Comprehensive Analysis of Genetic Ancestry and Its Molecular Correlates in Cancer

AACR Cancer Disparities Progress Report 2024: Achieving the Bold Vision of Health Equity

Advances in cancer prevention, early detection, and treatments have led to unprecedented progress against cancer. However, these advances have not benefited everyone equally. Because of a long history of structural inequities and systemic injustices in the United States, many segments of the US population continue to shoulder a disproportionate burden of cancer. The American Association for Cancer Research (AACR) Cancer Disparities Progress Report 2024 (CancerDisparitiesProgressReport.org) outlines the recent progress against cancer disparities, the ongoing challenges faced by medically underserved populations, and emphasizes the vital need for further advances in cancer research and patient care to benefit all populations.

Genetic ancestry is related to potential sources of breast cancer health disparities among Colombian women

Breast cancer health disparities are linked to clinical-pathological determinants, socioeconomic inequities, and biological factors such as genetic ancestry. These factors collectively interact in complex ways, influencing disease behavior, especially among highly admixed populations like Colombians. In this study, we assessed contributing factors to breast cancer health disparities according to genetic ancestry in Colombian patients from a national cancer reference center. We collected non-tumoral paraffin embedded (FFPE) blocks from 361 women diagnosed with breast cancer at the National Cancer Institute (NCI) to estimate genetic ancestry using a 106-ancestry informative marker (AIM) panel. Differences in European, Indigenous American (IA) and African ancestry fractions were analyzed according to potential sources of breast cancer health disparities, like etiology, tumor-biology, treatment administration, and socioeconomic-related factors using a Kruskal-Wallis test. Our analysis revealed a significantly higher IA ancestry among overweight patients with larger tumors and those covered by a subsidized health insurance. Conversely, we found a significantly higher European ancestry among patients with smaller tumors, residing in middle-income households, and affiliated to the contributory health regime, whereas a higher median of African ancestry was observed among patients with either a clinical, pathological, or stable response to neoadjuvant treatment. Altogether, our results suggest that the genetic legacy among Colombian patients, measured as genetic ancestry fractions, may be reflected in many of the clinical-pathological variables and socioeconomic factors that end up contributing to health disparities for this disease.

Comparison of germline and somatic structural variants in cancers reveal systematic differences in variant generating and selection processes

Although several recent studies have characterized structural variants (SVs) in germline and cancer genomes, the features of SVs in these different contexts have not been directly compared. We examined similarities and differences between 2 million germline and 115 thousand tumor SVs from a cohort of 963 patients from The Cancer Genome Atlas (TCGA). We found significant differences in features related to their genomic sequences and localization that suggest differences between SV-generating processes and selective pressures. For example, we found that transposon-mediated processes shape germline much more than somatic SVs, while somatic SVs more frequently show features characteristic of chromoanagenesis. These differences were extensive enough to enable us to develop a classifier - "the great GaTSV" - that accurately distinguishes between germline and cancer SVs in tumor samples that lack a matched normal sample.

A hybrid epithelial/mesenchymal state in head and neck cancer: A biomarker for survival with differential prognosis by self-reported race

BACKGROUND

Head and neck squamous cell carcinoma (HNSCC) is the 6th leading cause of cancer-related mortality, with racial disparities amplifying the challenges in treatment. Although the relationship between hybrid epithelial/mesenchymal (E/M) states and tumor progression is of interest, no studies have characterized the clinical relevance of hybrid E/M states in head and neck cancer outcomes among self-reported racial cohorts.

METHODS

Given the overlap in gene expression between hybrid E/M malignant cells and cancer-associated fibroblasts, we utilized deconvolution of bulk RNA sequencing data from oral cavity and laryngeal squamous cell carcinoma tumors from The Cancer Genome Atlas. We utilized our previously collected single-cell profiles to generate inferred malignant profiles and then scored these for hybrid E/M. We then conducted a survival analysis on overall and disease-free survival among self-reported Black and White Americans.

FINDINGS

The hybrid E/M state was differentially associated with head and neck cancer survival by self-reported race and ethnicity, with a stronger association in non-Hispanic Black patients. Black patients with a high hybrid E/M score had a higher risk of death or recurrence (hazard ratio [HR]: 4.18 [95% confidence interval (CI): 2.06, 8.49]) than White patients with a high hybrid E/M score (HR: 1.58 [95% CI: 1.11, 2.26]).

CONCLUSION

Our results suggest a complex interplay of social structure, racism, and genetic diversity. We implore researchers to consider the social and biological context contributing to disparities.

FUNDING

A.L.M. received support from the National Institute of Minority Health and Health Disparities (K01MD013897 [principal investigator (PI), A.L.M.]). S.V.P. received support from the National Institute of Dental and Craniofacial Research (R01DE032865 [PI, S.V.P.] and R01DE032371 [PI, S.V.P.]).

Germline variation contributes to false negatives in CRISPR-based experiments with varying burden across ancestries

Reducing disparities is vital for equitable access to precision treatments in cancer. Socioenvironmental factors are a major driver of disparities, but differences in genetic variation likely also contribute. The impact of genetic ancestry on prioritization of cancer targets in drug discovery pipelines has not been systematically explored due to the absence of pre-clinical data at the appropriate scale. Here, we analyze data from 611 genome-scale CRISPR/Cas9 viability experiments in human cell line models to identify ancestry-associated genetic dependencies essential for cell survival. Surprisingly, we find that most putative associations between ancestry and dependency arise from artifacts related to germline variants. Our analysis suggests that for 1.2-2.5% of guides, germline variants in sgRNA targeting sequences reduce cutting by the CRISPR/Cas9 nuclease, disproportionately affecting cell models derived from individuals of recent African descent. We propose three approaches to mitigate this experimental bias, enabling the scientific community to address these disparities.

Cis- and trans-eQTL TWASs of breast and ovarian cancer identify more than 100 susceptibility genes in the BCAC and OCAC consortia

Transcriptome-wide association studies (TWASs) have investigated the role of genetically regulated transcriptional activity in the etiologies of breast and ovarian cancer. However, methods performed to date have focused on the regulatory effects of risk-associated SNPs thought to act in cis on a nearby target gene. With growing evidence for distal (trans) regulatory effects of variants on gene expression, we performed TWASs of breast and ovarian cancer using a Bayesian genome-wide TWAS method (BGW-TWAS) that considers effects of both cis- and trans-expression quantitative trait loci (eQTLs). We applied BGW-TWAS to whole-genome and RNA sequencing data in breast and ovarian tissues from the Genotype-Tissue Expression project to train expression imputation models. We applied these models to large-scale GWAS summary statistic data from the Breast Cancer and Ovarian Cancer Association Consortia to identify genes associated with risk of overall breast cancer, non-mucinous epithelial ovarian cancer, and 10 cancer subtypes. We identified 101 genes significantly associated with risk with breast cancer phenotypes and 8 with ovarian phenotypes. These loci include established risk genes and several novel candidate risk loci, such as ACAP3, whose associations are predominantly driven by trans-eQTLs. We replicated several associations using summary statistics from an independent GWAS of these cancer phenotypes. We further used genotype and expression data in normal and tumor breast tissue from the Cancer Genome Atlas to examine the performance of our trained expression imputation models. This work represents an in-depth look into the role of trans eQTLs in the complex molecular mechanisms underlying these diseases.

Multiomic analysis of uterine leiomyomas in self-described Black and White women: molecular insights into health disparities

BACKGROUND

Black women are at an increased risk of developing uterine leiomyomas and experiencing worse disease prognosis than White women. Epidemiologic and molecular factors have been identified as underlying these disparities, but there remains a paucity of deep, multiomic analysis investigating molecular differences in uterine leiomyomas from Black and White patients.

OBJECTIVE

To identify molecular alterations within uterine leiomyoma tissues correlating with patient race by multiomic analyses of uterine leiomyomas collected from cohorts of Black and White women.

STUDY DESIGN

We performed multiomic analysis of uterine leiomyomas from Black (42) and White (47) women undergoing hysterectomy for symptomatic uterine leiomyomata. In addition, our analysis included the application of orthogonal methods to evaluate fibroid biomechanical properties, such as second harmonic generation microscopy, uniaxial compression testing, and shear-wave ultrasonography analyses.

RESULTS

We found a greater proportion of MED12 mutant uterine leiomyomas from Black women (>35% increase; Mann-Whitney U, P<.001). MED12 mutant tumors exhibited an elevated abundance of extracellular matrix proteins, including several collagen isoforms, involved in the regulation of the core matrisome. Histologic analysis of tissue fibrosis using trichrome staining and secondary harmonic generation microscopy confirmed that MED12 mutant tumors are more fibrotic than MED12 wild-type tumors. Using shear-wave ultrasonography in a prospectively collected cohort, Black patients had fibroids that were firmer than White patients, even when similar in size. In addition, these analyses uncovered ancestry-linked expression quantitative trait loci with altered allele frequencies in African and European populations correlating with differential abundance of several proteins in uterine leiomyomas independently of MED12 mutation status, including tetracoidpeptide repeat protein 38.

