RESUMEN
Objective: To examine the clinical value of routine contrast esophagram (RCE) for the diagnosis of anastomotic leakage (AL) after three-incision esophagectomy with cervical anastomosis. Methods: Clinical data of 1 022 patients with esophageal cancer who underwent McKeown three-incision esophagectomy with cervical anastomosis from January 2015 to December 2019 at Department of Minimally Invasive Esophageal Surgery, Tianjin Medical University Cancer Hospital and Institute were analyzed retrospectively. There were 876 males and 146 females, aging(M(IQR)) 48(16) years (range: 36 to 84 years). There were 253 patients (24.8%) with neoadjuvant therapy, and 817 patients (79.9%) with minimally invasive esophagectomy. According to the diagnosis and treatment habits of the attending surgeons, 333 patients were included in the RCE group, and RCE was performed on the 7th day postoperative, while 689 patients were included in the non-RCE group, and RCE was performed when the patients had suspicious symptoms. Taking clinical symptoms, RCE, CT, endoscopy and other methods as reference to the diagnosis of AL, the sensitivity and specificity were used to analyze and evaluate the efficacy of RCE for the diagnosis of AL. The data were compared by U test or χ² test between groups. Results: The incidence rate of AL after three-incision esophagectomy was 7.34% (75/1 022), including 30 cases in the RCE group and 45 cases in the non-RCE group (9.0%(30/333) vs. 6.5%(45/689), χ²=2.027, P=0.155). The diagnostic time of AL was 9(5) days postoperative (range: 4 to 30 days). Among them, 23 cases showed cervical leakages, 50 cases showed intro-thoracic leakages, and 2 cases both cervical and intro-thoracic leakages. The diagnostic time of patients with intro-thoracic leakages was longer than that of cervical leakages (10(4) days vs. 6(3) days, Z=-2.517, P=0.012). Among the 333 patients in the RCE group, 16 cases of RCE indicated leakages including 11 cases of true positive and 5 cases determined to be false positive, while 317 cases indicated no abnormalities including 19 cases developed leakages. The sensitivity and specificity of RCE to detect AL were 36.7%(11/30) and 98.3%(298/333), respectively. The Youden-index was 0.35, and the diagnostic accuracy was 92.8%(309/333). The positive and negative predictive value were 11/16 and 94.0%(298/317), respectively. Conclusions: Routine contrast esophagram after three-incision esophagectomy with cervical anastomosis has low sensitivity and high specificity in the diagnosis of AL. The diagnostic time of AL is the 9th day after surgery. It is necessary to prolong the observation time clinically, and combine RCE with CT, endoscopy and other inspection methods for diagnosis.