Impact of Endoscopic Eradication of Barrett's Esophagus on Esophageal Adenocarcinoma: A Comparative Modeling Analysis

Research Correspondence

The incidence of esophageal adenocarcinoma (EAC) has risen dramatically over the past four decades in the U.S. and much of the Western world.¹ Most clinical guidelines recommend patients with Barrett's esophagus (BE) undergo endoscopic surveillance with tissue biopsy to grade the severity of precursor lesions and detect curable neoplasia.² Additionally, techniques for endoscopic eradication treatment of BE, such as endoscopic mucosal resection (EMR) and radiofrequency ablation (RFA), have increasingly been employed to limit progression to EAC.³

The National Cancer Institute's (NCI) Cancer Intervention and Surveillance Modeling Network (CISNET) includes three esophageal cancer modeling groups who have independently developed population-based models for the natural history of BE and EAC;⁴ these models have been validated by calibration to NCI SEER data and by numerous comparative modeling exercises.⁵ The aim of the current study was to use the CISNET models to analyze perform a comparative modeling analysis to determine the effectiveness and cost-effectiveness of an RFA centered endoscopic eradication treatment strategy for management of a population of patients with BE. We sought to test and assess the impact of multiple strategies utilizing endoscopic eradication therapy on EAC incidence and mortality, and to estimate the number of surveillance endoscopies and treatments required to produce potential clinical benefits. In addition, we performed a cost-effectiveness analysis to assess the various strategies from a healthcare utilization perspective. Detailed technical profiles of each model are available on the NCI CISNET website and details regarding the methods are available in a downloadable pdf.⁶

In our base case analysis, the simulated cohort was comprised of males born in 1950 with BE diagnosed at age 60. Patients were tracked for EAC incidence and mortality until death by any cause or age 100. Endoscopic surveillance and eradication therapy were discontinued after age 80. Risk of progression to cancer was dependent on calendar year, birth cohort, age, and sex. Outcomes for each strategy analyzed included EAC incidence and mortality; total numbers of surveillance endoscopies and endoscopic eradicative treatments; numbers of treatments needed to avert one EAC death (NNT/death); unadjusted and quality-adjusted life years; number of complications from endoscopy and treatment; and total costs. The NNT/death was calculated as the total number of ablative treatments divided by the number EAC deaths averted by a given strategy. As many patients require multiple treatments, the total number of treatments required provides a better estimate of resource utilization than the number of patients that required treatment. Treatments included the total number of EMR and RFA treatments. Incremental results compared the NNT/death for a given strategy to the next-least invasive strategy by dividing the number of additional treatments by the additional EAC mortality reduction in the more invasive strategy. Cost-effectiveness analyses were conducted from a third party payer perspective. Quality of life (utility values) were derived from literature and used to convert absolute life-years of each strategy into quality-adjusted life-years (QALYs). We quantified the effectiveness of each strategy in terms of quality-adjusted life-years (QALYs), and associated costs, applying the conventional 3% discount rate to both. Incremental cost-effectiveness ratios (ICERs) were calculated between strategies as the ratio of incremental cost to incremental gain in quality-adjusted life-years. Comprehensive details of model inputs or parameter estimates including methods for derivations are available online.⁶

