This retrospective cohort study included patients diagnosed with esophageal squamous cell carcinoma between April 2019 and February 2023. Inclusion criteria were (1): histopathologically confirmed esophageal squamous cell carcinoma (2), classification as locally advanced stage, and (3) treatment with definitive concurrent chemoradiotherapy (CCRT). Exclusion criteria included (1): presence of secondary malignancies (2), incomplete radiotherapy, and (3) prior receipt of neoadjuvant or adjuvant therapies.

Pre-treatment evaluations comprised physical examinations, standard hematological and biochemical tests, barium esophagograms, esophagogastroduodenoscopy with endoscopic ultrasound, contrast-enhanced cervical, thoracic, and abdominal computed tomography scans, brain magnetic resonance imaging, bone scans, electrocardiograms, echocardiograms, and pulmonary function tests. 18-fluorodeoxyglucose positron emission tomography/computed tomography was performed in cases where distant metastasis was suspected, and bronchoscopies were conducted if bronchial or tracheal invasion was suspected.

All patients received intensity-modulated radiotherapy using 6-MV photons. The gross tumor volume (GTV) included the primary tumor (GTVp) and metastatic lymph nodes (GTVn), identified via endoscopy, contrast-enhanced thoracic computed tomography, and/or 18-fluorodeoxyglucose positron emission tomography/computed tomography. The primary clinical target volume (CTVp) included GTVp with an additional radial margin of 0.5-1.0 cm and longitudinal margins of 3 cm. The clinical target volume for metastatic lymph nodes (CTVn) was delineated based on involved field irradiation, excluding elective nodal irradiation (20, 21). The planning target volume (PTV) was defined by adding a margin of 0.5-1.0 cm to the CTV. Cone beam CT was used for daily verification during the first week of treatment, followed by weekly verification thereafter.

Radiotherapy was delivered five days per week, targeting a dose of either 60.0 Gy in 30 fractions or 50.4 Gy in 28 fractions. Dose constraints for organs at risk were set as follows (1): mean lung dose < 17 Gy, lung V20 < 30%, and lung V5 < 65% (2); mean heart dose < 25 Gy, heart V30 < 35%, and heart V40 < 25% (3); maximum spinal cord dose < 45 Gy.

Concurrent chemotherapy regimens comprised cisplatin plus fluorouracil or paclitaxel plus cisplatin administered during thoracic radiotherapy. The cisplatin plus fluorouracil regimen involved cisplatin at 75 mg/m² on day 1 and continuous intravenous fluorouracil at 1500 mg/m² over 72 hours starting on day 1, repeated every 3 weeks. The paclitaxel plus cisplatin regimen was administered in two options (1): paclitaxel at 175 mg/m² and cisplatin at 75 mg/m² on day 1, every 3 weeks (2); paclitaxel at 50 mg/m² and cisplatin at 25 mg/m² on day 1, weekly.

Two levels of dose reduction were permissible for concurrent chemotherapy. Chemotherapy was discontinued in cases of grade 4 peripheral nerve toxicity, grade 2 or higher renal toxicity, or other unacceptable adverse events.

Patients were administered ICIs, including tislelizumab, camrelizumab, toripalimab, and sintilimab, all of which are programmed death-1 (PD-1) inhibitors. During chemoradiotherapy, tislelizumab (200 mg), camrelizumab (200 mg), toripalimab (240 mg), and sintilimab (200 mg) were administered intravenously every 3 weeks.

Post-chemoradiotherapy, maintenance therapy with ICIs was continued every 3 weeks for up to 1 year, or until the occurrence of unacceptable toxicity, disease progression, patient refusal, or at the discretion of the treating oncologist. Dose reductions for ICIs were not permitted.

Laboratory tests were conducted weekly during chemoradiotherapy and every 3 weeks thereafter. Adverse events were evaluated according to the Common Terminology Criteria for Adverse Events (CTCAE, version 5.0).

Follow-up assessments were scheduled every 3 months for the first two years, every 6 months during the third to fifth year, and annually thereafter. These included physical examinations, esophagogastroduodenoscopy, contrast-enhanced cervical, chest, and abdominal computed tomography scans, electrocardiograms, echocardiography, and pulmonary function tests. Brain magnetic resonance imaging, bone scans, and 18-fluorodeoxyglucose positron emission tomography/computed tomography were performed optionally. Biopsies were conducted when recurrence was suspected.

Tumor response was independently evaluated by two experienced oncologists using the Response Evaluation Criteria in Solid Tumors (RECIST, version 1.1) post-radiotherapy completion. Treatment failures were categorized as either locoregional or distant disease, based on histological, cytological, or definitive radiological evidence. Locoregional recurrence was defined as any recurrence within the esophagus or regional lymph nodes, while distant metastasis referred to non-regional lymph node recurrence or systemic metastasis. Deaths were confirmed through official statements.

The primary endpoint was progression-free survival (PFS), defined as the time from treatment initiation to disease progression or death from any cause, or censored at the last follow-up. The secondary endpoint was OS, defined as the time from treatment initiation to death, or censored at the last follow-up. Post-radiotherapy clinical response and safety were also evaluated.

The continuous variable of age was categorized using its respective median value. The continuous variable of hemoglobin was categorized into anemia grades. The primary tumor length was categorized based on a receiver operating characteristic curve. The body mass index (BMI) was categorized into underweight (BMI < 18.5), normal weight (BMI 18.5-24.9), and overweight (BMI > 24.9). Categorical variables, including demographic factors, clinical stages, and treatment parameters, were analyzed using the χ2 test or Fisher’s exact test.

