Significance and limitations of routine p16/Ki-67 immunohistochemistry as a diagnostic tool for high-grade squamous intraepithelial lesions of the uterine cervix
Article information
Abstract
Objective
To evaluate the diagnostic utility and limitations of routine p16 and Ki-67 immunohistochemistry (IHC) in detecting high-grade squamous intraepithelial lesions (HSILs) in the uterine cervix.
Methods
We reviewed 2,061 cervical biopsy records, including 271 morphologically indeterminate squamous lesions, evaluated using p16/Ki-67 IHC for HSIL detection or exclusion. HSIL was diagnosed based on p16 positivity and a high Ki-67 labeling index (Ki-LI). In cases that remained inconclusive after IHC, follow-up histological and/or cytological outcomes were assessed.
Results
p16/Ki-67 IHC established a definitive diagnosis of either HSIL or non-HSIL in 74.2% (201/271) of morphologically indeterminate cases, whereas 25.8% (70/271) remained inconclusive. p16/Ki-67 IHC contributed to diagnosing 120 HSIL cases, representing 11.9% (120/1,011) of all HSILs cases and 44.3% (120/271) of morphologically indeterminate cases. Among the 70 inconclusive cases, 58 had available follow-up data, of which 22 were subsequently diagnosed with HSIL, including 12 within 1 month of the initial biopsy. HSIL outcomes were more frequent in cases with suspicious HSIL on the initial biopsy (66.7% [12/18]). Based on the p16/Ki-LI status observed in the initial biopsy, patients with HSIL outcomes were categorized into three groups: p16-positive/low Ki-LI (54.2% [13/24]), p16-negative/high Ki-LI (50.0% [5/10]), and p16-negative/low Ki-LI (16.7% [4/24]). Multiple comparisons revealed a significant difference between the p16-positive/low Ki-LI and p16-negative/low Ki-LI groups (Benjamini-Yekutieli adjusted P=0.0435), while other comparisons were not significant.
Conclusion
p16/Ki-67 IHC significantly improved the diagnostic performance for HSIL. In cases that remain inconclusive after IHC, IHC-based risk stratification offers a valuable approach for surveillance, thus mitigating delays in HSIL diagnosis.
Introduction
Cervical cancer ranks as the fourth most common malignancy and fourth leading cause of cancer-related deaths among women worldwide, with an estimated 604,000 new cases and 342,000 deaths reported in 2020 [1]. In Japan, cervical cancer occurs less frequently and is ranked ninth in cancer-related mortality, with an estimated 12,800 new cases and 4,200 deaths, based on updated registry data [2]. Owing to effective cervical cancer screening programs and the widespread adoption of human papillomavirus (HPV) vaccination, many socioeconomically developed countries have shown stable or decreasing trends in the incidence and mortality rates of cervical cancer over the past decade [3].
Most cervical cancers are associated with HPV, including squamous cell carcinoma, adenocarcinoma, and precursor lesions such as squamous intraepithelial lesions (SILs), cervical intraepithelial neoplasia (CIN), and adenocarcinoma in situ [4]. Early detection of precursor lesions, which is crucial for ensuring optimal patient management and prevention of cancer-related deaths, involves ancillary studies targeting HPV, such as HPV testing, in situ hybridization, and p16 immunohistochemistry (IHC). p16 is a cyclin-dependent kinase inhibitor that accumulates in the cells of HPV-driven neoplasms due to the integration of the viral genome, particularly the E7 gene, which disrupts the negative feedback mechanism [5]. Therefore, p16 could be used as a surrogate diagnostic marker for HPV-driven neoplasms. High-grade squamous intraepithelial lesions (HSILs), including CIN grade 2 (CIN2) and CIN grade 3 (CIN3), showed positive p16 staining, whereas reactive or metaplastic epithelia were negative. Based on its diagnostic significance, the lower anogenital squamous terminology (LAST) project recommended p16 IHC to distinguish HSILs from benign mimics [6]. Owing to its limited specificity, the LAST guidelines do not endorse p16 IHC to confirm straightforward HSILs or low-grade squamous intraepithelial lesions (LSILs), that is, CIN grade 1 (CIN1). This limitation has prompted the use of additional biomarkers, with extensive research demonstrating that combining p16 and Ki-67 IHC improves HSIL diagnostic accuracy [7-16]. However, to the best of our knowledge, only a few studies have systematically evaluated the limitations of p16/Ki-67 IHC, with a focus on outcomes of indeterminate cases during follow-up. Therefore, to elucidate the significance and limitations of the routine use of p16/Ki-67 IHC in real-world diagnostic practice, we retrospectively reviewed outcomes of indeterminate squamous lesions that remained inconclusive despite the use of IHC to diagnose or exclude HSILs.
