STANFORD, CA – January 8, 2024 – In a landmark study published today in Nature, researchers at Stanford Medicine have introduced a revolutionary framework for classifying breast cancers, moving beyond traditional protein-receptor definitions to a deeper understanding rooted in the structural variations of tumor DNA. This new system groups breast cancers into three main archetypes based on genomic anomalies, including the amplification of cancer-associated genes (oncogenes) and the presence of untethered DNA circles called extrachromosomal DNA (ecDNA). Crucially, these foundational genomic signatures are established remarkably early in cancer development and persist as the disease progresses and metastasizes, offering unprecedented insight into tumor aggressiveness and recurrence risk.
This pioneering work promises to transform clinical decision-making, enabling physicians to more accurately predict patient outcomes and paving the way for highly targeted therapeutic interventions. By distinguishing patients most likely to benefit from aggressive early treatment from those who may safely defer such intense approaches, the research underscores the critical importance of robust biomarkers and early, precise intervention in the fight against breast cancer.
"My lab has had a long-standing interest in understanding how aggressive breast tumors arise, why they are resistant to therapy and why they are prone to recur in distant organs," stated Christina Curtis, PhD, the RZ Cao Professor and a professor of oncology, of genetics, and of biomedical data science, and senior author of the study. "This research shows that breast tumors develop key structural variants that set the tumor on its course very early in its development. In short, some are born to be bad. It emphasizes the importance of robust biomarkers and of intervening early in the course of the disease."
The study’s lead authors include formal postdoctoral scholar Kathleen Houlahan, PhD, postdoctoral scholar Lise Mangiante, PhD, former research assistant Cristina Sotomayor-Vivas, and graduate student Alvina Adimoelja.
The Evolving Landscape of Breast Cancer Classification
For decades, breast cancer classification has relied on a relatively broad-stroke approach, primarily categorizing tumors based on the presence or absence of specific protein receptors on cancer cells. While foundational, this system has long faced limitations in fully capturing the intricate biology and diverse clinical trajectories of the disease.
Traditional Receptor-Based Classification
Historically, breast cancers have been categorized into three main types based on receptor expression:
- Hormone-Receptor Positive (HR+): These tumors express elevated levels of receptors that bind to estrogen or progesterone. They are the most common type and are often treated successfully with a combination of hormone therapy, chemotherapy, surgery, and radiation, aimed at lowering estrogen production, blocking estrogen binding, or degrading estrogen receptors.
- HER2-Positive (HER2+): Constituting about 15% to 20% of all cases, these cancers have elevated levels of the HER2 receptor. They are known to be aggressive but have seen dramatic improvements in patient outcomes with the advent of targeted therapies that block HER2 activity, such as trastuzumab (Herceptin).
- Triple-Negative Breast Cancer (TNBC): Accounting for roughly 10% of newly diagnosed cases, TNBCs do not express either of the hormone receptors (estrogen or progesterone) or HER2. These are often considered the most challenging to treat successfully due to a lack of specific therapeutic targets and tend to recur early.
While this traditional classification has guided treatment for many years, it often fails to predict the full spectrum of tumor behavior, particularly the nuanced risks of recurrence, even within seemingly well-managed groups. The limitations became increasingly apparent as clinicians observed varied responses to treatment and unexpected recurrences, prompting a deeper quest for more precise classification systems.
The Dawn of Molecular Subtyping
Dr. Curtis, director of artificial intelligence and cancer genomics at the Stanford Cancer Institute, has been at the forefront of this quest for over a decade. Her previous work sought to move beyond surface-level receptor status to a more granular, molecular understanding of breast cancer.
In 2012, her team pioneered the use of machine-learning techniques to analyze DNA and RNA sequences from both healthy cells and breast tumors. This allowed them to construct a comprehensive "molecular snapshot" of each tumor, detailing genetic alterations and their impact on gene expression. This groundbreaking study identified 11 clinically significant subgroups of breast cancer – a far greater resolution than previously achieved. These subgroups exhibited distinct prognoses, hinting at a more complex biological reality, but the immediate clinical application of this information remained somewhat unclear.
Unveiling Persistent Recurrence Risk
The true clinical significance of these molecular subgroups began to crystallize in subsequent research. A study involving 75,000 people with estrogen-receptor positive breast cancer revealed a troubling pattern: even after five years of hormone therapy, and even in groups initially considered low-risk, breast cancer recurrences continued years, sometimes decades, after initial diagnosis. This observation spurred Dr. Curtis and her colleagues to investigate whether their previously defined subgroups could better delineate this elusive, long-term risk.
