ImmunoRx in TNBC: Conflicting Results

Pam Harrison

December 13, 2019

SAN ANTONIO ― Two separate phase 3 neoadjuvant trials in which immunotherapy was added to chemotherapy for patients with early, triple-negative breast cancer (TNBC) theoretically should have had similar outcomes, but one was impressively positive and the other disappointingly negative.

The question now becomes: Why?

Peter Schmid, MD, PhD, Barts Cancer Institute, London, United Kingdom, reported that the pathologic complete response (pCR) rate among patients with previously untreated, nonmetastatic TNBC who received pembrolizumab (Keytruda, Merck) in addition to neoadjuvant chemotherapy was 64.8%, vs 51.2% for patients who received the same chemotherapy but no immunotherapy (P = .0005), which is an absolute difference of 13.6%.

He spoke at a press briefing here at the San Antonio Breast Cancer Symposium (SABCS) 2019.

On the other hand, Luca Gianni, MD, Fondazione Michelangelo, Milan, Italy, reported that the addition of atezolizumab (Tecentriq, Genentech) to chemotherapy in seemingly similar patients did little to affect the pCR rate, which was 43.5% for the patients who received immunotherapy and 40.8% for the patients who did not receive immunotherapy.

On closer scrutiny, the two trials, although conducted in similar patients who shared a generally poor prognosis, differed with regard to patient selection and the choice and duration of chemotherapy used. In addition, a different checkpoint inhibitor was used.

Some Differences in the Trials

In the KEYNOTE-522 trial, in which pembrolizumab was evaluated, a total of 1174 patients were enrolled. Key eligibility criteria included stage II and stage III breast cancer and no stage I disease. Notably, patients with bilateral or multifocal primary tumors as well as inflammatory breast cancers were allowed into the trial, and half of all patients were lymph node positive.

In the NeoTRIPaPDL1 Michelangelo trial, in which atezolizumab was evaluated, a total of only 280 patients were enrolled. At baseline, about half of patients had early high-risk disease, and about half had locally advanced disease. In addition, 87% of patients had lymph node involvement, which is a higher percentage than the total in the KEYNOTE study.

The chemotherapy regimens used in the two trials were also substantively different.

Patients in KEYNOTE-522 received a standard chemotherapy regimen. Specifically, patients were randomly assigned to receive either four cycles of pembrolizumab 200 mg every 3 weeks or placebo plus paclitaxel 80 mg/m2 every week plus carboplatin AUC 5 every 3 weeks or AUC 1.5 every week in the first 12 weeks.

This was followed by four cycles of either pembrolizumab or placebo plus doxorubicin 60 mg/m2 every 3 weeks or epirubicin 90 mg/m2 every 3 weeks plus cyclophosphamide 600 mg/m2 every 3 weeks in the subsequent 12 weeks prior to surgery.

After definitive surgery, patients could receive radiotherapy if indicated, but the patients in the experimental arm also received pembrolizumab or placebo every 21 days for nine cycles or until disease recurrence.

The pCR rate was 51.2% in the control arm (the patients who did not receive immunotherapy), which is an indicator that the regimen is intensive, Schmid noted.

In contrast, the neoadjuvant regimen used in the Italian study was not a standard chemotherapy regimen, Schmid said to Medscape Medical News. Patients received neoadjuvant carboplatin AUC 2 and nab-paclitaxel 125 mg/m2 on days 1 and 8, with or without atezolizumab 1200 mg on day 1.

"Both regimens were given every 3 weeks for eight cycles and were followed by surgery and by four cycles of an anthracycline regimen, as per investigator choice," the investigators explained.

Moreover, the primary endpoint of the KEYNOTE study is pCR, whereas the primary endpoint of the Italian study is event-free survival rate at 5 years. Gianni defended his trial's primary endpoint. At the time of trial design, it was well established that pCR was an "extremely powerful predictor" of long-term outcome in chemotherapy but not in immunotherapy.

