Roche Molecular Systems, Inc., Branchburg, NJ) which provides individual results for HPV 16 and HPV 18 genotypes and pooled results for the rest 12 HR-HPV types: 31,33,35,39,45,51,52,56,58,59,66 and 68 in a single analysis. Briefly, target DNA primers amplify a sequence of approximately 200 nucleotides within the L1 region of the HPV genome of the sample with the use of polymerase chain reaction (PCR) to detect the 14 high-risk HPV types [18]. The whole process was performed in the Cobas 1 4800 fully automated System. CLART 1 HPV3 assay. HPV genotyping of the 12 HR-HPV types, that were not evaluated separately with Cobas technology, was performed with the CLART 1 HPV3 kit (Genomica S.A.U, Madrid, Spain), an in vitro diagnostic kit that detects 49 HPV genotypes including High Risk (16, 18, 31, 33, 35, 39,  Detection is based on CLART 1 technology which includes PCR amplification of a 450bp fragment within the L1 viral region using type-specific probes. The visualization of the amplified products is performed in a low-density microarray platform and interpreted in an automatic Genomica's reader (CAR 1 or Clinical array reader) [19]. For the purposes of the current study, we involved cases positive for the 14 HR-HPV genotypes.

p16/ki-67 dual stain (CINtec 1 PLUS Cytology kit)
The CINtec 1 PLUS Cytology Kit is an immunocytochemistry assay for the qualitative detection of p16 INK4a and ki-67 proteins in cervical cytology preparations. The proteins are detected using a ready-to-use cocktail of primary monoclonal antibodies: human p16 INK4a protein (clone E6H4™) and primary recombinant rabbit antibody directed against human ki-67 protein (clone 274-11 AC3) [20]. The slides for p16/ki-67 dual stain were performed using CINtec 1 PLUS kit (Roche mtm laboratories AG, Mannheim, Germany) on an automated slidestainer platform: Ventana BenchMark XT Slide Stainer.
All slides were evaluated by two experienced cytopathologists unaware of histology and HPV testing results. A case was interpreted as positive if one or more cervical epithelial cells were dual stained for both p16 INK4a (brown cytoplasmic stain) and ki-67 (red nuclear stain) regardless of their morphologic appearance. Negativity was established when no synchronous p16/ki-67 dual staining in the same cell was observed, in accordance to the manufacturer's criteria and instructions [21]. In Fig 1A and 1B below, examples of p16/ki-67 dual stained cells in cases of ASCUS and LSIL cytology are presented. A case was excluded or scored as invalid if no p16 INK4a and/or ki-67staining was visible or if there were less than 5,000 squamous cells in the slide [21].

Colposcopy and histology
The population enrolled in our study, underwent colposcopic evaluation of the cervix and the colposcopic findings were reported according to the ASCCP Terminology for Colposcopic Practice [22]. If no visible lesion was found, the colposcopic impression was classified as 'Negative', whereas positivity was then reported as 'LSIL' (low-grade abnormalities) or 'HSIL' (highgrade abnormalities). Biopsy from visible cervical lesions was performed by qualified, trained physicians. Histopathologic review of the cervical biopsies was performed from specialized pathologists blinded to all laboratory results. Cervical histopathological diagnoses included CIN� 1 (CIN1, CIN2, CIN3, or cancer).

Screening strategies and costs data
We designed and evaluated three different CC screening strategies based on the current algorithms used: Cytology (LBC), HPV test with 16/18 genotyping, Co-testing (Cytology & HPV test with 16/18 genotyping) with the addition of p16/ki-67 dual stain as a triage strategy.
The costs of the lab-tests for screening (Cytology, HPV test with 16/18 genotyping, p16/ki-67 dual stain) and the diagnostic methods (Colposcopy, Cervical biopsy and office visit to specialist) presented in Table 2 below were evaluated from a healthcare perspective, in Public Healthcare System or Private Practice. In Public Healthcare System, cost inputs were obtained in 2019, from the official price list of the National Organization for the Provision of Health  Services (EOPYY) in Greece. In the Private sector, costs were based on 2019 price lists of two different Private Hospitals in Greece (Private Practice 1&2) and the average costs were evaluated. In Public Healthcare System, the cost of p16/ki-67 dual stain would be covered by the healthcare system when used in practice. Therefore, its cost was assumed to be the same as the Private Practice's price.
