1. Introduction

Oral and oropharyngeal cancer are the most frequent malignant tumors affecting the oral and maxillofacial region. Due to their high mortality and detrimental functional consequences, these cancers have emerged as a pressing public health issue. An estimated 495,801 new cases and 240,498 deaths from oral squamous cell carcinoma (OSCC) and oropharyngeal squamous cell carcinoma (OPSCC) are predicted worldwide in 2022 [1]. Despite dedicated efforts from clinicians and researchers over the past 3 decades to refine the diagnosis and treatment of oral cancer, the incidence and mortality of oral cancer are continuously increasing, according to a Global Burden of Disease Study carried out in the 10 most populous nations [2]. Squamous cell carcinoma (SCC) is the most prevalent histological type of oral and oropharyngeal cancer [3]. Its highly invasive nature, coupled with its tendency to trigger cervical lymph node metastasis, leads to a critical challenge in therapy, ultimately limiting patient survival time. Numerous studies have searched for biomarkers for OSCC and OPSCC, such as p53, CD44, and VEGF [4–6]. However, due to limited sample sizes and other reasons, no definitive prognostic predictive biomarker has been established for clinical practice [7, 8].

A complicated, genetically encoded regulatory mechanism dictates the control of cell proliferation. In cancer, this regulatory mechanism breaks down and leads to uninhibited tumor cell growth. Ki-67, a macromolecular protein located within the cell nucleus, exhibits a presence during cell mitosis in the G1, S, G2, and M phases, reaching peak expression in the M phase. After completion of cell mitosis, Ki-67 diminishes rapidly and is absent in quiescent (G0) cells [9]. Therefore, Ki-67 is a valuable marker for rapidly and precisely identifying cell proliferation activity within both normal and tumor cell populations. Numerous studies have consistently demonstrated that the expression of Ki-67 is upregulated in cancerous tissues and is significantly associated with an unfavorable prognosis in various types of tumors, including but not limited to lung cancer, breast cancer, and bladder cancer [10–12]. Gadbail et al. [13] have revealed a significant correlation between elevated Ki-67 expression and the poor differentiation status of tumor cells in the context of oral submucosal fibrosis and OSCC. Moreover, high Ki-67 expression was associated with advanced clinical TNM stage, metastasis, and shorter survival duration (less than three years) in contrast to patients with early clinical TNM stage, no metastasis, and longer survival (more than three years). These findings suggest a potential association between Ki-67 and an unfavorable prognosis in OSCC and OPSCC [14].

Although several studies have suggested that Ki-67 is an adverse prognostic predictive marker for OSCC and OPSCC [14–16], conflicting perspectives have emerged, with some scholars suggesting no discernible correlation between Ki-67 expression and prognosis [17–19]. The conflict here may argue the limitations of sample size which may underscore the necessity for data validation from larger sample cohorts. To address this, the present study meticulously compiled clinical and pathological data from 469 OSCC/OPSCC patients hospitalized between 2012 and 2022. The patients were followed up until December 31, 2022, to evaluate the association between the expression of Ki-67 and the clinicopathological features of OSCC/OPSCC patients, as well as explore its potential impact on prognosis.

2. Materials and Methods

3. Results

3.3. Relationship Between Ki-67 Expression and Prognosis

As of December 31, 2022, patients’ overall survival and outcome were obtained through follow-up. The relationship between clinicopathological features and patient prognosis in OSCC and OPSCC was compared using univariate Cox regression analysis. All variables except sex were found to affect the prognosis. The Cox proportional hazard model was established after stepwise regression which eliminates the mediating variables, pathological grade, and clinical TNM stage, and it was finally determined that Ki-67 (p < 0.01), age (p < 0.05), recurrence (p < 0.001), metastasis (p < 0.001), and primary tumor site (p < 0.05) were independent indicators of patient’s prognosis. Patients with higher Ki-67 expression, older age, recurrence, metastasis, and tumor located in the oropharynx had a worse prognosis than patients with lower Ki-67 expression, younger age, no recurrence, no metastasis, and tumor located elsewhere in the oral cavity (HR > 1 for all five factors) (Table 3).

