Non-inferiority of minimally invasive oesophagectomy: an 8-year retrospective case series.

Abstract

The trend towards laparoscopic surgery seen in other specialties has not occurred at the same pace in oesophagectomy. This stems from concerns regarding compromised oncological clearance, and complications associated with gastric tube necrosis and anastomotic failure. We present our experience of minimally invasive oesophagectomy (MIO) compared to open and hybrid surgery. We aim to ascertain non-inferiority of MIO by evaluating impact on survival, oncological clearance by resection margin and lymph node harvest and post-operative complications. Data were sourced retrospectively 2008-2015. Three approaches were studied. MIO (3-stage Mckeown), hybrid (2-stage Ivor Lewis, laparoscopy, thoracotomy) and open (2-stage Ivor Lewis). Five-year survival was 54.2%. Surgical approach had no significant impact on survival at any stage of disease (Stage 0/I p=0.98; stage II p=0.2; stage III p=0.76). There was no statistically significant difference in oncological clearance by resection margins between procedures when compared by disease stage (p=0.49). A higher number of nodes were harvested in hybrid [median 27.5 (6-65)] and open surgeries [median 26 (4-54)] than in MIO [median 20 (7-44)] (p>0.01). Numbers of nodes resected did not impact risk of recurrence [recurrence, median 25 (6-54), no recurrence, 26 (4-65)] (p=0.25). Anastomotic strictures (22.4%) and potential leaks (17.9%) were more common in MIO (strictures p>0.01, leaks p=0.08), although associated morbidity was lower. Respiratory complications were less common in MIO (2.9%) versus hybrid (13.3%) (p=0.02). Wound infection and chyle leak were also lower (wound 1.5% MIO 3.5% open, p=0.6; chyle leak 1.5% MIO, 6.7% hybrid, p=0.2). Our results show no negative impact of MIO on survival or oncological clearance. Respiratory and wound complications are lower in MIO, but rates of anastomotic strictures and potential anastomotic leaks are increased. This may be due to the longer length of conduit and subclinical ischaemia at the anastomosis and merits further evaluation.