CONCLUSION

Our study shows that Black women have a higher prevalence of uterine leiomyomas harboring mutations in MED12 and that this mutational status correlates with increased tissue fibrosis compared with wild-type uterine leiomyomas. Our study provides insights into molecular alterations correlating with racial disparities in uterine leiomyomas and improves our understanding of the molecular etiology underlying uterine leiomyoma development within these populations.

Pan-Cancer Comparative and Integrative Analyses of Driver Alterations Using Japanese and International Genomic Databases

Using 48,627 samples from the Center for Cancer Genomics and Advanced Therapeutics (C-CAT), we present a pan-cancer landscape of driver alterations and their clinical actionability in Japanese patients. Comparison with White patients in Genomics Evidence Neoplasia Information Exchange (GENIE) demonstrates high TP53 mutation frequencies in Asian patients across multiple cancer types. Integration of C-CAT, GENIE, and The Cancer Genome Atlas data reveals many cooccurring and mutually exclusive relationships between driver mutations. At pathway level, mutations in epigenetic regulators frequently cooccur with PI3K pathway molecules. Furthermore, we found significant cooccurring mutations within the epigenetic pathway. Accumulation of mutations in epigenetic regulators causes increased proliferation-related transcriptomic signatures. Loss-of-function of many epigenetic drivers inhibits cell proliferation in their wild-type cell lines, but this effect is attenuated in those harboring mutations of not only the same but also different epigenetic drivers. Our analyses dissect various genetic properties and provide valuable resources for precision medicine in cancer.

SIGNIFICANCE

We present a genetic landscape of 26 principal cancer types/subtypes, including Asian-prevalent ones, in Japanese patients. Multicohort data integration unveils numerous cooccurring and exclusive relationships between driver mutations, identifying cooccurrence of multiple mutations in epigenetic regulators, which coordinately cause transcriptional and phenotypic changes. These findings provide insights into epigenetic regulator-driven oncogenesis. This article is featured in Selected Articles from This Issue, p. 695.

KRAS Allelic Variants in Biliary Tract Cancers

Importance

Biliary tract cancers (BTCs) contain several actionable molecular alterations, including FGFR2, IDH1, ERBB2 (formerly HER2), and KRAS. KRAS allelic variants are found in 20% to 30% of BTCs, and multiple KRAS inhibitors are currently under clinical investigation.

Objectives

To describe the genomic landscape, co-sequence variations, immunophenotype, genomic ancestry, and survival outcomes of KRAS-mutated BTCs and to calculate the median overall survival (mOS) for the most common allelic variants.

Design, Setting, and Participants

This retrospective, multicenter, pooled cohort study obtained clinical and next-generation sequencing data from multiple databases between January 1, 2017, and December 31, 2022. These databases included Princess Margaret Cancer Centre, MD Anderson Cancer Center, Foundation Medicine, American Association for Cancer Research Project GENIE, and cBioPortal for Cancer Genomics. The cohort comprised patients with BTCs who underwent genomic testing.

Main Outcome and Measure

The main outcome was mOS, defined as date of diagnosis to date of death, which was measured in months.

Results

A total of 7457 patients (n = 3773 males [50.6%]; mean [SD] age, 63 [5] years) with BTCs and genomic testing were included. Of these patients, 5813 had clinical outcome data available, in whom 1000 KRAS-mutated BTCs were identified. KRAS allelic variants were highly prevalent in perihilar cholangiocarcinoma (28.6%) and extrahepatic cholangiocarcinoma (36.1%). Thirty-six KRAS allelic variants were identified, and the prevalence rates in descending order were G12D (41%), G12V (23%), and Q61H (8%). The variant G12D had the highest mOS of 25.1 (95% CI, 22.0-33.0) months compared with 22.8 (95% CI, 19.6-31.4) months for Q61H and 17.8 (95% CI, 16.3-23.1) months for G12V variants. The majority of KRAS-mutated BTCs (98.9%) were not microsatellite instability-high and had low tumor mutational burden (ranging from a median [IQR] of 1.2 (1.2-2.5) to a mean [SD] of 3.3 [1.3]). Immune profiling through RNA sequencing of KRAS and NRAS-mutated samples showed a pattern toward a more immune-inflamed microenvironment with higher M1 macrophage activation (0.16 vs 0.12; P = .047) and interferon-γ expression compared with wild-type tumors. The G12D variant remained the most common KRAS allelic variant in all patient ancestries. Patients with admixed American ancestry had the highest proportion of G12D variant (45.0%).

Conclusions and Relevance

This cohort study found that KRAS allelic variants were relatively common and may be potentially actionable genomic alterations in patients with BTCs, especially perihilar cholangiocarcinoma and extrahepatic cholangiocarcinoma. The findings add to the growing data on genomic and immune landscapes of KRAS allelic variants in BTCs and are potentially of value to the planning of specific therapies for this heterogeneous patient group.

Genomic ancestry and cancer among Latin Americans

Latin American populations, characterized by intricate admixture patterns resulting from the intermingling of ancestries from European, Native American (NA) Asian, and African ancestries which result in a vast and complex genetic landscape, harboring unique combinations of novel variants. This genetic diversity not only poses challenges in traditional population genetics methods but also opens avenues for a deeper understanding of its implications in health. In cancer, the interplay between genetic ancestry, lifestyle factors, and healthcare disparities adds a layer of complexity to the varying incidence and mortality rates observed across different Latin American subpopulations. This complex interdependence has been unveiled through numerous studies, whether conducted on Latin American patients residing on the continent or abroad, revealing discernible differences in germline composition that influence divergent disease phenotypes such as higher incidence of Luminal B and Her2 breast tumors, EGFR and KRAS mutated lung adenocarcinomas in addition to an enrichment in BRCA1/2 pathogenic variants and a higher than expected prevalence of variants in colorectal cancer associated genes such as APC and MLH1. In prostate cancer novel risk variants have also been solely identified in Latin American populations. Due to the complexity of genetic divergence, inputs from each individual ancestry seem to carry independent contributions that interplay in the development of these complex disease phenotypes. By understanding these unique population characteristics, genomic ancestries hold a promising avenue for tailoring prognostic assessments and optimizing responses to oncological interventions.

CHEK2 Founder Variants and Thyroid Cancer Risk

Background: Germline pathogenic variants in CHEK2 are associated with a moderate increase in the lifetime risk for breast cancer. Increased risk for other cancers, including non-medullary thyroid cancer (NMTC), has also been suggested. To date, data implicating CHEK2 variants in NMTC predisposition primarily derive from studies within Poland, driven by a splice site variant (c.444 + 1G>A) that is uncommon in other populations. In contrast, the predominant CHEK2 variants in non-Polish populations are c.1100del and c.470T>C/p.I157T, representing 61.1% and 63.8%, respectively, of all CHEK2 pathogenic variants in two large U.S.-based commercial laboratory datasets. To further delineate the impact of common CHEK2 variants on thyroid cancer, we aimed to investigate the association of three CHEK2 founder variants (c.444 + 1G>A, c.1100del, and c.470T>C/p.Ile157Thr) on NMTC susceptibility in three groups of unselected NMTC patients. Methods: The presence of three CHEK2 founder variants was assessed within three groups: (1) 1544 NMTC patients (and 1593 controls) from previously published genome-wide association study (GWAS) analyses, (2) 789 NMTC patients with germline exome sequencing (Oncology Research Information Exchange Network [ORIEN] Avatar), and (3) 499 NMTC patients with germline sequence data available in The Cancer Genome Atlas (TCGA). A case-control study design was utilized with odds ratios (ORs) calculated by comparison of all three groups with the Ohio State University GWAS control group. Results: The predominant Polish variant (c.444 + 1G>A) was present in only one case. The proportion of patients with c.1100del was 0.92% in the GWAS group, 1.65% in the ORIEN Avatar group, and 0.80% in the TCGA group. The ORs (with 95% confidence intervals [CIs]) for NMTC associated with c.1100del were 1.71 (0.73-4.29), 2.64 (0.95-7.63), and 2.5 (0.63-8.46), respectively. The proportion of patients with c.470T>C/p.I157T was 0.91% in the GWAS group, 0.76% in the ORIEN Avatar group, and 0.80% in the TCGA group, respectively. The ORs (with CIs) for NMTC associated with c.470T>C/p.I157T were 1.75 (0.74-4.39), 1.52 (0.42-4.96), and 2.31 (0.58-7.90), respectively. Conclusions: Our analyses of unselected patients with NMTC suggest that CHEK2 variants c.1100del and c.470T>C/p.I157T have only a modest impact on thyroid cancer risk. These results provide important information for providers regarding the relatively low magnitude of thyroid cancer risk associated with these CHEK2 variants.

Generative Adversarial Networks Accurately Reconstruct Pan-Cancer Histology from Pathologic, Genomic, and Radiographic Latent Features

Artificial intelligence models have been increasingly used in the analysis of tumor histology to perform tasks ranging from routine classification to identification of novel molecular features. These approaches distill cancer histologic images into high-level features which are used in predictions, but understanding the biologic meaning of such features remains challenging. We present and validate a custom generative adversarial network - HistoXGAN - capable of reconstructing representative histology using feature vectors produced by common feature extractors. We evaluate HistoXGAN across 29 cancer subtypes and demonstrate that reconstructed images retain information regarding tumor grade, histologic subtype, and gene expression patterns. We leverage HistoXGAN to illustrate the underlying histologic features for deep learning models for actionable mutations, identify model reliance on histologic batch effect in predictions, and demonstrate accurate reconstruction of tumor histology from radiographic imaging for a 'virtual biopsy'.

