Introduction
Esophageal cancer is the eighth most common cancer worldwide and the sixth leading cause of cancer-related mortality. Even in cases where the tumor is superficial, it exhibits a high metastatic potential with poor disease prognosis (1). Esophageal squamous cell carcinoma (ESCC) is the predominant type (2). ESCC remains refractory to multimodal treatment, with a 5-year survival rate of 20% in patients with stage IVb disease who are ineligible for surgery (3). Although neoadjuvant or definitive chemoradiotherapy (CRT) is the standard treatment for locally advanced ESCC, patient outcomes remain unsatisfactory, with recurrence rates as high as 30-50% (4). Recent studies have suggested a synergistic effect between CRT and immune checkpoint inhibitors (ICIs), highlighting the need for reliable biomarkers to predict the tumor response and survival outcomes (4).
Immunotherapy with ICIs has shown promise. However, the response rate to nivolumab, an anti-programmed cell death protein 1 (PD-1) antibody targeting the immunosuppressive auxiliary signal PD-1/PD-1 ligand 1 (PD-L1), in previously treatment-refractory ESCC was 19% in the ATTRACTION-3 trial (5). The CheckMate 648 trial (ClinicalTrials.gov, NCT03143153) (6) and KEYNOTE 590 trial (ClinicalTrials.gov, NCT03189719) (7) also demonstrated a certain degree of therapeutic efficacy, underscoring the urgent need to enhance immunotherapy response rates in ESCC.
Several studies have reported correlations between tumor immunity and inflammation, and the therapeutic effects of chemotherapy and immunotherapy. Nutrition-related factors, such as the prognostic nutritional index (PNI), which is based on serum albumin and lymphocytes (8,9), modified Glasgow prognostic score (mGPS), calculated from albumin and C-reactive protein (CRP) (10), CRP-to-albumin ratio (CAR) using albumin and lymphocytes (11), and controlling nutritional status (CONUT), which incorporates albumin, lymphocytes and total cholesterol (12,13), are widely used as prognostic indicators in cancers, including ESCC. These indices highlight the prognostic importance of albumin and C-reactive protein.
Predicting treatment effects is difficult; however, several reports have identified factors associated with treatment outcomes. Makino et al reported objective response and disease control rates of 17.0 and 47.9%, respectively, for nivolumab monotherapy in patients with unresectable advanced ESCC. They identified performance status, number of metastatic organs, CAR, neutrophil-lymphocyte ratio (NLR), and the psoas muscle index (PMI) as relevant influencing factors (14). Kosumi et al found that the waist circumference was significantly associated with the disease control rate at the initial evaluation and the PNI (15). These reports suggest that systemic factors, including the nutritional and cancer immunity status, influence the therapeutic efficacy of ICIs.
However, there are few reports on the factors that influence the therapeutic efficacy of nivolumab after definitive CRT. Given the availability of alternative therapies, such as paclitaxel, predicting nivolumab's efficacy is crucial, and a comprehensive evaluation, including assessment of systemic conditions and immune status, is needed.
Therefore, this study aimed to identify the factors influencing the therapeutic efficacy of nivolumab after CRT in ESCC and examine the prediction methods for treatment success.
Discussion
In this study, we examined the association of patient background characteristics and blood parameters with nivolumab efficacy after CRT. We identified BW, WBC, percentage of basophil fraction, PLT, ALB, and CRP as the significant factors influencing nivolumab treatment after CRT.
The overall response and disease control rates with nivolumab treatment were 22.72 and 40.91%, substantially similar to those observed in the ATTRACTION-3 trial (5). In recent years, an abscopal effect has been associated with the post-CRT status in gastrointestinal cancers, such as esophageal, gastric, pancreatic, hepatobiliary, and colorectal cancers, (17). These findings, together with the results of this study, suggest that ICIs may be more selective after radiation therapy, especially in patients meeting the conditions indicated in this study.
BW and ALB are important indicators of the nutritional status. Although the PNI did not show a significant difference in this study, our findings underscore the importance of nutritional status. Previous studies in patients with non-small cell lung cancer have demonstrated that a higher BMI correlates with better prognosis in patients treated with ICIs, with BMI categorized as <18.5, 18.5-22.9, 23-24.9, and ≥25(18). When considered alongside reports on waist circumference (15), it appears that body size, including body weight, is important in immunotherapy. However, additional studies are required to elucidate the underlying mechanism.
Interestingly, BW showed a significant association with several inflammation- and immunity-related parameters, including WBC, basophils, platelets, and CRP. These associations suggest that BW may reflect not only the long-term nutritional status but also systemic inflammatory and immune conditions. Given that WBC, basophils, and CRP are involved in tumor-associated immune modulation, BW could serve as an integrative marker of the host-tumor-immune axis.
Notably, in this analysis, BW did not show a significant association with ALB levels. While ALB is widely used as a nutritional marker, it is also a negative acute-phase reactant and can decrease rapidly in response to inflammation, due to cytokine-mediated suppression of hepatic synthesis (19). In contrast, BW may be less influenced by short-term inflammatory changes and could more accurately reflect the patient's baseline nutritional and immunological state. These findings warrant further investigation to determine whether BW functions as a surrogate marker or has a mechanistic role in the context of cancer-related inflammation and immunotherapy. WBC and CRP are direct inflammation-related factors. Cancers often induce both acute and chronic inflammation. In ESCC, high preoperative CRP levels have been associated with tumor progression and poor survival in patients (20). In environments with pre-existing systemic inflammation, tumor immunity may be impaired.
