Abstract
Background/Aim: For patients with unresectable locally advanced pancreatic cancer (LAPC), carbon-ion radiotherapy (C-ion RT) can safely deliver higher doses than conventional photon therapy, increasing the potential for long-term survival. However, achieving meaningful improvements in survival rates requires reliable prognostic biomarkers to identify patients likely to benefit from treatment. Patients and Methods: In this study, we measured plasma levels of soluble interleukin-6 receptor (sIL-6R) before C-ion RT and examined their association with the risk of distant metastasis (DM), local recurrence (LR), and overall survival (OS). Results: Results showed that patients with higher plasma sIL-6R levels had a lower risk of DM [hazard ratio (HR)=0.53; p=0.033] and improved OS (HR=0.55; p=0.037). No significant association was observed between LR and plasma sIL-6R levels (HR=1.47; p=0.273). Conclusion: These findings suggest that pretreatment plasma sIL-6R levels may serve as a prognostic marker for C-ion RT in LAPC.
Pancreatic cancer (PC) is a common malignancy with a particularly challenging prognosis in cases of locally advanced PC (LAPC), even amidst significant medical advancements (1-3). Pancreatic radiotherapy is used as an adjuvant therapy both before and after surgery, as well as for managing locally advanced or recurrent pancreatic tumors (4, 5). Carbon-ion radiotherapy (C-ion RT) offers several biological advantages over conventional photon irradiation. A unique feature of the C-ion beam is its lower entrance dose with most energy concentrated at the end of its path, resulting in a dose peak known as the “Bragg peak” (6). This allows for improved dose localization, targeting the tumor with substantial energy while minimizing damage to surrounding normal tissues. Additionally, C-ion beams have biological benefits due to their high linear energy transfer and dense energy deposition per unit length (7).
Over recent decades, C-ion RT has emerged as a promising treatment option for inoperable PC (8-10). Multicenter studies have shown that C-ion RT provides superior clinical outcomes in terms of overall survival (OS) and local control compared to conventional chemoradiation or chemotherapy alone (9). Shinoto et al. reported a 2-year survival rate of 53% with combined C-ion RT and chemotherapy (10), primarily attributed to the targeted dose distribution and enhanced biological effects due to higher linear energy transfer (11, 12).
Interleukin-6 (IL-6) is an inflammatory cytokine involved in immune response, hematopoiesis, bone metabolism, development, and tumor growth (13). Clinically, patients with untreated PC exhibit significantly higher plasma IL-6 levels compared to healthy individuals, and IL-6 levels have been shown to positively correlate with tumor stage (14, 15), making IL-6 a valuable diagnostic biomarker for PC. IL-6 binds to a specific receptor, IL-6R, which exists in two forms: membrane-bound or soluble (sIL-6R). In the blood, sIL-6R binds to IL-6, and the sIL-6R/IL-6 complex initiates intracellular trans-signaling by binding to the gp130 receptor present on various cell surfaces (16). Research suggests that IL-6 trans-signaling primarily promotes pro-inflammatory responses and tumor formation (17).
The clinical relevance of serum and plasma sIL-6R levels has been reported across various cancers. sIL-6R levels are higher in patients with multiple myeloma than in healthy individuals (18). In prostate cancer, patients with metastasis demonstrate significantly elevated plasma sIL-6R and TGF-β1 levels before radical prostatectomy (19). Elevated pretreatment serum sIL-6R levels also predict poor prognosis in patients with esophageal squamous cell carcinoma undergoing preoperative chemoradiation (20). Additionally, high plasma sIL-6R levels at diagnosis are linked to increased glioma risk (21). Based on these findings, this study investigated the role of pretreatment plasma sIL-6R levels in the clinical outcomes of C-ion RT.
Patients and Methods
Patients. The National Institutes for Quantum Science and Technology Institutional Review Board approved this study (registration number 18-004). All experiments were conducted following approval guidelines, and all participants provided written informed consent. Patients with advanced PC were enrolled at QST Hospital between 2016 and 2020. Plasma samples from 103 patients in a retrospective cohort were obtained from the QST Hospital medical data bank (MDB project number #2017-02). These samples were collected prior to the initiation of C-ion RT and analyzed for this study. Primary exclusion criteria included the presence of inflammatory or autoimmune diseases and any evidence of other malignant tumors.
