RESULTS
A total of 26 consecutive patients were identified with a specimen consistent with intramucosal carcinoma. Patient allocation is described in Figure 1. One patient was excluded from the esophagectomy group as he was identified on final pathology to have node positive disease. Two patients were excluded from the EMR group when initial EMR showed invasive deeper than T1sm1 and were referred for consideration of oncologic resection. Therefore, 12 patients were analyzed in the esophagectomy group and 11 patients were analyzed in the EMR group.
At baseline, patients in both groups were similar with respects to age and gender. There was a trend towards patients receiving EMR having more co-morbidities than patients undergoing esophagectomy. In the esophagectomy group, the majority received an Ivor-Lewis (67%) versus a transhiatal esophagectomy (33%). Most (75% vs 91%) intramucosal adenocaricomas were found as part of a short (<3cm) segment of Barrett’s in both groups and there were no differences in grade of differentiation. Patient characteristics are presented in Table 1.
Of the 11 patients being treated by esophagectomy, one patient (1/12, 9%) experienced recurrence of disease 1.6 years after Ivor-Lewis resection, culminating in mortality 5 years after initial surgery. In the EMR group, one patients (9%) experienced progression of disease after treatment 8 months after initial EMR for a T1m3 lesion and was not a surgical candidate when nodal disease was discovered. One patient (9%) was found to have recurrence of adenocarcinoma after initial clearance by EMR on the 6th surveillance endoscopy, 4 years after his initial treatment. He was treated with a transhiatal esophagectomy at that time. He had long segment of Barrett’s with several biopsies showing high grade dysplasia. Successful eradication by EMR was therefore accomplished in 10/11 (91%) patients with one long-term recurrence treated with salvage esophagectomy in the EMR group.
The morbidity of esophagectomy was significantly greater than with EMR. Most patients undergoing esophagectomy (8/12 (75%)) had a post-operative complication, two of which required re-intervention or intensive care stays (17%). Median length of stay after esophagectomy was 15 [10-22] days. Further postoperative outcomes by type of operation are described in Table 2. EMR, on the other hand, was performed as an outpatient procedure with low morbidity. In the EMR group, no patient had 30-day complications related to the procedure. One patient was admitted for 1 day after endoscopic control of bleeding during EMR to ensure no rebleeding occurred. All other EMRs were performed on an outpatient basis. A median of 1 [1-2] endoscopies were required to achieve eradication. There was one major complication in the EMR group: an anastomotic leak after transhiatal esophagectomy performed upon recurrence of disease. Clinical outcomes are presented in Table 2.
Estimated survival functions for 5 year mortality and disease-free survival are presented in Figures 2 and 3. There was one mortality in each group, leading to similar survival curves between esophagectomy and EMR (p=0.62). Disease-free survival curves are parallel between both groups, though one additional recurrence in the EMR group required salvage esophagectomy for disease clearance. Average follow-up period was 3.5 (1.6) years and 2.7 (1.1) years respectively (p=0.16).