Advancements in Data Science and Precision Medicine: Implications for Cancer Prevention, Treatment and Survivorship in Latinos

Share On Social!

This content is from the “Advancing the Science of Cancer in Latinos: 2024 Conference Proceedings.”

Integrating Clinical and Genomic Data for Tumor Molecular Profiling in Latino Patients with Cancer

Dr. Enrique I. Velazquez-Villarreal is Assistant Professor in the Department of Integrative Translational Sciences at Beckman Research Institute and the City of Hope Comprehensive Cancer Center.

Multi-omics and colorectal cancer in Latinos

Dr. Velazquez-Villarreal’s presentation discussed how multi-omics can be used to understand colorectal cancer disparities in Latino individuals and how global and local ancestry impacts these disparities. He also described clinical and genomic data integration in colorectal cancer disparities and recent technologies used to understand the tumor microenvironment. Multi-omics is a powerful tool in characterizing tumor cells in Latinos, and it works by combining structural genomics (whole exome sequencing, whole genome sequencing), functional genomics (RNA sequencing), epigenomics, and spatial transcriptomics. By integrating this information with clinical data and epidemiological findings, the dream of precision medicine may be realized. However, the amounts of data involved in these analyses may be overwhelming and may require innovative tools such as machine learning and artificial intelligence.

Dr. Velazquez-Villarreal and colleagues have recently used this aggregation of data to try to discover why Latinos have increased rates of early onset colorectal cancer. Global genomic ancestry analysis of 82 Latino individuals with colorectal cancer revealed significant population admixtures, with high proportions of Admixed American (70%) and European (48%) components, followed by South Asian (26.1%), East Asian (24.5%), and African (7.4%) components. This data was integrated with information about each patient’s ancestry, age, collection site, tumor content, gender, and more to generate a full multi-omic picture.

Some outliers were observed. One patient with mostly East Asian ancestry self-identified as having Spanish/Hispanic/Latino ethnicity and chose “other” to describe their race. Another patient identified as White racially, with a Cuban ethnicity, but had a largely African ancestry. These outliers underline the importance of multi-omics to provide a full picture of each individual.

The data gathered by Dr. Velazquez-Villarreal and colleagues revealed important patterns in colorectal cancer presentation. For example, patients diagnosed at an earlier age tended to have a higher Admixed American component. Patients with a larger Admixed American component had a higher risk of colorectal cancer compared with patients with higher European, East Asian, African, and South Asian components. Integrating genomic data with clinical data, including treatment type and survival time, has yet to reveal a discernable pattern, although Dr. Velazquez-Villarreal and colleagues are continuing this analysis.

New Technologies

One new technology that may be effective in understanding colorectal cancer is spatial transcriptomics. The spatial transcriptomics workflow begins by fixing a tissue sample on a slide, staining with hematoxylin and eosin, and imaging the sample. Through this process, the tissue is permeabilized, allowing the RNA to diffuse onto the slide surface and bind to spatially barcoded oligonucleotides. Copy DNA (cDNA) is then synthesized from the immobilized RNA, and this is used to generate sequencing libraries. In this way, the cancer component and immune component of each sample can be identified, as well as the detection of specific tumor suppressor genes like APC and TP53, and oncogenes such as KRAS. The interactions of each component of the tissue samples are assessed, ultimately revealing the story of the tissue sample. For example, immune cells can be seen surrounding the cancer cells, carrying out an attack on the rapidly reproducing cells.

Multi-omics approaches are transforming colorectal cancer disparities research in Latinos. Whole exome sequencing, in combination with global and local ancestry data, has already revealed important associations in colorectal cancer risk, and promises many more insights. Innovative technologies such as spatial transcriptomics are also promising in their characterization of tissue samples and the resulting understanding of the important interactions therein. The integration of clinical and genomic data has enabled us to uncover intricate molecular factors contributing to disparities in colorectal cancer among Latinos.

Breast Cancer Polygenic Risk Scores in Hispanic/Latinx/Latina Individuals

Dr. Laura Fejerman is Professor in the Department of Public Health Sciences at the University of California Davis Comprehensive Cancer Center.