Genomic ancestry in kidney cancer: Correlations with clinical and molecular features

INTRODUCTION

Genetic ancestry (GA) refers to population hereditary patterns that contribute to phenotypic differences seen among race/ethnicity groups, and differences among GA groups may highlight unique biological determinants that add to our understanding of health care disparities.

METHODS

A retrospective review of patients with renal cell carcinoma (RCC) was performed and correlated GA with clinicopathologic, somatic, and germline molecular data. All patients underwent next-generation sequencing of normal and tumor DNA using Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets, and contribution of African (AFR), East Asian (EAS), European (EUR), Native American, and South Asian (SAS) ancestry was inferred through supervised ADMIXTURE. Molecular data was compared across GA groups by Fisher exact test and Kruskal-Wallis test.

RESULTS

In 953 patients with RCC, the GA distribution was: EUR (78%), AFR (4.9%), EAS (2.5%), SAS (2%), Native American (0.2%), and Admixed (12.2%). GA distribution varied by tumor histology and international metastatic RCC database consortium disease risk status (intermediate-poor: EUR 58%, AFR 88%, EAS 74%, and SAS 73%). Pathogenic/likely pathogenic germline variants in cancer-predisposition genes varied (16% EUR, 23% AFR, 8% EAS, and 0% SAS), and most occurred in CHEK2 in EUR (3.1%) and FH in AFR (15.4%). In patients with clear cell RCC, somatic alteration incidence varied with significant enrichment in BAP1 alterations (EUR 17%, AFR 50%, SAS 29%; p = .01). Comparing AFR and EUR groups within The Cancer Genome Atlas, significant differences were identified in angiogenesis and inflammatory pathways.

CONCLUSION

Differences in clinical and molecular data by GA highlight population-specific variations in patients with RCC. Exploration of both genetic and nongenetic variables remains critical to optimize efforts to overcome health-related disparities.

Germline determinants of aberrant signaling pathways in cancer

Cancer is a complex disease influenced by a heterogeneous landscape of both germline genetic variants and somatic aberrations. While there is growing evidence suggesting an interplay between germline and somatic variants, and a substantial number of somatic aberrations in specific pathways are now recognized as hallmarks in many well-known forms of cancer, the interaction landscape between germline variants and the aberration of those pathways in cancer remains largely unexplored. Utilizing over 8500 human samples across 33 cancer types characterized by TCGA and considering binary traits defined using a large collection of somatic aberration profiles across ten well-known oncogenic signaling pathways, we conducted a series of GWAS and identified genome-wide and suggestive associations involving 276 SNPs. Among these, 94 SNPs revealed cis-eQTL links with cancer-related genes or with genes functionally correlated with the corresponding traits' oncogenic pathways. GWAS summary statistics for all tested traits were then used to construct a set of polygenic scores employing a customized computational strategy. Polygenic scores for 24 traits demonstrated significant performance and were validated using data from PCAWG and CCLE datasets. These scores showed prognostic value for clinical variables and exhibited significant effectiveness in classifying patients into specific cancer subtypes or stratifying patients with cancer-specific aggressive phenotypes. Overall, we demonstrate that germline genetics can describe patients' genetic liability to develop specific cancer molecular and clinical profiles.

An immunogenetic basis for lung cancer risk

Cancer risk is influenced by inherited mutations, DNA replication errors, and environmental factors. However, the influence of genetic variation in immunosurveillance on cancer risk is not well understood. Leveraging population-level data from the UK Biobank and FinnGen, we show that heterozygosity at the human leukocyte antigen (HLA)-II loci is associated with reduced lung cancer risk in smokers. Fine-mapping implicated amino acid heterozygosity in the HLA-II peptide binding groove in reduced lung cancer risk, and single-cell analyses showed that smoking drives enrichment of proinflammatory lung macrophages and HLA-II+ epithelial cells. In lung cancer, widespread loss of HLA-II heterozygosity (LOH) favored loss of alleles with larger neopeptide repertoires. Thus, our findings nominate genetic variation in immunosurveillance as a critical risk factor for lung cancer.

Novel molecular hepatocellular carcinoma subtypes and RiskScore utilizing apoptosis-related genes

Hepatocellular carcinoma (HCC) is the third leading cause of global cancer-related deaths. Despite immunotherapy offering hope for patients with HCC, only some respond to it. However, it remains unclear how to pre-screen eligible patients. Our study aimed to address this issue. In this study, we identified 13 prognostic genes through univariate Cox regression analysis of 87 apoptosis-related genes. Subsequently, these 13 genes were analyzed using ConsensusClusterPlus, and patients were categorized into three molecular types: C1, C2, and C3. A prognostic model and RiskScore were constructed using Lasso regression analysis of 132 significant genes identified between C1 and C3. We utilized quantitative polymerase chain reaction to confirm the model's transcript level in Huh7 and THLE2 cell lines. Both molecular subtypes and RiskScores effectively predicted patients benefiting from immunotherapy. Cox regression analysis revealed RiskScore as the most significant prognosis factor, suggesting its clinical application potential and providing a foundation for future experimental research.

A comprehensive analysis of clinical and polygenic germline influences on somatic mutational burden

Tumor mutational burden (TMB), the total number of somatic mutations in the tumor, and copy number burden (CNB), the corresponding measure of aneuploidy, are established fundamental somatic features and emerging biomarkers for immunotherapy. However, the genetic and non-genetic influences on TMB/CNB and, critically, the manner by which they influence patient outcomes remain poorly understood. Here, we present a large germline-somatic study of TMB/CNB with >23,000 individuals across 17 cancer types, of which 12,000 also have extensive clinical, treatment, and overall survival (OS) measurements available. We report dozens of clinical associations with TMB/CNB, observing older age and male sex to have a strong effect on TMB and weaker impact on CNB. We additionally identified significant germline influences on TMB/CNB, including fine-scale European ancestry and germline polygenic risk scores (PRSs) for smoking, tanning, white blood cell counts, and educational attainment. We quantify the causal effect of exposures on somatic mutational processes using Mendelian randomization. Many of the identified features associated with TMB/CNB were additionally associated with OS for individuals treated at a single tertiary cancer center. For individuals receiving immunotherapy, we observed a complex relationship between PRSs for educational attainment, self-reported college attainment, TMB, and survival, suggesting that the influence of this biomarker may be substantially modified by socioeconomic status. While the accumulation of somatic alterations is a stochastic process, our work demonstrates that it can be shaped by host characteristics including germline genetics.

A Multiethnic Germline-Somatic Association Database Deciphers Multilayered and Interconnected Genetic Mutations in Cancer

Inherited germline and acquired somatic alterations can both promote human tumor development. Elucidating the cooperation between somatic and germline genetic alterations that drive tumorigenesis could help inform precision cancer prevention and treatment strategies. Here, leveraging genomic genotyping and sequencing data from 9,029 patients with cancer with European, East Asian, and African ancestry, we performed a pan-cancer analysis to evaluate the associations between germline SNPs and somatic alterations, including single-nucleotide variant and small insertion/deletion mutations, copy-number variation, tumor mutational burden, and mutational signatures. Genome-wide significant germline-somatic pairs were abundant, and most of the associations were observed in one cancer type and one ancestry group. A user-friendly interactive Multiethnic Germline-Somatic Association (MGSA) database (http://wanglab-hkust.cn:3838/MGSA/) was developed, which can be used to query, browse, and download the results of the association analyses. Moreover, the MGSA database offers additional survival analysis and functional annotation. Together, this work provides a resource for uncovering the clinical and biological roles of associations between germline variants and somatic alterations in human cancer.

SIGNIFICANCE

Comprehensive analysis of connections between germline variants and somatic events in cancer offers a resource for investigating the functional significance of genetic mutations and exploring genetic factors contributing to racial disparities.

