P16INK4a/ki67 Immunocytochemistry in Improving the Predictive Value for High Grade Cervical Intraepithelial (≥CIN2) Neoplasia in Pap Smear.

INTRODUCTION: Cervical cytology has limited sensitivity to detect cervical pre-cancerous lesions. High-risk human papillomavirus (hr-HPV) DNA testing has high sensitivity but its specificity is limited. This study was done to assess the utility of p16INK4a/ki-67 dual stained cytology in improving the predictive value for high-grade cervical (CIN2+) lesions. AIM/
OBJECTIVE: To assess the significance of P16/Ki-67 immunocytochemistry in improving the predictive value for high-grade cervical intraepithelial (≥CIN 2+) lesions on Pap smear.
MATERIAL AND METHODS: This was a prospective diagnostic study that included 93 patients with ASC-US/LSIL/ASC-H and HSIL on thin prep cervical smears and who also underwent hr-HPV DNA test and colposcopy-guided biopsy. Biopsy was the gold standard against which the performance of P16INK4a/Ki-67 and hr-HPV results were compared.
RESULTS: In women of all ages, sensitivity of (96.8%) hr-HPV test and p16/Ki-67 dual immunocytochemistry (≥1 positive cell) were similar and negative predictive value (NPV) was (97.1% vs. 97.9%) but the latter test showed better specificity (69.4% vs. 53.2%) and positive predictive value (PPV, 61.2% vs. 50.8%) for ≥CIN 2 lesions. A higher cut off of at least 10 positive cells gives a higher specificity and PPV, with slightly decreased sensitivity and NPV.
CONCLUSION: Because high-risk HPV test has a high sensitivity and NPV, whereas P16/Ki-67 dual immunocytochemistry (≥10 positive cells) has a high specificity and PPV, the latter can be recommended as an ancillary test in hr-HPV-positive women to reduce the number of women going for colposcopy and biopsies. Copyright:

INTRODUCTION

Cervical cancer remains an important problem, most importantly in developing countries where it is associated with a high incidence and increased mortality.[1] In India, cervical cancer is the second most common cancer.[2] Pap smear is considered as one of the best screening tests in medicine, which has reduced the incidence of cervical cancer significantly by more than 50%. However, Pap smear has limitations. The sensitivity of one Pap smear to identify cervical intraepithelial lesions ranges from 30 to 87% and specificity from 86% to 100%.[3] Most of the abnormal cervical cytology are composed of atypical squamous cells of undetermined significance (ASC-US) and low-grade squamous intraepithelial lesions (LSIL).[4]

The 5-year risk for developing CIN3 or more among cases of ASC-US was 2.6% and for LSIL was 5.2%.[5] Hence, there is a need for triaging ASC-US/LSIL patients for colposcopy and biopsy. High-risk human papilloma virus (hr-HPV) has an important role in the development of cervical cancer.[67] The hr-HPV test has shown better sensitivity (96.1% vs. 53.0%) when compared to Pap smear screening.[8] This test identifies all women infected with hr-HPV, irrespective of the transient nature of infection, especially in women <30 years of age. Most often, the infection clears within 1 to 2 years.[9] However, all women with positive hr-HPV are referred for colposcopy and biopsy, which might have been largely unnecessary in younger women with transient infection. This unnecessary testing also discourages the women from participating in screening programs.

A few host cell biomarkers have been evaluated to increase the specificity in cervical screening.[10] One such biomarker that has been recently identified is dual P16INK4A/ki67.[11] P16 INK4A (also referred to as p16) is a tumor suppressor protein that causes cell cycle arrest by inhibiting cyclin D–cyclin-dependent kinase 4 complex formation. p16 immunohistochemistry is used for the identification of p16 overexpression in cervical biopsies, which indicates oncogenic transformation caused by persistent hr-HPV infections.[12] Ki-67 is used as proliferative index for the assessment of growth fraction of the cell cycle.[13]

