Radiotherapy (RT) is an effective treatment for advanced oropharyngeal cancer (OPC), especially when combined with chemotherapy if surgical treatment is not feasible, preserving functions, such as speech and swallowing (1). However, the effectiveness of radiotherapy can vary among individuals. Factors such as tumor stage, status of HPV infection (2), the patient’s nutritional status, and tumor immunity have all been associated with treatment outcomes (3,4). Namely, a poor nutritional status may increase the risk of infections and complications during treatment, making it difficult to complete the course of treatment. Nutritional status also has been associated with tumor immunity, and it has been reported that an individual’s immune system function can affect their response to treatment (5-7). Therefore, assessment of nutritional status and tumor immunity prior to treatment is an important prognostic factor in predicting outcome of radiation therapy for oropharyngeal cancer.

The controlling nutritional status (CONUT) score was developed as a scoring system to evaluate the nutritional status of patients using three indices: serum albumin, total lymphocyte count, and total cholesterol level (8), there have been reports of the prognostic utility of the CONUT score in the surgical treatment of several types of cancers (9-11). However, little has been reported on its usefulness in predicting the outcome of radiotherapy. Therefore, this study examined the pretreatment nutritional status and factors involved in tumor immunity (especially tumor-infiltrating CD8 or FoxP3-positive T lymphocytes) in patients with OPC undergoing chemoradiotherapy (CRT) at our institution. The goal was to develop a risk score to identify patients who would benefit from radiation therapy in combination with these factors.

Patients. This retrospective study was approved by the Ethics Committee of Sapporo Medical University Hospital (No. 312-168). All patients provided opt-out consent for the use of their clinical information and biopsy specimens. The study included 84 patients who underwent definitive RT, CRT, or bioradiotherapy (BRT) at our institution between May 2005 and December 2022, and who met the following criteria: [1] pathologically confirmed squamous cell carcinoma of the oropharynx; [2] received a radiation dose of 66 Gy or higher; [3] availability of sufficient clinical data, including clinical outcomes; and [4] availability of archived biopsy specimens (formalin-fixed paraffin-embedded tissue).

The diagnostic workup included visual inspection, palpation, fiberscopy, and head and neck imaging using computed tomography (CT) and/or magnetic resonance imaging (MRI). Clinical tumor staging was determined according to the American Joint Committee on Cancer (AJCC) TNM staging system, version 8 (12). The baseline characteristics of patients are shown in Table I. The CONUT score was calculated to assess nutritional status (8). Briefly, nutritional status was evaluated based on serum albumin, lymphocyte count, and total cholesterol levels, with a total score ranging from 0 to 12. A score below 3 was defined as good nutritional status, while a score above 3 indicated poor nutritional status.

Treatment. All patients were treated with curative intent. The radiation dose ranged from 66 Gy to 72 Gy in standard fractionation. Radiation therapy was delivered using 4-MV or 6-MV X-rays. Thirty-four patients were treated with three-dimensional conformal radiation therapy (3D-CRT), while fifty patients received intensity-modulated radiation therapy (IMRT). Among those who underwent chemotherapy, seven patients received induction chemotherapy prior to RT, and sixty-two were treated with concurrent CRT. The chemotherapy agents used were cisplatin, 5-FU, or S-1, either as single agents or in combination. Five patients underwent BRT with cetuximab. Seventeen patients received RT alone.

Immunohistochemical staining. Immunohistochemical staining was carried out using the methods described in a previous study (13). In brief, formalin-fixed, paraffin-embedded specimens from pretreatment endoscopic biopsy were sliced into 3-mm thick sections and mounted on glass slides. Primary antibodies against the following antigens were used: p16 (Ventana Medical Systems, Tucson, AZ, USA), CD8 (clone C8/144B; Dako, Glostrup, Denmark), FoxP3 (clone 236A/E7; Abcam, Cambridge, UK), and HLA class II [clone EPR3692 (ab92511); Abcam]. Positive p16 expression was defined as strong and diffuse nuclear and cytoplasmic staining in 70% or more of the tumor cells (2). The number of CD8- and FoxP3-positive cells was manually counted at 400× magnification under a microscope. In this study, CD8 positivity was defined as more than 270 positive cells, and FoxP3 positivity was defined as more than 60 positive cells. HLA class II positivity in the tumor was evaluated across the entire slide, and a tumor was considered HLA class II positive if the expression was 30% or higher. The immunohistochemical staining results were evaluated by two researchers (including one pathologist) who were blinded to treatment outcomes. Representative images are shown in Figure 1.