Divergent HLA variations and heterogeneous expression but recurrent HLA loss-of- heterozygosity and common HLA-B and TAP transcriptional silencing across advanced pediatric solid cancers

The human leukocyte antigen (HLA) system is a major factor controlling cancer immunosurveillance and response to immunotherapy, yet its status in pediatric cancers remains fragmentary. We determined high-confidence HLA genotypes in 576 children, adolescents and young adults with recurrent/refractory solid tumors from the MOSCATO-01 and MAPPYACTS trials, using normal and tumor whole exome and RNA sequencing data and benchmarked algorithms. There was no evidence for narrowed HLA allelic diversity but discordant homozygosity and allele frequencies across tumor types and subtypes, such as in embryonal and alveolar rhabdomyosarcoma, neuroblastoma MYCN and 11q subtypes, and high-grade glioma, and several alleles may represent protective or susceptibility factors to specific pediatric solid cancers. There was a paucity of somatic mutations in HLA and antigen processing and presentation (APP) genes in most tumors, except in cases with mismatch repair deficiency or genetic instability. The prevalence of loss-of-heterozygosity (LOH) ranged from 5.9 to 7.7% in HLA class I and 8.0 to 16.7% in HLA class II genes, but was widely increased in osteosarcoma and glioblastoma (~15-25%), and for DRB1-DQA1-DQB1 in Ewing sarcoma (~23-28%) and low-grade glioma (~33-50%). HLA class I and HLA-DR antigen expression was assessed in 194 tumors and 44 patient-derived xenografts (PDXs) by immunochemistry, and class I and APP transcript levels quantified in PDXs by RT-qPCR. We confirmed that HLA class I antigen expression is heterogeneous in advanced pediatric solid tumors, with class I loss commonly associated with the transcriptional downregulation of HLA-B and transporter associated with antigen processing (TAP) genes, whereas class II antigen expression is scarce on tumor cells and occurs on immune infiltrating cells. Patients with tumors expressing sufficient HLA class I and TAP levels such as some glioma, osteosarcoma, Ewing sarcoma and non-rhabdomyosarcoma soft-tissue sarcoma cases may more likely benefit from T cell-based approaches, whereas strategies to upregulate HLA expression, to expand the immunopeptidome, and to target TAP-independent epitopes or possibly LOH might provide novel therapeutic opportunities in others. The consequences of HLA class II expression by immune cells remain to be established. Immunogenetic profiling should be implemented in routine to inform immunotherapy trials for precision medicine of pediatric cancers.

Tumor sequencing of African ancestry reveals differences in clinically relevant alterations across common cancers

Cancer genomes from patients with African (AFR) ancestry have been poorly studied in clinical research. We leverage two large genomic cohorts to investigate the relationship between genomic alterations and AFR ancestry in six common cancers. Cross-cancer type associations, such as an enrichment of MYC amplification with AFR ancestry in lung, breast, and prostate cancers, and depletion of BRAF alterations are observed in colorectal and pancreatic cancers. There are differences in actionable alterations, such as depletion of KRAS G12C and EGFR L858R, and enrichment of ROS1 fusion with AFR ancestry in lung cancers. Interestingly, in lung cancer, KRAS mutations are less common in both smokers and non-smokers with AFR ancestry, whereas the association of TP53 mutations with AFR ancestry is only seen in smokers, suggesting an ancestry-environment interaction that modifies driver rates. Our study highlights the need to increase representation of patients with AFR ancestry in drug development and biomarker discovery.

Cis- and trans-eQTL TWAS of breast and ovarian cancer identify more than 100 risk associated genes in the BCAC and OCAC consortia

Transcriptome-wide association studies (TWAS) have investigated the role of genetically regulated transcriptional activity in the etiologies of breast and ovarian cancer. However, methods performed to date have only considered regulatory effects of risk associated SNPs thought to act in cis on a nearby target gene. With growing evidence for distal (trans) regulatory effects of variants on gene expression, we performed TWAS of breast and ovarian cancer using a Bayesian genome-wide TWAS method (BGW-TWAS) that considers effects of both cis- and trans-expression quantitative trait loci (eQTLs). We applied BGW-TWAS to whole genome and RNA sequencing data in breast and ovarian tissues from the Genotype-Tissue Expression project to train expression imputation models. We applied these models to large-scale GWAS summary statistic data from the Breast Cancer and Ovarian Cancer Association Consortia to identify genes associated with risk of overall breast cancer, non-mucinous epithelial ovarian cancer, and 10 cancer subtypes. We identified 101 genes significantly associated with risk with breast cancer phenotypes and 8 with ovarian phenotypes. These loci include established risk genes and several novel candidate risk loci, such as ACAP3, whose associations are predominantly driven by trans-eQTLs. We replicated several associations using summary statistics from an independent GWAS of these cancer phenotypes. We further used genotype and expression data in normal and tumor breast tissue from the Cancer Genome Atlas to examine the performance of our trained expression imputation models. This work represents a first look into the role of trans-eQTLs in the complex molecular mechanisms underlying these diseases.

Population-specific Mutation Patterns in Breast Tumors from African American, European American, and Kenyan Patients

Women of African descent have the highest breast cancer mortality in the United States and are more likely than women from other population groups to develop an aggressive disease. It remains uncertain to what extent breast cancer in Africa is reminiscent of breast cancer in African American or European American patients. Here, we performed whole-exome sequencing of genomic DNA from 191 breast tumor and non-cancerous adjacent tissue pairs obtained from 97 African American, 69 European American, 2 Asian American, and 23 Kenyan patients. Our analysis of the sequencing data revealed an elevated tumor mutational burden in both Kenyan and African American patients, when compared with European American patients. TP53 mutations were most prevalent, particularly in African American patients, followed by PIK3CA mutations, which showed similar frequencies in European American, African American, and the Kenyan patients. Mutations targeting TBX3 were confined to European Americans and those targeting the FBXW7 tumor suppressor to African American patients whereas mutations in the ARID1A gene that are known to confer resistance to endocrine therapy were distinctively enriched among Kenyan patients. A Kyoto Encyclopedia of Genes and Genomes pathway analysis could link FBXW7 mutations to an increased mitochondrial oxidative phosphorylation capacity in tumors carrying these mutations. Finally, Catalogue of Somatic Mutations in Cancer (COSMIC) mutational signatures in tumors correlated with the occurrence of driver mutations, immune cell profiles, and neighborhood deprivation with associations ranging from being mostly modest to occasionally robust. To conclude, we found mutational profiles that were different between these patient groups. The differences concentrated among genes with low mutation frequencies in breast cancer.

SIGNIFICANCE

The study describes differences in tumor mutational profiles between African American, European American, and Kenyan breast cancer patients. It also investigates how these profiles may relate to the tumor immune environment and the neighborhood environment in which the patients had residence. Finally, it describes an overrepresentation of ARID1A gene mutations in breast tumors of the Kenyan patients.

Molecular Characterization of Endometrial Carcinomas in Black and White Patients Reveals Disparate Drivers with Therapeutic Implications

Although the incidence of endometrial carcinoma (EC) is similar in Black and White women, racial disparities are stark, with the highest mortality rates observed among Black patients. Here, analysis of 1,882 prospectively sequenced ECs using a clinical FDA-authorized tumor-normal panel revealed a significantly higher prevalence of high-risk histologic and molecular EC subtypes in self-identified Black (n = 259) compared with White (n = 1,623) patients. Clinically actionable alterations, including high tumor mutational burden/microsatellite instability, which confer benefit from immunotherapy, were less frequent in ECs from Black than from White patients. Ultramutated POLE molecular subtype ECs associated with favorable outcomes were rare in Black patients. Results were confirmed by genetic ancestry analysis. CCNE1 gene amplification, which is associated with aggressive clinical behavior, was more prevalent in carcinosarcomas occurring in Black than in White patients. ECs from Black and White patients display important differences in their histologic types, molecular subtypes, driver genetic alterations, and therapeutic targets.

SIGNIFICANCE

Our comprehensive analysis of prospectively clinically sequenced ECs revealed significant differences in their histologic and molecular composition and in the presence of therapeutic targets in Black versus White patients. These findings emphasize the importance of incorporating diverse populations into molecular studies and clinical trials to address EC disparities. This article is featured in Selected Articles from This Issue, p. 2293.

Ancestry and self-reported race in Brazilian breast cancer women

OBJECTIVE

This study aimed to evaluate the association between self-reported race/color and ancestry in Brazilian patients with breast cancer.

METHODS

This was an observational, transversal, epidemiological study, evaluating race and ancestry in 1,127 patients with breast cancer. For genetic ancestry, a 46-AIM-INDEL panel was used. The ancestral profile was evaluated with the Structure v.2.3.3 software. Descriptive statistics were performed. To assess differences between race and ancestry, an analysis of variance with Bonferoni adjustment was used.

RESULTS

The race distribution was 77.7% white, 17.6% brown, 4.1% black, 0.4% yellow, and 0.3% cafuse. The African ancestry proportion was significantly (p<0.001) more evident in black [0.63±0.21 (0.17-0.96)], followed by brown [0.25±0.16 (0.02-0.70)], and less frequent in white skin color. The European ancestry proportion was significantly (p<0.001) higher in white [0.72±0.17 (0.02-0.97)], followed by brown [0.57±0.19 (0.12-0.92)], yellow [0.27±0.31 (0.12-0.620], and black [0.24±0.19 (0.02-0.72)]. The Asiatic ancestry proportion is significantly (p<0.001) higher in yellow [0.48±0.51 (0.04-0.93)]. The Amerindian ancestry proportion frequency was the least frequent in all groups, and cafuse patients did not express differences between all race groups. The brown race group presented differences in African and European ancestry.

CONCLUSION

Although we found many similarities between white European ancestry, black African ancestry, and yellow Asian ancestry, there is great miscegenation between patients. Although they can be labeled as having one race, they do present many ancestral genes that would allow their inclusion in another race group.