Breast cancer risk scores

Dr. Fejerman began by emphasizing that although the focus of the presentation was on polygenic risk scores and advancing the science of breast cancer in Latinas, an equally important endeavor is community outreach and engagement, which is another aspect of Dr. Fejerman’s work. The main body of the presentation included a review of genome-wide association studies (GWAS) of breast cancer in Latinas, a discussion of polygenic risk scores (PRS) for breast cancer and their transferability from original studies to diverse populations, and the results of a study testing breast cancer PRS in US Latinas and Latin American individuals.

Breast cancer is the most common cancer in women worldwide. There is variation in breast cancer incidence in the US based on race/ethnicity, with Latina individuals having lower incidence than other groups. Reported incidence rates also vary by country in Latin America and within US Latinas by nationality. For example, genetic ancestry studies performed by Dr. Fejerman and colleagues have revealed that greater percentage of European ancestry is associated with higher risk of breast cancer.^,

Breast cancer genetic studies are informative not only for understanding cancer biology and improving drug development, but also for risk prediction, leading to improved prevention and early detection. Although certain genetic variants are associated with increased risk of breast cancer, most cases are not associated with high penetrance mutations. Instead, variants of high frequency have been identified, each of which only contribute minimally to risk, but which may be combined to form a PRS.^,,,, This PRS can then be used to inform patients how often they should be screened, and at what age screening should begin.

Currently, however, PRS used to assess risk in Europeans are only “satisfactory” in differentiating breast cancer cases and control cases, leaving much room for improvement. Furthermore, non-cancer PRS predictions used in Europeans have been shown to have limited usefulness in assessing Latina, Asian, and African individuals. Factors influencing PRS transferability include differences in heritability, differences in the effect sizes of causal variants, differences in the frequency of causal variants, differences in linkage disequilibrium (LD) patterns, and the fact that correlation between PRS and ancestry could lead to correlation between social determinants of health (SDOH) and the PRS.

A step forward in PRS for breast cancer in Latinas

Dr. Fejerman is part of a consortium for breast cancer in Latina women that is studying 5,700 cases and 12,750 controls from the US and Latin America. A range of ancestry is represented in these cases, with Peru showing the highest percentage of Indigenous American ancestry, and the US showing higher percentages of European ancestry. Latina women in this study showed similar allele frequency and odds ratios to women in European studies.

However, Dr. Fejerman and colleagues have discovered 2 polymorphisms that are associated with Latina women of Indigenous American ancestry only, called rs140068132 and rs3778609.^, When these polymorphisms are added to the PRS, they make a huge difference in the area under the curve, or efficacy, of the PRS. Specifically, the European-based PRS model is less effective when considering women from Peru, a country with high Indigenous American ancestry. However, when the 2 new polymorphisms are added, the PRS becomes much more effective. In fact, adding the 2 new polymorphisms increased the efficacy of the PRS across the board, even in Latinas with more European ancestry, indicating the usefulness of this approach in the Latina population.

Summary

The PRS used in European breast cancer studies can be useful in Latinas, but it is improved by adding Indigenous American ancestry polymorphisms. Furthermore, adding Indigenous American ancestry variants does not seem to affect prediction in highly European Latina individuals, indicating that additional Indigenous American ancestry variants are likely to improve the PRS.

One limiting factor is the dearth of patients included in the Latina GWAS studies when compared to the European studies. Because larger Latina GWAS are required, the Latin America Genetics and Genomics Breast Cancer Consortium (LAGENO-BC) has been established to share knowledge and materials to facilitate breast cancer genetics research relevant to the diverse populations of Latin America and its diaspora. With this collaboration and increase in participation, many more discoveries may be made in the field of breast cancer in Latina individuals.

Using Molecular Epidemiology for the Study of Helicobacter pylori, Premalignant and Malignant Gastric Lesions

Dr. Maria Constanza Camargo is Earl Stadtman Investigator in the Division of Cancer Epidemiology and Genetics at the National Cancer Institute (NCI).