"If you move to endocrine therapy, there is no relationship [between pCR and long-term outcomes] at all," Gianni emphasized. In other words, the value of the pCR measure is in question.

Different Inhibitors

Pembrolizumab is a programmed cell death protein–1 (PD1) inhibitor, whereas atezolizumab is a programmed cell death ligand–1 (PD-L1) inhibitor.

"At the end of the day, both drugs interfere with the axis between PD1 and PD-L1, and it is largely assumed that mechanistically, they will have the same effect," Gianna said.

"But there may be subtle differences between the two drugs that may not be so subtle in the end," he suggested.

"My personal theory is that the dynamics of tumor biology are very high in early disease, and by giving patients chemotherapy first, patients who had PD-1-negative tumors before they started immunotherapy turned PD-1-positive through the process of chemotherapy," Schmid explained.

This process gets lost over time, he added, such that advanced tumors are no long dynamic and thus immune cells can no longer penetrate into the tumor as they can in early disease.

"You can have the best checkpoint inhibitor in the world and it may activate immune cells, but if the immune cells aren't there, they can't kill the cancer," Schmid added.

Still Early

Commenting on the two studies, press session moderator Kent Osborne, MD, Baylor College of Medicine, Houston, Texas, suggested that the use of immunotherapy in breast cancer is still in its infancy relative to other cancers, but its potential use is intriguing, especially in TNBC.

"I don't think we should give up yet on atezolizumab, as it may be an important component to our treatment for breast cancer eventually," Osborne said.

"But the pembrolizumab study looks very promising, and I think that it is probably going to be practice changing for patients with TNBC," he predicted.

Asked for comment, Aditya Bardia, MD, MPH, Massachusetts General Hospital Cancer Center, Boston, said that interest in the use of immunotherapy for breast cancer is increasing, especially in TNBC.

"TNBC has an aggressive tumor biology, so clinically, it's an unmet need, and we need better therapies for it,” he said.

"Unfortunately, we do not have a biomarker that can yet predict who would not benefit from immunotherapy, and even the PD-1-negative subgroup in KEYNOTE-222 benefited from treatment," Bardia acknowledged.

"But what we have consistently seen in breast cancer is that response to immune therapy is much higher when treatment is given as first-line treatment, and this speaks to the immune microenvironment, where it is harder to elicit an immune response, at least to single-agent immunotherapy in more advanced disease," he added, pointing out that both studies were in the first-line setting.

The KEYNOTE-222 study was funded by Merck. The NeoTRIPaPDL1 Michelangelo study was supported by Hoffman–La Roche and Celgene. Schmid receives grants and editorial support from F. Hoffmann–La Roche; institutional research support from AstraZeneca, Roche/Genentech, Oncogenex, Novartis, and Astellas; and honoraria from Pfizer, AstraZeneca, Novartis, Roche, Merck, Boehringer Ingelheim, Bayer, Eisai, Celgene, and Puma. Gianni receives consulting fees from ADC Therapeutics, AstraZeneca, Celgene, Eli Lilly, G1 Therapeutics, Genentech, Genomic Health, Merck Sharp & Dohme, Oncolytics Biotech, Odonate Therapeutics, Onkaido, Roche, Pfizer, and Tahio Pharmaceuti and has contracted research with Zymeworks, Daiichi Sankyo Zymeworks, Daiichi Sankyo, and Revolution Medicine. Osborne has served on the advisory boards for Genentech and AstraZeneca, is a consultant for Tolmar Pharma, and holds stock in GeneTex. Bardia has served as a consultant for or on the advisory boards of Genetech, Roche, Pfizer, Novartis, Merck, Daiichi, Immunomedics, and Sanofi. The Barts Cancer Institute has received research funding from the same companies.

San Antonio Breast Cancer Symposium (SABCS) 2019: Abstract GS3-03 and GS3-04. Presented December 12, 2019.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.