The tumor multi-omic landscape of endometrial cancers developed on a germline genetic background of adiposity

High body mass index (BMI) is a causal risk factor for endometrial cancer but the tumor molecular mechanisms affected by adiposity and their therapeutic relevance remain poorly understood. Here we characterize the tumor multi-omic landscape of endometrial cancers that have developed on a background of lifelong germline genetic exposure to elevated BMI. We built a polygenic score (PGS) for BMI in women using data on independent, genome-wide significant variants associated with adult BMI in 434,794 women. We performed germline (blood) genotype quality control and imputation on data from 354 endometrial cancer cases from The Cancer Genome Atlas (TCGA). We assigned each case in this TCGA cohort their genetically predicted life-course BMI based on the BMI PGS. Multivariable generalized linear models adjusted for age, stage, microsatellite status and genetic principal components were used to test for associations between the BMI germline PGS and endometrial cancer tumor genome-wide genomic, transcriptomic, proteomic, epigenomic and immune traits in TCGA. High BMI germline PGS was associated with (i) upregulated tumor gene expression in the IL6-JAK-STAT3 pathway (FDR=4.2×10-7); (ii) increased estimated intra-tumor activated mast cell infiltration (FDR=0.008); (iii) increased single base substitution (SBS) mutational signatures 1 (FDR=0.03) and 5 (FDR=0.09) and decreased SBS13 (FDR=0.09), implicating age-related and APOBEC mutagenesis, respectively; and (iv) decreased tumor EGFR protein expression (FDR=0.07). Alterations in IL6-JAK-STAT3 signaling gene and EGFR protein expression were, in turn, significantly associated with both overall survival and progression-free interval. Thus, we integrated germline and somatic data using a novel study design to identify associations between genetically predicted lifelong exposure to higher BMI and potentially actionable endometrial cancer tumor molecular features. These associations inform our understanding of how high BMI may influence the development and progression of this cancer, impacting endometrial tumor biology and clinical outcomes.

Chromatin Accessibility Landscape of Human Triple-negative Breast Cancer Cell Lines Reveals Variation by Patient Donor Ancestry

African American (AA) women have an excessive risk of developing triple-negative breast cancer (TNBC). We employed Assay for Transposase-Accessible Chromatin using sequencing to characterize differences in chromatin accessibility between nine commonly used TNBC cell lines derived from patients of European and African ancestry. Principal component and chromosome mapping analyses of accessibility peaks with the most variance revealed separation of chromatin profiles by patient group. Motif enrichment and footprinting analyses of disparate open chromatin regions revealed differences in transcription factor activity, identifying 79 with ancestry-associated binding patterns (FDR < 0.01). AA TNBC cell lines exhibited increased accessibility for 62 transcription factors associated with epithelial-to-mesenchymal transition, cancer stemness/chemotherapeutic resistance, proliferation, and aberrant p53 regulation, as well as KAISO, which has been previously linked to aggressive tumor characteristics in AA patients with cancer. Differential Assay for Transposase-Accessible Chromatin signal analysis identified 1,596 genes located within promoters of differentially open chromatin regions in AA-derived TNBC, identifying DNA methyltransferase 1 as the top upregulated gene associated with African ancestry. Pathway analyses with these genes revealed enrichment in several pathways, including hypoxia. Culturing cells under hypoxia showed ancestry-specific stress responses that led to the identification of a core set of AA-associated transcription factors, which included members of the Kruppel-like factor and Sp subfamilies, as well as KAISO, and identified ZDHHC1, a gene previously implicated in immunity and STING activation, as the top upregulated AA-specific gene under hypoxia. Together, these data reveal a differential chromatin landscape in TNBC associated with donor ancestry. The open chromatin structure of AA TNBC may contribute to a more lethal disease.

SIGNIFICANCE

We identify an ancestry-associated open chromatin landscape and related transcription factors that may contribute to aggressive TNBC in AA women. Furthermore, this study advocates for the inclusion of diversely sourced cell lines in experimental in vitro studies to advance health equity at all levels of scientific research.

Comparative Genomic Landscape of Urothelial Carcinoma of the Bladder Among Patients of East and South Asian Genomic Ancestry

BACKGROUND

Despite the low rate of urothelial carcinoma of the bladder (UCB) in patients of South Asian (SAS) and East Asian (EAS) descent, they make up a significant portion of the cases worldwide. Nevertheless, these patients are largely under-represented in clinical trials. We queried whether UCB arising in patients with SAS and EAS ancestry would have unique genomic features compared to the global cohort.

METHODS

Formalin-fixed, paraffin-embedded tissue was obtained for 8728 patients with advanced UCB. DNA was extracted and comprehensive genomic profiling was performed. Ancestry was classified using a proprietary calculation algorithm. Genomic alterations (GAs) were determined using a 324-gene hybrid-capture-based method which also calculates tumor mutational burden (TMB) and determines microsatellite status (MSI).

RESULTS

Of the cohort, 7447 (85.3%) were EUR, 541 (6.2%) were AFR, 461 (5.3%) were of AMR, 74 (0.85%) were SAS, and 205 (2.3%) were EAS. When compared with EUR, TERT GAs were less frequent in SAS (58.1% vs. 73.6%; P = .06). When compared with non-SAS, SAS had less frequent GAs in FGFR3 (9.5% vs. 18.5%, P = .25). TERT promoter mutations were significantly less frequent in EAS compared to non-EAS (54.1% vs. 72.9%; P < .001). When compared with the non-EAS, PIK3CA alterations were significantly less common in EAS (12.7% vs. 22.1%, P = .005). The mean TMB was significantly lower in EAS vs. non-EAS (8.53 vs. 10.02; P = .05).

CONCLUSIONS

The results from this comprehensive genomic analysis of UCB provide important insight into the possible differences in the genomic landscape in a population level. These hypothesis-generating findings require external validation and should support the inclusion of more diverse patient populations in clinical trials.

Unique Spectrum of Activating BRAF Alterations in Prostate Cancer

PURPOSE

Alterations in BRAF have been reported in 3% to 5% of prostate cancer, although further characterization is lacking. Here, we describe the nature of BRAF alterations in prostate cancer using a large cohort from commercially available tissue and liquid biopsies subjected to comprehensive genomic profiling (CGP).

EXPERIMENTAL DESIGN

Tissue and liquid biopsies from patients with prostate cancer were profiled using FoundationOne CDx and FoundationOne Liquid CDx CGP assays, respectively. Tissue biopsies from non-prostate cancer types were used for comparison (n = 275,151). Genetic ancestry was predicted using a single-nucleotide polymorphism (SNP) based approach.

RESULTS

Among 15,864 tissue biopsies, BRAF-activating alterations were detected in 520 cases (3.3%). The majority (463 samples, 2.9%) harbored class II alterations, including BRAF rearrangements (243 samples, 1.5%), K601E (101 samples, 0.6%), and G469A (58 samples, 0.4%). BRAF-altered prostate cancers were enriched for CDK12 mutations (OR, 1.87; 9.2% vs. 5.2%; P = 0.018), but depleted in TMPRSS2 fusions (OR, 0.25; 11% vs. 32%; P < 0.0001), PTEN alterations (OR, 0.47; 17% vs. 31%; P < 0.0001), and APC alterations (OR, 0.48; 4.4% vs. 8.9%; P = 0.018) relative to BRAF wild-type (WT) disease. Compared with patients of European ancestry, BRAF alterations were more common in tumors from patients of African ancestry (5.1% vs. 2.9%, P < 0.0001) and Asian ancestry (6.0% vs. 2.9%, P < 0.001).

CONCLUSIONS

Activating BRAF alterations were detected in approximately 3% of prostate cancers, and most were class II mutations and rearrangements; BRAF V600 mutations were exceedingly rare. These findings suggest that BRAF activation in prostate cancer is unique from other cancers and supports further clinical investigation of therapeutics targeting the mitogen-activated protein kinase (MAPK) pathway.

Unique Genomic Alterations and Microbial Profiles Identified in Patients With Gastric Cancer of African, European, and Asian Ancestry: A Novel Path for Precision Oncology

OBJECTIVE

Here, we characterize differences in the genetic and microbial profiles of GC in patients of African (AFR), European, and Asian ancestry.

BACKGROUND

Gastric cancer (GC) is a heterogeneous disease with clinicopathologic variations due to a complex interplay of environmental and biological factors, which may affect disparities in oncologic outcomes..

METHODS

We identified 1042 patients with GC with next-generation sequencing data from an institutional Integrated Mutation Profiling of Actionable Cancer Targets assay and the Cancer Genomic Atlas group. Genetic ancestry was inferred from markers captured by the Integrated Mutation Profiling of Actionable Cancer Targets and the Cancer Genomic Atlas whole exome sequencing panels. Tumor microbial profiles were inferred from sequencing data using a validated microbiome bioinformatics pipeline. Genomic alterations and microbial profiles were compared among patients with GC of different ancestries.

RESULTS

We assessed 8023 genomic alterations. The most frequently altered genes were TP53 , ARID1A , KRAS , ERBB2 , and CDH1 . Patients of AFR ancestry had a significantly higher rate of CCNE1 alterations and a lower rate of KRAS alterations ( P < 0.05), and patients of East Asian ancestry had a significantly lower rate of PI3K pathway alterations ( P < 0.05) compared with other ancestries. Microbial diversity and enrichment did not differ significantly across ancestry groups ( P > 0.05).