Materials and methods
1. Case selections
A total of 2,381 consecutive cervical biopsies performed over 7 years (January 1, 2015 to December 31, 2021) were retrieved from the database of the Department of Diagnostic Pathology at Kumamoto University Hospital. Pathology records were reviewed using the pathology diagnostics system EXpath (Laboratory Information Systems for Pathology; INTEC Inc, Tokyo, Japan). The data included histological diagnoses, histological findings, and p16/Ki-67 IHC results from the initial biopsy. Patients with invasive carcinomas, including microinvasive carcinomas, and those who underwent chemoradiation and/or surgical therapy were excluded. Of the 2,381 cases initially retrieved, 320 met at least one of the exclusion criteria, leaving 2,061 cases eligible for inclusion in the study. Additionally, we obtained available data regarding follow-up histology and/or cytology in 58 cases in which the initial biopsy results were inconclusive despite p16/Ki-67 IHC assistance.
2. Histopathology
In all cases, hematoxylin and eosin (HE)-stained slides were examined by four board-certified pathologists under the supervision of the chief pathologist (Y.M.) specializing in gynecological pathology and oncology. Prior to the release of the 5th edition of the World Health Organization (WHO) classification of female genital tract tumors in 2020 [17], LSIL and HSIL diagnoses were based on the 4th edition, published in 2014 [18]. Since the publication of the 5th edition in 2020, LSIL and HSIL have been diagnosed according to the latest editions. The diagnostic criteria for LSIL and HSIL did not differ significantly between the 4th and 5th editions. Therefore, the original pathological data for all cases were analyzed without slide re-evaluation. Other diagnostic terms included the following: negative for intraepithelial lesion or malignancy (NILM) for reactive atypia, squamous metaplasia, and otherwise normal stratified squamous epithelium, and atypical immature metaplasia (AIM) for immature squamous metaplasia showing mild nuclear enlargement or heterogeneity in nuclear size and shape, which has been interchangeably called immature condyloma or metaplastic LSIL. In cases of morphologically inconclusive squamous lesions interpreted as suspicious but not diagnostic of HSIL, both block-positive staining for p16 and high Ki-67 labeling index (Ki-LI) confirmed the diagnosis (Supplementary Figs. 1, 2).
3. IHC
p16/Ki-67 IHC was performed on formalin-fixed, paraffin-embedded tissues using anti-human p16 mouse monoclonal antibodies (cloneE6H4, 1:1; Roche Diagnostics International AG, Rotkreuz, Switzerland) and anti-human Ki-67 monoclonal mouse antibodies (cloneMIB-1, 1:50; DakoCytomation, Glostrup, Denmark). The protocol was as follows: a fully automated system using an automated slide stainer (Benchmark ULTRA; Ventana Medical Systems, Tucson, AZ, USA) with cell conditioning 1 (CC1) solution (ethylenediaminetetraacetic acid buffer [Ventana Medical Systems]).
When using the automated slide stainer, the slides were deparaffinized and buffered in CC1 solution at 95°C for 64 minutes. The slides were then incubated with the primary antibody at 36°C for 16 minutes, followed by amplification using an UltraView Polymer Detection Kit (Ventana Medical Systems) with diaminobenzene as the chromogen and hematoxylin counterstaining, following the manufacturer’s protocol.
4. Interpretation of p16 IHC results
Strong and diffuse block-positive results were interpreted as positive. For example, HSILs exhibited continuous staining of abnormal cells. Nuclear staining, with or without cytoplasmic staining, was observed at the basal side of the epithelium at a minimum and extended into the upper portion depending on the lesion grade. CIN3 is typically characterized by full-thickness positive staining, while CIN2 shows a gradient of negative or weakly positive staining in the upper portion of the epithelium, reflecting superficial differentiation. The following staining patterns were considered negative: absence of staining, rare singly dispersed cell staining, patchy but strong staining, and discontinuous staining in the basal cells. Cytoplasmic staining was considered negative.
5. Ki-LI
Based on previous studies evaluating Ki-67 overexpression in LSILs and HSILs [7-9,11,13], a cutoff value of 50% on the basal side and 5% on the superficial layer was utilized in the present study to define high Ki-LI. For example, p16-positive atypical squamous epithelium with Ki-LI exceeding 50% on the basal side and/or 5% on the superficial layer was considered indicative of HSIL.