A combination of p16 and ki-67 detects the HPV-infected cells undergoing transformed and uncontrolled proliferation. Many studies have been done on the utility of p16 and Ki-67 dual immunocytochemistry (ICC) for the diagnosis of CIN2+ lesion.[141516] These studies have shown that the specificity of dual immunocytochemistry is higher as compared to that of the hr-HPV test in the detection of underlying high-grade cervical intraepithelial lesion. In all the above studies, the co-expression of cytoplasmic p16 and nuclear Ki-67 in at least one cell was considered as positive ICC test, indicating a transformed infection. According to the study by Ziemke, when a higher cut off of 10 cells was used, the probability of underlying high-grade lesions (CIN2+) also increased in a sample.[17] p16/Ki-67-dual immunocytochemistry expression is assessed irrespective of the morphology.[18] These studies showed the importance of dual immunostaining in precancerous minor cervical abnormalities. The purpose of this study was to assess the significance of dual p16/ki67 immunocytochemistry in improving the efficiency of the screening system for cervical cancer.

MATERIALS AND METHODS

After approval from the Institutional Review Board (min no: 10175), this prospective diagnostic study was conducted in the Department of Pathology, in conjunction with the Department of Virology and Department of Gynaecologic Oncology. A total of 175 patients from gynecology/gynecologic oncology, whose cervical smears were reported by the Department of General Pathology as ASC-US/LSIL/ASC-H and HSIL, were identified for the study over a period of 8 months. Each case was reviewed by a pathologist for inclusion in the study. Consent form and information sheet were given to all patients willing to participate in this study. All patients, who were less than 25 years, pregnant women, patients unwilling for colposcopy/biopsy, seropositive patients, and patients previously treated for CIN were excluded from this study. The residual thin prep sample material of all study patients (Hologic vials) was stored at 2 to 4°C for up to 6 weeks, and slides were prepared for ICC staining using Thin Prep® 2000 Processor (Hologic™ Inc.). Immunocytochemistry analysis, based on the CINtec PLUS kit (Roche MTM Lab-oratories, Heidelberg, Germany), was done on these smears according to the manufacturer's instructions. All slides were interpreted by two pathologists. The presence of at least one cervical epithelial cell with co-localization of brown color cytoplasmic staining for p16 and red color nuclear staining for Ki-67 within the same cell was regarded as a positive CINtec PLUS test result [Figure 1a]. If there was no co-localization of the immunostain, the CINtec PLUS test result was considered negative [Figure 1b].

Figure 1

(a–f) p16/Ki-67 dual immunostaining. (a) Positive control-squamous cell carcinoma showing positivity for p16/Ki-67 (400×). (b) Negative control-negative for intraepithelial lesion showing negativity for p16/Ki-67 (400×). (c) HSIL showing positivity for p16/Ki-67 (400×). (d) ASC-H showing positivity for p16/Ki-67 (400×). (e) LSIL showing positivity for p16/Ki-67 (400×). (f) ASC-US showing positivity for p16/Ki-67 (400×)

The presence of cervical epithelial cells showing a single immunoreactivity only for one of the two markers is not considered as a positive test result for the CINtec PLUS kit. Strict criteria for positive and negative tests were followed to avoid discrepancy. Smears from a known squamous cell carcinoma were used as a positive control, with colocalization of both brown cytoplasmic immunostaining (p16) and red nuclear immunostaining (Ki-67). Different cell types present in representative cervical cytology specimens, which are known to be negative for the expression of the p16 and Ki-67 antigens (such as superficial cells), may serve as an internal negative control. We also further evaluated the utility of p16/Ki-67-dual immunocytochemistry, with a cut off for positivity as more than 10 cells.