CONCLUSIONS

Distinct patterns of genomic alterations and variations in microbial profiles were identified in patients with GC of AFR, European, and Asian ancestry. Our findings of variation in the prevalence of clinically actionable tumor alterations among ancestry groups suggest that precision medicine can mitigate oncologic disparities.

Translational Epidemiology: Genetic Ancestry in Breast Cancer: What Is the Role of Genetic Ancestry and Socioeconomic Status in Triple-Negative Breast Cancer?

Racial/ethnic and socioeconomic disparities seen in triple-negative breast cancer (TNBC) have prompted questions regarding the role of genetic ancestry in breast cancer (BC) subtype development, tumor biology, and ultimately prognosis. The causes of disparities in TNBC are influenced greatly by both sociopolitical factors and genetic ancestry, and now, the potential genomic underpinnings of social factors. To comprehensively understand disparities in TNBC, it is critical to take a translational epidemiologic approach that takes into account genomic and non-genomic factors. Understanding the interplay between genetic ancestry and social genomics and their proportional influence on outcomes can guide our priorities for screening, diagnosis, and interventions for this aggressive BC subtype.

Tackling the lack of diversity in cancer research

Despite the clear benefit of studying biological samples from diverse genetic backgrounds and geographical locations, our current knowledge of disease is mostly derived from the study of European-descent individuals. In the cancer field, this is reflected in the poor representation of African and Amerindian/Latino samples in most large public data repositories. This lack of diversity is due to several reasons, but here we focus on (1) the lack of support for studies on non-European populations that are performed in low- and middle-income countries (LMICs), and (2) unequal partnerships between scientists in LMICs and those in high-income countries. We argue that expanding access to research funding, increasing the participation of underrepresented scientists in editorial boards and international conferences, facilitating the publication of studies conducted in these countries, and properly acknowledging LMIC researchers' contributions in publications and grant applications will promote equity for scientists working in LMICs. We envisage that this will translate to more impactful research in these countries, which will include more samples from diverse populations. For the cancer field, this will broaden our understanding of pathomechanisms and may help to improve the treatment of patients from all backgrounds.

Deregulated miRNA Expression in Triple-Negative Breast Cancer of Ancestral Genomic-Characterized Latina Patients

Among patients with triple-negative breast cancer (TNBC), several studies have suggested that deregulated microRNA (miRNA) expression may be associated with a more aggressive phenotype. Although tumor molecular signatures may be race- and/or ethnicity-specific, there is limited information on the molecular profiles in women with TNBC of Hispanic and Latin American ancestry. We simultaneously profiled TNBC biopsies for the genome-wide copy number and miRNA global expression from 28 Latina women and identified a panel of 28 miRNAs associated with copy number alterations (CNAs). Four selected miRNAs (miR-141-3p, miR-150-5p, miR-182-5p, and miR-661) were validated in a subset of tumor and adjacent non-tumor tissue samples, with miR-182-5p being the most discriminatory among tissue groups (AUC value > 0.8). MiR-141-3p up-regulation was associated with increased cancer recurrence; miR-661 down-regulation with larger tumor size; and down-regulation of miR-150-5p with larger tumor size, high p53 expression, increased cancer recurrence, presence of distant metastasis, and deceased status. This study reinforces the importance of integration analysis of CNAs and miRNAs in TNBC, allowing for the identification of interactions among molecular mechanisms. Additionally, this study emphasizes the significance of considering the patients ancestral background when examining TNBC, as it can influence the relationship between intrinsic tumor molecular characteristics and clinical manifestations of the disease.

Potential pathogenic germline variant reporting from tumor comprehensive genomic profiling complements classic approaches to germline testing

Existing guidance regarding clinically informed germline testing for patients with cancer is effective for evaluation of classic hereditary cancer syndromes and established gene/cancer type associations. However, current screening methods may miss patients with rare, reduced penetrance, or otherwise occult hereditary risk. Secondary finding of suspected germline variants that may confer inherited cancer risk via tumor comprehensive genomic profiling (CGP) has the potential to help address these limitations. However, reporting practices for secondary finding of germline variants are inconsistent, necessitating solutions for transparent and coherent communication of these potentially important findings. A workflow for improved confidence detection and clear reporting of potential pathogenic germline variants (PPGV) in select cancer susceptibility genes (CSG) was applied to a research dataset from real-world clinical tumor CGP of > 125,000 patients with advanced cancer. The presence and patterns of PPGVs identified across tumor types was assessed with a focus on scenarios in which traditional clinical germline evaluation may have been insufficient to capture genetic risk. PPGVs were identified in 9.7% of tumor CGP cases using tissue- and liquid-based assays across a broad range of cancer types, including in a number of "off-tumor" contexts. Overall, PPGVs were identified in a similar proportion of cancers with National Comprehensive Cancer Network (NCCN) recommendations for germline testing regardless of family history (11%) as in all other cancer types (9%). These findings suggest that tumor CGP can serve as a tool that is complementary to traditional germline genetic evaluation in helping to ascertain inherited susceptibility in patients with advanced cancer.

Addressing the Challenge of Biomedical Data Inequality: An Artificial Intelligence Perspective

Artificial intelligence (AI) and other data-driven technologies hold great promise to transform healthcare and confer the predictive power essential to precision medicine. However, the existing biomedical data, which are a vital resource and foundation for developing medical AI models, do not reflect the diversity of the human population. The low representation in biomedical data has become a significant health risk for non-European populations, and the growing application of AI opens a new pathway for this health risk to manifest and amplify. Here we review the current status of biomedical data inequality and present a conceptual framework for understanding its impacts on machine learning. We also discuss the recent advances in algorithmic interventions for mitigating health disparities arising from biomedical data inequality. Finally, we briefly discuss the newly identified disparity in data quality among ethnic groups and its potential impacts on machine learning.

Profiling the Somatic Mutational Landscape of Breast Tumors from Hispanic/Latina Women Reveals Conserved and Unique Characteristics

Somatic mutational profiling is increasingly being used to identify potential targets for breast cancer. However, limited tumor-sequencing data from Hispanic/Latinas (H/L) are available to guide treatment. To address this gap, we performed whole-exome sequencing (WES) and RNA sequencing on 146 tumors and WES of matched germline DNA from 140 H/L women in California. Tumor intrinsic subtype, somatic mutations, copy-number alterations, and expression profiles of the tumors were characterized and compared with data from tumors of non-Hispanic White (White) women in The Cancer Genome Atlas (TCGA). Eight genes were significantly mutated in the H/L tumors including PIK3CA, TP53, GATA3, MAP3K1, CDH1, CBFB, PTEN, and RUNX1; the prevalence of mutations in these genes was similar to that observed in White women in TCGA. Four previously reported Catalogue of Somatic Mutations in Cancer (COSMIC) mutation signatures (1, 2, 3, 13) were found in the H/L dataset, along with signature 16 that has not been previously reported in other breast cancer datasets. Recurrent amplifications were observed in breast cancer drivers including MYC, FGFR1, CCND1, and ERBB2, as well as a recurrent amplification in 17q11.2 associated with high KIAA0100 gene expression that has been implicated in breast cancer aggressiveness. In conclusion, this study identified a higher prevalence of COSMIC signature 16 and a recurrent copy-number amplification affecting expression of KIAA0100 in breast tumors from H/L compared with White women. These results highlight the necessity of studying underrepresented populations.

SIGNIFICANCE

Comprehensive characterization of genomic and transcriptomic alterations in breast tumors from Hispanic/Latina patients reveals distinct genetic alterations and signatures, demonstrating the importance of inclusive studies to ensure equitable care for patients. See related commentary by Schmit et al., p. 2443.

Epidemiology of Bladder Cancer in 2023: A Systematic Review of Risk Factors

CONTEXT

Bladder cancer (BC) is common worldwide and poses a significant public health challenge. External risk factors and the wider exposome (totality of exposure from external and internal factors) contribute significantly to the development of BC. Therefore, establishing a clear understanding of these risk factors is the key to prevention.

OBJECTIVE

To perform an up-to-date systematic review of BC's epidemiology and external risk factors.

EVIDENCE ACQUISITION

Two reviewers (I.J. and S.O.) performed a systematic review using PubMed and Embase in January 2022 and updated it in September 2022. The search was restricted to 4 yr since our previous review in 2018.

EVIDENCE SYNTHESIS

Our search identified 5177 articles and a total of 349 full-text manuscripts. GLOBOCAN data from 2020 revealed an incidence of 573 000 new BC cases and 213 000 deaths worldwide in 2020. The 5-yr prevalence worldwide in 2020 was 1 721 000. Tobacco smoking and occupational exposures (aromatic amines and polycyclic aromatic hydrocarbons) are the most substantial risk factors. In addition, correlative evidence exists for several risk factors, including specific dietary factors, imbalanced microbiome, gene-environment risk factor interactions, diesel exhaust emission exposure, and pelvic radiotherapy.