In each CC screening strategy we included the direct medical costs of the procedures per screened woman from a healthcare perspective. Specifically, direct medical costs consist of screening tests used (Cytology, HPV test with 16/18 genotyping, p16/ki-67) and diagnostic methods (Colposcopy and Cervical biopsy) plus office visit (physician's time for patient's evaluation). The costs for the annual follow-up screening tests as well as any treatment needed for CIN were not included.
First strategy: Cytology as primary screening test (Fig 2). Women with Normal cytology return to Routine Screening.
Second strategy: HPV test as primary screening test (Fig 3). Women with HR-HPV test (-) return to Routine Screening.
Women with HR-HPV test (+): a) Triage with cytology. Women with an HR-HPV test (+) and cytology � ASCUS are referred to colposcopy.

Statistical analysis
Data were expressed as mean ± SD for continuous variables and as percentages for categorical data. The Kolmogorov-Smirnov test was utilized for normality analysis of the parameters. To evaluate the risk of CIN2+ lesions, according to HPV genotype, the Chi-square test (χ 2 ) and Odds Ratio (O.R.) with 95% Confidence Interval (CI) were calculated. Diagnostic accuracies of HR-HPV test, Colposcopy and p16/ki-67 method for the detection of CIN2+ were evaluated using Sensitivity, Specificity, Positive Predictive Value (PPV) and Negative Predictive Value (NPV). Comparison between them was performed using the Paired Samples t-test. A Receiver Operating Curve (ROC) analysis was conducted to compare the prognostic abilities of HR-HPV test and p16/ki-67 for the detection of CIN2+ calculating the respective areas under the curve (AUC) with 95% CI, using the maximum likelihood estimation method, which has the advantage of being free of assumption about the Gaussian distribution of underlying variables. All tests were two-sided and statistical significance was set at p < 0.05. All statistical analyses were conducted using SPSS vr 21.00 (IBM Corporation, Somers, NY, USA).

Results
A total of 200 liquid-based cytology samples with CIN � 1 on histological evaluation and HR-HPV (+) were analyzed. Among those, 93 cases had ASCUS cytology, 103 had LSIL and 4 cases were excluded as invalid due to low cellularity. The mean age of all 196 finally enrolled women was 37.5 years old, with participants < 30 years old (27%) and � 30 years old (73%). Only a small percent (12.8%) of the women were HPV vaccinated, the majority (90.8%) had a pap test performed during the previous years and 26.0% had other vaginal infections. Most women (91.8%) had over 2 sexual partners during their lifetime and approximately half the population was nulliparous. Smoking habits of our population were equally distributed ( Table 1).
All patients were selected to be HR-HPV positive and the HR-HPV test results were classified as HPV 16/18 (+) or HPV 16/18 (-) when no HPV 16 or HPV 18 was detected. In our population, 80/196 (40.8%) cases were HPV 16/18 positive and 116/196 (59.2%) had some other of the 12 HR-HPV types. In cases of multiple infections, patients were classified according to the highest risk type for its constituents. For example, if a case distributed both HPV 16 (+) and HPV 31 (+), it was then classified into the HPV 16/18 (+) group. The percentages of HR-HPV genotypes in our study are presented in Fig 5. HPV 16 was the most frequent high-risk type in 66 out of 196 cases (33.7%) followed by HPV 31 (17.3%) and HPV 51 (13.8%). The HPV 56, HPV 58 and HPV 66 distributed similar percentages (11.2%-11.7%), followed by HPV 18 (8.7%). The HPV 52 and HPV 59 presented a percentage of 7.7% and the rest HR-HPV types: 33, 35, 39, 45, 68 presented lower frequencies.
In order to evaluate the risk of CIN2+ according to HPV genotype, patients were sorted into fourteen groups depending on HR-HPV genotype's positivity as shown in Table 3. Among patients with HPV16 (+) or HPV18 (+), either with single or multiple infections, the odds ratio for CIN2+ lesions were 3.2 (95% CI 1.74-6.00) and 5.2 (95% CI 1.62-16.53) respectively. These results are statistically significant higher compared with the odds ratio of the rest 12 HR-HPV genotypes.
