NEW YORK – Agendia has amassed tumor samples, as well as genomic and clinical data, from 6,000 breast cancer patients it has tested within the FLEX Registry, a resource that the molecular diagnostics firm is hoping will be a rich resource for studying groups that have been historically underrepresented in research.
FLEX, which began enrollment in 2017, is already two-thirds of the way to reaching its goal of involving at least 10,000 early-stage breast cancer patients. All the enrolled patients have consented to sharing their tumor samples and clinical information, which are linked to results from Agendia's MammaPrint and BluePrint tests, and full transcriptome analysis, as well as 10 years of follow-up data. Investigators from 85 participating sites can use the data in this growing registry to propose their own sub-studies, which a central committee reviews and approves.
So far, 59 sub-study proposals have been submitted and 27 approved projects are either completed or underway. Nearly half of these approved sub-studies are exploring questions related to biological, genetic, and socioeconomic cancer disparities. The FLEX registry has allowed researchers to delve into research topics that were historically hard to study due to small cohorts in other datasets, such as the differential gene expression patterns between male and female breast cancer patients; the molecular profiles and clinical-pathologic features of Asian breast cancer patients; and the molecular features of African American and Caucasian patients with basal-type breast cancer and obesity.
MammaPrint, BluePrint, and the full genome
As a requirement for enrollment into FLEX, patients must have their breast cancers tested with Agendia's MammaPrint and BluePrint tests. The MammaPrint test uses a 70-gene signature to characterize patients' risk of breast cancer recurrence as either high or low risk, while the BluePrint test uses an 80-gene array to classify tumors into subtypes, including luminal A, luminal B, HER2-type, or basal-type. Together, MammaPrint and BluePrint are commonly used to help guide treatment decisions for early-stage breast cancer patients and have been particularly informative when it comes to guiding treatment escalation and de-escalation, as well as long-term use of endocrine therapy.
While MammaPrint and BluePrint are valuable tools for clinicians, they don't tell the whole story regarding the molecular differences in patients' tumors. According to William Audeh, Agendia's chief medical officer, this was one of the driving factors behind launching the FLEX Registry, within which Agendia also performs transcriptomic analysis on patients' tumor samples and links it to more than 800 clinical data points.
"FLEX allows us to capture that whole-transcriptome data," Audeh said. "Technology has progressed to the point that we can look at all 20,000-plus genes in a cancer based upon the expression of those genes. And that's vital information that was simply being lost every day."
According to Audeh, the same microarray-based technology used to assess the expression of 150 genes collectively in MammaPrint and BluePrint is capable of analyzing patients' whole-transcriptome data, and the potential utility of the information gleaned from this broader analysis has attracted many clinicians to participate in the FLEX Registry. Participating sites now include at least eight National Cancer Institute-designated cancer centers and countless community-based sites from diverse geographic areas of the country.
The diversity of the patient population, as well as the sheer numbers of patients enrolled from these widespread participating sites, is part of what makes FLEX uniquely suited to enable genomics research in underrepresented patient groups. (Audeh suspects that Agendia will not only meet the 10,000-patient minimum enrollment benchmark but exceed it threefold.)
"In many cases, we have specifically gone after certain geographic locations where we know, for example, there will be a better representation of African American women or Latin women," Audeh said. "We know that Latin women in South Florida actually have a different genetic ancestry than the Latin women in California, even though they might all be self-described as Latin. We have gone after these specific geographic locations to make sure that … we have quite a significant number of [minority patients]."
For now, the majority of patients enrolled in FLEX are still white women. However, Audeh said that the registry is currently large and diverse enough to facilitate research into groups that are not white or female.
Genomics research in male breast cancer
Indeed, one approved and completed sub-study enabled by the FLEX Registry looked into differential gene expression patterns between male and female breast cancer patients, and investigated whether Agendia's MammaPrint test, which was developed in female patients, could predict breast cancer recurrence equally accurately for male breast cancer patients, and inform treatment decisions.
Jennifer Crozier, an oncologist with Baptist MD Anderson Cancer Center in Jacksonville, Florida, who led this specific sub-study, said that she has been interested in studying male breast cancer for a while, but did not have access to a large enough sample of these patients prior to her center's participation in FLEX. "It's kind of a dream to be able to have this database and tissue to look at and answer these questions," she said.
Less than one percent of all breast cancer patients are male. "It's a population of patients where we just don't have a lot of knowledge," Crozier said. "[Genomics research] is hard to do when your practice is only seeing up to three male breast cancer patients a year… It's something where we needed this sort of collective registry that FLEX is providing in order to be able to study populations like this."
Once Crozier and her co-investigators proposed their sub-study to the FLEX review committee and received approval, they were able to access data collected from patients enrolled in FLEX, a significant number of whom are male. Crozier and her team specifically analyzed 480 samples — 400 from female patients and 80 from male patients — for which whole-transcriptome microarray data were available.
Within this data, the researchers identified 143 genes that were differentially expressed between male and female breast cancer patients. Eighty-nine of these differentially expressed genes were in patients whose cancers had been characterized by MammaPrint as being at high risk of recurrence, while 248 genes were differentially expressed between MammaPrint low-risk patients.
Pathway analysis showed that genes consistently downregulated in male patients compared to female patients were associated with innate and adaptive immunity, while genes that were upregulated in male patients were associated with hormone metabolic processes such as estrogen response by way of MTORC1 signaling and MYC targets. Crozier believes these findings could affect treatment strategies for male breast cancer patients down the line — potentially in relation to the use of endocrine therapy — though it's too early to say exactly how.