Table 1 summarizes the five strategies for the management of BE. Figure 1 presents the highlighted modeling results for the strategies for each of the three modeling groups with the endpoints of EAC incidence and mortality. The model analysis found that the impact of the different treatment strategies, measured relative to a baseline of surveillance alone, was consistent across all three models. HGD treatment resulted in an average decrease in EAC incidence of 51% (range 46%-54%) and an EAC mortality reduction of 44% (range 39%-49%). In terms of NNT/death, HGD treatment was the most efficient, with a mean of 44 (range 30-56). In this strategy, relatively few treatments were required to achieve a substantial reduction (range 39%-49%). In contrast, the incremental NNT/death for LGD compared to HGD treatment was 346 and 166 in the MGH and ERASMUS/UW models, respectively. The LGD treatment strategy (simulated by the MGH and ERASMUS/UW models only) resulted in a decrease in EAC incidence by 63% (range 58%-67%) and EAC mortality by 58% (range 53%-62%). Treating all BE patients at age 60 decreased the number of EAC cases by 71% (range 68%-79%) and the number of EAC deaths by 68% (range 58%-81%). This strategy was resource intensive, with NNT/death of 350 (range 253-518) compared to HGD treatment alone, and an ICER of $182,093-$422,256/QALY, which is above a $100,000/QALY willingness-to-pay threshold. Model predictions diverged on the cost-effectiveness of treatment for low grade dysplasia (LGD). Additional results of our modeling analyses including unadjusted life expectancy, costs per strategy, and sensitivity analyses on key parameter inputs and surveillance assumptions are available online.⁶

Table 1. Characteristics of simulated interventions on Barrett's esophagus patient cohort.

Strategy	NDBE patients	LGD patients	HGD patients
Natural History (NH)	No intervention	No intervention	No intervention
Surveillance without RFA treatment (S Strategy)	Surveillance endoscopy with biopsies every 3 years	Surveillance endoscopy with biopsies every 6 months in the first year, thereafter every year	Surveillance endoscopy with biopsies every 3 months
BE surveillance with treatment for high-grade dysplasia only (HGD Strategy)	Surveillance endoscopy with biopsies every 3 years	Surveillance endoscopy with biopsies every year	RFA therapy followed by surveillance*
BE surveillance with treatment for all dysplasia (LGD Strategy)	Surveillance endoscopy with biopsies every 3 years	RFA therapy followed by surveillance*	RFA therapy followed by surveillance*
Treatment for all BE patients (BE Strategy)	RFA therapy followed by surveillance*	RFA therapy followed by surveillance*	RFA therapy followed by surveillance*

NDBE: Barrett's esophagus with no dysplasia; LGD: low-grade dysplasia; HGD: high-grade dysplasia

^*All post-treatment surveillance intervals can be found in E-table 5.⁶

Figure 1.

The upper part of the figure shows the EAC incidence per 1000 BE patients per model and strategy (no discounting). The lower part of the figure shows the EAC deaths per 1000 BE patients. The EAC incidence and mortality reductions are shown for the endoscopic eradicative treatment strategies compared to the strategy including only surveillance and no endoscopic eradicative treatment. The range in model estimates reflects differences in model structures and assumptions on BE prevalence and time to development of malignancy. Abbreviations: NH= Natural History; S=Surveillance; HGD=high-grade dysplasia; LGD=low grade dysplasia; BE=Barrett's esophagus without dysplasia

In conclusion, our results strongly confirm the current guidelines that endorse endoscopic eradication therapy for patients with high grade dysplasia.² Our divergent results for LGD, highlights the need for a better understanding of the uncertainties surrounding the LGD health state.⁷ Benefits are predicted to be achieved for all BE endoscopic eradication strategies; however, the efficiency of eradication is substantially reduced if patients with LGD and no dysplasia are treated, and substantially more resources are required to avert a cancer death in these settings. These findings were consistent across all three CISNET esophageal cancer models and were robust to sensitivity analyses of RFA efficacy and durability. Our results add further support for endoscopic eradication therapy for BE patients with HGD, and suggest that strategies targeting less severe disease will require close scrutiny for cost-effectiveness. Efficiency of care would be greatly enhanced through improved methods to stratify risk of cancer in lesser forms of dysplasia and, therefore to better identify individuals who would benefit most from endoscopic therapy.

Abbreviations

BE

Barrett's Esophagus

CE-D

Complete eradication of dysplasia

CE-IM

Complete eradication of intestinal metaplasia

EAC

Esophageal adenocarcinoma

EET

Endoscopic Eradicative Therapy

HGD

High grade dysplasia

LGD

Low grade dysplasia

ND

No dysplasia

NNT/death

Number of treatments needed to prevent one death

RFA

Radiofrequency ablation