Introduction
Endoscopic submucosal dissection (ESD) and endoscopic mucosal resection (EMR) are minimally invasive procedures to remove the dysplastic Barrett’s esophagus (BE) lesion. The optimal resection technique for dysplastic BE lesions remains unclear as while EMR is safe, effective and easy to perform, recurrence rates are lower with ESD. We aim to compare the cost-effectiveness of EMR and ESD for resection of BE neoplasia at different lesion size cutoffs.
Aims & Methods
An incremental cost-effectiveness analysis was performed using a Markov decision tree model including parameters available from the literature review (including 1267 patients who underwent ESD and 4410 who underwent EMR) for BE neoplasia. Monte Carlo simulations were conducted for randomly sampled values for each model estimate to predict the probability of cost-effectiveness at various willingness-to-pay (WTP) thresholds. The base case was a patient with high-risk BE lesions (high-grade dysplasia (HGD) or T1a esophageal adenocarcinoma (EAC)). Outcomes data for BE patients with HGD or T1a esophageal adenocarcinoma were analyzed. The median follow-up period was 31 months (18 months to 5 years). The average lesion size before intervention was higher for ESD compared to EMR (22.7mm vs 13.6mm). We divided the BE HGD lesions into three categories: <10mm, 10-20mm, and >20mm to compare the cost effectiveness of EMR and ESD at different lesion sizes. The probability of overall post-resection recurrence was analyzed, and the incremental cost of the ESD was compared to that of EMR. Effectiveness was defined as the number of esophagectomies or repeat endoscopic procedures avoided per 1000 cases in the simulation. The costs of EMR and esophagectomy were derived through CPT codes and the existing literature, and the cost of ESD was derived through the billing department of our institution. The probabilistic sensitivity analysis reflects the effect of uncertainties of the model variables on the final model outcomes. All our analysis was conducted using TreeAge Pro Software (2021 R1 Healthcare, Win 64-bit).
Results
ESD resulted in higher rates of R0 resection (74.5% vs 40.82%, p=0.001) but higher rates of periprocedural adverse events such as bleeding (2.79% vs 1.13%, p=0.01) and perforation (0.46% vs 0.04%, p=0.01) compared to EMR. ESD yielded a significantly lower overall local recurrence compared EMR (2.92% vs 11.27%, p=0.01). The proportion of recurrence requiring repeat endoscopic treatment following ESD or EMR was 1.84% and 8.96%, respectively. ESD prevented an additional 71.21 per 1000 cases of repeat endoscopic resection compared to EMR. ESD would prevent approximately 12 additional esophagectomies compared to EMR for every 1000 patients undergoing the procedure at an incremental cost per surgery avoided of $5362. Based on our analysis, ESD was more cost effective than EMR and yielded higher Quality of Life (QoL) for lesion size >20mm and for lesion size 10-20mm. For lesion size <10mm, EMR was more cost effective than ESD and yielded higher QoL for lesion size <10mm. Limitations of the study include: (a) Utility data is based on the literature review and may not reflect individual patient preferences. (b) It is assumed in the model that all patients with T1a EAC are eligible for both EMR and ESD at baseline.
Conclusion
ESD is associated with higher upfront costs for the patient, but is more cost-effective in the long term by lower overall recurrence and lower probability of repeat endoscopic or surgical treatment for lesion size 10-20mm and >20mm.