"Something we'll be proposing [next] is, if the patient does have a recurrence, we're going to want to take a biopsy of that [recurrence site] and do genomic testing on it, and look at the difference from before and after the patient has been exposed to endocrine therapy," she said. "We'll learn a lot from that."
This longer follow-up and analysis of male and female patients' tumors at the time of recurrence could lead to additional FLEX sub-studies, Crozier said, because as of now, male and female patients are treated similarly, if not identically, even though most evidence supporting treatment strategies stem from female patients. Learning that genes associated with endocrine resistance are upregulated in male breast cancer patients could, for instance, affect the use of estrogen receptor modulators like tamoxifen in male patients. With additional research in bigger cohorts, downregulated innate and adaptive immunity genes could similarly affect the use of immunotherapy strategies in male patients.
Crozier and colleagues also performed supervised hierarchical clustering of MammaPrint and BluePrint genes, and were reassured to find that there were no distinct gene clusters specific to male and female patients, and there wasn't a difference in average MammaPrint indexes between male and female breast cancer patients classified as low risk and high risk. This indicated that the performance of both Agendia assays is valid in male and female patients.
"We were hoping to find that the test would be equivalent between males and females as far as risk stratification," Crozier said. "That's always our fear when we're doing these genomic tests for male patients … [and] we're extrapolating these data when nearly all of [the validation] has been performed in females."
Exploring genomics by patient ancestry
In another sub-study facilitated by the Flex Registry data, investigators led by Dipali Sharma of Johns Hopkins looked into differential gene expression patterns between African American versus white patients with obesity who had been diagnosed with breast cancers that were basal-type according to BluePrint and high-risk according to MammaPrint. While this study is still ongoing, Sharma and co-authors are hoping to shed light on the influence of metabolic factors and ancestry on patient outcomes.
The data will need time to mature for the researchers to assess longer-term outcomes, but as of now, Sharma and colleagues have used FLEX to discover that there are six genes that are consistently expressed to a higher degree in breast tumors from African American patients versus white patients. The researchers suspect that these highly expressed genes suggest upregulation of Notch-associated aggressiveness — a factor that may have implications for comorbidities like obesity — as well as upregulation of pathways associated with stemness, metastasis, and chemotherapy resistance.
The analysis to date broadly points to additional disparities in African American versus Caucasian patients beyond those attributable to clinical and social factors. According to Sharma, identifying these other disparities may help advance unique treatment approaches.
Another sub-study — this one led by Margaret Chen of ProHEALTH Care — found distinct gene set enrichment patterns in tumors of Asian breast cancer patients compared with Caucasian and African American patients, which may influence clinical outcomes. Myriad other FLEX-enabled sub-studies interrogating genomic differences between patients of different backgrounds are ongoing as well. Agendia's Audeh did acknowledge, however, that one limitation of these studies is the fact that ancestry data in the FLEX Registry are derived from self-reported information.
"Although one can do simple genetic tests to determine what your genetic ancestry is, we haven't been doing that with gene expression," he said, sharing that one future direction for FLEX will be to "get a handle on true genetic ancestry through expression."
Additional sub-studies, future directions
Beyond facilitating genomics research into underrepresented groups, the FLEX Registry has already proven useful as a resource for breast cancer research broadly. For example, FLEX data have allowed investigators to characterize the molecular heterogeneity of triple-negative breast cancers, as well as differential gene expression in luminal-type invasive lobular carcinoma and invasive ductal carcinoma by MammaPrint risk stratification. And another FLEX sub-study presented in June demonstrated that breast cancer subtype classification based on BluePrint analysis can differ significantly from information gleaned from traditional tests, such as immunohistochemistry-based classification.
Going forward, Audeh said that the FLEX Registry could also incorporate data from genomic testing patients receive at the time of recurrence, or after they fail to experience complete pathologic responses to treatment. "We have one sub-study now where we're analyzing the tissue at diagnosis and then whatever cancer is left afterward to look at what caused it to resist treatment and what we can learn from that," he said. And although FLEX is currently limited to patients with stage I-III breast cancer, Audeh said the sample and data collection approach could be emulated for metastatic patients.
Finally, Audeh shared that information gleaned from liquid biopsies may soon be incorporated into FLEX. Blood-based testing is currently being piloted at some of the FLEX sites, allowing researchers to analyze information from plasma, including circulating tumor DNA or the immune cells in the blood. "We're looking to integrate occasional blood samples," he said. "So far, we've only incorporated that into a number of sites, but we also realize that blood gets drawn on patients with breast cancer all the time, and there's a great deal of information contained there. If we can combine it with what we learn from the tumor itself, that would be even more powerful."
Ultimately, both Audeh and Crozier underscored that the value of this registry will grow as it enrolls more patients and a more diverse cohort.
"Our ultimate goal is 30,000 patients in FLEX, and we do think that over the course of five to ten years we will have numbers like that," Audeh said, admitting that patient enrollment fell off slightly due to the COVID-19 pandemic, but that the registry is still on track to reach 10,000 patients within a year. "It's extraordinary to have those kinds of numbers because it really does let us dig into information in subsets that we wouldn't get otherwise," he said.
Beginning in 2021, the FLEX Registry will include three years' worth of follow-up data for many of the patients enrolled, which will allow investigators like Crozier to look at outcomes and ask questions of the data beyond what was possible in the observational sub-studies reported to date.