In 2019, they demonstrated that by overlaying the receptor status of breast tumors with their molecular subgroup classification, they could powerfully predict which hormone-receptor positive tumors were prone to recur many years after initial diagnosis and treatment. Specifically, four of the eight estrogen-receptor positive subgroups were found to be significantly more likely than others to return, even 10 or 20 years post-diagnosis. When these four high-risk groups were combined, the findings were stark: one-quarter of women with hormone-receptor positive, HER2-negative breast tumors faced a nearly 50% chance of their cancer recurring decades after initial diagnosis. This elevated recurrence risk was profound, surpassing even that of people with triple-negative breast cancer and mirroring the dire prognosis of HER2-positive breast cancers before the approval of trastuzumab, which revolutionized outcomes for that group.
This refined approach also offered critical insights for triple-negative tumors, identifying patients who were unlikely to experience recurrence more than five years after diagnosis and treatment, as well as those who were highly susceptible. Such patient stratification proved invaluable, enabling clinicians to pinpoint individuals who might require aggressive early treatment or more intensive long-term monitoring, while also identifying those who could potentially avoid harsher, unnecessary therapeutic approaches.
Despite these significant advances, a fundamental question persisted: what underlying biological mechanisms were driving these profound differences among the subgroups? What was the deeper genomic architecture dictating these varied clinical destinies?
A Deep Dive into Genomic Architecture: The Three New Archetypes
The lingering mystery spurred Dr. Curtis and her team to "take a step back," as she described, to investigate the foundational genomic alterations. Their goal was to move beyond simply identifying patterns of gene expression or copy number variations, and instead, to deconstruct the processes that give rise to these characteristic events. This involved a comprehensive assessment of the genomic architecture – the mutations and structural variations within a cancer cell’s DNA – of nearly 2,000 breast cancers, spanning all stages from ductal carcinoma in situ (stage 0) to advanced metastatic disease (stage 4). This meticulous analysis led to the identification of three distinct genomic archetypes.
Archetype 1: Complex Amplifications and Extrachromosomal DNA (ecDNA)
This archetype is characterized by a distinctive genomic landscape featuring complex, yet localized, amplifications of cancer-associated genes (oncogenes) on chromosomes. A critical accompanying feature is the abundant presence of small, circular DNA structures known as extrachromosomal DNA (ecDNA). Unlike chromosomal DNA, ecDNA exists untethered to the rest of the genome, often "chock-full of oncogenes."
Recent studies have highlighted ecDNAs as potent drivers of cancer growth and evolution, largely because they frequently bypass normal cellular regulatory mechanisms that control gene copy number and expression. Their independent existence allows for rapid amplification and expression of oncogenes, conferring a significant growth advantage to tumor cells and contributing to therapy resistance.
Significantly, the researchers found that the high-risk hormone-receptor positive subgroups identified in their earlier work strongly overlapped with the HER2-positive subgroup within this ecDNA-rich archetype. This observation is profound, as Dr. Curtis noted: "Here we have two different molecular subtypes, which we treat differently in the clinic but that strongly overlap in their patterns of chromosomal instability." This shared underlying genomic instability, driven by focal amplifications and ecDNA, suggests a common vulnerability despite their distinct protein receptor profiles.
Archetype 2: Global Genomic Instability and DNA Repair Deficiency
In contrast to the localized amplifications seen in the first archetype, this group of tumors exhibited widespread genomic instability. Their genomes displayed alterations across vast regions, accumulating numerous mutations and structural rearrangements, giving the appearance of "scars" across the entire genome. As Dr. Curtis explained, "It’s not limited to particular oncogenes." A significant subset of these globally unstable tumors also showed clear signs of being deficient in their ability to repair DNA damage. This impaired DNA repair machinery allows for the rampant accumulation of mutations, fueling rapid tumor evolution and increasing complexity.
This archetype strongly correlated with triple-negative breast cancers, providing a deeper understanding of why these aggressive tumors are often so challenging to treat and prone to early recurrence. Their chaotic genomes, often coupled with a compromised ability to correct errors, create a highly adaptable and resilient cancer.
Archetype 3: Relatively Stable Genomes
The third archetype represents the "garden-variety" hormone-receptor positive, HER2-negative breast cancers, typically associated with lower, more predictable risks of recurrence. These tumors are characterized by relatively stable genomes, meaning they exhibit fewer large-scale structural variations or widespread genomic alterations compared to the other two archetypes. Their genomic integrity suggests a less aggressive biological underpinning, aligning with their generally more favorable prognosis and better response to standard therapies.
Early Establishment and Pervasive Influence
A pivotal finding of the study is that these defining structural variations and genomic architectures are not late-stage developments. Instead, they are present in the earliest stages of the disease, including ductal carcinoma in situ (stage 0), and are meticulously maintained as the tumors grow, advance, and spread throughout the body. This early establishment underscores the idea that some cancers are "born to be bad," with their aggressive trajectory set from the very beginning. Furthermore, these genomic variations were found to correlate with how immune cells infiltrate and respond to the tumor, highlighting their pervasive influence on the tumor microenvironment and immune evasion strategies.
