This study protocol was approved by the ethics investigation committee of The First Affiliated Hospital of Zhengzhou University. Ethical approval code: SS-2018-22. Written informed consent was obtained from each patient during questionnaire administration for the collection and analysis of applicable clinical data.

The inclusion criteria were: (i) NEAS confirmed by imaging and pathology findings; (ii) patients were treated with interventional chemotherapy; (iii) Age >18 yr and Age <85yr, gender unlimited; (iv) all patients refused the second operation. The exclusion criteria were: (i) a benign cicatricial stricture of an esophagogastric anastomosis; (ii) AIC was not used to treat NEAS; (iii) AIC for unresectable esophageal cancer.

We retrospectively analyzed the data of patients with NEAS after resection of an esophageal tumor treated by AIC in the Interventional Therapy Center of The First Affiliated Hospital of Zhengzhou University from October 2014 to December 2019. These data comprised medical records, imaging findings, interventional procedures, and follow-up information.

Dysphagia was assessed using Stooler grading criteria (14). Patients with an esophageal fistula and severe stricture of an esophagogastric anastomosis were treated with a nasojejunal nutrition tube. The degree of stricture was determined by DSA.

According to the classification standard for dyspnea set by the American Thoracic Association (ATA), all patients with a stricture in their airways were evaluated for improvement of respiratory function 1–7 days before and after stent implantation in the airways, or after AIC for esophageal cancer.

Clinical staging of all patients before and after treatment was evaluated according to the criteria for clinical staging set by the American Joint Committee on Cancer (15). The decrease of clinical stage after AIC represents the better effect of AIC. The clinical efficacy of drugs used to treat esophageal cancer was evaluated according to the complete response (CR), partial response (PR) and whether the patient had stable disease (SD) or progressive disease (PD) (16, 17). CR+PR was defined as the objective response rate (ORR). Disease control rate (DCR) was defined as CR+ PR + SD.

If the curative effect of the lesion reached a CR, then conversion to radiotherapy was indicated. If the curative effect of the lesion reached a PR or SD, pulse perfusion chemotherapy was undertaken. If the lesion was evaluated as PD, other types of palliative treatment were indicated.

Adverse reactions to the chemotherapy drugs, changes in tumor size, and laboratory test results were recorded. The duration of survival was documented. The toxicity and side-effects of chemotherapy drugs were evaluated according to National Cancer Institute Common Toxicity criteria and classification of anticancer-drug toxicity (0–IV).

Data were analyzed using Prism 5 (GraphPad, San Diego, CA, USA). An independent two-sample t-test was used to compare quantitative data. Data are the mean ± SD. The Student’s t-test was used to compare continuous variables. The chi-square test was employed to compare categorical data. P < 0.05 was considered significant.

Fifty patients with NEAS after esophagectomy (32 men and 18 women; mean age, 63.6 ± 8.8 years) formed the study cohort. Of these 50 cases, two had hypertension, three had diabetes mellitus (DM), and one had hypertension and DM. Thirty-eight cases were complicated with different degrees of airway stricture: tracheal stricture (30 cases), carina-region stricture (one), stricture of the right-main bronchus (one) and stricture of the left-main bronchus (six; including three cases of stricture of the left-main bronchus combined with atelectasis). Among the 50 cases, 10 cases were complicated with an esophageal fistula (eight cases of an esophagotracheal fistula and two cases of esophagogastric anastomotic fistula) ( Table 1 ).

The basic information of all patients before and after esophagectomy and before recurrence is shown in Table 2 .

Among the 38 patients with different degrees of an airway stricture before procedure, 13 patients had tolerable dyspnea without the requirement of an airway stent, and 25 patients with a severe stricture of the airways who could not tolerate AIC were treated with an airway stent. Among these 25 patients, four cases underwent implantation with a Y-type metal stent, and 21 cases had a tubular metal stent. Eleven patients with grade-IV dysphagia due to a neoplastic anastomotic stricture were treated with a nasal nutrition tube under fluoroscopy guidance. Ten patients with NEAS (including 5 cases of IV dysphagia) and an esophageal fistula had conservative treatment (e.g., nutrition-tube placement) and one patient with an esophageal fistula was treated with a covered airway stent.

