New Data in NSCLC Bolster Support for Perioperative Immunotherapy

CHICAGO — A randomized trial testing the addition of perioperative pembrolizumab (Keytruda) to chemotherapy in operable non-small cell lung cancer (NSCLC) hit one of its two primary endpoints in an interim analysis, with a favorable trend for the other.

All patients in the study received cisplatin-based neoadjuvant chemotherapy, and at 2 years, the event-free survival (EFS) rate reached 62.4% in those given pembrolizumab before and after surgery, as compared with 40.6% for those assigned to placebo, reported Heather Wakelee, MD, of Stanford Cancer Institute in California.

That difference amounted to a 42% reduction in the risk for disease recurrence or death (HR 0.58, 95% CI 0.46-0.72, P<0.001), and this EFS benefit in KEYNOTE-671 was consistent across all subgroups, including histology and PD-L1 status, she detailed at the American Society of Clinical Oncology (ASCO) meeting here.

But on overall survival (OS), “we have not yet crossed the significance boundary,” said Wakelee. “However, at this time we do see the curves starting to separate.”

At 2 years, an estimated 80.9% of patients were alive in the pembrolizumab arm versus 77.6% in the placebo arm (HR 0.73, 95% CI 0.54-0.99), according to findings from the study, which were published simultaneously in the New England Journal of Medicine.

Despite that, Wakelee concluded that the data support the perioperative use of the PD-1 checkpoint inhibitor in resectable stage II-III NSCLC, and the findings further reinforce the role of checkpoint blockade in operable patients.

Pembrolizumab is already approved by the FDA in the adjuvant setting, as is atezolizumab (Tecentriq). And last year, the agency approved neoadjuvant nivolumab (Opdivo) in resectable NSCLC based on the CheckMate-816 trial.

In the last few months, two other perioperative trials testing PD-1/L1 inhibitors in early NSCLC — namely investigational toripalimab and durvalumab (Imfinzi) — showed benefit in the Neotorch and AEGEAN trials.

These perioperative immunotherapy studies, including KEYNOTE-671, represent “a major advance as new standards of care for the treatment of resectable lung cancer,” said ASCO-invited discussant Mark Awad, MD, PhD, of the Dana-Farber Cancer Institute in Boston.

But “should we give it pre-op, post-op, or both?” he asked.

Although Awad cautioned that the data are immature, “there appears to be an encouraging signal for overall survival benefit favoring a neoadjuvant plus or minus adjuvant immunotherapy approach.” He pointed out that interim analyses from adjuvant-only phase III trials — IMpower010 and KEYNOTE-091/PEARLS — have yet to show a clear separation in OS curves.

On the additive role of adjuvant immunotherapy when patients have already received neoadjuvant chemoimmunotherapy, Awad suggested that further study is needed. “For now, I think it’s reasonable to continue adjuvant immunotherapy in discussion with our patients,” he said.

The double-blind KEYNOTE-671 trial enrolled NSCLC patients with stage II-IIIB (N2) disease deemed fit for surgery. Overall, 797 participants were randomized 1:1 to either the perioperative pembrolizumab strategy or to the control arm.

All patients received four cycles of neoadjuvant chemotherapy (cisplatin plus either gemcitabine or pemetrexed [Alimta] depending on histology) with IV pembrolizumab (200 mg) or placebo every 3 weeks. Following surgery, patients in the two arms received up to 13 cycles of adjuvant pembrolizumab at the same dose and schedule, or placebo.

Stratification factors included disease stage, PD-L1 expression, histology, and geographic region.

Patients in the study had a median age of 63-64, and over 70% were men. Most were white (60%), and 31% were Asian. A majority of the participants had a smoking history.

For disease characteristics, 57% had nonsquamous histology, and 70% had pathologic stage III disease at baseline. About a third had ≥50% PD-L1 expression, a third had 1-49% PD-L1 expression, and a third had <1% PD-L1 expression.

Three-fourths of patients completed all four cycles of neoadjuvant treatment, with about 85% going on to surgery. In the adjuvant setting, 73% of patients in the pembrolizumab arm received at least one dose versus 67% of those in the control group (40% and 35% completed all 13 planned cycles).

EFS and OS in the intent-to-treat population constituted the dual primary endpoints, with secondary endpoints including major pathological response (MPR), pathologic complete response (pCR), and safety.

Following surgery, rates of MPR and pCR both significantly favored the investigational arm versus the control arm (P<0.0001 for both):

• MPR: 30.2% vs 11%, respectively
• pCR: 18.1% vs 4.0%

Toxicity was “as expected,” said Wakelee, and reflective of a patient population who had chemotherapy and went on to surgery.

Grade ≥3 treatment-related adverse events (AEs) were more frequent in the pembrolizumab arm (45% vs 37%) and a higher proportion of patients in that arm discontinued all treatment due to AEs (13% vs 5%). There were four and three treatment-related deaths, respectively.

As expected, immune-mediated AEs were higher with the checkpoint inhibitor, with grade ≥3 immune-mediated AEs in 5.8% versus 1.3% of the control arm.