Esophageal cancer is the sixth leading cause of cancer-related mortality globally (1). Esophageal squamous cell carcinoma (ESCC) is the predominant histological subtype in Asia, accounting for approximately 90% of all cases (2). Definitive chemoradiotherapy (dCRT) is the standard treatment for patients with unresectable locally advanced esophageal squamous cell carcinoma (LA-ESCC) (3). However, the three-year survival rate under this treatment regimen ranges between 26.9% and 55.4% (4–8), with approximately half of the patients experiencing recurrence within three years (4–6). Therefore, it is imperative to develop novel therapeutic strategies to improve patient outcomes and reduce recurrence rates.

Recently, immune checkpoint inhibitors (ICIs) have demonstrated promising applications in the treatment of esophageal cancer. In patients with advanced or metastatic ESCC, the latest findings from the Keynote 590 (9) and ESCORT-1st (10) trials have demonstrated improved clinical outcomes, establishing ICIs as part of the standard treatment regimen. Moreover, in unresectable LA-ESCC, several studies (10, 11) have also reported encouraging efficacy of ICIs when combined with dCRT concurrently, although these findings are based on relatively small patient cohorts. These studies demonstrate potential clinical benefits with improved overall survival (OS) and progression-free survival (PFS) outcomes compared to chemotherapy alone, with manageable treatment-related toxicity. Taken together, current evidence indicates that concurrent ICIs combined with dCRT may represent a promising therapeutic strategy for patients with unresectable LA-ESCC.

However, large-scale randomized phase III trials evaluating ICIs combined with dCRT concurrently for unresectable LA-ESCC remain under investigation. In this context, we conducted a retrospective analysis of LA-ESCC patients treated with this combination therapy at Jiangsu Cancer Hospital to evaluate its efficacy and safety profile.

This retrospective study evaluated the strategy of dCRT combined with concurrent ICIs in a relatively large cohort of patients with unresectable LA-ESCC. The one-year and two-year OS rates were 83.0% (95% CI, 77.5%–89.0%) and 64.3% (95% CI, 57.2%–72.3%), respectively. The corresponding PFS rates were 69.4% (95% CI, 62.6%–76.8%) at one year and 50.2% (95% CI, 42.9%–58.8%) at two years. This study also showed manageable treatment-related toxicity with no treatment-related deaths. Additionally, the induction therapy prior to dCRT demonstrated no survival benefit. While its application in real-world settings requires further investigation and should be guided by clinical judgment. In conclusion, the combination of dCRT with concurrent ICIs demonstrates potential as a therapeutic strategy for unresectable LA-ESCC, though its efficacy needs to be validated through large-scale randomized controlled trials.

Our findings suggest that combining dCRT with concurrent ICIs may improve survival outcomes, representing a promising therapeutic strategy for LA-ESCC patients. In our study, the addition of ICIs to concurrent dCRT resulted in a two-year OS rate of 64.3% (95% CI, 57.2%–72.3%) and a two-year PFS rate of 50.2% (95% CI, 42.9%–58.8%). Consistent with two-year OS rates ranging from 50% to 66.7% and PFS rates between 46.4% and 54.4% for standard dCRT regimens (4, 5, 17). The EC-CRT-001 phase II trial (11) evaluating concurrent ICIs with dCRT for LA-ESCC reported a promising 1-year OS rate of 78.4%, though the 1-year PFS rate (54.5%) did not exceed standard dCRT regimen. The investigators attributed this unexpected PFS outcome to their higher enrollment of stage IVA patients (38%) compared to previous studies. Notably, our cohort included a substantially greater proportion of stage IVA disease (49.7%) than reported in prior studies (8.5%-38.7%) (4, 5, 11, 17), yet still demonstrated encouraging survival outcomes, suggesting potential therapeutic benefit across advanced disease stages. Consistent with our findings, Zhang et al. (18) conducted a phase 1b study of camrelizumab combined with concurrent dCRT and reported 2-year OS and PFS rates of 69.6% and 65.0% respectively. Additional supportive evidence includes a durvalumab plus tremelimumab regimen, which achieved 2-year OS and PFS rates of 75% and 57.5% (19). A recent multicenter real-world study involving 102 ESCC patients treated with concurrent ICIs and dCRT reported 1- and 2-year OS rates of 86.7% and 66.9%, and PFS rates of 66.7% and 47.3%, respectively (20), which are generally consistent with our findings. Several ongoing phase III trials, such as KEYNOTE-975 (pembrolizumab), RATIONALE-311 (tislelizumab), KUNLUN (durvalumab), and ESCORT-CRT (camrelizumab), are expected to provide high-level evidence to further define the role of concurrent ICIs combined with dCRT in this setting. It is also important to note that the present study was conducted in a real-world clinical setting, where treatment adherence and patient management may vary from the strictly regulated conditions of prospective clinical trials. Despite these potential sources of variability, the survival outcomes observed in our study underscore the clinical value of integrating ICIs into dCRT regimens for unresectable LA-ESCC patients, even for patients with advanced-stage disease. These results provide further support for the application of immunotherapy in this population and contribute to the growing body of real-world evidence on its effectiveness.

