Early Triple-Negative Breast Cancer: Are Checkpoint Inhibitors Ready for Neoadjuvant or Adjuvant Use?

Get Permission

Recent clinical trials have been encouraging for the neoadjuvant or adjuvant use of immune checkpoint inhibitors in triple-negative breast cancer, but is this approach ready for the clinic? This question was addressed at PER’s Miami Breast Cancer Conference, held virtually this year, by Adam M. Brufsky, MD, PhD, Professor of Medicine at the University of Pittsburgh School of Medicine and Co-Director of the Comprehensive Breast Cancer Center.1

Adam M. Brufsky, MD, PhD

Adam M. Brufsky, MD, PhD

“Immunotherapy is a hot topic now in triple-negative breast cancer. However, although immune checkpoint inhibitors do increase the rates of pathologic complete response, are they unequivocally doing enough for us to integrate them into clinical practice?” he asked.

The results of two trials in early-stage disease have provided some support for this, whereas another yielded disappointing results. Aside from key findings, other interesting observations emerged in these studies, spawning new research questions, according to Dr. Brufsky.

KEYNOTE-522: Chemotherapy Plus Pembrolizumab

KEYNOTE-522 assigned 1,174 women with newly diagnosed stage II or III triple-negative breast cancer to receive neoadjuvant chemotherapy (carboplatin plus paclitaxel followed by anthracycline plus cyclophosphamide) plus pembrolizumab or chemotherapy alone.2 After surgery, the pembrolizumab cohort received nine cycles of adjuvant pembrolizumab.

There were two planned interim analyses; if the first showed improved pathologic complete response rates, the second was performed approximately 24 months after the first subject enrolled. A third interim analysis was performed, but only the U.S. Food and Drug Administration (FDA) was given those data. The second interim analysis was based on 784 patients in the pembrolizumab arm and 390 in the chemotherapy-alone arm. Median follow-up was 15 months.

At the first interim analysis, regardless of PD-L1 expression, more patients in the pembrolizumab arm achieved a pathologic complete response: 64.8% vs 51.2%, representing an absolute difference of 13.6% (P = .00055). In patients with PD-L1–positive disease (combined positive score [CPS] ≥ 1%), however, rates improved even more—interestingly, not only in the pembrolizumab arm (68.9%) but also in the chemotherapy-alone arm (54.9%). These rates were much higher than those seen in PD-L1–negative patients (45.3% and 30.3%, respectively). The approximately 25% greater rate of pathologic complete response in the PD-L1–positive population seems to point to PD-L1 status as a biomarker of response to chemotherapy, in Dr. Brufsky’s opinion.

“Event-free survival also appeared to be improved. We saw a 6% benefit with pembrolizumab, but the follow-up was fairly short,” he noted. Since relapses in triple-negative disease tend to occur within the first 3 years, “we’re only about halfway to that point, and more follow-up is needed.”

The second interim analysis showed event-free survival at 18 months to be 91.3% and 85.3%, respectively (hazard ratio [HR] = 0.63; 95% confidence interval [CO] = 0.43–0.93). Although the Kaplan-Meier curves separated, the difference did not meet the stringent predesignated P value for significance.

IMpassion031: Chemotherapy Plus Atezolizumb

The phase III IMpasion031 trial assigned 333 women to neoadjuvant chemotherapy—nab-paclitaxel followed by doxorubicin and cyclophosphamide—with and without atezolizumab; the experimental arm also received 1 year of adjuvant atezolizumab.3,4 The results were similar to those of KEYNOTE-522: pathologic complete response rates were 57.6% with atezolizumab and 41.1% with chemotherapy alone, a difference of 16.5% (P = .0044).

Also reflecting KEYNOTE-522, the PD-L1–positive population was more likely than the PD-L1–negative cohort to achieve a pathologic complete response: 68.8% with atezolizumab vs 49.3% with chemotherapy alone, compared with 47.7% and 34.4%, respectively, in the PD-L1–negative cohort. Also, similarly, even in the control arm, the PD-L1–positive patients fared better than PD-L1–negative patients, he noted.

Event-free survival was marginally improved, with rates of 10.3% in the atezolizumab arm vs 13.1% in the chemotherapy-alone arm (HR = 0.76). However, this was based on few events.

Hypothyroidism With Pembrolizumab

Pembrolizumab was generally well tolerated in the trials, though there was a clear increase in hypothyroidism. This side effect was seen in about 15% of patients in KEYNOTE-522 and in 7% of those in IMpassion031 (another 3% or so developed other thyroid abnormalities in IMpassion031).

“We think what may be happening is that pembrolizumab is turning subclinical hypothyroidism (ie, related to aging, etc) into a condition that is clinically significant. We know that in some patients, this will be long term,” commented Dr. Brufsky.

NeoTRIPaPDL1: Chemotherapy Plus Atezolizumab

Outcomes were different, however, in the phase III NeoTRIPaPDL1 trial of 280 patients, with no benefit observed with the addition of atezolizumab to nab-paclitaxel plus carboplatin.5,6 The rates of pathologic complete response in the intent-to-treat population were 43.5% with atezolizumab plus chemotherapy and 40.8% with chemotherapy alone (P = .066).

“We saw very little difference in pathologic complete response, which is interesting, and also not much difference in the PD-L1–positive group. The question is why did we not see this in NeoTRIP but we did see it in other trials?” he said.

One hypothesis is that the baseline imbalance of tumor-infiltrating lymphocytes—more tumor-infiltrating lymphocytes in the control arm—may have resulted in a better performance by chemotherapy alone.6 When patients were stratified by the concentration of tumor-infiltrating lymphocytes, among the 32% with a concentration of at least 40%, the pathologic complete response rate was 71.4% with pembrolizumab and 63% with chemotherapy alone. In patients with a concentration of less than 40%, these rates dropped to 34% and 28%, respectively.