CONCLUSIONS

We present a contemporary overview of the epidemiology of BC and the current evidence for BC risk factors. Smoking and specific occupational exposures are the most established risk factors. There is emerging evidence for specific dietary factors, imbalanced microbiome, gene-external risk factor interactions, diesel exhaust emission exposure, and pelvic radiotherapy. Further high-quality evidence is required to confirm initial findings and further understand cancer prevention.

PATIENT SUMMARY

Bladder cancer is common, and the most substantial risk factors are smoking and workplace exposure to suspected carcinogens. On-going research to identify avoidable risk factors could reduce the number of people who get bladder cancer.

Equitable machine learning counteracts ancestral bias in precision medicine, improving outcomes for all

Gold standard genomic datasets severely under-represent non-European populations, leading to inequities and a limited understanding of human disease [1-8]. Therapeutics and outcomes remain hidden because we lack insights that we could gain from analyzing ancestry-unbiased genomic data. To address this significant gap, we present PhyloFrame, the first-ever machine learning method for equitable genomic precision medicine. PhyloFrame corrects for ancestral bias by integrating big data tissue-specific functional interaction networks, global population variation data, and disease-relevant transcriptomic data. Application of PhyloFrame to breast, thyroid, and uterine cancers shows marked improvements in predictive power across all ancestries, less model overfitting, and a higher likelihood of identifying known cancer-related genes. The ability to provide accurate predictions for underrepresented groups, in particular, is substantially increased. These results demonstrate how AI can mitigate ancestral bias in training data and contribute to equitable representation in medical research.

Distinct shared and compartment-enriched oncogenic networks drive primary versus metastatic breast cancer

Metastatic breast-cancer is a major cause of death in women worldwide, yet the relationship between oncogenic drivers that promote metastatic versus primary cancer is still contentious. To elucidate this relationship in treatment-naive animals, we hereby describe mammary-specific transposon-mutagenesis screens in female mice together with loss-of-function Rb, which is frequently inactivated in breast-cancer. We report gene-centric common insertion-sites (gCIS) that are enriched in primary-tumors, in metastases or shared by both compartments. Shared-gCIS comprise a major MET-RAS network, whereas metastasis-gCIS form three additional hubs: Rho-signaling, Ubiquitination and RNA-processing. Pathway analysis of four clinical cohorts with paired primary-tumors and metastases reveals similar organization in human breast-cancer with subtype-specific shared-drivers (e.g. RB1-loss, TP53-loss, high MET, RAS, ER), primary-enriched (EGFR, TGFβ and STAT3) and metastasis-enriched (RHO, PI3K) oncogenic signaling. Inhibitors of RB1-deficiency or MET plus RHO-signaling cooperate to block cell migration and drive tumor cell-death. Thus, targeting shared- and metastasis- but not primary-enriched derivers offers a rational avenue to prevent metastatic breast-cancer.

Proteogenomics of clear cell renal cell carcinoma response to tyrosine kinase inhibitor

The tyrosine kinase inhibitor (TKI) Sunitinib is one the therapies approved for advanced renal cell carcinoma. Here, we undertake proteogenomic profiling of 115 tumors from patients with clear cell renal cell carcinoma (ccRCC) undergoing Sunitinib treatment and reveal the molecular basis of differential clinical outcomes with TKI therapy. We find that chromosome 7q gain-induced mTOR signaling activation is associated with poor therapeutic outcomes with Sunitinib treatment, whereas the aristolochic acid signature and VHL mutation synergistically caused enhanced glycolysis is correlated with better prognosis. The proteomic and phosphoproteomic analysis further highlights the responsibility of mTOR signaling for non-response to Sunitinib. Immune landscape characterization reveals diverse tumor microenvironment subsets in ccRCC. Finally, we construct a multi-omics classifier that can detect responder and non-responder patients (receiver operating characteristic-area under the curve, 0.98). Our study highlights associations between ccRCC molecular characteristics and the response to TKI, which can facilitate future improvement of therapeutic responses.

Cancer Outcomes Among Prediabetes and Type 2 Diabetes Populations With Dietary and Physical Activity-based Lifestyle Interventions

CONTEXT

People with type 2 diabetes (T2D) have higher risks of cancer incidence and death.

OBJECTIVE

We aimed to evaluate the relationship between dietary and physical activity-based lifestyle intervention and cancer outcomes among prediabetes and T2D populations.

METHODS

We searched for randomized controlled trials with at least 24 months of lifestyle interventions in prediabetes or T2D populations. Data were extracted by pairs of reviewers and discrepancies were resolved by consensus. Descriptive syntheses were performed, and the risk of bias was assessed. Relative risks (RRs) and 95% CIs were estimated using a pairwise meta-analysis with both a random-effects model and a general linear mixed model (GLMM). Certainty of evidence was evaluated using the Grading of Recommendations Assessment, Development, and Evaluation framework, and trial sequential analysis (TSA) was conducted to assess if current information is enough for definitive conclusions. Subgroup analysis was performed by glycemic status.

RESULTS

Six clinical trials were included. Among 12 841 participants, the combined RR for cancer mortality comparing lifestyle interventions with usual care was 0.94 (95% CI, 0.81-1.10 using GLMM and 0.82-1.09 using random-effects model). Most studies had a low risk of bias, and the certainty of evidence was moderate. TSA showed that the cumulative Z curve reached futility boundary while total number did not reach detection boundary.

CONCLUSION

Based on the limited data available, dietary and physical activity-based lifestyle interventions had no superiority to usual care on reducing cancer risk in populations with prediabetes and T2D. Lifestyle interventions focused on cancer outcomes should be tested to better explore their effects.

Inferring genetic ancestry from cancer sequencing data

Genetic ancestry is an important biological determinant of cancer health disparities that is not captured by self-identified race and ethnicity (SIRE). Belleau et al. recently developed a systematic computational approach to infer genetic ancestry from cancer-derived molecular data from different genomic and transcriptomic profiling assays, creating opportunities to interrogate population-scale data sets.

Establishment and Validation of a Novel Risk Score for Hepatocellular Carcinoma Based on Bile Acid and Bile Salt Metabolism-Related Genes

Liver cancer is a public disease burden with an increasing incidence rate globally. Bile acid and bile salt's metabolic pathways participate in liver tumorigenesis and regulate the tumor microenvironment. However, there still remains a lack of systematic analysis of the genes related to bile acid and bile salt metabolic pathways in hepatocellular carcinoma (HCC). The mRNA expression data and clinical follow-up information of patients with HCC were obtained from public databases, including The Cancer Genome Atlas, Hepatocellular Carcinoma Database, Gene Expression Omnibus, and IMvigor210. The bile acid and bile salt metabolism-related genes were extracted from Molecular Signatures Database. Univariate Cox and logistic least absolute shrinkage and selection operator regression analyses were conducted to establish the risk model. Single sample gene set enrichment analysis, Estimation of STromal and Immune cells in MAlignant Tumour tissues using Expression data, and Tumor Immune Dysfunction and Exclusion were adopted to analyze immune status. The efficiency of the risk model was tested using a decision tree and a nomogram. We determined two molecular subtypes based on bile acid and bile salt metabolism-related genes, with the prognosis of the S1 subtype being markedly superior to the S2 subtype. Next, we established a risk model based on the differentially expressed genes between the two molecular subtypes. The high-risk and low-risk groups showed significant differences in the biological pathways, immune score, immunotherapy response, and drug susceptibility. Our results demonstrated the good predictive performance of the risk model in immunotherapy datasets and established that it could be an essential factor affecting the prognosis of HCC. In conclusion, we identified two molecular subtypes based on bile acid and bile salt metabolism-related genes. The risk model established in our study could effectively predict the prognosis of patients with HCC and their immunotherapeutic response, which may contribute to targeted immunotherapy in HCC.

TONSL Is an Immortalizing Oncogene and a Therapeutic Target in Breast Cancer

Study of genomic aberrations leading to immortalization of epithelial cells has been technically challenging due to the lack of isogenic models. To address this, we used healthy primary breast luminal epithelial cells of different genetic ancestry and their hTERT-immortalized counterparts to identify transcriptomic changes associated with immortalization. Elevated expression of TONSL (Tonsoku-like, DNA repair protein) was identified as one of the earliest events during immortalization. TONSL, which is located on chromosome 8q24.3, was found to be amplified in approximately 20% of breast cancers. TONSL alone immortalized primary breast epithelial cells and increased telomerase activity, but overexpression was insufficient for neoplastic transformation. However, TONSL-immortalized primary cells overexpressing defined oncogenes generated estrogen receptor-positive adenocarcinomas in mice. Analysis of a breast tumor microarray with approximately 600 tumors revealed poor overall and progression-free survival of patients with TONSL-overexpressing tumors. TONSL increased chromatin accessibility to pro-oncogenic transcription factors, including NF-κB and limited access to the tumor-suppressor p53. TONSL overexpression resulted in significant changes in the expression of genes associated with DNA repair hubs, including upregulation of several genes in the homologous recombination (HR) and Fanconi anemia pathways. Consistent with these results, TONSL-overexpressing primary cells exhibited upregulated DNA repair via HR. Moreover, TONSL was essential for growth of TONSL-amplified breast cancer cell lines in vivo, and these cells were sensitive to TONSL-FACT complex inhibitor CBL0137. Together, these findings identify TONSL as a regulator of epithelial cell immortalization to facilitate cancer initiation and as a target for breast cancer therapy.