The material for this test was collected at the time of colposcopy using the digene HC2 DNA collection device, which consists of a cervical brush and digene Specimen Transport Medium (STM). An in vitro nucleic acid test with signal amplification by Hybrid Capture2 assay (HC2; Digene Corp., Gaithersburg, MD) was used. The hr-HPV DNA data were interpreted based on the results (positive/negative) obtained from Hybrid Capture (HC2) in the Department of Virology. Serial amplification assay for hr-HPV DNA in the cervical specimen can detect up to 13 different hr-HPV DNA types in cervical cells. The hr-HPV DNA detection in Hybrid Capture (HC2) technique is done using microplate chemiluminescence. The digene analysis hr-HPV DNA test cut off of 1 pg/mL is equivalent to 1,00,0000 HPV copies/mL or 5,000 copies per test. An STM specimen with relative light unit (RLU)/cut off Value ratios equal or greater than 1 are considered “positive.” Specimens with RLU/cut off value ratios <1 were considered “Negative” or “None detected” for the 13 HPV types tested.

Colposcopy and biopsy were also done on all study patients in the Gynecology Outpatient Department according to the accepted diagnostic protocol. Histopathological slides (4 μm) were cut on the microtome from paraffin blocks, stained with hematoxylin and eosin (H and E) and assessed by histopathologists who were blinded to the ICC and HPV results of the study patients. Biopsy was considered as the gold standard. The new dual immunocytochemistry and hr-HPV DNA results were correlated with those of gold standard biopsy. The clinical details of these patients were obtained from the charts retrieved from the Medical Records Department. The clinical features that were analyzed included age, indication for cervical screening, marital status, parity, and menopause status.

Data entry and analysis were done using the EpiData software. Statistical analysis was performed using SPSS statistics 16.0. For the comparison of positivity of dual immunocytochemistry test and high-risk HPV test with biopsy, the Pearson χ2 test was used, as well as the Fisher exact test. To assess the statistical significance, we used P value. A P value < 0.05 was considered statistically significant. Diagnostic statistics were used to assess the performance of hr-HPV DNA test and p16/Ki-67 dual immunocytochemistry. The sensitivity, specificity, and positive predictive value (PPV) and negative predictive value (NPV) were calculated with corresponding 95% confidence intervals.

RESULTS

A total of 175 cases were reported as ASC-US, LSIL, ASC-H, and HSIL over 8 months. Thirty-two patients did not come for follow-up, 11 patients did not have undergo biopsy, and 8 patients did not have HPV. Fifty-six cases were excluded, leaving a total of 123 samples for further testing. All slides were kept in 100% ethyl alcohol fixative and stored at 2 to 3°C. All tests were done in three batches in different time periods. Among 123 samples, only 76 cases showed adequate cellularity, 17 cases were paucicellular. Thirty cases did not have enough material to process and hence were excluded from the study, leaving a total of 93 women for inclusion in the study. ASC-US: 42 (46.2%), LSIL: 16 (17.02%), ASC-H: 13 (13.9%), HSIL: 22 (23.6%) were included for final analysis [Table 1]. The most common intraepithelial lesion identified in all age groups of women was ASC-US. The median age was 42.55 years (standard deviation of 11.02 years) for overall abnormal pap smears. The youngest woman was 24 years old and the oldest patient was 71 years old. Table 2 shows baseline characteristics of patients. There were 12 patients (12.09%) less than 30 years of age and 81 patients (87.09%) above 30 years of age. Among 93 women, 75 were symptomatic (80%), whereas 18 were asymptomatic (20%) who were referred for routine screening. Symptomatic women presented with following complaints: bleeding per vagina (60%), lower abdominal pain (15%), white discharge per vagina (11%), itching (7%), and burning micturition (4%), post coital bleeding (3%). Bleeding per vagina was the most common clinical presentation of women who presented to gynecology/gynecologic oncology department. Among 93 women, 25 were in postmenopausal age group (27%) and 68 were in the reproductive age group (73%) Figure 1 shows positive cases of HSIL(1c), ASC-H(1d), LSIL(1e) and ASC-US(1f). Table 3 shows distribution of p16/Ki-67 and hr-HPV-positive results in patients with abnormal cytology and biopsy. Histopathology results revealed 51 (54.8%) cases without CIN, 11 (11.8%) CIN 1, 13 (14%) CIN2, and 18 (19.4%) CIN3 respectively. P16/Ki-67 test was positive in 49 (52.7%) and negative in 44 cases (47.3), whereas 59 (63.4%) were positive for hr-HPV and 34 (36.6%) were negative. Thus, the overall positive rate for p16/Ki-67 (52.7%) was less than that for hr-HPV test (63.4%). Table 4 shows that the p16/Ki-67 positivity was more in hr-HPV-positive cases as compared to that in hr-HPV-negative cases (P value < 0.001).