The ROC curve analysis of the two methods, HR-HPV test and p16/ki-67, is shown in Fig  6. For the ASCUS group the areas under the curve (AUC) for the HR-HPV test and p16/ki-67 were 0.644 (95% CI 0.503-0.784) and 0.938 (95% CI 0.862-1.000) and for the LSIL group, the AUC of HR-HPV test and p16/ki-67 were 0.685 (95% CI 0.580-0.790) and 0.952 (95% CI 0.903-1.000) respectively. In order to evaluate the costs of the three different proposed CC screening strategies (Figs  2-4) we estimated, from a healthcare perspective, either in Public Healthcare System or in Private Practice, the total direct medical costs of the procedures during our clinical trial. The total direct medical costs (screening tests and diagnostic methods plus physician's time during office visit) for our population were calculated as the number of cases in each of the three different screening strategies multiplied by the direct medical costs per case ( Table 6). The calculations presented in Table 6 were conducted using the data of the costs in Table 2 and the resource use from the clinical data of our population in Table 4.
Regarding the 1 st Cytology screening strategy (Fig 2), when triaging with a) HPV test plus 16/18 genotyping the total direct medical costs in our study population would be €21.389,92 in Public Healthcare System and €83.020,00 in Private Practice, whereas with b) p16/ki-67 dual stain triage the costs would be €17.508,66 and €68.022,50 respectively. Concerning the 2 nd HPV test screening strategy approach (Fig 3), we calculated that by using a) cytology triage, the direct medical costs for Public Healthcare System and Private Practice would be €25.883,76 and €105.350,00 respectively, while with the b) p16/ki-67 triage test, the total direct

Discussion
In the present study, we evaluated the clinical performance of p16/ki-67 in comparison to the existing strategy, i.e. HPV testing with colposcopy referrals in a group of Greek population with ASCUS and LSIL cytology. These methods are already applied in the primary screening in most organized health care systems worldwide [5,7]. Although HPV DNA test appears to be highly sensitive in multiple studies for the detection of precancerous lesions, it lacks specificity. This leads to the detection of a large number of HPV positive women, who are referred to unnecessary colposcopies since HPV DNA test cannot discriminate an active/persistent infection from a transient one [23].
The p16 INK4a protein is a cyclin-dependent kinase (CDK) inhibitor that plays an important role in cell cycle regulation. The over expression of E7 oncoprotein, in persistent HR-HPV infections, leads to p16 INK4a upregulation which might serve as a biomarker for the detection of cervical precancerous lesions [24,25]. ki-67 is another marker which is strictly associated with cell proliferation and is expressed normally during cell mitosis [26]. The co-expression of p16 INK4a and ki-67 in the same cell, does not physically exist in normal cells and this coexpression is considered a valuable marker for cell-cycle deregulation and cell transformation due to HPV infection [27]. Therefore, a positive p16/ki-67 result could indicate a higher-grade dysplasia on the cervix and provide information about CIN2+ lesions [28,29].
In our study, besides p16/ki-67, we classified HPV positive patients into two categories, as HPV 16/18 (+) and HPV 16/18 (-) according to algorithms for primary HPV screening [30]. In concordance with the well documented bibliography facts, our results show that HPV 16/18 are the most prevalent types for developing CIN2 or higher lesions, in comparison to the other HR-HPV types [31][32][33].
We also identified the distribution of the 14 HR-HPV types in our Greek population. Briefly, our results are fully consistent with previous studies in Greece, where HPV 16 is the most prevalent type followed by HPV 31 and HPV 51 [34][35][36]. This observation shows that the distribution of HR-HPV genotypes kept the same pattern despite the population habit alterations and mixed-race interaction due to refugee's immigration in the country over these years.
In order to evaluate the accuracy of the methods, we compared the p16/ki-67 dual stain with the HPV testing and colposcopy, which are currently proposed by the screening guidelines. In our study, the results showed that p16/ki-67 method had both high sensitivity and specificity as triage test, both in the two cytology groups (ASCUS/LSIL), improving the diagnosis of CIN2+ lesions.
