Among all patients who had a pCR, regardless of regimen, the 5-year EFS rate was 90.3%, compared with 55.7% for those who did not receive a pCR, translating into a hazard ratio of 0.19 for pathologic complete responses (P < .0001).
Looking at treatment-related adverse events occurring following surgery – that is, after atezolizumab therapy had stopped – they did not detect any new safety signals. The most common grade 3 or greater toxicity in each arm was neutropenia, followed by leukopenia, and in the atezolizumab arm there was one case of a grade 3 myocarditis that occurred 2 weeks after surgery and before the patient started on an anthracycline.
The authors also conducted a mass cytometry analysis of potential predictors of response to checkpoint inhibitors in TNBC, and reported the results in Nature.
“Basically, baseline density of several activated immune cells predicted for higher probability of pCR with atezolizumab but not with chemotherapy without atezolizumab,” Dr. Gianni said.
Specifically, they saw that high density of CD8-positive, TCF1-positive and Ki-67-positive markers were associated with increased pCR and EFS rates with the addition of atezolizumab.
Why no benefit to the ICI?
Invited discussant Kevin Kalinsky, MD, MS, from the Winship Cancer Institute at Emory University in Atlanta said that at least two possible explanations floated for the lack of either a pCR or EFS benefit in NeoTRIP don’t hold water.
For example, the theory that a difference in efficacy between PD-1 and PD-L1 inhibitors might explain the lack of benefit in NeoTRIP is undermined by IMpassion031, which showed pCR and EFS benefits with atezolizumab.
In addition, although NeoTRIP did not include an anthracycline in the neoadjuvant regimen, neither did the NeoPACT trial, in which patients received six cycles of neoadjuvant taxane, carboplatin, and pembrolizumab, and had a 58% pCR rate, with a high 2-year EFS rate among patients who had a pCR.
“As we know, triple-negative breast cancer is a heterogeneous disease. Could it be by chance that there are differences in tumor biology reflected across the various neoadjuvant trials, including molecular subtypes? Also in NeoTRIP, previously reported we’ve seen that there were a higher rate of TILs in the chemo-alone arm; high TILs can be associated with chemosensitivity, and maybe this influenced the results,” he said.
Predictive markers to immunotherapy in TNBC are still needed, he said, because neither PD-L1 expression, tumor mutational burden, or TILs have proven to be reliable biomarkers for this subtype.
The NeoTRIP Michelangelo trial was supported by Hoffman-La Roche and Celgene. Dr. Gianni disclosed financial interests with Roche and others including advisory board activity, consulting, and personal fees. Dr. Kalinsky reported advisory/consulting activities for various companies, including Genentech/Roche.