In 50 patients, the feeding artery of the tumor was identified and perfused with a chemotherapy drug. For each patient, 1–4 feeding arteries were perfused, including the bilateral inferior thyroid artery (11 cases), unilateral inferior thyroid artery (six), bilateral bronchial artery (16), unilateral bronchial artery (30), right gastroepiploic artery (27), thyroid carotid artery (five) and right gastric artery (three). When the 5F catheter could not enter the tumor feeding artery, the microcatheter was used to superselect the tumor feeding artery. When intercostal artery and bronchial artery are involved, microcatheter should be used to bypass the potential spinal artery branch and be as close as possible to the tumor feeding artery. Intraoperatively, a microcatheter was used for super-selective intubation 63 times Thirty-seven patients received one course of AIC, 10 patients received two courses of AIC, and three patients received three courses of AIC.

Fifty patients who underwent esophagectomy had different degrees of stricture of an esophagogastric anastomosis before AIC: 34 cases had grade III and 16 cases had grade IV. According to Stooler grading criteria for dysphagia classification, after AIC, one patient had grade I, 40 cases had grade II, and nine cases had grade III. The degree of stricture of the esophagogastric anastomosis was relieved to different degrees after AIC. Eight of the 11 patients who underwent nasogastric intubation due to an esophagogastric anastomotic stricture had the nutrition tube removed and oral feeding was initiated after 3 weeks of AIC. Among them, three cases continued to have grade-III dysphagia. Twenty-five patients had dyspnea before surgery. According to the ATA classification of dyspnea, all 25 patients were grade IV before stenting. After tracheal stenting, oxygen saturation was >95% without oxygen inhalation, and 11 cases had grade-II and 14 cases had grade-III dyspnea. Thirteen patients with an airway stricture could tolerate dyspnea without implantation of an airway stent. According to the ATA classification of dyspnea, all 13 patients had grade-III dyspnea preoperatively. One week after surgery, oxygen saturation was >95% without oxygen inhalation. Among them, five cases had grade-I and eight cases had grade-II dyspnea. In one patient with a fistula in the neck portion of the esophagus, closure was achieved 3 weeks after placement of a nasojejunal nutrition tube. One patient with an esophagotracheal fistula was treated with a covered airway stent and nutrition tube, and the fistula healed completely after one course of AIC.

The clinical stage of all patients before treatment was T3 (six cases) and T4b (44). After 1–3 courses of treatment, 50 patients were followed up, and the clinical stage was T1 (four cases), T2 (12), T3 (16) and T4b (18 cases). After treatment, the clinical stage decreased significantly ( Table 3 ).

After the first course of AIC, a CR was noted in four cases, PR in 31 cases and SD in 15 patients and the ORR was 70.0%. Thirteen patients received a second course of AIC for esophageal cancer. After the second course, a CR was noted in two cases, PR in 10 patients and SD in one case, and the ORR was 92.3%. Three patients received a third course of AIC for esophageal cancer, after which a CR was noted in two cases, SD in one patient, and the ORR was 66.7%. After 1–3 courses of treatment, a CR was documented in eight cases, PR in 28 patients, and SD in 14 cases, and the ORR was 72.0%, and the DCR was 100% ( Table 4 ).

Adverse reactions of grade I–III occurred after AIC for esophageal cancer ( Table 5 ). Nausea and vomiting (42.0%) were noted in 21 cases, thrombocytopenia was documented in seven patients (14.0%), fever in three cases (6.0%), and leukopenia in eight cases (16.0%). These and other common adverse reactions were relieved in a short time after symptomatic treatment.

During follow-up, the median duration of survival was 8.5 months. Twenty five patients died of systemic organ failure at the end of the tumor. The duration of survival (in months) of seven patients was 12.0 ± 1.8, and that of five patients was 24.0 ± 1.2. Thirteen patients died of tumor-related respiratory failure, and five patients survived for 12.0 ± 2.1 months. Seven patients died of gastrointestinal bleeding, and one patient survived >12 months after AIC. One patient died of systemic bone metastasis 12 months after procedure.

Currently, four patients are alive. One patient is receiving postoperative radiotherapy, has survived for 36 months, and is consuming a liquid diet. Two patients have received immune-targeted therapy: one patient has survived for >36 months and the other patient has survived for 24 months without dysphagia/dyspnea symptoms. The other patient has survived the longest (>42 months) without dyspnea symptoms. For these four cases, 1-year survival is 46.0%, and the median duration of survival is 8.5 months.