Follow-up schedule. The median follow-up duration was 54 months (range=3-191 months). Post-treatment evaluations included physical examinations, laboratory tests for tumor markers, fiberscopy, CT and/or MRI, and positron emission tomography/CT when necessary. The follow-up intervals were 1 to 2 months during the first and second years after treatment, and 3 to 6 months thereafter. The overall survival (OS) rate was calculated from the date radiation therapy ended to the date of death from any cause or the date of the last follow-up visit. The progression-free survival (PFS) rate was calculated from the date radiation therapy ended to the date of death from any cause, the date of recurrence detection, or the date of the last follow-up visit. The cut-off date for this analysis was February 2024.

Statistical analysis. The OS and PFS curves were estimated using the Kaplan-Meier method, and the log-rank test was used to compare survival curves. Comparisons between clinical and demographic characteristics with immunostaining were performed using Fisher’s exact test. The Cox proportional hazards model was used for univariate and multivariate analyses. Multivariate analyses were conducted on variables with a p-value of less than 0.05 in the univariate analysis. All values were two-sided, with a significance level set at 0.05. All statistical analyses were performed using EZR software, version 1.33 (14).

At the time of analysis, 21 of the 84 patients had died of their primary disease and 18 had died of other diseases, and the 5-year OS rate for all patients was 57.9%. The 5-year OS rates according to stage were 89.1% for Stage I, 72.2% for Stage II, 35.7% for Stage III, and 24.6% for Stage IV (Figure 2). The 5-year PFS rates by Stage were 66.7% for Stage I, 62.4% for Stage II, 37.5% for Stage III, and 9.2% for Stage IV.

Table II shows the association between immunostaining for proteins involved in tumor immunity and clinical factors and nutritional status. When comparing the expression of CD8-positive cells infiltrating around the tumor with clinical factors, the CD8-positive group had a significantly higher p16-positive rate (82.1% vs. 35.7%, p<0.001) and a higher percentage of patients in Stages I-III (89.3% vs. 60.7%, p=0.024). There was no clear correlation between the expression of FoxP3-positive cells infiltrating around the tumor and any clinical factors. In the group with positive HLA class II expression on the tumor surface, p16 positivity was significantly higher (75.9% vs. 38.2%, p=0.001), and the percentage of non-drinkers was higher (44.8% vs. 18.2%, p=0.019).

Figure 3A-C shows the OS rate stratified by the positivity of CD8, FoxP3, and HLA class II. Significantly better OS was observed in the CD8-positive group (Figure 3A; 5y-OS 78% vs. 42%, p=0.019); FoxP3 showed no significant difference between the positive and negative groups (Figure 3B; 5y-OS 58% vs. 56%, p=0.667). HLA class II showed a trend toward better OS in the positive group (Figure 3C;5y-OS 70% vs. 46%, p=0.082).

According to the univariate results presented in Table III, the factors significantly associated with OS were Stage (HR=3.608, p<0.001), p16 status (HR=2.736, p=0.003), Body mass index (BMI) (HR=1.986, p=0.044), CONUT score (HR=2.279, p=0.017), and CD8 expression (HR=2.445, p=0.019), whereas other factors were not significantly associated with OS. In the multivariate analysis, Stage, BMI, and CONUT score were identified as significant factors associated with OS (p=0.037, p=0.042, and p=0.012).

Finally, a combined risk score was generated by summing the status of Stage, p16, BMI, CONUT score, and CD8 expression as follows: Risk Score=(Stage IV)*1+(p16 negative)*2+(BMI<22)*2+(CONUT≥3)*1+(CD8-negative)*1.

The combined risk score ranged from 0 to 7. The low-risk group was defined as having a score of 0-2, while the high-risk group was defined as having a score of 3-7. OS was significantly better in the low-risk group compared to the high-risk group (p<0.001; 5-year OS: 80.0% vs. 36.5%; Figure 4). Figure 5 shows the predictive accuracy using Receiver Operating Characteristic (ROC) analysis: the area under the curve (AUC) was 0.696 for clinical factors, such as stage and p16 status alone. However, when nutritional and tumor immune status factors—such as BMI, CONUT score, and CD8—were added, the AUC increased to 0.765, indicating an improvement in predictive accuracy with borderline significance (p=0.079).

The standard curative treatment for OPC is surgical resection or RT combined with chemotherapy (1,15). However, the functional deficit caused by surgery leads to a decrease in quality of life (16), since the location of the oropharynx is involved in various functions, such as mastication, swallowing, and speech (16). On the other hand, CRT may be an effective treatment that can preserve these functions, but the outcome is not always better than surgery and sometimes causes a certain range of late adverse events (17). To solve this problem, we focused on the relationship between nutritional status, tumor immunity, and the outcomes from RT.

