Researchers and clinicians have known that breast cancer is a heterogeneous disease for more than a decade. They've also known that recognized breast cancer subtypes are associated with particular disease outcomes, treatment responses, and mortality rates. What's been more recently contested is whether triple-negative — that is, estrogen receptor-, progesterone receptor-, and HER2-negative — tumors are "effectively the same" as basal-like tumors in terms of disease outcome, according to Paul Pharoah, a senior clinical research fellow at the Cambridge Cancer Centre in the UK.
"Because triple-negative tumors are relatively uncommon, very large sample sizes are required to get statistically robust answers," Pharoah says. To get those numbers, he and his colleagues at the Breast Cancer Association Consortium initiated a large, multi-center collaboration, which spans half the globe — from the Netherlands to Australia — and brings together many principal investigators and extensive data sets. The dozen cohorts from this consortium supply data from 10,159 cases.
Pharoah and his collaborators deliberately analyzed their samples independent of any systemic adjuvant therapies, so that molecular markers assayed effectively determined the disease outcome status, rather than the choice and duration of therapies. "We believe it's important that the different subgroups are investigated within clinical trials," Pharoah says. In addition to better predicting treatment response, "you can use subtypes to estimate risk of recurrence more accurately than current models," he says.
Pharaoh's group has shown that a certain subgroup of the ER-positive disease has a constant mortality rate over time, which has implications for long-term clinical follow-up. Even though these women have a good prognosis, "they're never really 'cured' as such, because you never know that the recurrence rate is going to decline," he says. Conversely, some tumors that are commonly acknowledged as poorer-prognosis malignancies have mortality rates that fall over time. "In a sense, if you survive the initial high mortality, you'd almost be considered as being cured," Pharoah says.
He adds that one of the more prevalent criticisms of his study might also prove to be its greatest strength: the fact that "these markers were all carried out in different labs, at different times, using different reagents" isn't an indication of poor standardization across cohorts as much as evidence of the common markers' robustness. "If you get the same answer, even though they've been done in a whole lot of different labs, using slightly different standards, it suggests that it would be robust in clinical practice — where, of course, tests are also done in different labs using different standards, and so on," Pharoah says.
Going forward, Pharoah says that the team intends to expand the number of studies involved and mine patients' medical records for treatment-specific information.