Diversifying preclinical research tools: expanding patient-derived models to address cancer health disparities

Abstract

Cancer health disparities define a critical healthcare issue for racial/ethnic minorities in the USA. Key findings have led to cancer treatment improvements tailored to minority patients, but such successes have been rare. Here, we highlight how the use of patient-derived xenograft (PDX) and organoid models could resolve current blocks toward precision cancer health equity.

Figure 1:. Lack of ethnic/racial representation in cancer genome databases and in preclinical model cohorts.

An example for the disparity-associated gastric cancer (Left) U.S. self-reported racial/ethnic population proportions from the 2020 census. Racial groups are represented in the top pie chart, ethnic groups are represented in the bottom pie chart. (Middle) Expected proportion of gastric cancer (GC) patients by race and ethnicity. Relative risk compared to NLWs was calculated for each racial/ethnic minority group from 2018 NCI SEER age-adjusted incidence and expected proportions for each racial/ethnic group among GC patients calculated based on increased incidence over NLWs. Racial groups are represented in the top pie chart, ethnic groups are represented in the bottom pie chart. (Right) Racial/ethnic group representation in TCGA and COSMIC pan cancer cohorts and gastric cancer subset cohort. TCGA self-reported race/ethnicity (patient samples collected globally dictate differences in racial categories compared to US census) and genetic ancestry from admixture analysis. COSMIC genetic ancestry from admixture analysis. Racial group abbreviations: W = White, B = Black, AA = African American, As = Asian, AsA = Asian American; NH/PI = Native Hawaiian/Pacific Islander, AI/AN = American Indian/Alaskan Native. All racial groups can include individuals who identify as Latino. Ethnic group abbreviations: NL = Not Latino. Genetic ancestry abbreviations: SA = South Asian, NLA = Native/Latin American, Ad/Unk = Admixed or Unknown.

Figure 2:. Cancer patient-derived mouse xenograft (PDX) and organoid (PDO) pre-clinical model generation and application to address genetic ancestry-associated cancer risk.

PDX and PDOs are derived from racial/ethnic minority tumor and normal biopsies and can be maintained and banked for further use. PDX/PDOs provide regenerative pre-clinical models for genomic analysis and molecular characterization, including subtype classification and driver alteration identification. Characterized models can be further used for therapeutic screening and drug sensitivity and resistance genetic alteration discovery, creating a biobank of deeply characterized minority patient-derived pre-clinical models for future cancer research. Altogether, model biobanks and genomic data generated from racial/ethnic minority cancer patients provide the missing data for inclusion in genetic ancestry-associated risk studies and inform minority patient-focused clinical trials to contribute to minority cancer patient precision medicine and address cancer health disparities.