Hope S. Rugo, MD: There have been some studies looking at combinations of PI3K [phosphatidylinositol 3-kinase] inhibitors and other treatment, CDK4/6 [cyclin-dependent kinases 4 and 6] inhibitors, and so forth. It seems as though most of those have met with too much toxicity, so we'll see where that goes, but there is interest in using PI3 kinase inhibitors in other settings too. Priyanka, I know that you've led some of the work in that area.

Priyanka Sharma, MD: Yes, we evaluated the specific PI3 kinase inhibitor with nab-paclitaxel in metastatic HER2 [human epidermal growth factor receptor 2]-negative breast cancer. It also has encouraging efficacy in the phase 2 trial. Like Hope said, we recommended prophylaxis with an antihistaminic for most of our patients, and that led to a really low-grade manageable rash situation. We also asked our patients to have diabetic education right up front, at least in the context of that trial. I think that helped quite a bit.

If we can teach patients ahead of time how to control the carbohydrates that's helpful, and we also have some preclinical data that suggest that some type of ketogenic diet, low carbohydrate intake, might add to the efficacy of these inhibitors. Of course, we don't know that from clinical trials, but certainly in terms of being able to tolerate the drug, that helps.

Hope S. Rugo, MD: I know there is some concern that giving insulin might abrogate to some degree the effects, and certainly that's for PI3 kinase inhibitors, and although seen in animal models, we really don't have any idea about that in the human setting, I would say. You have to treat the diabetes, but there is interest in the oral agents like the SGLT2 [sodium-glucose cotransporter 2] inhibitors as well, so that'll be interesting to see. We want to get John involved and his comments in this area because it’s so important for all of us, and also some feedback from all of you when we hear what he has to say.

From a formulary or medical policy perspective, John, how are you differentiating breast cancer at various stages of disease progression? For example, how do you decide on CDK4/6 inhibitors, first in everybody, for PIK3CA mutations? What evidence is needed to support the use outside of the label? Also, when you have multiple agents in the same class like CDK4/6 inhibitors, how do you decide which one is OK to use?

John Fox, MD, MHA: There are a lot of questions in there. Let me first step back and say that I think most payers have agreed that they're going to follow NCCN [National Comprehensive Cancer Network] guidelines in choosing which therapies to cover. There are some payers in some states that don't follow NCCN guidelines, but most payers have agreed to use these guidelines. Almost 30 states have mandates to cover FDA-approved cancer therapies. Certainly Medicare and Medicare Advantage Plans have to follow the NCCN guidelines or another compendium.

Their drugs are in a protected class. Cancer drugs are in a protected class. So I think the question more commonly is not whether to cover therapies, but under what circumstances. I think the advent of molecular diagnostics has really helped us and increased the probability that patients get on the therapy that is most appropriate for their cancer type. One of the common questions is, can you get paid for doing molecular diagnostic testing? And interestingly, in patients who need IHC [immunohistochemistry] testing, for example, nobody's evaluating those, and most commonly payers aren't looking at single gene mutations.

If you're looking for a single gene mutation like a PIK3CA mutation, those are covered. It’s typically when you get into these very costly next-generation sequencing tests, for example, if you have to do a liquid biopsy because you don't have tissue and have to use a Guardant laboratory test or a Foundation Medicine test, those tests are prior authorized to determine that they're medically appropriate. Interestingly, I work in a regional delivery system. We don't perform these tests locally.

I think everyone here probably does tests in their own institution and may or may not be concerned about getting testing covered. But I think that's a significant issue and one that we have been on the forefront of trying to ensure that people get tested so that we can make sure that they're on the therapy that's most appropriate given their tumor mutation status and stage of disease. To come back to your questions, I think for patients who are PR [progesterone receptor]-positive, if a provider chooses to use a given CDK4/6 inhibitor, that's their prerogative to use, and it's also their prerogative in most delivery systems which one to use.

There was an indirect analysis performed last year published in Breast Cancer Research and Treatment that compared all 3 CDK4/6 inhibitors, and it showed that they had equal efficacy in the first-line and second-line setting. So a priori, we as a plan don't have a preference over which one is used. There are some differences in the toxicities. I think abemaciclib had higher diarrhea rates, and Kisqali [ribociclib] has higher rates of QTc [corrected QT interval] prolongation. In those circumstances it would be hard to tell a provider that you can’t use palbociclib first line when the others have toxicities.

There are many health plans today that have incentive programs or prior authorization programs around these drugs to try to steer providers to a select drug, but there aren’t many circumstances that I'm aware of where a provider is precluded from using their preferred CDK4/6 inhibitor. With that said, in the traditional Medicare population, which many of these patients are traditional Medicare beneficiaries, a provider is free to choose whichever regimen they prefer.

Hope S. Rugo, MD: That's actually really great to hear. Of course, we do know that there aren’t head-to-head data on the CDK4/6 inhibitors yet, and I hope there never are because I hope that money is put into studying new agents in new settings rather than to comparing the different agents. I just think that that would be such a huge study.

John Fox, MD, MHA: The cost of doing the study would be prohibitive. I agree with you that the findings may not change providers' opinions around which therapy is best. There can be real-world data that are examined. That has some challenges in the coding accuracy, but that might be the best place to identify toxicities and overall survival.

Hope S. Rugo, MD: I think that it may be for toxicities. I think it's really hard to look at clinical end points in real-world data in terms of overall survival, though it certainly had been a great interest. I think on the clinician side we've seen a lot of issues with that data.

Transcript Edited for Clarity