Breast and bowel cancers diagnosed in people ‘too young to have cancer’: A blueprint for research using family and twin studies

John L. Hopper, Shuai Li, Robert J. MacInnis, James G. Dowty, Tuong L. Nguyen, Minh Bui, Gillian S. Dite, Vivienne F. C. Esser, Zhoufeng Ye, Enes Makalic, Daniel F. Schmidt, Benjamin Goudey, Karen Alpen, Miroslaw K. Kapuscinski, Aung Ko Win, Pierre‐Antoine Dugué, Roger L. Milne, Harindra Jayasekara, Jennifer D. Brooks, Sue Malta, Lucas Calais‐Ferreira, Alexander C. Campbell, Jesse T Young, Tú Nguyen‐Dumont, Joohon Sung, Graham G. Giles, Daniel D. Buchanan, Ingrid Winship, Mary Beth Terry, Melissa C. Southey, Mark A. JenkinsGenetic 2024 · DOI

TL;DR

Young breast and bowel cancers are a growing concern with significant morbidity, yet they remain understudied. This paper proposes a research blueprint leveraging family and twin studies, introducing novel statistical methods like VALID, DEPTH, ICE FALCON, and ICE CRISTAL to disentangle genetic and environmental factors contributing to these early-onset cancers.

What Problem Does This Paper Address?

Breast and bowel cancers are common, but a significant minority are diagnosed at young ages (before 40-50), often at later stages and with higher mortality. These "too young" cancers are increasing in incidence and have a disproportionately large impact on years of life lost. Despite this, epidemiological research has historically focused on older individuals, leaving critical knowledge gaps regarding risk factors, causes, and prevention strategies for younger populations. While familial relative risks are higher at younger ages, the underlying genetic and environmental contributions remain largely unexplained, even with advances in genetic testing.

Methods

The paper introduces and discusses several novel statistical approaches designed to analyze complex family and twin data for cancer research:

  • VALID (Variance in Age-specific Log Incidence Decomposition): This model aims to quantify familial, genetic, genomic, and nonfamilial (environmental) variances in cancer risk, specifically focusing on age-specific log(incidence). It helps understand the sources of risk variation and aids in personalized risk prediction.
  • DEPTH (DEPendency of association on the number of Top Hits): This method is designed for analyzing genome-wide association study (GWAS) data. It uses a sliding window of multiple variants to identify genomic regions (rather than individual variants) implicated in risk, offering a more powerful approach for discovery.
  • ICE FALCON (Inference about Causation from Examining FAmiliaL CONfounding): This approach is specifically for analyzing data from related individuals (like twins and siblings). It aims to infer causation by comparing within-individual and across-pair regression coefficients, helping to distinguish true causal effects from familial confounding.
  • ICE CRISTAL (Inference about Causation from Examining Changes in Regression coefficients and Innovative STatistical AnaLysis): This method is applicable to unrelated individuals and is similar to ICE FALCON but uses family history as a proxy for familial confounding. It also infers causation by examining changes in regression coefficients.

These methods are presented as tools to overcome limitations in traditional epidemiological studies and to leverage the unique insights provided by family and twin data.

Datasets

The paper highlights the importance and availability of several large-scale family-based resources:

  • Breast Cancer Family Registries (Breast CFR): An international cohort study of multigenerational families with and without breast cancer, containing extensive genetic, biospecimen, and epidemiological data.
  • Colon Cancer Family Registries (Colon CFR): Similar to the Breast CFR, this registry focuses on colorectal cancer, collecting comprehensive data on families with and without the disease.
  • Nordic Twin Study: A population-based study of twins that has been combined with other resources.
  • FinnGen study: A Finnish study that includes genomic data, such as polygenic risk scores.
  • UK Biobank: Offers the potential for family studies due to the identification of genetically related participants.

The authors emphasize the harmonization of data from the Breast and Colon CFRs to create a comprehensive resource for studying both cancer types.

Key Insights and Ideas

  • Younger ages amplify familial risk: Familial relative risks and variances in risk are significantly higher at younger ages, making these ages a critical window for discovering the causes of cancer.
  • "Major genes" explain only a fraction: Even for young-onset cancers, high-risk mutations in well-known genes (like BRCA1/2) explain only a small proportion of cases, indicating the need to discover other "major genes."
  • "Heritability" can be misleading: The concept of heritability as a percentage of risk explained can be misleading for guiding genetic research, especially for binary outcomes. The paper advocates for focusing on absolute genetic variance and understanding risk gradients.
  • Genes and environment interact dynamically: Genetic risks are not immutable. The paper suggests that as genetic risk increases, the importance of non-genetic risk factors also increases, highlighting opportunities for risk reduction.
  • Causation is key, but hard to establish: Distinguishing true causal relationships from associations, especially in the presence of familial confounding, is a major challenge. The ICE FALCON and ICE CRISTAL methods are proposed to address this.
  • Early life exposures matter: The research blueprint emphasizes investigating early-life exposures (e.g., height, diet, physical activity, medication use) as potential contributors to young-onset cancers.

