Breast cancers are rare diseases—and must be treated as such

With over 1.5–1.7 million new diagnoses of breast cancer worldwide each year, and at a time when 25% of new cancers diagnosed in women annually are breast cancers, it seems completely counter to current understanding and dogma to represent breast cancers as rare diseases. Yet, from the earliest days of molecular analysis of breast cancer, 1 through the molecular subclassification of breast cancers using transcriptomics 2 to current multi-omics data, 3 scientific advances are constantly challenging our view of breast cancer (and other diseases) as single entities. Clearly today, there exist multiple disease entities, defined at a molecular level but combined under the site of origin in the breast, some if not all of which may represent rare diseases. What is a rare cancer? Definitions of “rare” cancers vary widely, even though many are based on a common statistical approach measuring annual incidence expressed per 100,000 individuals. Bodies including the World Trade Center health program ( http://www.cdc.gov/wtc/pdfs/WTCHP_PP_RareCancers05052014.pdf/ http://epi.grants.cancer.gov/events/rare-cancers/ ) suggest incidence rates of 15/100,000 population/annum would be sufficient to define cancers as rare, whilst Rare Cancers Europe ( http://www.rarecancerseurope.org/About-Rare-Cancers ) and the European Society for Medical Oncology are more stringent, setting the rate at 6/100,000/annum. Most stringent are definitions set by the IRCI ( http://ecancer.org/journal/editorial/20-international-rare-cancers-initiative-irci.php ), which suggest rare cancers are confined to those with incidences of 2–3/100,000. Other definitions of “rare cancers” are less well characterized including “difficult to treat” cancers as “rare”. Could breast cancers be rare? Breast cancer incidence in the US/Canada is around 155/100,000 women per annum, therefore if breast cancers were evenly subdivided in as few as 11 subgroups (using the broadest definition), these subgroups would individually represent “rare cancers”. Should the distribution of subtypes be uneven, as we would expect, the reality of rare breast cancers would become apparent with far fewer subdivisions. In fact, we can probably already identify rare breast cancer subtypes. Breast cancer subtypes–nothing new under the sun Since the work of Perou and Sorlie et al. over 15 years ago, 2 molecular subtyping of cancer has been both a driver of novel research approaches and an intense debating point in the clinical management of cancer. The identification of multiple subtypes presents a challenge both to researchers seeking to investigate breast cancer, and to clinicians seeking to treat this disease. Phrases like “one size does not fit all” or “divide and conquer” recognize this challenge intellectually but the question remains: Have we truly integrated the current knowledge base around multiple cancer subtypes and recognized the fundamental shift in both preclinical and clinical research strategy required for future progress in clinical management of breast cancer required by current evidence? We should note in passing that the work by Perou, Sorlie et al., however transformative, was far from the first to demonstrate the existence of “subtypes” in breast cancer. From the work of Beatson in 1890, through the development of targeted therapies such as anti-ER and anti-HER2 therapies from the 1960s–1990s, “subtyping” by oestrogen receptor (ER), PgR, and human epidermal growth factor receptor homologue 2 (HER2) measurement and other pathological features (e.g., classification of subtypes such as lobular carcinomas, etc) was central to both clinical management and preclinical understanding of breast cancer. This point is absolutely critical since it places Sorlie’s work correctly as a pivotal milestone in the continuous development of the understanding of the complex biology of breast cancer, rather than the final delineation of breast cancer subtypes. Recognizing this fundamental truth allows us to develop a broader molecular and functional mapping of the disease entities included under the umbrella of breast cancer and move towards a stratified and ultimately personalized medicine approach with the potential to rapidly impact patient management and survival. How many breast cancer sub-types? Work on subtyping has progressed rapidly over the past few years, increasingly recognizing the need to integrate prior knowledge with novel approaches. Illustrative of the molecular diversity within existing breast cancer subtypes is work investigating the molecular stratification of lobular breast cancers. Lobular cancers represent a subset of predominantly ER-positive breast cancers which are pathologically distinct from ductal carcinomas. They are also molecularly distinct from ductal ER-positive breast cancers, exhibiting in addition to CDH1 mutations/loss, markedly different mutational spectra to ER-positive ductal cancers. Pathologically and molecularly multiple subtypes of lobular breast carcinomas are identified. 4 This may reflect differential clinical behavior of lobular breast carcinomas both in comparison to ductal carcinomas of the breast and within subsets of pathologically defined lobular carcinomas. Likewise, broad molecular subgroups defined by Sorlie and others may be increasingly sub-divided with both clinical and molecular consequences. Amongst the 15–20% of newly diagnosed ER-negative breast cancers, approximately 50% are HER2-amplified and eligible for treatment with HER2 targeted agents. For the “triple negative” (ER, PgR, and HER2 negative) breast cancers, further sub-division into basal (CK5/6, EGFr expressing cancers) or 6 further molecular subgroups 5 , 6 also suggests a degree of complexity not fully explained by the first generation of molecular profiling experiments. Finally, comprehensive molecular profiling approaches across multiple tumor types are providing further striking insights into molecular subtyping of cancers, suggesting classification of breast and other cancers across more than 30 molecular subclasses. 5 , 7 , 8 , 9 This evidence shows that breast cancer represents a far more complex panel of diseases than previously understood. Whilst for many approaches, the clinical utility of subtyping may require additional developmental work further sub-division of breast cancer appears inevitable. Whether we ultimately arrive at 10, 20 or as many as 50–100 sub divisions of disease, some, if not of most, subtypes will surely satisfy the definition of “rare cancers”. Does it matter? The pivotal question is not how many subtypes of breast cancer exist, but rather what, if any, consequences the further subdivision of breast cancer might have on future clinical and preclinical research strategies? Put simply–are the strategies developed for evaluating novel treatment approaches when breast cancers was regarded as a single disease still viable and appropriate if breast cancer fragments into multiple cancer sub-types? Whether these cancers sub-types are rare or not may be secondary to the challenge of adapting research strategies to enable rapid progress towards a stratified and personalized medicine approach. In the late 20th century, innovations in the adjuvant clinical management of breast cancer relied on ever larger phase III clinical trials seeking to identify increasingly marginal survival improvements for patients. Pivotal breast cancer trials over the past decade recruited many thousands of patients and were powered to detect modest (often 3–5%) improvements in absolute risk across these patients. Clearly such large trials are neither logical nor feasible in small cancer sub-groups as defined above. Indeed the entire approach to breast cancer trials might require re-evaluation. To explore this question, let us accept that lobular cancers (ca 15% of all new breast cancers) represent at least 6 subtypes, that triple negative breast cancer represents a further 7 subtypes, and that 10–20 subtypes exist for ER positive breast cancers. 5 , 7 , 8 , 9 Let us further assume that these subtypes are evenly distributed within the cancer spectrum. Now, each of the 6 lobular carcinoma subtypes has an incidence of <4/100,000 per annum. For triple negative breast cancer (TNBC) subtypes incidence is <3/100,000. Only 13 subtypes of non-lobular luminal breast cancers are required for these to have incidences of <6/100,000 per annum. Each subtype would then represent between 4500–9000 new cases in the US in 2016 (based on statistics at http://www.breastcancer.org/ ), roughly equivalent to the number of new testicular cancers. Clearly in this context, existing trial designs and approaches would not be logical, unless we persist in ignoring the message of molecular diversity in breast cancer.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.