6. Statistical analysis
All statistical analyses were performed using EZR software (Saitama Medical Center; Jichi Medical University, Saitama, Japan) [19], which is a graphical user interface for R Foundation (The R Foundation for Statistical Computing, Vienna, Austria). It is a modified version of the R commander designed to add statistical functions frequently used in biostatistics. Fisher’s exact test was used to assess the statistical significance of differences between groups (significance level P<0.05). For significant comparisons, the Benjamini-Yekutieli procedure was applied to control for multiple comparison correction, and adjusted P-values of less than 0.05 were considered statistically significant. This procedure was chosen for the stringent control of the false discovery rate under test dependency, making it suitable for the conservative assessment in this study.
7. Ethical approval
The protocol for the present retrospective observational study was approved by the Ethics Committee of Kumamoto University Hospital (no. 2743) and was performed using the opt-out method on our hospital website (http://kuhpath.jp/uploads/1685422783_1.pdf).
Results
1. Histological examination
A total of 2,061 eligible uterine cervical biopsies were included in the analysis, as shown in Fig. 1. In 1,774 (86.1%) cases, a definite diagnosis was established based on HE-stained sections alone, categorized as follows: HSIL, 891 cases (758 HSIL/CIN3, 96 HSIL/CIN2, 37 HSIL/grade undetermined); LSIL/CIN1 62 cases; SIL (undetermined HSIL or LSIL), three cases; and others. HE staining alone failed to establish a definite diagnosis in 271 cases (13.1%) of squamous lesion, including the following: HSIL vs. benign mimics, 208 cases; HSIL/CIN2 vs. LSIL/CIN1, 50 cases; HSIL/CIN3 vs. HSIL/CIN2, 10 cases; and HSIL vs. AIM, three cases.

Flow chart showing the case selection. HSIL, high-grade squamous intraepithelial lesion; AIS, adenocarcinoma in situ ; SMILE, stratified mucin-producing intraepithelial lesion; CIN, cervical intraepithelial neoplasia; LSIL, low-grade squamous intraepithelial lesion; SIL, squamous intraepithelial lesion; AIM, atypical immature metaplasia; NILM, negative for intraepithelial lesion or malignancy; IHC, immunohistochemistry.
2. IHC examination
p16/Ki-67 IHC was performed in 271 cases of morphologically indeterminate squamous lesions identified on HE sections, accounting for 13.1% of the total eligible biopsy specimens. As shown in Fig. 1, a conclusive diagnosis was established in 74.2% (201/271) of the cases while 25.8% (70/271) remained inconclusive. The conclusive diagnoses made using IHC included the following: HSIL, 120 cases (85 HSIL/CIN3, 30 HSIL/CIN2, 5 HSIL/grade undetermined); LSIL/CIN1, 18 cases; immature condyloma, three cases; reactive atypia, 14 cases; and NILM, 46 cases. p16/Ki-67 IHC contributed to the definitive diagnosis of 120 HSIL cases, accounting for 11.9% of all 1,011 HSIL cases in this series, comprising 891 cases identified on HE sections alone and 120 cases identified with the aid of IHC. The diagnostic impact of p16/Ki-67 IHC for a definite diagnosis was significantly higher in the “HSIL vs. LSIL” than in the HSIL group (88% [44/50] vs. 69.2% [144/208]; P=0.007432) (Supplementary Table 1). In the “HSIL vs. benign mimics” group, the addition of p16/Ki-67 IHC led to a final diagnosis of HSIL/CIN3 in 54.2% (78/144) of the cases (Supplementary Table 2). Conversely, in the “HSIL and LSIL” groups, 38.6% (17/44) of the cases were ultimately diagnosed as LSIL/CIN1 using p16/Ki-67 IHC. In the “CIN3 vs. CIN2” group, adjunctive p16/Ki-67 IHC classified 9 of 10 cases as HSIL/CIN2, with only one as HSIL/CIN3.
