From May 2011 to July 2021, thirteen treatment centers enrolled 401 patients, who were then randomized to undergo either large-field irradiation (n = 210) or small-field irradiation (n = 191, Fig. 1). The demographic and baseline characteristics were generally well-balanced between two groups in the intention-to-treat population (Table 1). The study population was predominantly male, comprising 88.0%. The average number of lymph node dissections was comparable between the LFI and SFI groups, recorded at 25 ± 11 and 25 ± 12, respectively. A significant proportion of participants, precisely 63.3%, were diagnosed with moderately differentiated ESCC. In the LFI group and SFI group, 5.7% and 7.3% of patients, respectively, were diagnosed with stage T4a. Lymph node negativity was reported in 32.9% of the LFI group and 31.9% of the SFI group. Furthermore, a notable proportion of patients, 43.3% in the LFI group and 40% in the SFI group, did not receive adjuvant chemotherapy after radiotherapy.

CONSORT diagram of patients. ESCC, esophageal squamous cell carcinoma; LFI, large-field irradiation; SFI, small-field irradiation; AE, adverse event

With a median follow-up time of 41.8 (95%CI, 38.0 to 45.7) months for survivors, disease progression occurred in 46.6% (95%CI, 39.8% to 53.7%) of the LFI group and 47.6% (95%CI, 40.4% to 55.0%) of the SFI group. The one-year, three-year DFS rates for the entire cohort were 72.0% (95%CI, 68.9% to 77.8%) and 51.6% (95%CI, 46.4% to 56.4%), respectively. The median DFS was 47.9 (95%CI, 22.2 to 73.6) months in the LFI group, closely similar to the 48.1 (95%CI, 12.1 to 84.0) months of the SFI group (HR = 0.87, 95%CI, 0.65 to 1.16; p = 0.32, Fig. 2A). In the LFI group, the one-year and three-year DFS rates were 76.4% (95% CI, 70.7% to 82.5%) and 51.5% (95% CI, 44.4% to 63.9%), respectively. In the SFI group, these rates were 67.6% (95% CI, 62.5% to 76.0%) and 51.6% (95% CI, 44.5% to 59.1%), respectively. Multivariable analyses identified the primary esophageal cancer lesion (HR = 0.71, 95%CI, 0.52 to 0.98; p = 0.03), T stage (HR = 0.47, 95%CI, 0.28 to 0.81; p = 0.006), and the number of lymph node metastases (HR = 0.47, 95%CI, 0.35 to 0.63; p < 0.001) as significant prognostic factors for DFS (Additional file 6: Fig. S5A). Notably, patients in both groups had a consistent DFS benefit across stratified factors (Fig. 2C).

Disease-free survival and overall survival of the ITT population. Comparing rates of disease-free survival (A) and overall survival (B) between two groups. Panel (C-D) shows subgroup analysis. No, number; HR, hazard ratio; LN, lymph node

The median OS for the entire cohort was 74.6 (95%CI, 35.8 to 111.1) months. Specifically, the LFI group had a median OS of 133.0 (95% CI, 32.1 to 233.9) months, compared to 73.5 (95% CI, 32.3 to 114.7) months in the SFI group. Predominantly, mortality was attributable to ESCC, with the exception of one patient who died of primary lung cancer; three patients who died of bowel obstruction and two patients who died of malnutrition; and two patients who died suddenly due to cardiovascular ailments. For the entire cohort, the one-year and three-year OS rates were 88.0% (95%CI, 84.9% to 91.4%) and 62.0% (95%CI, 57.1% to 66.8%), respectively. The estimated one-year and three-year OS rates were 89.2% (95%CI, 84.4% to 93.2%) and 63.2% (95%CI, 56.4% to 69.8%) for the LFI group, and 86.6% (95%CI, 81.1% to 91.2%) and 60.7% (95%CI, 52.9% to 67.1%) for patients in the SFI group, respectively. The difference of OS between the groups was not statistically significant (HR = 0.86, 95%CI, 0.63 to 1.16; p = 0.35, Fig. 2B). Multivariable analyses showed that patients undergoing over three cycles of adjuvant chemotherapy (HR = 1.61, 95% CI, 1.07 to 2.10; p = 0.02), those with T3 stage (HR = 0.37, 95% CI, 0.22 to 0.62; p < 0.001), and those with fewer than three lymph node metastases (HR = 0.50, 95% CI, 0.36 to 0.69; p < 0.001) had extended OS with PORT (Additional file 6: Fig. S5B). However, subgroup analyses revealed no notable differences in OS relative to baseline characteristics and treatment groups (Fig. 2D).