This study evaluated OS, local recurrence (LR), and distant metastasis (DM) in 103 patients. LR was defined as lesions within the planning target volume as indicated by computed tomography (CT), magnetic resonance imaging, or 18 fluoro-2-deoxyglucose positron emission tomography (18FDG-PET). The absence of LR was determined by the lack of evidence of a >20% increase in tumor size on CT and no 18FDG-PET uptake 6 months or more post-treatment. Recurrence was diagnosed based on imaging alone; biopsy confirmation was not obtained.
Measurement of plasma sIL-6R concentration. Plasma sIL-6R levels were measured using an enzyme-linked immunosorbent assay (ELISA) with the DuoSet Human IL-6Rα ELISA kit from R&D Systems (Minneapolis, MN, USA; Cat. No. DY227), following the manufacturer’s instructions. All assays were performed by an investigator blinded to the clinical data.
Statistical analysis. Survival times to death, LR, and DM were defined as the period from the end of C-ion RT to the occurrence of each event. Survival curves for OS, LR, and DM were estimated using the Kaplan-Meier method. Predictor variables for each event included age, sex, stage, chemotherapy, CA19-9 levels, and sIL-6R levels. Differences in survival curves by predictor variables were compared using the log-rank test. All factors were included as covariates in a Cox regression model. Categorical variables (age, sex, stage, chemotherapy, and sIL-6R) and one continuous variable (CA19-9) were assessed to quantify their effects as predictors.
Preliminary analysis indicated that the Cox model fit improved when sIL-6R was dichotomized using a cutoff value rather than treated as a continuous variable. Consequently, sIL-6R was treated as a categorical variable, with the optimal cutoff value for OS determined using the Akaike information criterion (AIC). The cutoff value was explored in increments of five and subsequently applied to the other endpoints. CA19-9 was treated as a log-transformed variable throughout the study, as model fit was significantly improved with this transformation. All analyses were performed using the R statistical package (http://www.r-project.org/), with statistical significance set at p<0.05.
Results
Patient characteristics and comparisons of OS, LR, and DM for each characteristic, assessed using the log-rank test, are summarized in Table I. Between 2016 and 2020, a total of 103 patients with PC underwent definitive C-ion RT with a dose of 55.2 GyE in 12 fractions at our institution. The median follow-up duration after C-ion RT was 18.7 months (range=0.2-78.6 months) for all patients in this study.
As shown in Figure 1A, patients were categorized into four groups based on chemotherapy history: no chemotherapy (16 patients, sIL-6R=80.1±13.8 ng/ml), chemotherapy during or after C-ion RT (23 patients, sIL-6R=90.3±13.4 ng/ml), chemotherapy before C-ion RT only (15 patients, sIL-6R=84.8±12.6 ng/ml), and chemotherapy both before and during or after C-ion RT (45 patients, sIL-6R=83.6±13.4 ng/ml). No significant differences were found in sIL-6R levels across chemotherapy groups or timing of administration (one-way ANOVA, p=0.118). In Figure 1B, no significant difference in sIL-6R levels was observed between patients with postoperative recurrence (16 patients, 80.8±11.0 ng/ml) and those with primary PC (87 patients, 85.5±14.1 ng/ml; p=0.216). Therefore, the analysis proceeded without differentiating patients based on surgical history.
Kaplan-Meier survival curves for OS, LR-free survival (LRFS), and DM-free survival (DMFS) are displayed in Figure 2. Each panel includes Kaplan-Meier estimates for all patients (Figure 2A, C and E) and for patients grouped by the optimal sIL-6R cutoff value of 95 ng/ml, identified via Cox regression analysis (Figure 2B, D and F). OS, LRFS, and DMFS did not significantly differ based on sIL-6R levels at this cutoff (log-rank p=0.407, 0.255, and 0.383, respectively). The Cox regression analysis results are presented in Table II. Males demonstrated an overall trend toward increased risk, with significant elevations in risk for OS [hazard ratio (HR)=1.89; confidence interval (CI)=1.17-3.05; p=0.009] and DM (HR=1.78; CI=1.07-2.96; p=0.026). Carbohydrate antigen 19-9 (CA19-9) (22, 23), also known as sialyl Lewis A, is a recognized serum biomarker for PC and has shown reliability as a diagnostic and prognostic marker (24). In this study, CA19-9 (analyzed on a logarithmic scale) served as a robust predictor of OS (HR=1.22; CI=1.09-1.37; p=0.001) and LR (HR=1.33; CI=1.11-1.59; p=0.002).