3. Cases with morphologically and immunohistochemically indeterminate squamous lesions
In 70 cases (25.8%) of 271 morphologically indeterminate squamous lesion cases, p16/Ki-67 IHC failed to distinguish between HSILs and non-HSILs (Fig. 1). The diagnoses rendered for these 70 lesions included the following: suspicious HSIL, 22 cases; undetermined HSIL or reactive atypia, 12 cases; SIL (undetermined HSIL or LSIL), one case; AIM, 31 cases; and suspicious LSIL, four cases. Representative microscopic features and p16/Ki-67 IHC results for these indeterminate cases are shown in Fig. 2. In the case of suspicious HSIL (Fig. 2A-C), the morphology and block-positive staining for p16 suggested HSIL, whereas low Ki-LI did not support a definite diagnosis. In the case of undetermined lesions, that is, HSIL vs. reactive atypia (Fig. 2D-F), high Ki-LI favored HSIL, whereas p16 IHC did not. In the case of AIM (Fig. 2G-I), which was p16-positive and showed low Ki-LI, relatively moderate nuclear abnormalities failed to exclude HSIL. In another case of p16-negative AIM (Fig. 2J-L), nuclear enlargement was mild to moderate, but nuclear hyperchromasia was significant and Ki-LI was high; therefore, HSIL could not be excluded in this particular case.

Representative histologies and p16/Ki-67 IHC results of Indeterminate Lesions in high-magnification fields (×400). Cases of suspected HSIL: (A) HE; (B) p16; and (C) Ki-67. Cases of undetermined HSIL or reactive atypia: (D) HE; (E) p16; and (F) Ki-67. p16-positive AIM: (G) HE; (H) p16; and (I) Ki-67. p16-negative AIM: (J) HE; (K) p16; and (L) Ki-67. HSIL, high-grade squamous intraepithelial lesion; AIM, atypical immature metaplasia; IHC, immunohistochemistry; HE, hematoxylin and eosin.
4. Outcomes of follow-up for indeterminate squamous lesions after IHC-aided evaluation
Among the 70 indeterminate squamous lesions evaluated by p16/Ki-67 IHC, follow-up histologic and/or cytologic data were available for 58 cases, of which 22 (37.9%) were subsequently diagnosed as HSIL (Table 1). HSIL was diagnosed on histology alone in five cases, cytology alone in five cases, and both in 12 cases (Table 1, Supplementary Table 3) on follow-up. When stratified by initial biopsy diagnosis, patients with suspicious HSIL exhibited a significantly higher frequency of HSIL outcomes than those with AIM (66.7% [12/18] vs. 23.1% [6/26]; P=0.00559; Benjamini-Yekutieli adjusted P=0.01677) after applying multiple comparisons. In contrast, no significant differences were observed between cases with suspicious HSIL and those with undetermined HSIL or reactive atypia (66.7% [12/18] vs. 27.3% [3/11]; P=0.0604; Benjamini-Yekutieli adjusted P=0.09060) (Table 1). When stratified by p16 status, p16-positive cases showed higher HSIL outcomes than p16-negative cases (54.2% [13/24] vs. 26.5% [9/34]; P=0.0536) and lower NILM outcomes (45.8% [11/24] vs. 67.6% [23/34]; P=0.113); however, the differences were not significant (Table 2).

Worst grade of histologic and/or cytologic follow-up outcomes based on the initial diagnosis in 58 cases with inconclusive results on HE+p16/Ki-67 IHC

Worst grade of histologic and/or cytologic follow-up outcomes based on p16 status and p16/Ki-LI status in 58 cases with inconclusive results on HE+p16/Ki-67 IHC
Based on the p16/Ki-LI status observed in the initial biopsy, patients with HSIL outcomes were categorized into three groups: p16-positive/low Ki-LI (54.2% [13/24]), p16-negative/high Ki-LI (50.0% [5/10]), and p16-negative/low Ki-LI (16.7% [4/24]) (Table 2). Multiple comparisons revealed significant differences between the p16-positive/low Ki-LI and p16-negative/low Ki-LI groups (P=0.0145; Benjamini-Yekutieli adjusted P=0.0435), but not between the p16-negative/high Ki-LI and p16-negative/low Ki-LI groups (P=0.0847; Benjamini-Yekutieli adjusted P=0.12705). No multiple comparison analysis was conducted between p16-positive/low Ki-LI and p16-negative/high Ki-LI (P=1.0) owing to minimal differences.
NILM outcomes were observed in 45.8% (11/24) of p16-positive/low Ki-LI cases, 50.0% (5/10) of p16-negative/high Ki-LI cases, and 75.0% (18/24) of p16-negative/low Ki-LI cases (Table 2). Fisher’s test showed no significant differences between the p16-positive/low Ki-LI and p16-negative/low Ki-LI groups (P=0.0753) or between the p16-negative/high Ki-LI and p16-negative/low Ki-LI groups (P=0.232).