Table 1

Baseline characteristics

	n	Percentage
Total cases	93
<30 years of age	12	12.09
>30 years of age	81	87.10
Postmenopausal age group	25	27
Reproductive age group	68	73
Asymptomatic	18	80
Symptomatic	75	20
Bleeding per vagina	45	60
Lower abdominal pain	12	15
White discharge per vagina	8	11
Itching	5	7
Burning micturition	3	4
Post coital bleeding	2	3

n, number.

Table 2

Abnormal cytology results with age distribution

	n = 93	Overall mean	Over all median	SD
ACSUS	42	43.3	42	11.8
LSIL	16	38	36	9.8
ASC_H	13	40.7	41	11.08
HSIL	22	48.2	47	11.4
Average		42.55	41.5	11.02

ASC-US, atypical squamous cells of undetermined significance, LSIL: Low-grade squamous intraepithelial lesion, ASC-H: Atypical squamous cells – cannot exclude HSIL, HSIL: High-grade squamous intraepithelial lesion.

Table 3

Distribution of p16/Ki-67 positivity and hr-HPV positivity in patients with abnormal cytology and biopsy.

Category	n	P16/ki-67 (%)	HR-HPV (%)	P
Cytology
ASCUS	42	10 (20.40)	14 (23)
LSIL	16	7 (14.03)	13 (22)
ASC-H	13	10 (20.40)	12 (16.90)
HSIL	22	22 (44.90)	22 (37.30)
ALL	93	49 (52.70)	59 (63.40)
Biopsy
CIN-0	51	13 (26.50)	22 (37.30)	<0.001
CIN-1	11	6 (12.20)	7 (11.09)
CIN-2	13	12 (24.50)	12 (20.30)
CIN-3	18	18 (36.70)	18 (30.50)

CIN, cervical intraepithelial neoplasia

Table 4

Association of p16/Ki-67 and hr-HPV results with biopsy

	P16/ki-67+	Percentage	P16/ki-67-	Percentage	P
ALL
hr-HPV - (n=34)	3	8.80	31	91.20	<0.001
hr-HPV + (n=59)	46	78.00	13	22.00
<CIN 2
HR-HPV - (n=33)	3	9.01	30	90.9	<0.001
HR-HPV + (n=29)	16	55.22	13	44.8
≥CIN2
HR-HPV - (n=1)	0	0	1	100	<0.032
HR-HPV + (n=30)	30	100	0	0

Among the 59 cases positive for hr-HPV tests, 30 cases (32%) had ≥CIN2 lesions (true positive) and in 29 cases (30%) biopsies were negative (false positive). Among hr-HPV-test negative cases, 1 case (1%) had ≥ CIN2 lesion (false negative) and 33 cases (37%) were negative (true negative) on biopsy. Among the 49 cases positive for p16/Ki-67, 30 cases (32%) had ≥CIN2 lesion (true positive) and 19 cases (20%) biopsies were negative (false positive). Among negative cases, 1 case (1%) had ≥CIN 2 lesions (false negative) and in 43 cases (47%) biopsies were negative (true negative). Thus, p16/Ki-67 test also showed less false positivity (19/62 cases) compared to hr-HPV test (29/62 cases), among cases that showed CIN >2 in biopsy.