In the era of ICIs, induction immunochemotherapy before dCRT has not been incorporated into standard guidelines for LA-ESCC, due to insufficient high-level evidence and potential toxicity concern. Large primary tumor volume has been identified as a poor prognostic factor in ESCC treated with definitive CCRT (21). Tumor volume reduction before CCRT may improve local control and survival outcomes. Phase II studies (22, 23) of this novel induction immunochemotherapy regimen have demonstrated promising improvements in local control rates and survival. These clinical benefits are biologically supported by preclinical validation of related mechanisms, including tumor vascular normalization and hypoxia alleviation (24). However, these regimens carry toxicity burdens, a high proportion of patients (80%–90%) experienced grade 3 or higher treatment-related adverse events, and cases of grade 5 pneumonitis have been reported (23). In our study, while no treatment-related deaths occurred, induction therapy combined with dCRT did not confer a survival benefit in patients with ESCC, as measured by either OS (P = 0.770, Figure 3A ) or PFS (P = 1.000, Figure 3B ). This difference may be attributed to the relatively small sample size in our study, as only 72 patients (43.6%) received induction immunochemotherapy, potentially limiting the statistical power to detect the therapeutic benefits of induction therapy. In conclusion, while our findings suggest induction therapy may offer limited clinical benefit in esophageal cancer management, further investigations are warranted to definitively establish its therapeutic role and fully characterize potential toxicities.

Several studies have shown that patients achieving CR after dCRT have significantly better survival than those with residual disease (7, 25). Consistent with these findings, our results demonstrated that the CR + PR group exhibited a longer OS (HR = 2.25; 95% CI 1.39–3.65; P < 0.001; Figure 4A ) and PFS (HR = 2.51; 95% CI 1.59–3.95; P < 0.001; Figure 4B ) compared to the SD + PD group. These findings underscore that early tumor shrinkage following dCRT combined with ICIs represents not only a marker of treatment sensitivity but also an independent prognostic indicator. Clinically, this highlights the value of early response evaluation, as patients who fail to achieve CR or PR face substantially higher risks of disease progression and mortality, and may require closer surveillance or consideration of additional systemic therapy. From a mechanistic perspective, radiotherapy can trigger the release of tumor antigens and activate antigen-presenting cells, and when combined with PD-1/PD-L1 blockade, it enhances cancer cell immunogenicity and improves long-term antitumor responses (26, 27). When considering overall response, the ORR rate in our study was 70.9%, comparable with that reported in previous studies using dCRT alone (70.9% versus 65%–74.6%) (6, 7), despite the higher disease burden in our cohort. In earlier reports, the proportion of T3–T4 patients was 72.7% (7), whereas our study included 92.7%. Similarly, stage IV patients accounted for only 10% in previous studies (6), compared with 49.7% in our cohort. To conclude, despite enrolling a high proportion of patients with advanced disease, our study achieved promising ORR and survival outcomes. This suggests that combining concurrent ICIs with dCRT not only improves tumor response but also turns these responses into real survival benefits, supporting the therapeutic value of this combined approach.

In our study, a BMI ≥25 was identified as an independent protective factor for PFS (HR = 0.49, 95% CI: 0.27–0.89, P = 0.020; Figure 4B ), suggesting that patients with higher BMI may derive greater benefit from dCRT combined with immunotherapy. Although obesity is generally recognized as a risk factor for the development of several cancers (28), increasing evidence indicates that overweight (BMI 25–29) and obese (BMI ≥ 30) patients may experience improved outcomes with ICIs (29–31), a phenomenon known as the obesity paradox (32). This has been partially explained by previous preclinical studies showing that obesity-associated metabolic signalling and inflammatory cues cause tumour-associated macrophages (TAMs)to induce PD-1 expression, which then drives a TAM-specific feedback mechanism that impairs tumour immune surveillance. This may contribute to increased cancer risk yet improved response to PD-1 immunotherapy in obesity (33). Nevertheless, only 37 patients (22.4%) in our cohort had a BMI ≥25, and no cases of severe obesity (BMI >35) were included. Thus, while our findings highlight a potentially important prognostic role for BMI in this treatment setting, they should be interpreted with caution and warrant confirmation in larger prospective studies.

T4 stage was also identified as an independent adverse prognostic factor for OS (HR = 2.34, 95% CI: 1.31–4.16, P = 0.04; Figure 4A ), underscoring the prognostic importance of local invasion into adjacent structures. Chen et al. found that tracheal/bronchial invasion and ulcerative tumors were independent risk factors for esophagotracheal fistula after dCRT (34). Historically, T4b disease has been associated with poor outcomes, with a 3-year OS of only 22.2% and locoregional failure in 60% of cases (35). In the immunotherapy era, the EPOC1802 trial, which included nearly 95% of patients with cT4 LA-ESCC, reported that maintenance ICIs after dCRT improved response rates and intermediate-term survival, even among patients with a high T4 tumor burden (36). Notably, in our study, despite a high proportion of T4 cases (94/165, 57%), the combination of dCRT and concurrent ICIs yielded encouraging survival outcomes without an increased incidence of post-dCRT esophagotracheal fistula, suggesting that immunotherapy integration may confer benefit with acceptable safety in this high-risk population. Prospective validation is warranted.

