Joyce O’Shaughnessy, MD: Unfortunately, resistance occurs over time. I know we're going to talk about your BYLieve data in a moment and targeting the PI3 kinase pathway, but broadly speaking, what are the major mechanisms of resistance that people are thinking about in terms of designing the next iteration of clinical trials to try to overcome resistance to CDK4/6 [cyclin-dependent kinases 4 and 6] inhibitors?

Hope S. Rugo, MD, FASCO: This is such a great question, and there's been a huge amount of work on it. I think the advent of ctDNA [circulating tumor DNA] is helping us a lot, looking at cell-free DNA to try to have more accessible, real-time evolution of the genomic changes that predict resistance. I think it's important to keep in mind that we know there are also protein changes in mRNA [messenger RNA], not just the mutations in DNA. We really do not have great data on that, and that would require serial biopsies, which we really have a lot of trouble doing. Also, we believe that potentially one place you biopsy may not be totally reflective of the other place that you biopsy.

Our data are limited by this, but I think that as we're able to use big data and try to combine different mutations, we may be able to look at patterns of mutations and pathways that are affected, rather than single mutations. Albeit that we're limited right now by looking at single mutations rather than pathways, it's been fascinating to see that some patients have RB [retinoblastoma] loss that results in resistance. And it does seem like this group of patients who have loss of retinoblastoma may be severely resistant to CDK4/6 inhibitors, and maybe we should take another pathway in that particular group of patients.

A small group of patients have these mutations up front; they’re more likely to be acquired. We also know that patients who have PIK3CA mutations, for example, benefit to a similar extent from CDK4/6 inhibitors but have a shorter duration overall of progression-free survival [PFS]. We've seen upregulation of cyclin E, or the CDK2 pathway, as well that seems to result in resistance over time. Then there are a number of other mutations that have come up. It's really interesting.

You get this genomic chaos over time, and as you see more mutations accumulate, it has been correlated with more resistance. We saw that in BOLERO-2, where in our subpopulation of patients who had multiple mutations, they faired very poorly. We weren't looking at serial mutations or much ctDNA because it was a new technology, but I think that trying to reverse these mechanisms of resistance has been difficult.

For example, we know that FGFR can be mutated, and that's a resistance mechanism with very interesting data by Carlos Arteaga, MD, and a number of investigators have been looking at this. The drugs that target FGFR receptors are plagued by toxicity, and that's where there have been really nice preclinical data looking at the combination of CDK4/6 inhibitors and inhibitors of PIK3CA.

Again, those combinations have been plagued by toxicity. The only study that really showed that you could tolerably combine agents and see some benefit potentially was the TRINITI data, which were presented last year at ASCO [the American Society of Clinical Oncology annual meeting] by Aditya Bardia, MD, MPH. We participated in that trial. The majority of patients had received prior CDK4/6 inhibitors and had progression; they received ribociclib, a tiny dose of everolimus, 2.5 mg, and exemestane.

ESR1 mutations aside...and CDK4/6 inhibitors do not seem to care about ESR1 mutation so much, it's the hormone therapy, and they may overcome some degree of ESR1 mutation resistance. In these patients, 96 who had prior CDK4/6 inhibitors, the clinical benefit rate was 41%, and it was interesting that the progression-free survival was 5.7 months, much longer than we would have expected from giving a mini dose of everolimus and continuing a CDK4/6 inhibitor and giving exemestane. I thought that this was an intriguing triplet combination.

The problem is that the everolimus is off patent and people aren’t so excited because they want new agents. I think we're going to see a whole new class of drugs be included, the Akt inhibitors and these triplet combinations. There is also a lot of interest in just continuing a different CDK4/6 inhibitor after progression. You might want to be able to separate out the group of patients for example, who have RB loss, but it's such a tiny group, you could even do that retrospectively. There are a few ongoing trials looking at that, and there's been a lot of interest stimulated by a collection of data that, Joyce, you participated in where patients received a CDK4/6 inhibitor after progression. Those data were presented at ASCO last year, too.

Joyce O’Shaughnessy, MD: Aditya Bardia took the lead on that, and there were 5 institutions. We went through and did very in-depth reporting of the patient characteristics and the sequence of treatment. The bottom line was there were about 55 patients or so who had mainly had palbociclib, and upon progression whether they had benefited or not, at some point they had had abemaciclib subsequently, either immediately subsequent to palbociclib or at a later date. If you look at the overall population, again this is a case series, not a prospective trial, about 36% of patients had clinical benefit, so stable disease mainly, for 6 months or longer.

That's certainly been my practice experience. Interestingly, and I wouldn't have predicted this, the progression-free survival was longer with abemaciclib immediately following palbociclib, probably because it was an earlier line of therapy as opposed to getting it later after some subsequent chemotherapy, where there still was a group that benefited, but it was a shorter PFS. I guess that does make sense. I was always trying to separate them, similar mechanisms of action.

There is some degree of noncross-resistance there. Preclinically, abemaciclib has good activity against CDK2 and to some extent CDK1. There is some debate about whether the levels that are reached in humans will effectively block those other CDKs, and CDK9 as well. At any rate, it basically has a broader mechanism of action. It may be picking up some of these mechanisms of resistance that you were mentioning, Hope.

That's a case series, and there are other prospective trials going on looking at that a bit more formally.

Transcript edited for clairty.