The outcomes of this study presented that 26 (21.3%) patients underwent colposcopically-directed multiple punch cervical biopsy with an H&E pathological diagnosis of HSIL and above (HSIL+), 11 (9.0%) patients had cervical intraepithelial neoplasia (CIN)2, nine (7.4%) patients had CIN3, and six (4.9%) patients had infiltrating carcinoma. Cytology, colposcopy, and p16INK4a immunostaining had 52.4%, 38.5%, and 92.3% sensitivity, respectively, and 76.2%, 94.8%, and 99% specificity, respectively. The positive predictive value of cytology, colposcopy, and p16INK4a immunostaining was 31.4%, 66.7%, and 96%, respectively, and the negative predictive value was 88.5%, 85.1%, and 97.9%, respectively. Compared with H&E staining, the kappa of cytology, colposcopy, and p16INK4a immunostaining was 0.327, 0.323, and 0.926, respectively.
3.1 Cervical Cancer Screening Strategies For HPV-negative Patients
Numerous epidemiologic studies have proven that persistent HR-HPV infection is the main causative factor of HSIL and infiltrating carcinoma.3 HR-HPV infection can be detected in almost all patients with precancerous lesion of the cervix and invasive carcinoma of the cervix, and there would be positive clinical detection of an HR-HPV subtype. However, a certain proportion of patients with precancerous lesion of the cervix and invasive carcinoma of the cervix may have negative HR-HPV subtype detection, and the management of such patients should be investigated further in particular.
At present, HR-HPV detection enhances screening sensitivity when it is used as an objective index for cervical cancer screening (preliminary or combined cytological screening).2 However, there is currently a lack of more sensitive and objective screening diagnosis indexes for HR-HPV-negative women. Following the in-depth study of the natural history of HPV infection in recent years, biomarkers have been used more frequently for screening cervical carcinoma and improving pathological diagnosis. The objective of the present study was to investigate the value of p16INK4a immunostaining for detecting HSIL in HR-HPV-negative patients.
3.2 p16 INK4a Immunostaining Aids The Accuracy Of Cin2 Diagnosis
HR-HPV-negative patients with abnormal cytological examination results or suspected symptoms or signs of invasive cervical carcinoma may be required to undergo colposcopy. The biopsy is performed under colposcope in the most suspicious lesion. Most traditional pathological examinations use morphological diagnosis employing H&E staining. Many studies have found that the reproducibility of H&E staining morphological diagnoses by pathologists is poor and that the consistency of CIN2 diagnosis is < 50%.12 The reproducibility of CIN2 + morphological diagnosis in the ASCUS/LSIL Triage Study for Cervical Cancer (ALTS) was only 43%.13 In another study, two pathologists agreed with 84% and 81%, respectively, for CIN3 diagnosis; for CIN2, the agreement was 13% and 31%, respectively.14 There are vast differences among pathologists for H&E staining diagnosis of CIN2, and there are many false positive or false negative results. Recent studies have suggested that auxiliary use of immunohistochemical staining may aid the accuracy of CIN2 diagnosis.
The p16INK4a protein can compete with cyclin D1 to bind with CDK4, inhibiting CDK4 activity. The p16INK4a/CDK-cyclin D/Rb complex is the key factor when cells exit the G1 phase of the cell cycle and enter the S phase. Any gene abnormality in this molecular chain may result in the loss of control of molecular action on the cell cycle. The E7 protein expressed by HR-HPV interferes with the normal function of the RB gene. E7 binds with phosphorylated Rb (pRb), inactivating the function of the RB gene, eliminating the negative feedback inhibition of pRb on p16INK4a protein expression, and results in p16INK4a overexpression.4 This leads to disorder of the cell cycle of cervical epithelial cells, resulting in the characteristic of immortality and the initiation of a series of carcinogenesis processes.15–18
Li et al. from our department carried out p16INK4a and Ki-67 immunostaining on the pathological sections of patients with CIN who were aged < 35 years. These researchers found that p16INK4a and Ki-67 immunostaining had very good consistency with CIN grading.19 Galgano et al. studied the immunostaining of 1,455 samples and used the strongest staining results as the cut-off value. They found that sensitivity was 86.7% and specificity was 82.8% for CIN2 + and that p16INK4a immunostaining is a useful and reliable diagnostic adjunct for distinguishing biopsies with and without CIN2 + .20 Bergeron et al. studied the consistency of p16INK4a immunostaining. They used a team of three gynaecopathology experts to establish a gold standard; 12 pathologists rendered independent diagnoses. They found that p16INK4a-immunostained slides significantly increased the diagnostic accuracy for detecting high-grade CIN (P = 0004) as compared with