Key Results

The paper primarily presents a framework and methodological advancements rather than specific quantitative results from new analyses. However, it references existing studies and illustrates the application of its proposed methods:

  • VALID application: The paper shows how VALID can decompose variance in risk for breast cancer, highlighting "unknown" variance at younger ages and the limited contribution of current polygenic risk scores at those ages. Similar patterns are discussed for bowel cancer.
  • DEPTH application: DEPTH has been applied to prostate cancer, identifying 112 novel putative risk regions, and to bowel cancer, identifying 82 likely risk regions not detected by conventional approaches.
  • ICE FALCON/CRISTAL examples: These methods have been applied to studies of atopic march, mammographic density, bone architecture, and DNA methylation, providing evidence for causal relationships. For instance, ICE FALCON demonstrated that BMI causes DNA methylation, not vice versa, using a much smaller dataset than Mendelian randomization.
  • Breast cancer risk factors: The Breast CFR study found strong evidence that oral contraceptive use was associated with a substantial protective effect against breast cancer diagnosed before age 40 for BRCA1 carriers, but not for BRCA2 carriers.

Limitations and Future Work

  • Limitations of current knowledge: The paper acknowledges that despite extensive research, the causes of a significant portion of familial risk for young-onset breast and bowel cancers remain unknown. Current polygenic risk scores have limited impact, especially for younger individuals.
  • Challenges in family studies: Collecting large-scale family data with high response rates and epidemiological rigor is complex and resource-intensive.
  • Future Directions:
    • Discovering new genetic causes: Identifying "major genes" for young breast and bowel cancers.
  • Improving polygenic risk scores: Developing more effective polygenic risk scores, particularly for Lynch syndrome.
  • Investigating early-life exposures: Conducting detailed analyses of early-life environmental factors and their interactions with genetic predispositions.
  • Leveraging advanced technologies: Combining new genetic, epigenetic, and proteomic technologies with family and twin study designs.
  • Harmonizing and expanding resources: Continuing to harmonize and expand existing family registries and exploring new data sources like national genealogies and electronic health records.
  • Qualitative research: Incorporating qualitative studies to gain deeper insights into participants' experiences and potential early-life causes.

Summary updated

Abstract

Young breast and bowel cancers (e.g., those diagnosed before age 40 or 50 years) have far greater morbidity and mortality in terms of years of life lost, and are increasing in incidence, but have been less studied. For breast and bowel cancers, the familial relative risks, and therefore the familial variances in age-specific log(incidence), are much greater at younger ages, but little of these familial variances has been explained. Studies of families and twins can address questions not easily answered by studies of unrelated individuals alone. We describe existing and emerging family and twin data that can provide special opportunities for discovery. We present designs and statistical analyses, including novel ideas such as the VALID (Variance in Age-specific Log Incidence Decomposition) model for causes of variation in risk, the DEPTH (DEPendency of association on the number of Top Hits) and other approaches to analyse genome-wide association study data, and the within-pair, ICE FALCON (Inference about Causation from Examining FAmiliaL CONfounding) and ICE CRISTAL (Inference about Causation from Examining Changes in Regression coefficients and Innovative STatistical AnaLysis) approaches to causation and familial confounding. Example applications to breast and colorectal cancer are presented. Motivated by the availability of the resources of the Breast and Colon Cancer Family Registries, we also present some ideas for future studies that could be applied to, and compared with, cancers diagnosed at older ages and address the challenges posed by young breast and bowel cancers.
BibTeX
@inproceedings{hopper2024breast,
  author = {John L. Hopper and Shuai Li and Robert J. MacInnis and James G. Dowty and Tuong L. Nguyen and Minh Bui and Gillian S. Dite and Vivienne F. C. Esser and Zhoufeng Ye and Enes Makalic and Daniel F. Schmidt and Benjamin Goudey and Karen Alpen and Miroslaw K. Kapuscinski and Aung Ko Win and Pierre‐Antoine Dugué and Roger L. Milne and Harindra Jayasekara and Jennifer D. Brooks and Sue Malta and Lucas Calais‐Ferreira and Alexander C. Campbell and Jesse T Young and Tú Nguyen‐Dumont and Joohon Sung and Graham G. Giles and Daniel D. Buchanan and Ingrid Winship and Mary Beth Terry and Melissa C. Southey and Mark A. Jenkins},
  title = {{Breast and bowel cancers diagnosed in people ‘too young to have cancer’: A blueprint for research using family and twin studies}},
  booktitle = {Genetic},
  year = {2024},
  doi = {10.1002/gepi.22555},
}

Loading related content...

Figure