Balancing Genomic and Clinical Risk in HR+ Breast Cancer

Nina D'Abreo, MD

In order to effectively tailor therapy to women with early-stage, hormone receptor (HR)—positive, HER2-negative breast cancer, equal consideration of a patient’s genomic and clinical risk is required, explained Nina D’Abreo, MD.

“There are some data suggesting that traditionally high-risk patients who have a low genomic risk may not benefit substantially from chemotherapy. This [information] is incorporated into our assessment of women with node-positive disease,” said D’Abreo. “We’re learning that it’s not just about genomic risk or clinical risk, it's going to be a combination of both that will ultimately decide who receives what therapy.”

In an interview during the 2019 OncLive^® State of the Science Summit™ on Breast Cancer, D’Abreo, medical director of the Breast Program at NYU Winthrop Hospital, clinical assistant professor, Department of Medicine, NYU Langone Health, explained how treatment is being tailored to women with early-stage, HR-positive, HER2-negative breast cancer.

OncLive^®: Could you discuss the focus of research in early-stage HR-positive, HER2-negative disease?

D'Abreo: Patients with HR-positive, HER2-negative breast cancer are the most commonly seen group of patients in oncology. The holy grail has been to not overtreat or undertreat these patients. We're trying to find a balance and trying to personalize therapy for this group of women.

What is known so far about these strategies?

For escalation of treatment, we’re looking at additive approaches. We're particularly interested in doing this among young, premenopausal women. These women have long lives ahead of them. In this [population], disease recurrence is as important as overall survival (OS), so one of the strategies for these women is the addition of ovarian suppression to maximize the endocrine therapy they’re receiving.

I talked about the updated 8- and 9-year data from the SOFT and TEXT trials to [illustrate] where we should really be applying these strategies. We've seen from the SOFT/TEXT trials that there's a certain group of women—especially high-risk women who are less than 35 [years of age]—who benefit the most [from chemotherapy]. It's important to consider the adverse events (AEs) [of treatment], but the data [detailing a more aggressive approach in this population] are very compelling.

Conversely, there are women who we generally don't want to give chemotherapy to. We’re trying to personalize therapy and make the right treatment plan for each patient. In that context, I highlighted the TAILORx and the MINDACT trials. These trials are using genomic risk to determine who should receive chemotherapy versus who does not. It's important for a practicing clinician to know when to apply these genomic tests. The “devil” is always in the details. You may apply one genomic test in a certain population and not in another.

Extending therapy is a form of escalating therapy. [We’re trying to determine] whether postmenopausal women or women who have completed 5 years of treatment [require additional therapy], and for how long: 7.5 years or 10 years? Several trials have shown a disease-free survival (DFS) [benefit with this approach].

Could you expand on some of those trials?

The trials that are evaluating extended therapy have looked at several patient populations. The patients who started treatment with tamoxifen are the subject of the aTTom and ATLAS trials. These women received 5 years of tamoxifen. The trials were conducted to see who had benefit from an 5 additional years of tamoxifen. Generally, women who receive 10 years of tamoxifen seem to benefit from the [additional 5 years of treatment].

If you started with an aromatase inhibitor (AI) as in the MA.17 trial, or with tamoxifen as in the National Surgical Adjuvant Breast and Bowel Project trials, patients generally had a DFS benefit, but less of an OS benefit. Similarly, patients who start with an AI and continue with an AI may experience some benefit in terms of DFS but not OS. Lastly, there are patients who start on an AI, and we’re trying to figure out whether they need to continue treatment for 10 years. One of the trials in that space is the IDEAL trial, where patients receive tamoxifen or an AI to begin with and [are followed to see] how long they need to receive an AI. The trial data suggest that 7.5 years may be just as good as 10 years in that space.

How are genomics guiding escalation and de-escalation strategies?

OncotypeDX and MammaPrint are the 2 main platforms that are universally known. However, there are several genomic platforms available. OncotypeDX is the 21-gene assay and MammaPrint is the 70-gene assay. It's important to be aware that they were studied in 2 different settings. The 21-gene assay was the subject of the TAILORx trial, which looked at women with recurrence scores between 11 and 25; this is the so called "intermediate-risk" group. We found that women above the age of 50 with risk scores below 25 really didn't benefit from chemotherapy; the same didn't necessarily to apply to women under the age of 50.

What's interesting is the recent re-application of clinical risk factors to genomic risk. In doing so, we can further parse out who among these really high-risk women with intermediate scores benefit from chemotherapy. If women have scores between 16 and 21, but they have a low clinical risk, they may not need chemotherapy. This has helped us tailor treatment to the women who really benefit from chemotherapy.

In the MINDACT trial, MammaPrint was used. Again, we saw a head-on assessment of clinical risk and genomic risk. The trial looked at women who have discordance between traditional clinical risk factors and genomic risk factors. Here, women were randomized to receive either endocrine therapy or chemotherapy plus endocrine therapy. It was up to the investigator whether they followed the patient’s genomic risk score or their clinical risk score. The study showed that patients with high clinical risk and low genomic risk had a small benefit from chemotherapy. Again, the “devil” is in the details.

The rate of distant metastasis-free survival had to be [at least] 92%, which it was. The difference between the patients who received chemotherapy versus those who only received endocrine therapy was [approximately 1.5%]. Most practicing clinicians would agree that’s a small difference to subject patients to chemotherapy. Importantly, patients with 1 to 3 positive lymph nodes were included in the trial, albeit a small number; these women are the focus of the ongoing RxPONDER trial.

How are you selecting which assay to use and when to use it?

The studies give us pretty clear guidelines. Furthermore, ASCO’s recommendations are pretty clear that OncotypeDX is the genomic assay of choice for women with node-negative disease to determine the women who will benefit from chemotherapy versus those who will not.

The MammaPrint assay is to be applied in the context of clinical risk. It became quite clear from the trial that it was the patients we apply the high clinical risk [module] to who would benefit from knowing what their genomic risk is. It boils down to us knowing how to define clinical risk and then applying the molecular scores to that.

Are there any emerging strategies that you would like to highlight?

For women with high-risk disease, we’re evaluating whether we can optimize their endocrine therapy from the very beginning—specifically, by adding CDK4/6 inhibitors to endocrine therapy. If we do that early on, these patients may not need extended endocrine therapy. We have an opportunity to parse out genomic risk among clinically high-risk patients to decide whether to add [other agents] to endocrine therapy. There are also data with mTOR inhibitors we're excited to see. Those are going to be the next choices for us.

What is your take-home message regarding the treatment of patients with early-stage, HR-positive, HER2-negative disease?

It is an exciting and a confusing time for us. I don't believe we can clearly divide between genomic and clinical risk. [We need to focus on] a combination of both. It's great to have an idea of where to apply which test, but in the end, it boils down to knowing where to apply the right test—not just knowing the data. Patient preference is ultimately the most important thing. The one thing that I didn’t mention is that some of these escalated therapies have AEs. Adherence rates can be low. Patients with extended risk, for example, could have higher risk of bone fractures and cardiovascular [events]. Parsing out the risk to benefit ratio is the most important thing for patients in the clinic.