Table 5 shows sensitivity, specificity, PPV, and NPV of p16/Ki-67 and hr-HPV tests in all cases with abnormal cytology to detect CIN2+ on biopsy. The sensitivity of dual-stained cytology and hr-HPV DNA testing was 96.8% for ≥CIN2. The specificity of dual-stained cytology was 69.4%, whereas hr-HPV test had a specificity of 53.2%. Specificity for dual-stained cytology was higher than that of hr-HPV test. The PPV for dual-stained cytology was 61.2%, whereas that for hr-HPV test had 50.8%. The PPV for dual-stained cytology was higher than that of hr-HPV test. The NPV for dual-stained cytology was 97.9%, whereas that for hr-HPV was 97.1%. The P value for both tests was not significant (>0.05). An increase in threshold for dual immunostaining (i.e., cut off of at least 10 positive cells) showed a higher specificity (93.5% vs. 69.4%) with PPV (86.2% vs. 61.2%); however, there was decrease in sensitivity (80.6% vs. 96.8%) and NPV (90.6% vs. 97.9%).

Table 5

Sensitivity, specificity, PPV, and NPV for p16/Ki-67 and hr-HPV tests for CIN>2 lesion (gold standard)

	Sensitivity% (95%CI)	Specificity% (95%CI)	PPV% (95%CI)	NPV% (95%CI)
Hr-HPV test	96.8 (83.3, 99.9)	53.2 (40.1, 66.0)	50.8 (37.5, 64.1)	97.1 (84.7, 99.9)
P16/Ki-67≥1 positive cell	96.8 (83.3, 99.9)	69.4 (56.4, 80.4)	61.2 (51.9, 69.8)	97.9 (86.1, 99.7)
P16/Ki-67≥10 positive cells	80.6 (62.5, 92.5)	93.5 (84.3, 98.2)	86.2 (70.5, 94.2)	90.6 (82.5, 95.2)

DISCUSSION

Cervical cytology has limited sensitivity to detect cervical precancerous lesion. High-risk HPV DNA test had shown good sensitivity when compared to Pap smear screening; however, it had low specificity. Identifying better triaging for hr-HPV-positive women in minor cytological abnormalities is important to reduce the number of unwanted colposcopies and biopsies. p16 is a protein that triggers the cell cycle arrest under normal physiological state. p16Ink4a acts as a CDK inhibitor and blocks the cyclin D–CDK4/6-mediated phosphorylation of Rb, which induces cell cycle arrest. In HPV-related neoplasms, the molecular mechanism of p16INK4a can be explained by HPV oncoproteins E6 and E7. The integration of the viral DNA into the host DNA causes the overexpression of these viral oncoproteins. The E6 protein binds and causes degradation of p53, whereas E7 promotes the progression of cell cycle by displacing E2F form RB protein.[1920] This inactivation of the RB protein releases p16INK4a from its negative feedback control, which in turn causes a paradoxical increase in the levels of p16.

The Ki-67 antigen is identified in all active stages of cell cycle; it is expressed in G1, S, G2, and mitosis. Ki67 protein is absent during the resting phase (G0). Ki-67 protein expression is a must for cell division cycle and indicates a proliferating index of a cell. The co-expression of both p16 and Ki-67 in the same cell indicates dysregulated cell cycle due to incorporation of viral oncoprotein into the host genome.

Our study was a prospective diagnostic study that evaluated the performance of dual-stained immunomarker p16/Ki-67 as against the hr-HPV test in cervical screening for identification of a higher cervical intraepithelial lesion among abnormal Pap smears.

This study included 42 ASC-US, 16 LSIL, 13 ASC-H, and 22 HSIL cases of abnormal Pap smears for analysis. The median age was 42.55 years (standard deviation of 11.02 years) for abnormal pap smears and was comparable to a median age of 41 years (Korolczuk et al., 2015). ASC-US was identified as the most common group in the study, in keeping with the Bethesda system for reporting cervical cytology.[21] p16/Ki-67 tests showed less positivity [Table 3] in low-grade biopsy (P value < 0.001). p16/Ki-67 positivity was more in hr-HPV-positive cases [Table 4] as compared to hr-HPV-negative cases (P value < 0.001). This study results were consistent with those reported in the literature.[22]

In this study, the performance of dual immunocytochemistry for high-grade (≥CIN2) lesion was as follows: sensitivity of 96.8%, NPV of 97.9%, specificity of 69.4%, and PPV of 61.2%. Thus, for identifying higher cervical intraepithelial lesions (≥CIN2) among all abnormal Pap smears, sensitivity and NPV of dual immunocytochemistry was almost similar to hr-HPV testing, but with improved specificity and PPVs.

