PALOMA-3 itself was a phase III trial where patients with hormone receptor positive, HER2 negative advanced breast cancer whose disease was progressing on an aromatase inhibitor and who were allowed to receive any lines of endocrine therapy and up to one line of prior chemotherapy, were randomised 2:1 to receive palbociclib or placebo with fulvestrant. The primary endpoint of that study was progression free survival. Patients were stratified based on their sensitivity to endocrine therapy, menopausal status and the presence or absence of visceral metastases.

The primary goal of PALOMA-3 was to evaluate the benefit in terms of progression free survival of adding the CDK4/6 inhibitor palbociclib to fulvestrant in this study population. We actually did this analysis that was presented at EBCC in 2020 where we wanted to look at subgroups that might impact overall survival for the entire trial. So our objective in this specific study was to look at the variables in terms of treatment in this patient population to understand the benefit of palbociclib and the impact on overall survival.

What were the key results of this study?

What we had looked at when we evaluated progression free survival, the primary endpoint for PALOMA-3 was that the progression free survival in the control arm was shorter in patients who received fulvestrant alone with their placebo in the PALOMA-3 population, compared to the other two second line trials that looked at a CDK4/6 inhibitor with fulvestrant. If you just looked at the control populations the progression free survival was half.

So clearly the demographics or patient characteristics of the control population were impacting response to endocrine therapy with fulvestrant. When we went back and looked at what factors could be impacting progression free survival, the mean difference between the three trials where PALOMA-3 was the first trial that was done in this setting and in that way was really designed differently, it was designed to be more all-inclusive. So patients were allowed who had received up to one line of prior chemotherapy. The subsequent trials eliminated that population so they didn’t allow any prior treatment with chemotherapy in the metastatic setting in order to be eligible for the trial. We know that about a third of the patients in PALOMA-3 received chemotherapy.

The second thing that was different with PALOMA-3 was the number of lines of endocrine therapy, so PALOMA-3 allowed any line of endocrine therapy. But those differences were felt to be relatively modest because actually at the time these studies were done there weren’t a huge number of options after progression on an AI for endocrine therapy.
So we went back to look at this in a subgroup analysis. If we just looked at the patients who had prior chemotherapy we understood that this was about a third of the patients in the trial. There were different characteristics of those patients as well. So we understood that patients who got chemotherapy tended to have more other aggressive features of their cancer.

Then when the overall survival analysis was first presented and then published for PALOMA-3 we saw no difference based on the pre-specified p-value required in terms of overall survival with the addition of palbociclib. Although numerically the median overall survival in the placebo group was 28 months and in the palbociclib arm was 34.9 months, the p-value did not meet statistical significance.

However, when we looked at the endocrine sensitive population that represented 79% of the intent to treat population, in that situation the hazard ratio did not cross 1, the confidence intervals did not cross 1 and the unstratified hazard ratio was 0.72, very similar to the other trials. So it’s really interesting, the numerical difference went from 29.7 to 39.7 months.

So then we were really interested in why we saw this difference in the endocrine sensitive population but not in the intent to treat population. So when we went back to look at the group of patients who had received chemotherapy or who had not received chemotherapy what was interesting was that for patients who received chemotherapy, they were more likely to have had visceral disease and certainly these patients were less likely to have endocrine sensitive disease. They had more lines of therapy for advanced breast cancer. So, for example, the patients who had no prior chemotherapy, the majority had received only one prior line of therapy for advanced disease. But for patients who had received chemotherapy almost all of them, three-quarters, had received two to three or more lines of therapy for advanced disease. So much more heavily pre-treated and, as I mentioned, more patients had visceral disease as well and there were more involved sites of disease. So three or more sites, which is recognised as a poor prognostic factor was increased in the number of patients who received prior chemotherapy, a third of the patients.

In addition, if you looked at the group of patients who had endocrine sensitive disease, this was different. In other words, patients who received prior chemotherapy were less likely to meet the criteria of endocrine sensitive disease. So then we thought maybe really it’s not endocrine sensitivity, it was just the first thing that was looked at. Really, it has to do with chemotherapy before treatment generating overall resistance.

So we looked actually at the group of patients who had received chemotherapy versus not. So, to remind you, 34% of the intent to treat population had received prior chemotherapy for advanced disease. In that group of patients the overall survival was just 26 months and there was no difference between patients receiving palbociclib or placebo with fulvestrant. The hazard ratio actually was 0.91. However, in patients who had not received prior chemotherapy that represented 66% of the intent to treat population, the hazard ratio was 0.75 and the difference in overall survival was 10.2 months favouring those patients who received palbociclib.