Among the 22 cases with subsequent HSIL diagnoses, 54.5% (12/22) and 77.3% (17/22) had HSIL outcomes within 1 and 6 months, respectively (Table 3), and 22.7% (5/22) had HSIL outcomes in more than 6 months, including two cases between 6 and 12 months and three cases after 12 months. Among the p16-positive/low Ki-LI cases, the interval from initial biopsy to subsequent diagnosis of HSIL was less than 1 month in 53.8% (7/13) of cases (Table 3). The incidence of HSIL in AIM cases was higher in the p16-positive group compared to the p16-negative group; however, the difference was not statistically significant (42.9% [3/7] vs. 15.8% [3/19]; P=0.293) (Table 4). Similarly, the NILM rate did not differ significantly between p16-positive and p16-negative cases (57.1% [4/7] vs. 73.7% [14/19]; P=0.0635).

Shortest interval from initial biopsy to subsequent HSIL diagnosis in 22 cases with an initial inconclusive diagnosis
Discussion
The present study demonstrates the clinical utility of p16/Ki-67 IHC in confirming or excluding HSIL to improve diagnostic accuracy. The use of p16/Ki-67 IHC reduces the risk of overdiagnosis and under-diagnosis, thereby supporting appropriate patient management. This technique is valuable not only for differentiating HSIL from benign mimics or LSIL but also for distinguishing HSIL/CIN2 from HSIL/CIN3. This distinction is crucial because the American Society for Colposcopy and Cervical Pathology guidelines in the United States recommend similar management for HSIL/CIN2 and HSIL/CIN3, including ablation or excision, except in younger patients with HSIL/CIN2, where conservative management is preferred [20]. Conversely, HSIL/CIN2 is typically managed with conservative follow-up in Japan [21]. Our findings align with previous reports [9-11,13,15], reaffirming the clinical significance of p16/Ki-67 IHC in diagnosing morphologically inconclusive squamous lesions and supporting the LAST guidelines [6]. However, as mentioned in the LAST project, the accuracy of morphological evaluation may affect the impact of p16 IHC on HSIL detection. For example, approximately 30% of LSIL cases are positive for p16 and a subset of such cases may be erroneously diagnosed as HSIL/CIN2. However, it is unlikely that morphologically unequivocal HSIL/CIN3 would be diagnosed as a benign mimic because of its p16 negativity. These findings emphasize that IHC should be considered based on a critical evaluation of morphology and that the results should be interpreted in the context of morphology. The frequency of p16 IHC used for squamous lesions in our routine practice was 13.1%, which is comparable to previous studies, where rates ranged from 6.2% to 13.9% [22-24] and was lower than the maximum range (20-40%) proposed by the 2020 WHO classification of female genital tract tumors [17], thereby contributing to more appropriate patient care. It is conceivable that the appropriate use of p16 IHC for carefully selected indeterminate squamous lesions based on strict morphological evaluation of SIL may significantly enhance the impact of p16 IHC on HSIL detection, as shown in the present study.
Unlike p16 IHC, the LAST guidelines do not recommend the addition of Ki-67 IHC to p16 IHC due to insufficient evidence to support the accurate diagnosis of SIL [6]. Although studies have demonstrated overexpression of Ki-67 in cases of HSIL [7-13], with the LI ranging from 5% to 50%, the optimal cutoff value for Ki-LI remains undetermined because of the lack of standardized diagnostic criteria for SIL and/or study design. However, one study sheds light on the combination of Ki-67 and p16 IHC, demonstrating that, in cases showing block-positive staining for p16, this combination was slightly less sensitive but more specific for HSIL diagnosis than p16 IHC alone. This study emphasizes the importance of the distribution of Ki-67-positive cells: diffuse staining confined to the lower third, superficial staining with skipped areas between the parabasal and upper layers, or continuous staining extending beyond the lower third of the epithelium [8]. Moreover, one study showed that the presence of HSIL/CIN3 in conization specimens could be predicted by Ki-LI exceeding 50% in a prior biopsy diagnosed with CIN2 [11]. Based on these previous data, we utilized a cutoff value of 50% in the basal layer and/or 5% in the superficial layer (with skip areas usually between the parabasal and upper zones) to determine low or high Ki-LI values.