The prognostic impact of tumor location in the setting of dCRT remains understudied, and existing evidence is inconsistent. Gao et al (37). reported that tumor location was independently associated with OS and PFS in multivariate analysis, with upper thoracic tumors achieving the highest 5-year OS (46.0%). This suggests that middle and lower thoracic tumors may have worse outcomes, possibly due to their proximity to critical structures such as the heart, lungs, and great vessels, which can limit radiation dose delivery and increase the risk of severe cardiopulmonary complications. Other dCRT studies found no statistically significant association between location and survival but noted a trend toward poorer outcomes for upper/middle thoracic disease compared with lower thoracic disease (P = 0.055) (38). In our cohort, middle esophageal tumor location was independently associated with worse OS (HR = 2.18, 95% CI: 1.05–4.52, P = 0.036; Figure 4A ), supporting the clinical value of incorporating tumor location into individualized target delineation and dose optimization. However, the number of patients with middle esophageal tumors was small (18/165, 10.9%), which may limit the statistical power and highlights the need for validation in larger, multicenter cohorts.

In our study, 60% of patients developed grade ≥ 3 lymphopenia, including 20 cases (12.1%) of grade 4, with no grade 5 events observed. Lymphocytes are highly radiosensitive, with significant depletion occurring even at relatively low radiation doses (39). This suggests that the observed lymphopenia is more likely attributable to radiotherapy rather than ICIs. Thoracic radiation, especially when involving large planning target volumes or vertebral bone marrow, has been shown to induce profound lymphocyte loss (40). Mechanistic studies have further demonstrated that radiation not only reduces circulating lymphocytes and suppresses lymphopoiesis, but also upregulates immune inhibitory molecules such as PD-L1 and CTLA-4, contributing to systemic immunosuppression (41, 42). Similar patterns of lymphopenia have been reported in previous studies of dCRT without the use of ICIs. Davuluri et al. observed a 27% incidence of grade 4 lymphopenia in esophageal cancer patients treated with dCRT alone (43), and Deng et al. reported grade 3–4 lymphopenia in up to 50% of patients receiving thoracic IMRT (44). These findings support the safety and feasibility of concurrently integrating ICIs into dCRT without significantly increasing severe lymphopenia. Moreover, severe lymphopenia has been associated with inferior survival outcomes across multiple studies (43–45). In our univariate analysis, patients with grade ≥3 lymphopenia had significantly shorter overall survival (HR = 1.90; 95% CI: 1.18–3.06; P < 0.05; Figure 3C ) and progression-free survival (HR = 2.15; 95% CI: 1.36–3.40; P < 0.001; Figure 3D ). Although these associations were not statistically significant in the multivariate analysis, the findings raise the possibility that lymphopenia may negatively impact clinical outcomes. Importantly, the combination of ICIs with dCRT did not appear to worsen treatment-related lymphopenia, further supporting its potential as a safe therapeutic strategy.

Our study has certain limitations. As this was a single-center, retrospective study, the generalizability of our findings is inevitably limited. The sample size was relatively small, and the median follow-up duration was only 24 months, which may hinder a thorough evaluation of long-term survival outcomes and delayed toxicities, including the potential long-term impact of immune-related adverse events. Future larger prospective studies with extended follow-up are needed to provide a more robust evaluation of long-term efficacy and safety. Although PSM was performed to minimize confounding, the retrospective and single-center nature of this study may still introduce selection bias. Therefore, our findings should be interpreted with caution, and validation through prospective, multi-center randomized trials will be essential to confirm the reliability and generalizability of our findings. In addition, the inability to standardize the type of PD-1 inhibitors could have introduced variability in treatment response. The assessment of clinical response was primarily based on CT imaging, which may compromise the accuracy of short-term efficacy evaluation. Further use of EUS and PET/CT in future studies could be integral in identifying regional nodes and assessing the risk of metastasis. Since this study was conducted in a single hospital within a real-world clinical setting, treatment decisions were made by clinicians based on routine practice, potentially influencing the observed outcomes and mirroring the practical complexities inherent in everyday patient management. In addition to the primary endpoints of OS and PFS, further research should include a comprehensive analysis of PD-L1 expression and the role of the tumor microenvironment in predicting treatment efficacy.

Our study provides preliminary evidence that the combination of dCRT with concurrent ICIs offers promising survival benefits for patients with unresectable LA-ESCC, while maintaining a manageable safety profile. However, phase III randomized controlled trials are warranted to validate these findings and further establish the role of this treatment strategy.