Surmounting the Learning Curve of Three-field Mie: One Institution’s Experience

Jeffrey S Fronza, MD, Brett C Sheppard, MD, Paul H Schipper, MD, Brian Diggs, PhD, Miriam A Douthit, MA, John G Hunter, MD. Oregon Health and Sciences University

Introduction: Minimally-invasive esophagectomy (MIE) is a technically demanding operation and expertise is concentrated in a small number of high volume centers. Our institution has performed MIE for 10 years utilizing Ivor-Lewis, transhiatal, and 3-field techniques. Three-field has only supplanted transhiatal as our preferred approach since 2009. In this study we report our entire experience with 3-field MIE in order to better understand its learning curve at an established center of esophageal surgery.
 

Methods: The Natural History of Esophageal Cancer and Related Diseases database has prospectively accrued perioperative and long-term follow-up data on over 350 patients since 2004. After querying this database, 73 patients were identified as having undergone 3-field MIE between Oct 2005 and June 2011. This cohort was divided into an early and late group in terms of our experience with 3-field MIE. We hypothesized, based on existing esophagectomy literature, that a year in which greater than 20 3-field MIE were performed was necessary to surmount its learning curve. Thus the 35 patients who underwent 3-field MIE during the last 1.5 years of our experience were compared to the 38 patients who underwent MIE during previous years. Of specific interest were surgical morbidity, in-hospital mortality, and oncologic outcome.
 

Results: The groups were similar in all preoperative factors except ASA class (84% ASA III in early vs. 31% in late [p<.001]). Ninety-six percent underwent 3-field MIE for cancer, and 80% had adenocarcinoma or dysplasia of the gastroesophageal junction. Ninety-three percent of our cancer patients underwent neoadjuvant therapy. Complications occurred in 77% of patients (90% early vs. 63% late [p=.007]). All 7 conversions and all 3 mortalities were in the early group. There were 5 reoperations in the chest or abdomen, and all but one was in the early group. Anastomotic leak rate was 16% (26% early vs. 6% late [p=.018]). Two patients, both in the early group, required takedown of their conduit and cervical esophagostomy. Recurrent laryngeal nerve palsy occurred in 11% (18% early vs. 3% late [p=.033]). Chylothorax occurred in two patients in the late group, one of whom was successfully managed with IR coil emboilization while the other required thoractomy. Pneumonia occurred in 37% of the early group and in only 14% of patients in the late group (p=.028). Atrial fibrillation was the most common complication in both groups (53% early vs. 23% late (p=.009). Median LOS was 13.5 in the early group vs. 9 in the late (p<.001). An R1 resection was done on 5 patients in the early group, while only once in the late group. Median lymph node harvest was significantly better in the late group (16 early vs. 19 late [p=.014]).
 

Conclusion: Three-field MIE remains a challenging operation, even in a high volume center. An institutional experience of one year with greater than 20 cases is needed to surmount the severe learning-curve, and subsequently results significantly improve. It appears prudent for these cases to be performed at institutions with a significant commitment to esophageal surgery as volume does matter.
 

View Poster