Table 3. Univariate analysis and multivariate analysis of overall survival by Cox regression analysis.

Characteristic	Univariate analysis			Multivariate analysis
	HR	95% CI	p value	HR	95% CI	p value
Age	1.025	1.002–1.049	0.030	1.025	1.002–1.049	0.036
Sex
Male	1.000			—	—	—
Female	0.750	0.449–1.251	0.271	—	—	—
Grade
WDSCC	1.000			—	—	—
MDSCC	1.576	0.738–3.362	0.240	—	—	—
PDSCC	4.365	1.892–10.073	0.001	—	—	—
Stage
Early stage	1.000			—	—	—
Advanced stage	3.554	2.051–6.160	< 0.001	—	—	—
Ki-67	1.020	1.009–1.031	< 0.001	1.015	1.004–1.026	0.007
Recurrence
Nonrec	1.000
Recurrent	2.192	1.284–3.740	0.004	2.755	1.601–4.741	< 0.001
Metastasis
Nonmeta	1.000
Metastatic	3.939	2.401–6.465	< 0.001	3.978	2.373–6.671	< 0.001
Site
Oral cavity	1.000
Oropharynx	2.137	1.288–3.546	0.003	1.709	1.020–2.862	0.042

To obtain optimal Ki-67 grouping, this study implemented two different cutoff values of Ki-67 expression. Based on the data, these values were the most reasonable. One group utilized 40%, the median, as the cutoff value. The other group used the optimal cutoff value of 67.5%, as determined by the ROC curve (Figure 1). The threshold has a specificity of 0.822, a sensitivity of 0.394, and an area under the curve (AUC) of 0.633. Two Kaplan–Meier survival curves were drawn independently (Figure 2). The curves indicate that patients with high Ki-67 expression had significantly shorter survival time than those with low Ki-67 expression. Patients in the high-expression group exhibited a poorer prognosis than those in the low-expression group, with a statistically significant difference observed (p < 0.05). These findings suggest that Ki-67 is an adverse prognostic marker for OSCC and OPSCC. Furthermore, when comparing the two graphs, it has been found that grouping the expression of Ki-67 with the best cutoff value obtained from the ROC curve results in a more significant difference (p < 0.0001) in prognosis than the difference (p = 0.014) obtained by grouping the expression of Ki-67 with the median.

3.4. A Nomogram Incorporating Ki-67 Expression

A prognostic model was developed using the five independent risk factors of Ki-67 expression, age, recurrence, metastasis, and primary tumor site, which is presented in a nomogram (Figure 3). The nomogram corresponded to the patient’s Ki-67 expression, recurrence, metastasis, age, and primary site of the tumor to the scores and accumulated the scores to get the total score, and finally, the total score was used to correspond to get the prognosis of 1-, 3-, and 5-year survival rates. For example, there was a 60-year-old patient with OSCC with cervical lymph node metastasis but no recurrence after treatment with Ki-67 expression of 80%. By corresponding the patient’s condition to the score in the nomogram, it can be calculated that his total score is approximately 190, corresponding to the 1-year survival rate of 86%, the 3-year survival rate of 69%, and the 5-year survival rate of 55%. Also, the prediction accuracy of the model was validated. The C-index of the model was 0.768. The model was validated by the bootstrap method, and the calibration curve was plotted (Figure 4). From the calibration curve, it could be seen that the 3-year survival rate predicted by the nomogram is very close to the actual 3-year survival rate of the patients, indicating that the model has a good prediction efficacy. The nomogram underwent external validation as well. Upon validation using an external dataset comprising 70 patients, the C-index was determined to be 0.714. Additionally, a calibration curve was generated (Figure 5), which demonstrates that the nomogram maintains a robust predictive performance during external validation. This finding underscores the model’s reliability and generalizability.