We next examined the association between IL-6R expression and OS in 177 PC patients, using an integrated database of previously published transcriptomics datasets (KMplot, www.kmplot.com), which was developed as an online tool for real-time meta-analysis of solid tumor microarray datasets to identify biomarkers associated with survival (25). IL-6R expression was found to be positively correlated with OS [HR=0.45 (0.29-0.69), log-rank p<0.00017].
The optimal cutoff value for sIL-6R concerning OS was identified as 95 ng/ml. Patients with sIL-6R levels above this threshold demonstrated distinct trends for each endpoint, with a significantly reduced risk for OS (HR=0.55; 95% CI=0.32-0.96; p=0.037) and DM (HR=0.53; 95% CI=0.30-0.95; p=0.033). However, no significant difference was identified between LR and plasma sIL-6R levels (HR=1.47; p=0.273).
Discussion
In the present study, we investigated the relationship between plasma sIL-6R levels in patients with PC before C-ion RT and their prognosis post-treatment. The log-rank test for high sIL-6R levels showed no significant difference in OS, LR, or DM. However, Cox regression analysis, adjusted for multiple predictors, revealed significant differences in OS and DM. Dividing the sIL-6R values by cutoff points showed that patients with high sIL-6R levels had favorable OS and DM outcomes but exhibited a trend toward a higher risk for LR, though this was not statistically significant.
Inflammation may influence both the malignancy of PC and the effectiveness of radiation therapy (RT), with inflammatory cytokines potentially serving as prognostic markers for treatment outcomes (26). Research on the role of inflammation in PC cell malignancy has shown that high expression of intracellular C3, induced by inflammatory cytokines, activates the Akt/Smad pathway, promoting cell migration and invasion (27). Additionally, Caxali et al. identified IL1A as a biomarker for predicting the effectiveness of RT in PC through omics analysis of clinical samples, finding that higher IL1A levels before RT are associated with reduced survival rates in PC patients (28).
Higher levels of sIL-6R have been reported to correlate with poor prognosis. Elevated pretreatment serum/plasma sIL-6R levels have been linked to decreased survival rates in multiple myeloma (18) and esophageal carcinoma (20), increased DM in prostate cancer (19), and a higher risk of glioma (21). In contrast, our study found that high sIL-6R levels in patients with PC served as positive prognostic factors for improved survival. Similar findings have been reported in neuroblastoma, where high sIL-6R levels at diagnosis were significantly associated with reduced metastatic disease (29), and in colorectal cancer, where increased sIL-6R expression was linked to better prognosis and disease suppression (30, 31).
IL-6 contributes to cancer progression through two pathways (32-34). The first is classical signaling, which involves IL-6 binding to the gp130-IL-6R complex on the cell membrane. The second is trans-signaling, which occurs when the IL-6-sIL-6R complex binds to gp130 on the cell membrane and initiates intracellular signaling. sIL-6R has a high affinity for IL-6, with approximately 70% of IL-6 in the blood binding to sIL-6R, which amplifies its biological effects (35). The observed reduction of sIL-6R in the plasma may be due to decreased sIL-6R production by tumor cells or increased sIL-6R consumption due to the enhanced affinity of IL-6/sIL-6R within the tumor stroma (29, 30).
There are two possible reasons why high sIL-6R levels may be associated with improved survival. Firstly, an abundance of IL-6 at the tumor site increases sIL-6R consumption (36), which may lead to reduced overall blood IL-6 and sIL-6R levels, thus lowering DM risk and improving survival rates. Secondly, soluble gp130 (sgp130), a natural antagonist of the IL-6/sIL-6R complex, selectively inhibits IL-6 trans-signaling, mitigating its biological effects (37). Although sgp130 levels were not measured in this study, it is hypothesized that increased sgp130 release might lead to greater binding of the IL-6-sIL-6R complex to sgp130, resulting in decreased IL-6 and sIL-6R plasma levels. This could further contribute to a lower DM risk and improved survival.