Understanding gastric cancer

Dr. Camargo’s presentation began by discussing the incidence and mortality of gastric cancer, which is the leading cause of cancer mortality globally. North and South America contribute approximately 10% of global new cases of gastric cancer every year. While the highest mortality rates in the Americas are in Peru, Haiti, Costa Rica, Guatemala, Chile, Ecuador, and Colombia, the US is considered a low-risk country, with an age-standardized mortality rate of 3.3 per 100,000. However, this risk is heterogeneous across racial and ethnic populations. While non-Hispanic White patients have relatively low gastric cancer mortality rates, Black, Asian, and Latino mortality is much higher.

Gastric cancer may be categorized by anatomical location of the tumor. Cardia cancer, located at the proximal end of the stomach, makes up approximately 18% of all gastric cancer. Noncardia cancers, which may be located in the fundus, corpus, greater or lesser curvature, antrum, or pylorus, make up approximately 82% of all gastric cancers, and are the most common subsites in Latino populations. Noncardia cancers are also strongly associated with Helicobacter pylori (H. pylori), with at least 90% of noncardia cancers linked with this infection.

According to the model of noncardia gastric carcinogenesis developed by Dr. Pelayo Correa, H. pylori infection first leads to benign non-atrophic gastritis. In about 10% of these cases, multifocal atrophy is observed, with or without intestinal metaplasia, which may progress to dysplasia or cancer. The progression from atrophy to cancer takes about 20 years, however, tools to intervene in that window are currently insufficient.

One potential tool showing promising results is machine learning-based prediction of gastric cancer incidence using endoscopic and histologic features. According to some international guidelines, individuals with intestinal metaplasia should be assessed every two or three years for prevention of invasive cancer development. Unfortunately, in the Americas, such screenings are not used, and this international recommendation is not followed, mainly due to the lack of local data.

Dr. Camargo’s work

Dr. Camargo and colleagues recently evaluated the proteome of H. pylori, and identified the humoral response antibodies against H. pylori in individuals with intestinal metaplasia. Although some anti-H. pylori antibodies have been validated as discriminatory markers, including those against the outer membrane protein 27 (Omp27) and the cytotoxicity-associated immunodominant antigen (CagA), a combination of multiple markers is needed for risk stratification.

As part of the H. pylori Genome Project (HpGP), Dr. Camargo and over 200 colleagues collected 1011 well-characterized clinical strains from 50 countries and generated high-quality genome and epigenome sequences. The publicly available HpGP dataset and the corresponding strains are a major asset for studies of H. pylori pathogenesis.

A genome-wide association study of gastric cancer was also recently conducted by Dr. Camargo and colleagues in Latino populations in Brazil, Chile, Colombia, Guatemala, Honduras, Mexico, Peru, and the US. Approximately 3,500 cases and 4,500 controls with gastric cancer were collected. Ongoing work is underway to develop a polygenic risk score. Dr. Camargo’s group are also working to develop a multi-component biomarker approach for non-invasive diagnosis of intestinal metaplasia by combining the following: polygenic risk score, anti-H. pylori antibodies, atrophy-related markers, inflammation-related markers, microbiome, proteomics, metabolomics, and liquid biopsy.

Expanding their descriptive studies, Dr. Camargo and colleagues have also looked at recent trends in gastric cancer incidence, observing increases in young individuals, especially among females. Data from the Surveillance, Epidemiology, and End Results (SEER) program was used to show that noncardia gastric cancer rates have increased by 2.68% in non-Hispanic White women, by 1.63% in Latina women, and by 1.47% in non-Hispanic Black women. These findings were even more stark when looking at the corpus specifically, with increases of 6.57%, 4.83%, and 5.88% respectively.

Dr. Camargo ended by outlining a take home message. Gastric cancer is and will continue to be one of the top health problems for several populations in the Americas. A persistently high prevalence of H. pylori infection is present in some populations, and a rising incidence of noncardia cancer has been observed among younger individuals. Further research into the causes of gastric cancer and its evolving nature is warranted. Continued and targeted prevention efforts are needed to reduce the burden of gastric cancer, and there is an urgent need to identify discriminatory biomarkers to inform diagnosis, prognosis, and treatment.