4. Discussion

OSCC and OPSCC stand as two of the most pervasive tumors worldwide. The conventional treatment involves comprehensive and sequential therapy, with surgical resection as the preferred method, accompanied by preoperative and postoperative chemotherapy and radiotherapy [20]. These cancers significantly impact patients’ quality of life by hampering vital functions, such as swallowing, pronunciation, and appearance [21]. Despite significant improvements in treatment by various researchers in recent years, the 5-year survival rate for patients with OSCC or OPSCC has not been significantly improved. Patients often die from local recurrence or metastasis after surgery. Therefore, there is an urgent need to identify a biomarker capable of accurately and efficiently predicting the risk of metastasis, recurrence, and prognosis in OSCC and OPSCC.

Ki-67 was initially detected in the nuclei of Hodgkin lymphoma cells 4 decades ago [22]. As a biological marker, the expression level of Ki-67 is intimately linked to tumor development and is extensively utilized as a prognostic and predictive indicator in cancer diagnosis and treatment [23]. To begin with, as a marker of cellular proliferation, elevated Ki-67 expression typically signifies an accelerated rate of tumor cell division, which directly fuels cancer progression. Furthermore, Ki-67 exerts significant influence within the tumor microenvironment. The expression level of Ki-67 is closely associated with the invasive phenotype of tumor cells. Specifically, higher expression levels of Ki-67 typically indicate a greater degree of invasiveness in tumor cells [24]. In addition, it is also correlated with the aggregation and functionality of tumor-associated macrophages (TAMs). TAMs are capable of secreting immunosuppressive factors that diminish the immune system’s antitumor capabilities, thereby fostering an environment conducive to tumor growth and advancement [25]. In OSCC and OPSCC, there are two conflicting perspectives, potentially attributed to sample size limitations or variations in tumor sampling sites. To address this, the proposal suggests that performing multisite tumor sampling could enhance the detection of intratumor heterogeneity in OSCC/OPSCC [26]. This study included 469 patients hospitalized between 2012 and 2022, and Ki-67 was obtained as an independent adverse prognostic marker for OSCC/OPSCC according to univariate and multivariate Cox regression analyses. Patients with high Ki-67 expression demonstrated a poor prognosis and short overall survival. Additionally, age, recurrence, metastasis, and primary tumor site were identified as independent prognostic factors. Old age and the presence of recurrence/metastasis were recognized as risk factors for tumor prognosis [27]. The prognosis for SCC in the oropharynx was significantly worse than oral cavity. We speculate that there is a potential link between this difference and human papillomavirus infection [28].

This study comparatively analyzed Ki-67 expression, revealing its correlation with various factors including sex, site, clinical stage, pathological grade, and metastasis. The Ki-67 expression was significantly higher in males compared to females, indicating a generally elevated cell proliferation activity in males with OSCC and OPSCC. Previous study has suggested that this may be related to risk factors such as smoking and alcohol abuse [29]. There was no difference in Ki-67 expression between the different age groups, corroborating findings from earlier studies [12, 30, 31]. Ki-67 expression increased with clinical stage progression, consistent with earlier findings [32–35]. This suggests that tumors advance more rapidly as Ki-67 expression increases. Moreover, the expression of Ki-67 in the metastatic group was significantly higher than that in the nonmetastatic group, indicating that the increased cell proliferative activity reflected by elevated Ki-67 also promotes cancer metastasis, the biological mechanism of which needs to be further clarified. However, the study did not establish a significant association between local recurrence, a pivotal biological behavior closely linked to prognosis, and Ki-67 expression (p < 0.05).

In this study, we further investigated the effect of different cutoff values of Ki-67 expression on prognostic prediction. Upon a comprehensive review of pertinent literature, it was observed that a precise grouping method for Ki-67 remains currently elusive. Some researchers utilized the median, a common grouping method, to categorize Ki-67 into two groups of high expression and low expression [16], while others have directly used 50% as the cutoff value [29]. In order to account for the frequency distribution of Ki-67 expression, some investigators establish their own cutoff values to avoid creating very small and too many groups [15]. For instance, Liu et al. [31] utilized 25%, 50%, and 75% as their cutoff values to divide Ki-67 into four groups. To find the cutoff value of the expression of the marker, some researchers performed the ROC curve analysis and obtained a Ki-67 cutoff value of 27.75 [17]. In this article, the study utilized both the median and the optimal cutoff value obtained by the ROC curve, respectively, classifying Ki-67 into two groups. The results indicate that the optimal cutoff value identified through the ROC curve had better grouping performance, and the difference in prognosis between the two groups was more significant. Xie’s meta-analysis [14] revealed that using different antibodies in immunohistochemistry and different cutoff values in data analysis may affect identifying the prognostic role of Ki-67 expression. Thus, future research is necessary to establish a comprehensive set of standardized procedures for Ki-67 immunohistochemical identification, and the setting of cutoff value to conduct in-depth research on Ki-67.