“If you had 40% or more tumor-infiltrating lymphocytes in the tumor, you had a substantial improvement in pathologic complete response, whether you were in the atezolizumab arm or chemotherapy arm,” he noted.

Another possibility is that the “chemotherapy backbone matters,” Dr. Brufsky added. The addition of the anthracycline in KEYNOTE-522 and IMpassion031 may have improved responses.

New Biomarkers Emerging?

From these trials, it seems that PD-L1 status and lymphocytic infiltration with tumor-infiltrating lymphocytes are both biomarkers of response—not necessarily to checkpoint inhibition but to chemotherapy. In PD-L1–positive patients, the addition of the checkpoint inhibitor is a “bonus” that seems to enhance the odds of pathologic complete response, according to Dr. Brufsky.

“The mystery we hope to unravel is why PD-L1 positivity and tumor-infiltrating lymphocytes are associated with higher pathologic complete response rates. These findings are fascinating, and I believe they will turn out to be something scientifically important when we explore them more,” he commented.

FDA ODAC Decision

The FDA’s Oncologic Drugs Advisory Committee (ODAC) has paid attention to the disconnect demonstrated by having two positive trials and one negative one. On February 9, 2021, ODAC voted 10 to 0 against supporting the approval of pembrolizumab plus chemotherapy as neoadjuvant therapy for triple-negative breast cancer, despite the increase in pathologic complete response shown in KEYNOTE-522.7

The decision took into account not only the first two planned interim analyses, but also the third interim analysis, which has not been publicly reported. Although the pathologic complete response difference was 13.6% at the first interim analysis, it dropped to 7.5% at the third analysis, which included more patients. The 9.3% difference at the second interim analysis had already failed to meet the statistical threshold for significance, noted Dr. Brufsky. The Kaplan-Meier event-free survival curves did separate around 12 months, though the difference fell outside the boundaries for statistical significance.

ODAC concluded that the “event-free survival data are immature, the improvement in pathologic complete response rate is small, and the addition of pembrolizumab to neoadjuvant and adjuvant treatment adds toxicity, which may be severe, irreversible, or require lifelong medication.” In brief, the FDA considered the pathologic complete response improvement with pembrolizumab not “clinically meaningful.”

Dr. Brufsky said he looks to NSABP B-59/GBG 96-GeparDouze to be the “tie-breaker trial.” B-59 will randomly assign 1,520 women with early triple-negative breast cancer to dose-dense anthracycline/cyclophosphamide plus carboplatin and paclitaxel, with and without atezolizumab (and 1 year of adjuvant atezolizumab). Longer-term follow-up of KEYNOTE-522 and IMpassion031 will also be informative.

Thoughts for the Clinic

So, where do we go next? “Data supporting checkpoint inhibitors as neoadjuvant therapy for triple-negative breast cancer are close, but not quite there yet. Given the risk of hypothyroidism, the bottom line, for now is we need to be very cautious if we decide to use them in the clinic,” Dr. Brufsky advised.

“However, we do have a disease that’s tough to treat,” he acknowledged. “We know that with standard chemotherapy, a certain percentage of these patients will relapse. Oncologists will have to think of the individual patient in front of them when they make that treatment decision.” 

DISCLOSURE: Dr. Brufsky has served as a consultant or advisor to Abbvie, Agendia, Bayer, Biotheranostics, Daiichi Sankyo/Lilly, Eisai, Genentech/Roche, Immunomedics, Merck, Micheal J. Hennssey Associates, Myriad Pharmaceuticals, OncLive, Novartis, Pfizer, Puma Biotechnology, Seattle Genetics, and Tyme; and has provided expert testimony on behalf of Pfizer.


1. Brufsky A: Evolving strategies in neoadjuvant/adjuvant treatment of early-stage triple-negative breast cancer: Role of immunotherapy. 2021 Miami Breast Cancer Conference. Presented March 5, 2021.

2. Schmid P, Cortes J, Pusztai L, et al: Pembrolizumab for early triple-negative breast cancer. N Engl J Med 382:810-821, 2020.

3. Harbeck N, Zhang H, Barrios CH, et al: IMpassion031: Results from a phase 3 study of neoadjuvant atezolizumab plus chemotherapy in early triple-negative breast cancer. ESMO Virtual Congress 2020. Abstract LBA11. Presented September 20, 2020.

4. Mittendorf EA, Zhang H, Barrios CH, et al: Neoadjuvant atezolizumab in combination with sequential nab-paclitaxel and anthracycline-based chemotherapy versus placebo and chemotherapy in patients with early-stage triple-negative breast cancer (IMpassion031): A randomised, double-blind, phase 3 trial. Lancet 396:1090-1100, 2020.

5. Gianni L, Huang C, Egle D, et al: Pathologic complete response to neoadjuvant treatment with or without atezolizumab in triple-negative, early high-risk and locally advanced breast cancer. NeoTRIPaPDL1 Michelangelo randomized study. 2019 San Antonio Breast Cancer Symposium. Abstract GS3-04. Presented December 12, 2019.

6. Bianchini G, Huang C, Egle D, et al: Tumour infiltrating lymphocytes, PD-L1 expression and their dynamics in the NeoTRIPaPDL1 trial. ESMO Virtual Congress 2020. Abstract LBA13. Presented September 18, 2020.

7. U.S. Food and Drug Administration: Oncology Drugs Advisory Committee Meeting, February 9, 2021. BLA 125514, Supplement-089. Combined FDA and Applicant ODAC Briefing Document. Available at Accessed April 27, 2021.