Comparisons of clinical responses post-radiotherapy and adverse events between the CCRT alone and CCRT+ICI groups were conducted. PFS and OS were compared using Kaplan-Meier methods and log-rank test statistics. Multivariable Cox regression analysis, adjusted for demographic and clinical variables, was conducted to independent prognostic factors for PFS and OS, with results presented as hazard ratios (HRs) and 95% confidence intervals (CIs).

Statistical analyses were performed using SPSS Statistics Version 26.0 software (IBM Co., Armonk, NY, USA) and R software (version 4.2.1). A two-tailed P-value of less than 0.05 was considered statistically significant in all analyses.

As illustrated in Figure 1 , our initial cohort consisted of 152 patients who received definitive CCRT as initial treatment. After applying inclusion criteria, 101 patients were enrolled in the study: 58 in the CCRT alone group and 43 in the CCRT+ICI group. Patient characteristics are summarized in Table 1 .

Baseline clinical characteristics, such as age, sex, smoking history, anemia grade, tumor site, primary tumor length, BMI, clinical T stage, clinical N stage, chemotherapy regimen, staging according to the eighth edition of the American Joint Committee on Cancer, and chemotherapy regimen, were comparably distributed between the two groups. The CCRT alone group had a higher proportion of patients with an Eastern Cooperative Oncology Group performance status score of 1 (86.2%) and a history of alcohol use (75.9%). Conversely, a greater percentage of patients in the CCRT+ICI group received a radiotherapy dose of 60 Gy (97.7%).

The median follow-up time was 21.5 months (interquartile range [IQR]: 12-27 months) for the CCRT alone group and 9 months (IQR: 5.5-14.5 months) for the CCRT+ICI group, with the last follow-up conducted in August 2023. Four patients were lost to follow-up: three in the CCRT alone group and one in the CCRT+ICI group.

The majority of patients in both groups received the paclitaxel plus cisplatin regimen for concurrent chemotherapy (72.4% in the CCRT alone group and 88.4% in the CCRT+ICI group). The cisplatin plus fluorouracil regimen was administered to 27.6% of patients in the CCRT alone group and 11.6% in the CCRT+ICI group. The median number of concurrent chemotherapy cycles was 2 (IQR: 1-3 cycles) in the CCRT alone group and 3 (IQR: 2-4.5 cycles) in the CCRT+ICI group.

In the CCRT+ICI group, 10 (23.2%) patients received 1 cycle, 11 (25.6%) received 2 cycles, and 22 (51.2%) received 3 cycles of concurrent immunotherapy. Post-radiotherapy, 32 patients were under observation, primarily due to patient refusal (n=11) or investigator decision (n=21). For maintenance therapy, 11 (25.6%) patients received ICIs, with a median of 1 cycle (IQR: 1-3 cycles). Discontinuation of maintenance immunotherapy was due to disease progression (n=1), adverse events (n=5), patient refusal (n=3), or investigator decision (n=1).

Clinical response evaluation post-radiotherapy was feasible for 90 (89.1%) patients ( Table 2 ). In the CCRT alone group, 1 (1.7%) patient achieved a complete response, compared to 5 (11.6%) patients in the CCRT+ICI group (P = 0.037). Partial response rates were 44.8% in the CCRT alone group and 41.9% in the CCRT+ICI group. The objective response rates were 46.5% and 53.5% for the CCRT alone and CCRT+ICI groups, respectively. Non-evaluation of response post-radiotherapy was primarily due to patient refusal.

Kaplan-Meier curves and risk tables for PFS are depicted in Figure 2A . The median PFS was 20 months for the CCRT alone group and 14 months for the CCRT+ICI group, with 1-year PFS rates of 55.2% and 58.9%, respectively.

In unadjusted analysis, no significant difference in PFS was observed between the two groups (HR = 1.26, 95% CI: 0.70-2.26; P = 0.445). Multivariable proportional hazard regression analysis also indicated that CCRT+ICI was not an independent prognostic factor for PFS (HR = 0.60, 95% CI: 0.19-1.92; P = 0.388; Figure 3A ).

Figure 2B presents Kaplan-Meier curves and risk tables for OS. The median OS for the CCRT alone group was 40 months, whereas it was not reached in the CCRT+ICI group. The 1-year OS rates were 75.9% for the CCRT alone group and 70.8% for the CCRT+ICI group.

Similar to PFS, unadjusted analysis revealed no significant difference in OS between the groups (HR = 1.21, 95% CI: 0.58-2.53; P = 0.613). Multivariable analysis confirmed that CCRT+ICI was not an independent prognostic factor for OS (HR = 1.25, 95% CI: 0.46-3.37; P = 0.660; Figure 3B ).

Figures 4A, B illustrate PFS and OS across different concurrent immunotherapy cycles. For PFS, median rates were 20 months for 1 cycle, 9 months for 2 cycles, and 17 months for 3 cycles. The 1-year PFS rates were 80.0% for 1 cycle, 36.4% for 2 cycles, and 64.3% for 3 cycles. No significant differences in PFS were observed across the different immunotherapy cycles (P > 0.05 in pairwise comparisons).

Regarding OS, median rates were not reached for 1 cycle, 9 months for 2 cycles, and not reached for 3 cycles. The 1-year OS rates were 80.0% for 1 cycle, 45.5% for 2 cycles, and 80.4% for 3 cycles. While no significant differences were found between 1 cycle and 2 cycles (P = 0.240), or 1 cycle and 3 cycles (P = 0.595), patients receiving 2 cycles exhibited significantly worse OS compared to those receiving 3 cycles (P = 0.044).

Adverse events are detailed in Table 3 . The CCRT alone group exhibited a higher incidence of leukopenia of any grade (93.1% vs. 76.7%), whereas the CCRT+ICI group had a higher incidence of pneumonitis of any grade (36.2% vs. 65.1%). The incidence of adverse events of grade 3 or higher was comparable between the two groups.