From May 2019 to September 2021, we retrospectively recruited ESCC patients who received camrelizumab plus chemoradiotherapy as neoadjuvant therapy before surgery at the Second Affiliated Hospital of Shandong First Medical University (Shandong, China). The key inclusion criteria were the following: (1) Patients with histologically confirmed locally advanced ESCC, which was defined as cT1N1-3M0 or cT2-4aN0-3M0 (AJCC, 8th Edition); (2) Patients were aged 18 years or older; (3) Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 or 1; and (4) With adequate organ function for surgical resection. Patients were excluded when they had other anti-tumor treatments before or during the neoadjuvant treatment, immunodeficiency disease, and other significant concurrent malignant tumors. The study was approved by institutional review board of Second Affiliated Hospital of Shandong First Medical University, and carried out in accordance with the Helsinki Declaration (as revised in 2013). Written informed consent was obtained from all patients.

Camrelizumab was given 200 mg intravenously every 3 weeks (a cycle). Simultaneously, the paclitaxel was administered intravenously at a dose of 100 mg/m² of body-surface area on days 1 and 8, and the carboplatin was administered intravenously at an area under the curve of 5 mg/ml per minute on day 1. The intensity-modulated radiotherapy was given according to Chinese treatment guidelines for esophageal carcinoma (19), and was prescribed to cover 95% of the planning gross tumor volume (PGTV), given at 2.0 Gy per fraction, five fractions per week. The surgery was performed at the surgeon's decision after completion of at least 2 cycles of neoadjuvant therapy. Patients were re-evaluated with contrast-enhanced CT within 1 week before surgery. Standard minimally invasive esophagectomy (MIE) was performed for all patients, and the upper tumor mainly was treated with three-incision McKeown surgery (three fields or two fields). The middle and lower segment tumors with two-incision Ivor-Lewis surgery. Besides, a gastric tube was applied to reconstruct the digestive tract after esophagectomy. The surgical indicators, including operative time, blood loss, blood transfusion, hospital stay, and resection margin were first recorded. The postoperative complications, including pneumonia, chylothorax, pleural effusion, wound infection, and recurrent nerve paralysis were also recorded. Patients were followed postoperatively with routine CT scans every 3 months in the first year following the treatment, and every 6 months thereafter.

The primary outcome was overall survival (OS) and disease-free survival (DFS) at 12 months. OS was defined as the time from neoadjuvant treatment to date of death, and DFS was defined as the time from neoadjuvant treatment until disease recurrence or death. Secondary outcomes were radiologic response prior to surgery, pathological responses including major pathological response (MPR) and complete pathological response (pCR), and treatment-related adverse events (TRAEs). Radiologic responses were assessed according to Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. MPR was defined as residual tumor cells ≤10% at the time of surgery, and pCR was defined as tumors without any viable tumor cells. TRAEs were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 5.0. PD-L1 expression was assessed using the PD-L1 IHC 22C3 pharmDx assay (Agilent Technologies, CA, USA) and was expressed as combined positive score (CPS) by dividing the number of PD-L1-stained tumor and immune cells with the total number of viable tumor cells and multiplying by 100. The tumor PD-L1 expression was defined positive when the CPS ≥1%.

All statistical analyses were done using SPSS 20.0 (IBM SPSS Inc., Chicago, IL, USA). Continuous variables were presented with as a mean ± standard deviation when the normality was verified by Shapiron-Wilk test (P > 0.1), otherwise median and range. The categorical variables were expressed as counts and percentages. The Kaplan-Meier method was used to estimate DFS and OS and corresponding 95% CIs. Differences were considered to be significant when P < 0.05.

As shown in Table 1, a total of 38 patients with ESCC were included (31 males, 7 females; median age 59 years). 30 (78.9%) patients had ECOG PS 0 and 8 (21.1%) had ECOG PS 1. More than half of the patients (60.5%) were smokers. Tumors were located in the proximal third of the esophagus in 8 (21.1%) patients, the middle third in 16 (42.1%) patients, and the distal third in 14 (36.8%) patients. There were 7 (18.4%) patients at TNM stage II, 27 (71.1%) patients at stage III, and 4 (10.5%) patients at stage IVA. Moreover, 22 (57.9%) patients had negative PD-L1 expression and 12 (31.6%) patients had positive PD-L1 expression. Time from completion of neoadjuvant therapy to surgery was 31 days (18–46).

The TRAEs during neoadjuvant treatment are shown in the Table 2. A total of 34 (89.5%) patients experienced some form of TRAEs. Grade 3 adverse events occurred in 8 (21.1%) patients, and no grade 4 or 5 adverse events were observed. Grade 3 radiation esophagitis occurred in 3 (7.9%) patients. Other grade 3 adverse events included pain in 4 (10.5%) patients, decreased WBC count in 2 (5.3%) patients, and fatigue in 1 (2.6%) patient. No treatment-related deaths were reported. Immune-related adverse events occurred in 18 (47.4%) patients. The most common immune-related adverse events were reactive capillary endothelial proliferation in 17 (44.7%) patients, and hypothyroidism in 6 (15.8%) patients, all in grade 1. Other immune-related adverse events including myocarditis, hepatitis, and nephritis were not observed.

The surgical details are presented in the Table 3. 16 (42.1%) patients had the McKewon surgery and 22 (57.9%) patients had Ivor Lewis surgery. The median operation time was 315 min (296–419), the median amount of blood loss was 196 mL (134–431), and the median hospital stay was 16 days (12–24). 36 (94.7%) patients achieved a R0 resection and 2 (5.3%) patients had a R1 resection. During the postoperative periods, 6 (15.8%) patients had pneumonia, 5 (13.2%) had chylothorax, 4 (10.5%) had pleural effusion, 2 (5.3%) had wound infection, and 2 (5.3%) had recurrent nerve paralysis. There was no death in hospital or any other serious intraoperative complications.

The radiologic and pathologic responses are summarized in Table 4. Of all 38 patients, 3 (7.9%) patients had a complete response (CR), 21 (55.3%) patients had a partial response (PR), and 14 (36.9%) had a stable disease (SD). The objective response rate (ORR) was 63.2% and the disease control rate (DCR) was 100.0%. According to postoperative pathological results, 19 (50.0%) patients had a MPR, 15 (39.5%) patients had a pCR, and 4 (10.5%) patients were non-responders.

By January 1, 2022, the median follow-up period was 18.5 months (range 3.4–28.5) and 6 (15.8%) patients had died. As demonstrated in Figure 1, the median OS and DFS was not reached in all patients; the OS rate was 87.6% at 12 months, 77.2% at 18 months, and 77.2% at 24 months; the DFS rate was 78.7% at 12 months, 73.5% at 18 months, and 73.5% at 24 months. In subgroup analyses, patients with a MPR or pCR achieved better OS (all P < 0.001, Figure 2A) and DFS (all P < 0.001, Figure 2B) than non-responders. Despite no significant statistical difference, patients with positive tumor PD-L1 expression tended to be associated with longer OS (P = 0.170, Figure 3A) and DFS (P = 0.118, Figure 3B).