Among the WBC fractions, only basophils showed a significant difference; however, there are very few studies on the relationship between basophils and ICIs. In murine models, basophils have been shown to produce chemokines such as CCL3 and CCL4, thereby promoting the infiltration of CD8+ T cells into the tumor microenvironment (21). This suggests that basophil activation may enhance antitumor immunity by facilitating CD8+ T cell-mediated tumor rejection. Furthermore, an increase in CCL3 levels within the tumor microenvironment was shown to promote T cell infiltration and enhance the efficacy of PD-1 blockade therapy (22). Although the mechanism remains unclear, one study suggested that basophils affect the tumor microenvironment via M2 macrophages (23), warranting further investigation. In patients with melanoma treated with anti-PD1 immunotherapy with a combination of nivolumab and ipilimumab, the relative basophil count was reported to correlate with the treatment response, along with relative eosinophils, absolute monocytes, LDH, and NLR (24). Increased baseline basophil levels also predicted significantly improved progression-free survival (PFS) in patients with non-small cell lung carcinoma treated with nivolumab (25). In contrast, some studies have reported that patients with high basophil levels show poorer responses to ICI therapy. In gastric cancer, eosinophils have also been implicated in promoting M2 macrophage polarization, potentially contributing to an immunosuppressive microenvironment and resistance to immunotherapy (23). A recent mechanistic study revealed that IL-4 secreted by bone marrow-derived basophils drives the immunosuppressive phenotype of tumor-infiltrating monocyte-derived macrophages (26). This suggests that elevated basophil activity and IL-4 production may promote immune evasion and resistance to ICIs. These controversial results suggest that differences between cancer types and interactions with other factors may need to be considered for a comprehensive evaluation.
In the present study, the cutoff value for basophils was set at 0.6% based on the Youden Index. Given that the normal reference range for basophils is approximately 0-1%, this cutoff was considered appropriate for dividing patients into high and low groups. However, further studies are warranted to validate the utility of this parameter.
PLTs act as active immune-like cells that suppress T cell activity, promote cancer metastasis, and regulate the spliced mRNA content (27). Inoue et al reported that nivolumab treatment for recurrent esophageal cancer was significantly more effective in patients with a low platelet-lymphocyte ratio, in addition to low lymphocytes and CAR (28). Additionally, Zhang et al reported that a lower systemic immune inflammation index (SII=PLT x neutrophils/lymphocytes) was associated with a significantly higher resection rate after preoperative treatment, including radiotherapy in ESCC (29). Although PLTs were often evaluated in combination with other parameters in previous studies, they were also identified as a significant factor in this study. Thus, it is now considered appropriate to evaluate the systemic inflammatory status not only by assessing the WBC count and CRP but also by combining various factors such as the lymphocyte fractions and PLT, which was also supported by the findings in this study.
In this study, CRE was significantly elevated, whereas the eGFR was lower in the PR + SD group. Furthermore, nivolumab did not require dose adjustment based on renal function, as antibodies are too large to be excreted by the kidneys except in severe nonselective proteinuria (30). Given that body size parameters, such as height and weight, also showed significant differences, we evaluated renal function based on CoCr using the Cockcroft-Gault formula, which accounts for the BW. However, there was no significant difference in CoCr between the two groups, suggesting that the difference in creatinine clearance was due to differences in the body size.
In the present study, no significant association was found between PD-L1 expression and treatment efficacy. Moreover, data on other tumor-specific immunologic markers such as the tumor mutational burden were not available, and therefore not analyzed. These biomarkers are recognized as potentially important predictors of response to ICIs and should be investigated in future studies with a larger cohort.
The six variables identified in this study were considered valid when three or more variables were within the range in the PR + SD group. Based on this, we concluded that these variables may be useful for predicting the treatment efficacy.
However, this study has some limitations, including its small sample size and single-center design. The size of the cohort, consisting of 22 patients receiving nivolumab and 14 controls, limited the statistical power and may have introduced potential biases, thereby reducing the generalizability of the findings. Accordingly, the results should be interpreted with caution and regarded as hypothesis-generating rather than conclusive. Larger-scale, multicenter studies are needed to validate and extend these findings to broader clinical settings.
Although the scoring system using six variables demonstrated an apparent association with clinical outcomes, this should also be interpreted with caution given the exploratory nature of the study and limited sample size. Further validation in larger, independent cohorts will be required to confirm the predictive utility and clinical relevance of these parameters.
Moreover, according to the commonly used rule of thumb requiring at least 10 events per variable for logistic regression, including multiple variables in a multivariate model would have risked overfitting and unreliable estimates. Therefore, we limited our analysis to univariate methods. As the number of cases increases, the predictive value of these variables can be comprehensively examined, and the cut-off values will become more appropriate, enhancing the association between the number of applicable variables and treatment outcomes.
Finally, being a single-institution study, center-specific factors such as patient selection, treatment implementation, and healthcare infrastructure may have influenced the results. These findings should therefore be interpreted within the context of this limitation, and future multicenter investigations will be essential to establish broader applicability.
In conclusion, although this study was based on a small sample of 22 cases, the six identified variables (BW, WBC, percentage of basophil fraction, PLT, ALB, and CRP) may be valuable for predicting the efficacy of ICI treatment after CRT in ESCC. By accumulating more cases based on our study's methodology, more reliable prediction methods can be developed.