Skip to main content
. Author manuscript; available in PMC: 2021 Sep 3.
Published in final edited form as: Cancer Res. 2020 Dec 8;81(4):1123–1134. doi: 10.1158/0008-5472.CAN-20-0335

Figure 3. Validating predicted prevalence and optimizing screening ages.

Figure 3

(A) Model predictions for BE prevalence in general US population stratified by sex, with contributions of relative risk (RR) of BE from the prevalent GERD subpopulation assumed to be RR=5, solid lines (shaded areas, RR=2-6). Dashed lines are BE prevalence data from CORI [14]. (B) Model predictions for BE prevalence in GERD populations with fixed values assumed for range of RR since birth (RR=2-6). CORI data (dashed lines) corresponds with model’s expected predictions for true GERD-specific BE prevalence contributing to subpopulation in (A) that would lie within shaded regions for men and women due to heterogeneity in GERD onset ages. For analogous age-specific populations, bottom row depicts optimization results from screening objective function (see Materials and Methods) with w = 1 in (C) general populations and (D) symptomatic GERD populations, including single age results for optimal initial screening age t’s (blue diamonds). Abbreviation: esophageal adenocarcinoma (EAC)