Abstract 5907: Hot and cold spots of young-onset colorectal cancer mortality in U.S. counties, 1999-2019

Abstract

Abstract Colorectal cancer mortality rates among those diagnosed under age 50 have been rising. Geospatial patterns of young-onset colorectal cancer (yoCRC) mortality rates in the U.S. have received limited attention, and prior studies were limited by a lack of adjustment for demographic factors, a focus only on hot spots, and a lack of cluster-specific relative risks (RRs). Adjustment allows clusters to represent areas where modifiable factors may be driving anomalous mortality rates. Aggregated 1999-2019 yoCRC mortality data for 3,036 counties was obtained from CDC WONDER Underlying Cause of Death, and demographics were obtained from 2005-2019 census variables and 2015 Robert Graham Center data. Mortality rates were stabilized via spatial smoothing, then a quasi-Poisson model was fit with median age, sex, race/ethnicity, and social deprivation. Adjusted yoCRC death counts were utilized in a Poisson circular spatial scan to identify Gini hot/cold spots at a maximum cluster size of 6% of the population at risk. Resulting RRs signified clusters where adjusted deaths were higher or lower than expected based on population and adjusted total deaths. Three statistically significant hot spots and five statistically significant cold spots were identified. The cluster with the largest log-likelihood ratio was a southern hot spot region encompassing counties horizontally from eastern Texas to central Georgia and vertically from southern Louisiana to southern Kentucky (RR: 1.26; p<0.0001). Other notable clusters included hot spots centered in North Carolina (RR: 1.18; p<0.0001) and Ohio (RR: 1.18; p<0.0001), and large cold spots in western counties (Table 1). Our results reiterate southern and Appalachian hot spots from prior literature and provide new insights into a notable Ohio hot spot along with vast western cold spots. Future work is needed to identify established and potentially novel factors that may be driving yoCRC mortality clustering patterns, such as obesity, diet, or physician access. Table 1. Poisson spatial scan results adjusted for median age, sex, race, and social deprivation Gini cluster Classification States in cluster Relative risk Log-likelihood ratio Monte Carlo p-value 1 Hot spot Alabama, Arkansas, Florida, Georgia, Louisiana, Mississippi, Tennessee, Texas 1.28 139.03 p<0.0001 2 Hot spot Georgia, Kentucky, North Carolina, South Carolina, Tennessee, Virginia, West Virginia 1.18 63.31 p<0.0001 3 Cold spot Arizona, California, Colorado, Idaho, Nevada, New Mexico, Utah, Wyoming 0.83 61.95 p<0.0001 4 Hot spot Kentucky, Michigan, Ohio, Pennsylvania, West Virginia 1.16 40.14 p<0.0001 5 Cold spot California, Oregon, Washington 0.86 39.10 p<0.0001 6 Cold spot California 0.87 35.53 p<0.0001 7 Cold spot Texas 0.86 29.91 p<0.0001 8 Cold spot Colorado, Iowa, Kansas, Minnesota, Missouri, Montana, Nebraska, North Dakota, Wisconsin, Wyoming 0.90 19.56 p<0.0001 Citation Format: R. Blake Buchalter, Alok A. Khorana, Shimoli Barot, David Liska, Stephanie L. Schmit. Hot and cold spots of young-onset colorectal cancer mortality in U.S. counties, 1999-2019 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5907.