The results shown in Table III indicate that stage, p16 status, BMI, CONUT score, and CD8-positive cell infiltration were significant factors associated with OS in univariate analysis, and stage, BMI, and CONUT score remained significant factors in multivariate analysis. A risk score combining these factors was developed to effectively differentiate between groups with higher and lower responses to CRT.

Previous studies have reported that several nutritional indices are associated with treatment outcomes in head and neck cancer (10,18-21). Oh et al. have evaluated the nutritional risk index (NRI) in patients undergoing radical CRT for head and neck cancer and reported that CRT can significantly reduce nutritional scores and that pretreatment NRI risk categories were associated with OS and treatment-related complications, indicating that the NRI may be a useful prognostic factor (18). Brewczyński et al. have examined the impact of CRT in oropharyngeal cancer patients and reported that nutritional status had a strong impact on OS and PFS and concluded that OPC patients require adequate nutritional support during CRT (20). Furthermore, Matsuura et al. have reported that PNI (Prognostic Nutritional Index), one of the nutritional markers is useful in predicting the response to immunotherapy for advanced head and neck cancer (21). Our results are consistent with these studies, suggesting that the CONUT score and BMI, both of which reflect nutritional status, are significant predicting factors in CRT for OPC.

As shown in Table II, there was a significant correlation between p16 status, namely HPV infection, and CD8 positivity. This is consistent with our previous study (7), suggesting that p16-positive tumors caused by HPV infection might induce a more active host anti-tumor immune response than p16-negative tumors, thus leading to higher CD8 infiltration. Several studies on tumor immunity have also been reported showing that tumor-infiltrating lymphocytes of cytotoxic T cells are associated with favorable RT effects (7,22-24). Haist et al. have investigated the function of CD8+ T cells in the tumor microenvironment in OPC patients and have reported that p16 expression correlates with CD8+ T cell infiltration around the tumor and associated with better CRT efficacy and survival (22). Kawaguchi et al. have also reported that in patients with OPC and nasopharyngeal cancer, a higher density of CD8+ tumor-infiltrating lymphocytes and increased expression of PD-L1 were associated with a better survival. Notably a higher CD8+ TIL density was a strong predictor of improved PFS and OS (23). Similar to these studies, our results suggest that factors related to tumor immunity, such as CD8+ cell infiltration, are associated with OS, and that the pretreatment tumor immunity affects the response to CRT.

As shown in Figure 4 and Figure 5, a risk score combining the significant factors found in the univariate analysis appeared to better predict treatment outcomes in OPC patients. There have been several attempts to distinguish outcomes by assigning risk scores (25). Lo et al. have identified low pretreatment hemoglobin and high systemic immunoinflammatory index levels as important indicators for predicting risk of death within one year after CRT for OPC and have concluded that nomograms created on this basis are useful for determining treatment strategy. Similarly, our study is based on relatively simple clinical data, such as Stage, p16 status, BMI, CONUT score, and CD8+ infiltration, and therefore, our scoring was relatively simple and clinically accessible.

Important limitations of this study were the relatively small sample size and the heterogeneity of patient profiles and treatments, which are drawbacks of retrospective studies and require further study for validation for more accurate prediction. Another limitation was that protein expression related to tumor immunity, such as CD8-positive cell infiltration and HLA expression, could vary by biopsy site and require evaluation by a skilled pathologist. However, our recent study has reported the usefulness of the Qupath automated measurement software for the evaluation of immunostained specimens (26), and its introduction may improve the accuracy of risk score evaluation.

In conclusion, this study proposed a combined risk scoring system based on an analysis of the association of clinical, nutritional, and immune factors with the outcome of patients undergoing radical CRT for OPC. Stratifying patients using this risk score may enable personalized treatment plans, helping to determine the optimal therapy and potentially improving outcomes for OPC.

This work was supported by JSPS KAKENHI Grant Number 23K14923, 23K07161 and 24K10913.

The Authors declare that they have no conflicts of interest in relation to this study.

Conceptualization, M.K. and J.K.; Methodology, M.S.; Software, S.M.; Validation, T.H., T.T. and Y.F.; Formal Analysis, S.M.; Investigation, M.S. and J.K; Data Curation, M.S., T.K., Y.H., Y.I., R.O., A.O., A.A, Y.T., and T.T.; Writing – Original Draft Preparation, M.S.; Writing – Review & Editing, M.S. and K.M.; Visualization, T.G. and R.O.; Supervision, T.T. and K.S.; Project Administration, K.S.; Funding Acquisition, M.S.