In this study, no significant difference was found between sIL-6R levels and LR, though patients with high sIL-6R levels showed a trend toward increased LR risk. Evaluating LR is challenging, and this result requires cautious interpretation. Firstly, the incidence of LR may have been underestimated in short-lived patients. This underestimation may have occurred because potential recurrences were not observed before patient death, leading to a lower LR rate. In contrast, LR may have been more accurately assessed in long-term survivors, resulting in a higher recurrence rate. Secondly, accurately measuring LR presented challenges in this study. Radiological evaluation of PC post-surgery is complex, particularly in distinguishing between local tumor recurrence, postoperative fibrosis, and inflammation (38-40). Similar challenges are encountered after C-ion RT (41). Pretreatment CA19-9 levels have demonstrated good prognostic value in predicting survival in patients with PC (42). In this study, CA19-9 was also found to be a risk factor for OS and LR, with higher levels correlating with poorer outcomes. One limitation of this study could be the discrepancy in plasma sIL-6R levels measured here (84.9±13.9 ng/ml) compared to those in patients with colorectal cancer (0.84±0.15 ng/ml) (31) and patients with neuroblastoma (36.0±4.4 ng/ml in low- and intermediate-risk patients) (29). Although a consensus on sIL-6R levels in patients with PC is still lacking, this discrepancy may be due to cancer cell type and differences in the sensitivity of the ELISA detection kit used.
In this study, we determined the cutoff value for sIL-6R levels using model fit criteria. While some studies determined cutoff values with similar methods (43, 44), others used receiver operating characteristic (ROC) analysis (45). Rooted in statistical decision theory, ROC analysis is commonly used in radiology to distinguish between healthy and diseased subjects via diagnostic tests and to compare test efficacy (46). However, ROC analysis cannot account for multiple covariates. In the presence of strong prognostic factors such as CA19-9, the effect of sIL-6R levels might not be fully assessed without adjusting for these factors. Consequently, cutoff values determined via ROC analysis are expected to be less precise than those used in this study. ROC analysis of sIL-6R for OS yielded a cutoff value of 84.79; however, the model fit was poorer than at the cutoff of 95. Additionally, the HR for sIL-6R was 0.998 (95% CI=0.629-1.585, p=0.995), further obscuring the relationship between sIL-6R levels and OS.
Study limitations. Firstly, it is a retrospective and non-randomized study, and these characteristics must be taken into account. Although sIL-6R levels were not significantly affected by chemotherapy or surgical history (Figure 1), chemotherapy protocols were not standardized, and both primary and post-surgical recurrence patients were included, leading to treatment variability. This study was also conducted at a single institution with a limited patient population, and the absence of post-treatment sIL-6R data is another limitation. Nonetheless, the findings are promising, and future data collection on IL-6 and sgp130 may help clarify the biological mechanisms linking sIL-6R to survival and distant metastasis.
Conclusion
This study found that higher pretreatment plasma sIL-6R levels were associated with improved DM control and OS in patients with PC who underwent C-ion RT. These findings suggest that plasma sIL-6R levels may serve as a valuable predictive biomarker to help identify patients with PC who are more likely to benefit from C-ion RT. Further research is warranted to confirm these results and explore the biological mechanisms underlying the association between sIL-6R levels and treatment response in C-ion RT for pancreatic cancer.
Acknowledgements
The Authors thank Dr. Kazutoshi Murata for medical data anonymization, Dr. Atsushi B. Tsuji and the staff of the medical data bank at the National Institutes for Quantum Science and Technology for supporting medical data bank usage, Dr. Toshio Hirano for his encouragement in this study, and Editage (www.editage.jp) for English language editing.
Footnotes
Authors’ Contributions
Conceptualization, T.I. (Takashi Imai), S.H., and T.T.; Data curation, M.S., T.I. (Tetsuro Isozaki), T.A., and T.T.; Formal analysis, K.D.; Funding acquisition, S.H.; Investigation, K.D. and T.T.; Methodology, K.D., M.S., T.I. (Takashi Imai), S.H., and T.T.; Project administration, S.H. and T.T.; Resources, K.D., M.S., S.H., and T.T.; Software, K.D.; Supervision, S.H., and T.T.; Roles/Writing - original draft, K.D., S.H., and T.T.; Writing - review & editing, K.D., S.H., and T.T. All Authors have read and agreed to the published version of the manuscript.
Funding
This work was supported by a research fund from the National Institutes for Quantum Science and Technology, Japan.
Conflicts of Interest
The Authors declare no conflicts of interest.
- Received October 27, 2024.
- Revision received November 6, 2024.
- Accepted November 18, 2024.
- Copyright © 2025 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).