Breast Cancer - Episode 14
Hope Rugo, MD, FASCO: Before we talk about how we would treat this patient, let’s review some of the data that we have. The major data is from the IMpassion130 trial in the metastatic setting with immune checkpoint inhibitors.
This trial randomized patients—who had metastatic breast cancer that was confirmed triple negative in the first-line setting and were at least 12 months from their adjuvant taxane therapy—to receive nab-paclitaxel with atezolizumab or placebo in a 1:1 ratio. The coprimary endpoints were progression-free survival [PFS] and overall survival [OS] in the overall intent-to-treat population and in the [PD-L1–positive subgroup] based on the Ventana PD-L1 (SP142) Assay that is a 1% or greater positive in immune cells [IC].
As everyone knows, this data led to the first regulatory approval of a checkpoint inhibitor for patients with triple-negative breast cancer. Patients who had PD-L1–positive disease, which represented 41% of the population, had a modest improvement in progression-free survival—going from 5.3 months to 7.5 months—but had a 7-month absolute improvement in overall survival. The statistical design had been based on whether or not you saw an overall survival benefit in the overall population.
You could look at the PD-L1–positive population, but because of the PFS benefit in this population, we did look at the overall survival benefit, which was still clinically important despite not being true to the statistical design. Interestingly, in the patients whose tumors were PD-L1–IC negative, there was absolutely no benefit from adding the atezolizumab. We also looked to see whether or not there was a difference in PD-L1 status in the primary versus metastatic tissues, and it’s difficult to know.
Most patients had tumor collected within 2 months of starting the study’s therapy, so a lot of patients had de novo metastatic disease. It appeared that primary tissue was more likely to be positive—62%—versus metastatic, but it didn’t matter which tissue you used to determine whether a patient was PD-L1 positive. We still saw a benefit from the addition of atezolizumab. Interestingly, the PD-L1 status by anatomic site varied some, and in other tumors too, liver appears less likely to be positive. Although we had a small number of samples, that was true here as well.
We did look at exploratory immune biomarkers to see if there was a difference in patients based on whether their tumor cells were positive, whether the stromal TILs [tumor-infiltrating lymphocytes] were more important. We also looked at patients based on their BRCA status. It turned out that PD-L1 was the most potent predictor of benefit from atezolizumab and that the other factors mattered only if the tumor was also PD-L1–IC positive.
Assessing immune-related adverse events is also a very important part of assessing any new immune targeted therapy. As you can see, rash was the most common toxicity and required systemic corticosteroids. There are other toxicities, however, including almost any organ. They’re just much less common—with about a 1% risk of hepatitis or adrenal insufficiency and about a 2% risk of pneumonitis. Hypothyroidism is very common, but we don’t usually treat it with steroids. That’s why it looks relatively less common here. But when I say “very common,” these rates of toxicity all are at less than 5% overall, although some studies have reported a higher rate of adrenal insufficiency in early-stage breast cancer.
Colitis is also seen, and it’s important to identify this early to avoid serious toxicities. Then recently, some neurologic complications have been reported as well, both central and in the spinal cord. They weren’t seen very frequently, however.
We also looked at concordance between different assays because the 22C3 assay and SP263 assay are used with other checkpoint inhibitors. It turns out that more tumor samples are positive based on these other 2 assays, and they looked at tumor cells and immune cells a little bit differently based on which test was being done. They’re not very concordant, although you include almost all the SP142 cases if you look at 22C3 positivity or SP263 positivity. When we looked again in a retrospective subset analysis in this group of patients who were biomarker evaluable, it still appeared that being SP142 positive was more likely to predict benefit in both PFS and OS with the addition of atezolizumab. You can see that really striking difference in the top curves versus the curves that looked at 22C3 and SP263, where there didn’t appear to be a benefit in overall survival based on positivity.
Then there are a number of studies that are actually looking at immune checkpoint inhibitors in metastatic triple-negative breast cancer. A number of trials are looking at alternate chemotherapy partners, such as paclitaxel. And then in patients who have relapsed earlier after their adjuvant or neoadjuvant therapy, there are studies exploring the use of platinum gemcitabine. There are also a couple studies looking at pembrolizumab.
One trial looked at pembrolizumab as a single agent compared to a menu of different chemotherapy options and has been reported and showed no difference, but there was a suggestion of a benefit in patients who had very high scores for PD-L1—20% or greater. KEYNOTE-355 is a phase 3 trial, and a press release from February 12 reported that pembrolizumab plus chemotherapy significantly improved PFS compared to chemotherapy alone, but only in the population of patients whose tumors expressed PD-L1 with a CPS [combined positive score] of 10 or greater. We were looking at 1 or greater as a standard positive test.
This trial is the first of its kind to be reported and looked at alternate chemotherapy partners in patients who relapsed 6 months or longer after their adjuvant taxane. It does potentially open up a new window for treatment options in patients who relapse early, and we expect to see this data presented at an upcoming meeting.
Transcript edited for clarity.