340P - Novel application of spatial analyses to investigate environmental factors and hormone receptor-positive breast cancer

Background

Incidence of hormone receptor-positive (HR+) breast cancer (BC) is increasing. With environmental factors associated with HR+ tumors, we used spatial analysis to identify regions with excess HR+ BC.

Methods

Surveillance data for 38,944 women with a primary diagnosis of BC with reported HR status during 2000-18 and no prior cancers were obtained from the Iowa Cancer Registry. HR+ tumors were defined as those that were estrogen receptor-positive or progesterone receptor-positive. Spatial patterns in HR+ BC/1,000 women and as a proportion of all women with BC with known HR status were examined by county (n=99) for premenopausal aged (15-49 years; n=7,935) and post-menopausal aged (≥50 years; n=31,009) women; proportions were analyzed using logistic regression with spatial correlation via intrinsic conditional auto-regressive models at the county level and general additive models (GAMs) at the residence level (premenopausal, n=7,930; postmenopausal, n=30,988).

Results

HR status was classified for 38,944 women (premenopausal: 6,188 HR+, 1,747 HR-; postmenopausal: 26,190 HR+, 4,819 HR-). Across all counties, rates of HR+ BC ranged from 4.9-17.1/1,000 premenopausal women, representing 57.6-96.8% of BC cases among these women and a state average of 78.0%. Respective values were 29.3-61.1/1,000 postmenopausal women, 75.6-92.6%, and 84.5%. The top 21 counties with the highest estimated proportions of HR+ BC after spatial smoothing within each age group represented 25.3% of all premenopausal women and 21.2% of all postmenopausal women in Iowa. Twelve (57.1%) of these counties overlapped between age groups; two counties were elevated above the respective state averages for both age groups, with 95% probability. GAMs indicated similar areas of increased probability of HR+ BC.

Conclusions

Findings suggest some spatial overlap in proportions of incident HR+ BC between pre- and postmenopausal women. This overlap may indicate shared environmental risk factors for HR+ BC between age groups in these regions. Our analytical approach provides a model that can be applied across broader geographic regions. Future analyses will explore temporal differences in these incident cancers.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

University of Iowa.

Funding

University of Iowa.

Disclosure

A. Thomas: Financial Interests, Personal, Advisory Board, One time Ad Board: AstraZeneca, Genentech; Financial Interests, Personal, Stocks/Shares: Pfizer, Bristol Myers Squibb, Gilead Sciences, Johnson and Johnson, Doximity; Financial Interests, Personal, Royalties, Husband: UpToDate; Financial Interests, Institutional, Local PI: Sanofi, Merck. All other authors have declared no conflicts of interest.