These results were comparable to those of Yu et al. study.[23] Yu et al. had highlighted the high sensitivity and NPV for hr-HPV test as compared to dual immunocytochemistry, whereas our study showed similar results for sensitivity and NPV for both the tests [Table 5] The present study showed better specificity and PPV for dual immunocytochemistry as compared to hr-HPV assay, which is also in keeping with the results of Yu et al. Ziemke, in 2017,[17] proposed that false positivity in older women was less than in women less than 30 years old, due to the longer infective period in older women. It could also be a good indicator of reparative and productive function of younger women. Ziemke further demonstrated in the same study, increased specificity and predictive value when the cut off for dual immunocytochemistry positivity was more than 10 cells. Using a score of 10 p16/Ki-67 marked cells as a positive result, instead of one marked cell, led to a significantly higher specificity (89% vs. 70.2%) and PPV (55.7% vs. 46.7%) among women in the LSIL group.

In our study, including all age groups, when we increased the cut off for positive interpretation to 10 cells, there was a significant increase in the specificity (93.5% vs. 69.4%) and PPV (86.2% vs. 61.2%), whereas the sensitivity (96.8 vs. 80.6) and NPV (97.9% vs. 90.6%) relatively decreased [Table 5].

Thus, increasing the cut off for dual immunocytochemistry resulted in significantly increased specificity and PPV and a relative decrease in sensitivity and NPV. These results were similar to those described by Ziemke although he had included only the LSIL cases.

In our study, one case (01%), which was reported as ASC-H, had negative hr-HPV and dual immunocytochemistry, but was reported as CIN2 on biopsy. In the study by Yu et al. 21 cases of ≥CIN2 lesions (9.1%) were p16/Ki-67-negative. It indicates that not all precancerous lesions progress to cervical cancer.[23] There was evidence that approximately 40% of undiagnosed CIN-2 will regress over 2 years. This could possibly explain the negative p16/Ki-67 immunocytochemistry. This case was also negative for hr-HPV test. Waldstrøm et al.[24] in 2012 showed that four cases of CIN2+ lesion (01%), were negative for both hr-HPV test and p16/Ki-67.

The present study performed in 93 women with abnormal Pap cytology showed similar sensitivity and NPV but increased specificity and PPV for dual immunocytochemistry as compared to the hr-HPV test. These data are concordant with those reported in the literature. Because p16/Ki-67 has better specificity and PPV than hr-HPV test, it can be used as a triage for hr-HPV-positive women. However, when one cell is used as a cut off for positive interpretation, the specificity and PPV were less than those at a cut off of 10 cells. This could be due to increased false positivity, particularly in younger women. By increasing the threshold of detection to at least 10 positive cells, there was increased specificity and PPV but with a decrease in sensitivity and NPV.

In women of all ages, sensitivity of hr-HPV test and p16/Ki-67 dual immunocytochemistry (≥1 positive cell) was similar (96.8%), and NPV was comparable (97.1% vs. 97.9%). However, dual immunocytochemistry showed better specificity (69.4% vs. 53.2%) and PPV (61.2% vs. 50.8%) for ≥CIN 2 lesions. A higher cut off of at least 10 positive cells gives a higher specificity and positive predictive value but a slightly decreased sensitivity and NPV. Because hr-HPV test has a high sensitivity and NPV, whereas p16/Ki-67 dual immunocytochemistry (≥10 positive cells) has a high specificity and PPV, the latter can be recommended as an ancillary test in hr-HPV-positive women to reduce the number of women going for colposcopy and biopsies. Thus, p16/Ki-67 dual immunocytochemistry can be used for risk stratification and also for appropriate management in patients with ASC-US and LSIL.

Ethical Clearence

Ethical committee approval was obtained on 16-8-2016

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.