SIGNIFICANCE

The chr.8q24.3 amplicon-resident gene TONSL is upregulated during the initial steps of tumorigenesis to support neoplastic transformation by increasing DNA repair and represents a potential therapeutic target for treating breast cancer.

Genetic Ancestry of 1127 Brazilian Breast Cancer Patients and Its Correlation With Molecular Subtype and Geographic Region

PURPOSE

Breast cancer molecular subtypes show significant differences in different ethnic groups in the United States, but no study has evaluated genetic ancestry in breast cancer in Brazilian women.

METHODS

Breast cancer patients from distinct parts of Brazil were evaluated. Molecular subtypes were determined by immunohistochemistry. Genetic ancestry was evaluated using a panel of 46 AIMs (ancestry informative markers), which classified genetic ancestry as European, African, Asian, and Amerindian. PCR products were subjected to capillary electrophoresis and analyzed using GeneMapper 4.0 software. Ancestry was evaluated with Structure v.2.3.3 software. Ancestry was tested for correlations with geographic region and molecular subtype. The chi-square test and ANOVA with Bonferroni adjustment were applied.

RESULTS

Genetic ancestry and clinical data were evaluated in 1127 patients. Higher rates of self-reported white ethnicity, European ancestry, and HER-2^- luminal tumors were identified in the South region, which may influence age at diagnosis and result in a higher rate of early tumors. Conversely, higher rates of African ancestry in the North and Northeast regions, self-reported nonwhite ethnicity, HER-2⁺ tumors, and triple-negative tumors were noted. Triple-negative and HER-2⁺ tumors were associated with higher advanced and metastatic disease rates at diagnosis, with triple-negative tumors being more frequent in young women.

CONCLUSION

Differences in genetic ancestry, self-reported ethnicity, and molecular subtype were found between Brazilian demographic regions. Knowledge of these features may contribute to a better understanding of age at diagnosis and the molecular distribution of breast cancer in Brazil.

Molecular Subtypes of Head and Neck Cancer in Patients of African Ancestry

PURPOSE

The purpose of this study was to better understand the complex molecular biomarkers and signatures of head and neck cancer (HNC) among Black patients and identify possible molecular changes associated with HNC disparities.

EXPERIMENTAL DESIGN

Molecular subtypes and genomic changes in HNC samples from patients of African and European ancestry in The Cancer Genome Atlas, Memorial Sloan Kettering Cancer Center, Broad Institute, MD Anderson Cancer Center, and John Hopkins University were identified. Molecular features (genomic, proteomic, transcriptomic) associated with race and genomic alterations associated with clinical outcomes were determined. An independent cohort of HNC tumor specimens was used to validate the primary findings using IHC.

RESULTS

Black patients were found to have a younger age at diagnosis, more aggressive tumor types, higher rates of metastasis, and worse survival compared with White patients. Black patients had fewer human papillomavirus-positive tumor types and higher frequencies of laryngeal subtype tumors. Higher frequencies of TP53, MYO18B, KMT2D, and UNC13C mutations and a lower frequency of PIK3CA mutations were observed in Black patients. Tumors of Black patients showed significant enrichment of c-MYC and RET-tyrosine signaling and amplifications. A significant increase in tumor expression of c-MYC in Black patients was observed and was associated with poor survival outcomes in the independent cohort.

CONCLUSIONS

Novel genomic modifications and molecular signatures may be related to environmental, social, and behavioral factors associated with racial disparities in HNC. Unique tumor mutations and biological pathways have potential clinical utility in providing more targeted and individualized screening, diagnostic, and treatment modalities to improve health outcomes.

Opportunities and Challenges with Artificial Intelligence in Genomics

The development of artificial intelligence and machine learning algorithms may allow for advances in patient care. There are existing and potential applications in cancer diagnosis and monitoring, identification of at-risk groups of individuals, classification of genetic variants, and even prediction of patient ancestry. This article provides an overview of some current and future applications of artificial intelligence in genomic medicine, in addition to discussing challenges and considerations when bringing these tools into clinical practice.

Group-shrinkage feature selection with a spatial network for mining DNA methylation data

Identifying disease-related biomarkers from high-dimensional DNA methylation data helps in reducing early screening costs and inferring pathogenesis mechanisms. Good discovery results have been achieved through spatial correlation methods of methylation sites, group-based regularization, and network constraints. However, these methods still have some key limitations as they cannot exclude isolated differential sites and only consider adjacent site ordering. Therefore, we propose a group-shrinkage feature selection algorithm to encourage the selection of clustered sites and discourage the selection of isolated differential sites. Specifically, a network-guided group-shrinkage strategy is developed to penalize weakly-correlated isolated methylation sites through a network structure constraint. The spatial network is constructed based on spatial correlation information of DNA methylation sites, where this information accounts for the uneven site distribution. The experimental simulations and applications demonstrated that the proposed method outperforms the advanced regularization methods, especially in rejecting isolated methylation sites; hence this study provides an efficient and clinical-valuable method for biomarker candidate discovery in DNA methylation data. Additionally, the proposed method exhibits enhanced reliability due to introducing biological prior knowledge into a regularization-based feature selection framework and could promote more research in the integration between biological prior knowledge and classical feature selection methods, thus facilitating their clinical application. Our source codes will be released at https://github.com/SJTUBME-QianLab/Group-shrinkage-Spatial-Network once this manuscript is accepted for publication.

Prognostic evaluation of polygenic risk score underlying pan-cancer analysis: evidence from two large-scale cohorts

BACKGROUND

Polygenic risk score (PRS) has been demonstrated to be effective in identifying individuals at high risk of developing cancer, but its prognostic value remains unclear.

METHODS

We constructed site-specific PRSs by aggregating the risk effect of independent variants derived from previous genome-wide association studies (GWASs) across 17 cancer types. The Cox proportional hazards model was used to evaluate the association of each PRS with cancer survival, leveraging data from two prospective European cohorts, namely the UK Biobank involving 19,628 incident cases and The Cancer Genome Atlas involving 7079 prevalent cases. The combined PRS (CPRS), determined by merging site-specific PRSs, was further used to assess the prognostic effect of PRS on overall cancer in a sex-specific manner.

FINDINGS

We discovered that the cancer risk-related PRS was associated with neither overall survival (OS) nor cancer-specific survival (CSS) of each site-specific cancer with an underlying false discovery rate (FDR) P > 0.05, as evidenced by consistent findings from the two cohorts. Furthermore, the fixed-effect meta-analysis of the two cohorts provided no evidence to support for an association between CPRS and overall cancer survival in both males [OS: hazard ratio (HR)_meta = 1.00, P_meta = 0.760; CSS: HR_meta = 1.01, P_meta = 0.447] and females (OS: HR_meta = 0.97, P_meta = 0.067; CSS: HR_meta = 0.96, P_meta = 0.054). Similar results were observed across multiple sensitivity analyses.

INTERPRETATION

Our findings indicate that the risk-specific PRS might not be a clinically useful tool in cancer prognosis prediction and further studies focusing on the development of polygenic prognostic score are warranted.

FUNDING

This project was funded by the National Natural Science Foundation of China (82173601 and 82073631), and Priority Academic Program Development of Jiangsu Higher Education Institutions (Public Health and Preventive Medicine).

A systematic review and meta-analysis of germline BRCA mutations in pancreatic cancer patients identifies global and racial disparities in access to genetic testing

BACKGROUND

Germline BRCA1 and BRCA2 mutations (gBRCAm) can inform pancreatic cancer (PC) risk and treatment but most of the available information is derived from white patients. The ethnic and geographic variability of gBRCAm prevalence and of germline BRCA (gBRCA) testing uptake in PC globally is largely unknown.

MATERIALS AND METHODS

We carried out a systematic review and prevalence meta-analysis of gBRCA testing and gBRCAm prevalence in PC patients stratified by ethnicity. The main outcome was the distribution of gBRCA testing uptake across diverse populations worldwide. Secondary outcomes included: geographic distribution of gBRCA testing uptake, temporal analysis of gBRCA testing uptake in ethnic groups, and pooled proportion of gBRCAm stratified by ethnicity. The study is listed under PROSPERO registration number #CRD42022311769.

RESULTS

A total of 51 studies with 16 621 patients were included. Twelve of the studies (23.5%) enrolled white patients only, 10 Asians only (19.6%), and 29 (56.9%) included mixed populations. The pooled prevalence of white, Asian, African American, and Hispanic patients tested per study was 88.7%, 34.8%, 3.6%, and 5.2%, respectively. The majority of included studies were from high-income countries (HICs) (64; 91.2%). Temporal analysis showed a significant increase only in white and Asians patients tested from 2000 to present (P < 0.001). The pooled prevalence of gBRCAm was: 3.3% in white, 1.7% in Asian, and negligible (<0.3%) in African American and Hispanic patients.

CONCLUSIONS

Data on gBRCA testing and gBRCAm in PC derive mostly from white patients and from HICs. This limits the interpretation of gBRCAm for treating PC across diverse populations and implies substantial global and racial disparities in access to BRCA testing in PC.