Among women with ASCUS and LSIL, sensitivities of p16/ki-67 dual stain were 90.4% and 95.0% respectively, higher than that of the HR-HPV test (52.3% and 65.5%). Previous studies have shown that p16/ki-67 can be a reliable tool for risk stratification of HPV positive women with mild cervical lesions, reporting high sensitivity values similar to that of HPV test [27,[37][38][39][40].
In our National Health Care System, colposcopy is widely used to detect cervical lesions and biopsy under colposcopy guidance, confirms or excludes CIN2+ lesions. We retrospectively collected colposcopy reports of the population we studied in our Department of Obstetrics & Gynecology which is a referral Department for gynecological cancer. We compared these reports to the histological findings, in order to evaluate the prognostic accuracy of colposcopy.
There are many factors that could explain this observation such as, training and experience of the clinicians involved, inadequate colposcopy due to lack of objective criteria to classify the lesions macroscopically and co-existing pathologies which tamper with the appearance of the normal mucosa.
Furthermore, we presented the ROC curve to show the diagnostic performance of p16/ki-67 and HPV test. The AUC of p16/ki-67 was greater in both cytology groups (0.938/0.952) compared to the HPV test (0.644/0.685). These results are comparable to those of a recent study were AUC of p16/ki-67 dual stain was (0.778), also significantly higher than that of HPV test (0.503) [41].
Additionally, in both cytology groups, we evaluated the combination of the two methods, HR-HPV test with 16/18 genotyping and p16/ki-67, in order to identify how many cases of CIN2+ will be missed. In the LSIL cases, when HPV16/18 and p16/ki-67 were both positive or negative, no women were misdiagnosed. In case of conflicting results between HPV16/18 and p16/ki-67, taking into consideration the p16/ki-67 positivity alone, we were able to diagnose 91.3% of women with CIN2+ that would otherwise escape diagnosis and only 8.7% were missed. On the other hand, we observed that if only HPV 16/18 was positive we could detect only 20% of CIN2+ and the rest were missed.
Our study showed that p16/ki-67 presented both high sensitivity and specificity values, with only a few CIN2+ being missed. Thus, it appears to be an excellent triage tool for accurate stratification of women with mild cervical lesions. Our results are in line with other articles where the effectiveness of the screening strategies may be improved by the addition of p16/ki-67 dual stain in the triage [42,43].
Regarding the analysis of the total direct medical costs of the three different CC screening scenarios, from a healthcare perspective, in our clinical trial, in the 1 st strategy, Cytology primary screening, the utilization of p16/ki-67 as an optional triage tool has reduced the costs by €3.881,26 in the Public Healthcare System and by €14.997,50 in Private Practice compared to HPV test triage. Additionally, a cost reduction of €21.647,50 in Private Practice was reported in the HPV test-based screening scenario by adapting dual stain in the triage, whereas an elevation of €4.694,18 was shown in the Public Healthcare System in comparison to cytology triage.
Finally, the costs for the Co-testing screening strategy were proved to be the highest among all screening scenarios.
It seems that the model Cytology as primary screening strategy and triage with p16/ki-67 dual stain may not only have high clinical and diagnostic impact, avoiding over-diagnosis and over-treatment, but also budgetary impact lowering annual costs for screening and seems to be the most realistic and safe scenario for CC screening.
As for the limitations of our study, the number of cases was small, as well as the fact that the follow-up of the women enrolled was not evaluated and is yet to be accomplished. Further evaluation of the follow-up is necessary to examine the application of the p16/ki-67 biomarker in triaging women with mild cervical lesions, as shown in many previous scientific studies. The role of p16/ki-67 negativity in combination with HR-HPV positivity remains to be fully interpreted and understood.
Summarizing our results, p16/ki-67 dual stain presented both higher sensitivity and specificity, providing good clinical outcomes, while lowering the annual direct medical costs of CC screening. Cytology in combination with p16/ki-67 dual stain is a powerful, safe and affordable tool to triage women with mild cervical lesions, especially in low-and middle-income countries where accurate but also budgetary adaptations of health care services are required.