At the data cutoff for this analysis, twenty-seven patients (12.9%, 95%CI, 9.0% to 18.1%) in the LFI group and 20.4% (95%CI, 16.0%-28.1%) in the SFI group experienced locoregional relapse. The median LRFS of both groups has not been reached. The one-year and three-year LRFS rates were 92.3% (95% CI: 88.3%-95.8%) and 83.7% (95% CI: 78.0%-88.4%) in the LFI group, and 87.0% (95% CI: 81.7%-91.6%) and 74.7% (95% CI: 68.0%-80.7%) in the SFI group. Patients in the LFI group had better locoregional control than those in the SFI group (HR = 0.54, 95%CI, 0.34—0.87; p = 0.01; Fig. 3A).

Locoregional-recurrence-free survival and distant-metastasis-free survival. Comparing rates of locoregional-recurrence-free survival (A) and distant-metastasis-free survival (B) between two groups. No, number; HR, hazard ratio

Distant disease was observed in 33.3% of the LFI group and 29.3% of the SFI group, with the median DMFS not yet attained in either group. The one-year and three-year DMFS rates were 83.5% (95% CI: 77.4%–88.0%) and 63.6% (95% CI: 56.9%-70.2%) in the LFI group, compared to 76.3% (95% CI: 69.7%-82.1%) and 66.8% (95% CI: 59.3%-73.0%) in the SFI group. No significant difference in DMFS was found between the groups (HR = 1.05, 95%CI, 0.74–1.49; p = 0.78, Fig. 3B).

Fewer patients in the LFI group experienced locoregional recurrence compared to the SFI group (12.9% [27/210] vs. 20.4% [39/191]), with this difference being statistically significant (p = 0.013). While distant disease was more prevalent, encompassing 25.7% (103/401) of cases across both groups, compared to locoregional recurrence at 11.0% (44/401). However, no significant difference in the incidence of distant metastasis between the groups was found (p = 0.326). Concurrent locoregional recurrence and distant metastasis were documented in 4.3% (9/210) in the LFI group and 6.8% (13/191) of patients in the SFI group (Table 2).

The lung was the most common site of metastasis in both groups, followed by the bone and liver. A higher incidence of supraclavicular lymph node recurrences was observed in the SFI group compared to the LFI group, involving 15 and 4 patients, respectively. Similarly, celiac lymph node recurrence manifested in twelve patients from the LFI group and eleven from the SFI group (Additional file 7: Fig. S6).

Among patients with local recurrences, a higher proportion of individuals in the SFI group experienced out-of-field failure compared to those in the LFI group (14.1% [27/191] vs. 4.8% [10/210], p < 0.05). In-field recurrence was observed in 8.1% (17/210) of patients in the LFI group and 5.8% (11/191) of patients in the SFI group (p < 0.05, Table 2).

A total of 186 patients (46.4%) experienced adverse events related to the treatment. The most prevalent toxicity observed was grade 2 radiation esophagitis, affecting 22.9% (48/210) of patients in the LFI group and 16.8% (32/191) in the SFI group. This was followed by grade 2 hematologic toxicity, occurring in 14.3% (30/210) of the LFI group and 10.5% (20/191) of the SFI group. In the LFI group, 6.7% (14/210) of patients experienced grade 3 or higher adverse events, including four cases of grade 3 hematologic toxicity, seven cases (3.3%) of grade 3 esophagitis, and two cases of grade 3 pneumonitis. In contrast, the SFI group had one patient (0.5%) with grade 3 hematologic toxicity, two patients of grade 3 esophagitis, and two patients (1.0%) with grade 3 nausea or vomiting. There were no occurrences of grade 4 or 5 toxicities in either treatment group. No significant differences in treatment-related toxicities were observed, and all toxicities were manageable in this study (Table 3).