Transforming the Future of Breast Cancer Care
The profound insights gleaned from this genomic architectural classification hold immense promise for revolutionizing breast cancer diagnosis, prognosis, and treatment. By understanding the foundational importance of these structural variations, researchers and clinicians can begin to unlock new therapeutic avenues and refine existing strategies.
Personalized Prognosis and Treatment Stratification
The ability to categorize breast cancers into these three archetypes based on their intrinsic genomic architecture offers an unparalleled opportunity for personalized medicine. Physicians will be equipped with a more robust tool to:
- Refine Prognosis: More accurately predict the likelihood of recurrence, even decades after initial diagnosis, especially for HR+ cancers previously considered low-risk.
- Guide Treatment Intensity: Identify patients who are most likely to benefit from aggressive early interventions (e.g., intensive chemotherapy, novel targeted therapies) versus those who may be able to safely de-escalate treatment, avoiding harsh side effects for minimal benefit. For instance, patients with Archetype 3 tumors might benefit from less aggressive approaches, while those with Archetype 1 or 2 would warrant more intensive, tailored therapies.
- Tailor Monitoring Strategies: Implement more intensive long-term surveillance for patients with high-risk genomic profiles, while potentially reducing the burden of frequent check-ups for those with stable genomes.
This precise stratification allows for a more nuanced and patient-centric approach, ensuring that individuals receive the right treatment at the right time, optimizing efficacy while minimizing unnecessary toxicity.
Novel Therapeutic Avenues
Understanding the specific genomic vulnerabilities of each archetype opens exciting new frontiers for drug development and repurposing existing therapies.
- Targeting DNA Repair Deficiencies: The finding that a subset of globally unstable tumors, particularly within the triple-negative and some HR+ groups, exhibit DNA repair deficiencies is highly actionable. Researchers speculate that existing drugs designed to target impaired DNA repair pathways, such as PARP inhibitors used in patients with BRCA1 and BRCA2 mutations (which lead to inherited forms of breast cancer), might also benefit the approximately 13% of people with DNA repair-deficient, estrogen-receptor positive breast cancers. This represents a significant opportunity to extend the utility of effective targeted therapies to a broader patient population.
- Attacking Focal Amplifications and ecDNA: For tumors belonging to Archetype 1, which rely on focal amplifications and the presence of ecDNA, new therapeutic strategies could emerge. Compounds that specifically target the mechanisms driving ecDNA formation, replication, or expression, or those that exploit the "replication stress" often induced by highly amplified oncogenes, could prove effective. This approach aims to disrupt the fundamental genomic machinery that fuels aggressive tumor growth in this subtype.
- Interrupting Mutational Processes: Beyond specific genomic features, the research points to the potential of directly targeting the mutational processes that propagate these catastrophic events. This could involve therapies that prevent the accumulation of structural variations or promote genomic stability, thereby hindering tumor evolution.
Dr. Curtis emphasized the profound implications: "Despite the complexity of their genomes, there are constraints and only so many evolutionary paths for a tumor to follow. We now have an understanding of how and when these complex alterations arise and their accompanying vulnerabilities." This knowledge provides clear targets for future drug discovery efforts.
The Promise of Early Intervention and Prevention
The finding that these defining genomic architectural variations are established "decades prior to the diagnosis of the tumor" underscores an extraordinary opportunity for earlier interventions. This not only points to the potential for earlier detection through highly sensitive biomarkers but also raises the tantalizing prospect of intervening in pre-cancerous lesions or even in individuals at high genetic risk before invasive cancer fully develops. By identifying these "born to be bad" tumors at their nascent stages, clinicians might one day be able to neutralize their aggressive potential, fundamentally altering the trajectory of the disease.
Broader Impact on Cancer Research
This study exemplifies the power of integrating advanced genomic sequencing with sophisticated computational biology and machine learning to unravel the complexities of cancer. The methodological rigor and the depth of insight gained from analyzing nearly 2,000 tumors across all stages set a new standard. The principles uncovered here regarding genomic architecture and tumor evolution may also have broader applicability to other cancer types, suggesting a universal framework for understanding cancer aggressiveness. Future research will undoubtedly focus on translating these genomic archetypes into clinically applicable diagnostic tests, validating new therapeutic targets in clinical trials, and continuously refining our understanding of tumor biology.
Christina Curtis is also a member of Bio-X and of the Stanford Cancer Institute and is a Chan Zuckerberg Biohub investigator, reflecting the interdisciplinary nature of this groundbreaking work. The study received vital funding from the National Institutes of Health (grants CA261719 and CA252457) and the Breast Cancer Research Foundation, highlighting the collaborative effort required to push the boundaries of cancer research. This new classification system marks a significant stride towards a future where breast cancer treatment is not just personalized, but truly predictive, offering renewed hope for millions of patients worldwide.
About the Author
