Single-agent Atezolizumab Improves Survival in PD-L1-High NSCLC

Atezolizumab monotherapy improved overall survival compared with platinum-based chemotherapy as a first-line treatment in patients with wild-type non–small cell lung cancer with PD-L1 expression ≥50%.

Atezolizumab (Tecentriq) monotherapy improved overall survival (OS) compared with platinum-based chemotherapy as a first-line treatment in patients with wild-type non–small cell lung cancer (NSCLC) with PD-L1 expression ≥50%, according to an interim analysis presented at Society for Immunotherapy of Cancer’s 34th Annual Meeting & Pre-Conference Programs (SITC 2019).1

“Atezolizumab represents a promising first-line treatment option in patients with PD-L1–high NSCLC,” said study author Giuseppe Giaccone, MD, PhD, an interim assistant professor of medicine at Weill Cornell Medicine.

The phase III IMpower110 trial preselects chemotherapy-naïve patients with stage IV squamous or nonsquamous NSCLC based on PD-L1 status, then randomizes patients 1:1 to receive either atezolizumab monotherapy at 1200 mg once every 3 weeks (arm A; n = 277) or 4 to 6 cycles of a platinum-based chemotherapy regimen based on histology (arm B; n = 277). No crossover was permitted.

Survival was the primary endpoint in the wild-type population, secondary endpoints were progression-free survival (PFS) and response rate, and duration of response (DOR).

In the statistical testing plan, patients were stratified by PD-L1 expression on tumor cells (TC) or immune cells (IC) to 1 of 3 groups: TC3/IC3 wild-type, high expression (≥50% TC; ≥10%I IC); TC2/3 or IC2/3 wild-type (high and moderate expression, ≥5% TC or IC); or TC1/2/3 or IC1/2/3 wild-type (entire population, ≥1%). Patients with EGFR and/or ALK mutations were excluded from the interim analysis. In total 554 patients (277 in each arm) were evaluated.

“[According to] the statistical testing plan if TC3 or IC3, the high expressor, wild-type patients were positive, then the second group of intermediate expressors would be tested, and if this was also positive in terms of survival, then the overall population would be tested,” explained Giaccone.

At a median follow-up of 15.7 months (range, 0-35), the median OS for 552 patients was 20.2 months (95% CI, 16.5-not estimable [NE]) in the atezolizumab arm compared with 13.1 months (95% CI, 7.4-16.5) in the chemotherapy arm (HR, 0.59; 95% CI, 0.40-0.89; P = .0106).

Improvement was seen in almost all subgroups of the high expressors, with the exception of the never smokers (n = 24), who did better with chemotherapy, demonstrating a median OS of 15.9 months versus 8.0 months with atezolizumab.

In the TC2/3 or IC2/3 wild-type population, the survival curve also favored atezolizumab, however, it did not reach statistical significance as the prespecified alpha boundary was not crossed in the analysis. “It cannot be considered statistically significant although, numerically, there is clearly an advantage of atezolizumab versus chemotherapy,” noted Giaccone.

As a result, OS was not formally tested in the TC1/2/3 and IC1/2/3 populations.

Median PFS in the TC3 or IC3 wild-type population was 8.1 months (95% CI, 6.8-11.0) in the atezolizumab arm and 5.0 months (95% CI, 4.2-5.7) in the chemotherapy arm (HR, 0.63; 95% CI, 0.45-0.88; P = .007). The confirmed overall response rate was 38.3% versus 28.6%, respectively; and the median DOR was not reached (range 1.8-29.3+) versus 6.7 months (range, 2.6-23.9+), respectively.

The PFS was also not formally tested in the TC2/3, IC2/3, or TC1/2/3, IC1/2/3 populations as testing was contingent on a positive primary endpoint being achieved in all 3 populations.

Although not formally tested, data were presented on the populations who did not cross the prespecified OS boundary. In the TC 2/3 or IC 2/3 wild-type subgroup, median OS was 18.2 months (95% CI, 13.3-NE) in the atezolizumab arm versus 14.9 months (95% CI, 10.8-16.6) in the chemotherapy arm (HR, 0.72; 95% CI, 0.52-0.99; P = .0416). 

Atezolizumab was also favored numerically in the TC 1/2/3 or IC 1/2/3 wild-type population with a median OS of 17.5 months (95% CI,12.8-23.1) when compared with chemotherapy (14.1 months; 95% CI, 11.0-16.6) but was not significant (HR, 0.83; 95% CI, 0.65-1.07; P = .1481).

In terms of safety, there were no new safety concerns for atezolizumab which demonstrated similar adverse events (AEs) seen in prior trials, including decreased neutrophil count, pruritus, and hypothyroidism. Treatment-related adverse events (TRAEs) occurred in 60.5% of patients in arm A and 85.2% of patients in arm B, and grade 3/4 TRAEs occurred in 12.9% and 44.1%, respectively.

“Chemotherapy had more grade 3 and 4 toxicity in general compared to atezolizumab,” said Giaccone. “In particular, it had more toxicities that caused interruption of the treatment.” Eighteen patients (6.3%) in the atezolizumab arm experienced AEs leading to treatment withdrawal compared with 43 patients (16.3%) in the chemotherapy arm. Grade 5 AEs were similar in both arms.

The proportion of patients who received different classes of subsequent cancer therapies was similar across the PD-L1 subgroups, however, Giaccone pointed out that “the chemotherapy arm had more patients who received at least 1 further line of treatment after progression and this was immunotherapy in almost 30% of the patients.”

Patients in arm B with nonsquamous histology received cisplatin, 75 mg/m2, or carboplatin, area under the curve (AUC) 6, plus pemetrexed, 500 mg/m2 intravenously every 3 weeks. Patients in arm B with squamous histology received cisplatin, 75 mg/m2, plus gemcitabine, 1250 mg/m2, or carboplatin, AUC 5, plus gemcitabine, 1000 mg/m2 intravenously, every 3 weeks.

Additional biomarker analyses will include comparison between monoclonal antibodies for immunohistochemistry as well as blood-based tumor mutational burden will be presented at an upcoming meeting.



Spigel DR, De Marinis F, Giaccone G, et al. IMpower110: Interim overall survival (OS) analysis of a Phase III study of atezolizumab (atezo) vs platinum-based chemotherapy (chemo) as first-line (1L) treatment (tx) in PD-L1–selected NSCLC. Presented at: 34th Annual Meeting & Pre-Conference Programs of the Society for Immunotherapy of Cancer; November 7-10, 2019; National Harbor, MD. LBAO81.