Immunohistochemical staining has become a routine method for tumor pathological examination, and most patients’ surgical specimens will be subjected to immunohistochemical staining. In this paper, we analyzed Ki-67 expression as a crucial component of the prognostic model. Specifically, we included the Ki-67 expression as a continuous variable in the nomogram. This approach enhanced the applicability of the nomogram and avoided the impact of varying samples and cutoff values. In clinical practice, the nomogram allows clinicians to predict a patient’s 1-, 3-, and 5-year survival rates based on essential clinical information and Ki-67 expression. The survival rates can be obtained through simple calculations and used as a reference for treatment planning.

Nevertheless, this study failed to adequately consider the impact of different treatment modalities on patient prognosis. On the one hand, with the gradual implementation of integrated sequential therapy, which emphasizes the necessity of postoperative complementary treatments such as radiotherapy and chemotherapy for patients with locally advanced disease to prolong their survival, the survival of patients with locally advanced disease and high Ki-67 expression has been extended. On the other hand, due to various factors, including intolerance to postoperative radiochemotherapy and refusal of such treatments, some patients did not receive treatment according to the standard treatment protocol. This deviation from the standard treatment plan may have had the opposite effect on the study conclusions. Therefore, the treatment modality may be a confounding factor in the correlation between Ki-67 and prognosis.

Our study confirmed a correlation between increased expression of Ki-67 and worse prognosis in patients with OSCC/OPSCC. Furthermore, the potential use of Ki-67 as a therapeutic target for OSCC/OPSCC treatment is also being investigated. Numerous studies have utilized diverse techniques to inhibit Ki-67 expression in tumor cells for treatment. For instance, one study highlighted the efficacy of trichosanthin, in combination with granzyme B, which effectively downregulated Ki-67 expression in tongue cancer, leading to the inhibition of tumor cell proliferation and the acceleration of tumor cell apoptosis, consequently achieving a therapeutic effect [36]. Another study by Listiyana et al. [37] employed an ethanolic extract from C. cinerariifolium leaf to decrease the expression of Ki-67 in tumor cells and the degree of oral epithelial dysplasia. Further comprehensive studies are needed to explore the potential of Ki-67 as a promising therapeutic target for OSCC/OPSCC.

In conclusion, our results strongly support that Ki-67 is a negative prognostic marker for OSCC and OPSCC. Our investigation also explored the association between Ki-67 and a range of clinicopathological features. A prognostic prediction model that incorporated Ki-67 expression was developed, providing a new tool for predicting patient prognosis in clinical practice.

Ethics Statement

This study was approved by the West China Hospital of Stomatology Institutional Review Board (ethics approval number: WCHSIRB-D-2023-437). And the Declaration of Helsinki protocols were followed throughout the study. All patients provided written consent for the use of their specimens for research purposes; none were identifiable.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Author Contributions

Tianyi Wang led the conceptualization, methodology, formal analysis, investigation, and data curation and wrote the original draft. Bing Yan collaborated with Tianyi Wang in conceptualization, methodology, formal analysis, and investigation and was responsible for validation, manuscript review and editing, supervision, and project administration. Haoran Ding and Shiyong Zhuang contributed to formal analysis, investigation, and data curation. Yi Li was responsible for validation, review and editing, and supervision. All authors participated in the manuscript review.

Funding

This study was funded by the Science and Technology Program of Tibet Autonomous Region (grant no. XZ202402ZY0004).