If we further went down and looked at endocrine sensitivity, now we’re now looking, obviously, at a smaller group. So if you looked at the endocrine sensitive group this difference was still maintained, which was interesting. There was no difference in the patients who had received prior chemotherapy but a difference of over 10 months in those who had not received prior chemotherapy.

If you looked at patients who had visceral or non-visceral disease, in patients who had visceral disease with no prior chemotherapy and received up to two prior regimens actually the unstratified hazard ratio was 0.74 although the number of patients now is quite small and the confidence interval does cross 1. It’s actually very interesting that even in patients who had visceral disease there was a hint of a difference. If you look at the patients who had non-visceral disease and no prior chemotherapy with less than two prior regimens the difference is even more striking. Unfortunately because of the inclusion of prior chemotherapy it’s hard to differentiate the differences between visceral and non-visceral disease.

Nonetheless I think that this is a very powerful predictor of benefit from CDK4/6 inhibitors and probably correlates with multiple mechanisms of resistance. So the patient group who needed prior chemotherapy, whether or not they needed it or not is hard for us to evaluate but obviously the treating physician made that determination, those patients seem to have developed multiple mechanisms of resistance that means that the benefit of the CDK4/6 inhibitor is reduced in terms of overall survival. We already saw that the control arm also didn’t do as well when they had received prior chemotherapy with half the progression free survival.

So it’s a lesson to us that patients who have metastatic disease should receive as many sequential lines of endocrine therapy as possible before we go on to chemotherapy. In particular, targeted therapy is more likely to provide benefit for patients earlier in the course of their treatment and before they receive chemotherapy. This is important because in general CDK4/6 inhibitors have not resulted in an impact on quality of life whereas chemotherapy clearly does. So patients should be treated with endocrine therapy with a CDK4/6 inhibitor before they go on to chemotherapy except in exceptional cases where patients have immediately life-threatening disease with organ dysfunction.

One last thing to mention in summary in terms of what we evaluated was we also looked at some of the other subgroups of patients and we saw this benefit in the non-chemotherapy treated patients again, regardless of visceral or non-visceral disease. But also what was interesting was there were some differences between the premenopausal, perimenopausal versus post-menopausal patients. Premenopausal patients tend to have more visceral disease and meet more criteria for endocrine resistance. But, importantly, more of those patients actually receive chemotherapy because the treating physicians have believed that those patients may have more benefit or more need of chemotherapy. This is no longer true. We’ve seen that CDK4/6 inhibitors result in an overall survival benefit in the first line setting and it’s very important for us to keep this in mind and not to use an induction with chemotherapy except for in those isolated cases where we have no choice. In general, patients will benefit more from receiving endocrine therapy and a CDK4/6 inhibitor up front.

What can the future impact of these results be on the treatment of HR /HER2- breast cancer?

We learned a lot, actually, from this analysis and from the comparison of the control arm of PALOMA-3 to the other second line studies. One is that probably giving a CDK4/6 inhibitor as early as possible in the treatment paradigm is likely to result in the best benefit to our patients.

The second thing we learned is that the administration of chemotherapy before endocrine therapy and a targeted agent, in this case the CDK4/6 inhibitor palbociclib, impacts the benefit that we can expect to see from adding the CDK4/6 inhibitor. This is the only trial that allowed prior chemotherapy before randomising to use of a CDK4/6 inhibitor or not. So it is our only dataset and there is no way to do cross-trial comparisons because of this difference in eligibility that clearly impacted the benefit of the CDK4/6 inhibitor in terms of overall survival.

If a patient has received prior chemotherapy they still benefit in terms of progression free survival and there may be subgroups that benefit from overall survival. But what we see here is that the addition of palbociclib to fulvestrant in patients who received any number of lines of endocrine therapy and chemotherapy in a third of the patients demonstrated an overall survival benefit when we look at the subset of patients who did not receive chemotherapy.
That really does direct our treatment for patients with hormone receptor positive, HER2 negative metastatic breast cancer. We should be using the targeted agent in combination with endocrine therapy as early as possible in the treatment paradigm in order to get the best benefit in overall survival and the best quality of life for our patients.