H&E slides and that the reproducibility of p16INK4a immunostaining interpretation was excellent (kappa = 0.899)21. A systematic review and meta-study analysed the consistency of CIN diagnosis between different pathologists. Five eligible articles were systemically reviewed and meta-analysed. The results showed strong agreement for pathologists’ interpretation of cervical biopsy specimens as p16INK4a-positive or -negative (pooled kappa = 0.90) and significantly higher agreement for a CIN2 + diagnosis with H&E morphology in conjunction with p16INK4a-positive immunostaining (kappa = 0.73) compared with diagnosis using H&E morphology alone (kappa = 0.41). The investigators suggested that the conjunctive use of H&E morphology with p16INK4a immunostaining improved inter-observer agreement of the CIN2 + diagnosis.22
The LAST Project working group of the College of American Pathologists and ASCCP proposed that although the currently available evidence is insufficient for confirming that any biomarker can be used to replace the basic pathological diagnosis, the addition of p16INK4a immunostaining in some cases may provide a more reliable and consistent pathological interpretation.10,23 However, can p16INK4a be used as a molecular biological substitute for pathological HSIL of the cervical tissues in HR-HPV-negative patients? Zhang et al. summarised 46 cases of initial biopsies with histopathologically diagnosed HSIL (CIN2 and CIN3) and found that all 28 cases with diffuse p16 immunostaining were confirmed as HSIL (CIN2 and CIN3) via cervical LEEP, and all seven cases with negative p16 immunostaining had no detectable HSIL as determined via cervical LEEP. Of the seven cases with focal or patch p16 immunostaining, cervical LEEP revealed that four had HSIL (CIN2) and three had LSIL (CIN1). The conclusion is that diffuse p16 immunostaining is the hallmark of HSIL because it correlates 100% with CIN2 and CIN3 lesions between initial biopsy and cervical LEEP specimens, whatever the HPV status.24 Solano et al. found that p16INK4a immunostaining had more diagnostic benefits, where their retrospective study of 596 patients revealed HSIL/CIN2-3 not found in the initial H&E staining.25
There was very strong consistency between positive p16INK4a immunostaining and H&E staining pathological diagnosis in our study (kappa = 0.926). The PPV and NPV of high-grade lesion diagnosis was 92.3% and 97.9%, respectively. Almost 100% of the cervical HSIL or infiltrating carcinoma could be excluded for patients with negative p16INK4a immunostaining. For the follow-up of the pathological results, we found that the PPV of p16INK4a-positive staining for pathological HSIL + after cervical LEEP was 87.5%, which is higher than that reported by Clinton et al., who found that HSIL detection increased from 48–76% (P < 0.05) after the wide clinical application of p16INK4a immunostaining.11
In our study, two patients with H&E-diagnosed CIN2 had focal positive p16INK4a immunostaining. The diagnoses of these two patients would be classified as LSIL according to the recommendations of the 2012 LAST guidelines, and follow-up may be conducted for management. The two patients also underwent cervical LEEP in the next menstrual cycle after cervical biopsy, and the results were CIN1 and CIN2, respectively.
3.3 Accurate management of women with a high risk of cervical cancer using p16 INK4a immunostaining
The following are preliminary suggestions for the clinical management of HR-HPV-negative patients from the current limited number of available cases. Firstly, when the results are HPV- and p16INK4a-negative, the possibility that the biopsy pathology is ≤ LSIL is high, and the NPV of p16INK4a for HSIL + is almost 100% thus, follow-up is suggested. Secondly, when the results are HPV-negative and p16INK4a–focal positive, and biopsy pathology is CIN2, there is the possibility of CIN2 + pathology after cervical LEEP, and continuous follow-up is suggested. Thirdly, when the results are HPV-negative and p16INK4a-positive, and biopsy pathology is CIN2, it indicates an actual high-grade precancerous lesion of the cervix, and intervention is advised.
Limitations. This study had several limitations. Firstly, this trial was not a randomised controlled trial. Secondly, this study was only a single-centre trial, and the sample size was limited. Thirdly, the clinical follow-up was short, and it was necessary to observe the clinical long term prognosis. Fourthly, diagnostic cervical LEEP can be conducted for elderly women, women with persistent CIN2 for > 2 years, or patients with other risk factors. As there were only two such patients in this study, more cases should be gathered in the future for an in-depth study to facilitate suggestions for suitable management.
Conclusion: Immunostaining for p16INK4a can be used as a molecular biological substitute for evaluating pathological HSIL of cervical tissue in HR-HPV-negative patients and can be used to aid HSIL detection.