Society: AGA

Introduction:
A family history of gastric cancer (GC) is a risk factor for developing this malignancy and portends an increased risk of precursor lesions such as gastric intestinal metaplasia (GIM). Recent American Gastroenterological Association guidelines cite having a family history of GC as an indication for surveillance of GIM; however, evidence on the threshold to initiate screening and surveillance in these individuals and the intervals at which it should continue is lacking. The aim of this study was to assess the effectiveness and cost-effectiveness of screening and surveillance strategies in individuals with a family history of GC.

Methods:
A computer-based Markov simulation model of the natural history of progression from normal gastric mucosa to gastric cancer in a population of Americans with a family history of gastric cancer was developed and calibrated in TreeAge (TreeAge Pro 2020, Williamstown, MA). The model start age was 18 and it had a time horizon until age 100 or death. Cycle length was one month. Model inputs were derived from published literature. The model was used to compare five screening and surveillance strategies: one-time upper endoscopy bundled with colonoscopy at age 45, with surveillance every 3 years or every 5 years if GIM is diagnosed; surveillance at 3- or 5-year intervals irrespective of pathology; and no screening/surveillance. The primary outcomes were total cost, quality-adjusted life years (QALYs), and incremental cost-effectiveness ratios (ICERs). Secondary outcomes were cancer mortality and unadjusted life-years gained. A willingness to pay threshold of $100,000 per QALY was used to determine whether a strategy was cost-effective.

Results:
The cost-effective strategy was one-time upper endoscopy at age 45 with surveillance every 5 years if GIM is diagnosed. This corresponded to an ICER of $13,780/QALY, with an incremental gain of 0.23 QALYs compared to no screening/surveillance. The lifetime cancer mortality rate with this strategy was 1.93%, compared to 3.84% with no screening/surveillance. All other intervention strategies were dominated because they had higher costs and lower QALYs gained compared to the cost-effective strategy. Although cancer mortality was lowest with surveillance at 3-year intervals irrespective of pathology (0.29%), this was offset by lower incremental QALYs gained (0.031) compared to no screening/surveillance, and higher costs ($15,078).

Conclusions:
Based on our Markov model, one-time screening with surveillance every 5 years if GIM is diagnosed was effective and cost-effective in individuals with a family history of gastric cancer in the US context. While this can inform considerations regarding screening and surveillance recommendations, clinical studies will be vital to validating the potential benefits of preventative interventions in this high-risk population.

Table 1. Base-case analysis of a Markov model comparing screening and surveillance strategies in individuals with a family history of gastric cancer.

Table 1. Base-case analysis of a Markov model comparing screening and surveillance strategies in individuals with a family history of gastric cancer.

Ann Zauber

Memorial Sloan Kettering Cancer Center Department of Epidemiology & Biostatistics

Appearances