Another finding of the present study is the outcome of inconclusive cases following the use of p16/Ki-67 IHC. Findings of p16 block positivity and high Ki-LI supported the final diagnosis in 11.9% of all 1,011 HSIL cases in this series. In contrast, indeterminate cases were followed up by histology and/or cytology after the initial biopsy using p16/Ki-67 IHC. However, 45.8% of the p16-positive/low Ki-LI cases were regarded as indeterminate for HSIL based on low Ki-LI and had a subsequent NILM outcome, indicating that p16 positivity alone might be a source of overinterpretation, whereas the addition of Ki-LI using the aforementioned cut-off values contributed to preventing overestimation. Conversely, 54.2% of the p16-positive/low Ki-LI cases also had a subsequent diagnosis of HSIL, suggesting a potential underestimation and delayed diagnosis of HSIL by Ki-67 IHC. Nevertheless, in more than half of the p16-positive/low Ki-LI cases with subsequent HSIL outcomes, the diagnostic delay for HSIL was less than 1 month. Therefore, appropriate and strict histological and/or cytological follow-ups appear to be helpful in avoiding delays in or missing HSIL diagnoses in the indeterminate group. This study has limitations in terms of the absence of data regarding Ki-LI or the distribution pattern of Ki-67-positive cells in cases of morphologically straightforward HSIL, LSIL, and NILM. This is because the study was based on the analysis of information retrieved from the pathology database without a thorough slide review, although the diagnostic criteria have been controlled for routine diagnostic practice at this single institution. Further studies may elucidate the optimal cutoff value of Ki-LI and/or the distribution pattern of Ki-67-positive cells for the diagnosis of HSIL or evaluation of the risk for subsequent HSIL diagnoses.
The predictive value of p16 for subsequent HSIL diagnoses among LSIL cases has been suggested in some studies [25-28], while others have argued that the correlation between p16 status and subsequent HSIL outcomes is insignificant [29]. Regarding Ki-67 IHC, very few studies have explored its predictive value for subsequent HSIL outcomes [27,28]. The present study showed that the combination of p16 and Ki-67 IHC was informative because HSIL was diagnosed significantly more frequently in p16-positive/low Ki-LI cases than in p16-negative/low Ki-LI cases. However, inconclusive squamous lesions after p16/Ki-67 IHC encompass a variety of conditions such as bona fide HSIL, LSIL, AIM, reactive atypia, and atrophy. Therefore, alternative diagnostic markers are required to distinguish between HSIL and benign conditions. Previously suggested diagnostic markers include cytokeratin 7 and ProExC, which may be used as progression markers for LSIL-HSIL [26-28]. However, the available data are limited and further studies are needed to justify the routine use of these potential markers as substitutes or adjuncts to p16/Ki-67 IHC.
From a clinical perspective, a management strategy based on risk stratification after an equivocal biopsy diagnosis is beneficial. The present study’s follow-up results demonstrated that patients with suspicious HSIL on the initial biopsy or those with a p16-positive/low Ki-LI status frequently received subsequent HSIL diagnoses within a short period, often just 1 month. Therefore, in cases of p16-positive/low Ki-LI indeterminate lesions, it might be advisable to strongly recommend re-examination by histology or cytology. However, factors influencing the re-examination interval, such as colposcopy results, were not investigated in the present study. The p16/Ki-67 IHC may become a powerful tool adjusted for the HPV genotype (HPV16/18 or HPV others), which could overcome the discrepancy between cytology and colposcopic examination with biopsy [30-34]. Among women over 45 years of age, biopsy underestimation is relatively common compared to cytology because the squamous columnar junction is deeper inside the cervical canal after menopause [30]. Under such circumstances, a constellation of findings, including age, HPV genotype, and results of p16/Ki-67 IHC, should be considered to determine the interval of follow-up biopsy or cytology.
In summary, the combination of p16 and Ki-67 IHC improved the diagnostic performance of cervical biopsies and contributed to the diagnosis of HSIL in a large proportion of morphologically indeterminate cases. Although p16/Ki-67 IHC has some limitations, it provides information for determining optimal management after the initial biopsy by stratifying the risk of bona fide HSIL. Particularly for p16-positive/low Ki-LI indeterminate lesions, shortening the follow-up interval may help avoid delays in subsequent HSIL diagnosis. The findings of this single-center retrospective study may also provide beneficial clinical information for the development of a more appropriate pre-cancer screening strategy. Further research on the natural course of indeterminate squamous lesions after p16/Ki-67 IHC is required to establish a more powerful pre-cancer screening strategy.
Supplementary Information
Notes
Conflict of interest
The authors have no conflicts of interest to declare.
Ethical approval
The protocol for the present retrospective observational study was approved by the Ethics Committee of Kumamoto University Hospital (no. 2743).
Patient consent
Not applicable.
Funding information
None.