Why This Study Matters
Immune checkpoint inhibitors targeting PD-L1/PD-1 — with or without chemotherapy — are the standard first-line treatment for non-small cell lung cancer (NSCLC) that lacks actionable genomic alterations. But survival benefit remains modest: more than 90% of patients experience disease progression within five years. Combining immunotherapies that hit complementary immune checkpoints, such as pairing an anti-TIGIT antibody with an anti-PD-L1 antibody, has been proposed as a way to deepen and extend antitumor responses.
SKYSCRAPER-01 tested this hypothesis directly: does adding tiragolumab (an anti-TIGIT monoclonal antibody) to atezolizumab improve outcomes over atezolizumab alone in patients with locally advanced unresectable, recurrent, or metastatic NSCLC with high tumor PD-L1 expression? The authors describe this as the first phase 3, randomized, global study in NSCLC to report clinical outcomes for an anti-TIGIT plus anti-PD-L1 combination.
Study Design
SKYSCRAPER-01 was a randomized, double-blind, placebo-controlled phase 3 trial. Patients were recruited between March 4, 2020, and August 31, 2021. Eligible patients were randomly assigned 1:1 to receive either tiragolumab 600 mg plus atezolizumab 1,200 mg IV, or placebo plus atezolizumab 1,200 mg IV, on day 1 of each 21-day cycle. Treatment continued until radiographic disease progression per RECIST v1.1 or loss of clinical benefit, in the absence of unacceptable toxicity. Site personnel and patients were blinded to treatment assignment.
Co-primary endpoints were investigator-assessed progression-free survival (INV-PFS) and overall survival (OS), evaluated in the primary analysis set (PAS) — all randomly assigned patients with PD-L1 tumor proportion score (TPS) of 50% or higher by the IHC 22C3 pharmDx assay.
Patient Population
Of 1,052 patients assessed for eligibility, 534 were randomly assigned (266 to tiragolumab plus atezolizumab; 268 to placebo plus atezolizumab). The primary analysis set (PAS) comprised 521 patients (262 vs. 259). Baseline characteristics were balanced between arms: median age 66.0 years, 74.7% male, 69.1% White, 63.7% previous tobacco use, 73.1% nonsquamous histology, with roughly 17% of patients in each arm having liver or brain metastases. Patients with EGFR mutations or ALK rearrangements were excluded.
Primary Endpoint Results
At the final INV-PFS analysis (data cutoff March 12, 2022; median follow-up 9.9 months), 337 PFS events had occurred. Median INV-PFS was 7.0 months with tiragolumab plus atezolizumab versus 5.6 months with placebo plus atezolizumab (HR, 0.78 [95% CI, 0.63 to 0.97]; P=.02). Because INV-PFS in the PAS did not cross the statistical significance boundary, OS was subsequently tested at a two-sided alpha level of .049, per the statistical analysis plan.
At the final OS analysis (data cutoff September 24, 2024; median follow-up 17.9 months), 345 OS events had occurred (170 deaths with tiragolumab plus atezolizumab; 175 with placebo plus atezolizumab). Median OS was 23.1 months versus 16.9 months (HR, 0.87 [95% CI, 0.71 to 1.08]; P=.22). The OS endpoint also failed to cross the statistical boundary for significance.
| Endpoint | Tiragolumab + Atezolizumab | Placebo + Atezolizumab | HR (95% CI) | P value |
|---|---|---|---|---|
| Median INV-PFS | 7.0 mo (5.6–9.8) | 5.6 mo (4.4–7.0) | 0.78 (0.63–0.97) | .02 (nonsignificant) |
| Median OS | 23.1 mo (17.7–28.8) | 16.9 mo (14.6–21.3) | 0.87 (0.71–1.08) | .22 (nonsignificant) |
| Confirmed ORR | 45.8% (CR 1.5%, PR 44.3%) | 35.1% (CR 1.2%, PR 34.0%) | — | — |
| Median DOR | 18.0 mo (13.6–24.4) | 14.6 mo (9.7–18.6) | — | — |
Both co-primary endpoints failed to reach statistical significance under the trial’s prespecified boundaries, despite numerically favoring the tiragolumab-containing arm.
Subgroup Analyses
The published article reports subgroup findings only in qualitative terms; exact hazard ratios, confidence intervals, and p-values for individual subgroups were not reported in the main text.
| Subgroup Analysis | Reported Finding |
|---|---|
| Key demographic/disease subgroups (INV-PFS) | An improvement in INV-PFS was observed across key subgroups with tiragolumab plus atezolizumab versus placebo plus atezolizumab; exact per-subgroup values were not reported in the source. |
| Key demographic/disease subgroups (OS) | Subgroup analyses of OS did not show a significant treatment effect favoring tiragolumab plus atezolizumab over placebo plus atezolizumab. |
| High TIGIT expression (≥10% and ≥15%) biomarker subgroup | A trend toward improved OS was noted, but the authors caution these subgroups had small sample sizes and broad confidence intervals that limit interpretation; high TIGIT expression was not prognostic or predictive of a PFS or OS benefit in the full or secondary analysis sets. |
Safety Profile
The safety-evaluable population comprised 267 patients receiving tiragolumab plus atezolizumab and 263 receiving placebo plus atezolizumab (median safety follow-up 10.6 vs. 7.8 months).
| Event | Tiragolumab + Atezolizumab (n=267) | Placebo + Atezolizumab (n=263) |
|---|---|---|
| All-cause adverse events, any grade | 95.9% | 91.3% |
| Grade 3–4 adverse events | 41.2% | 33.8% |
| Grade 5 adverse events | 10.9% | 9.9% |
| Treatment-related adverse events | 75.7% | 60.1% |
| Grade 3–4 treatment-related adverse events | 19.9% | 9.5% |
| Treatment-related deaths | 1.5% (n=4) | 0.8% (n=2) |
| Serious adverse events (any cause) | 43.1% | 41.1% |
| Adverse events leading to treatment withdrawal | 16.1% | 6.5% |
| Immune-mediated adverse events | 70.0% | 50.6% |
| Grade 3–4 immune-mediated adverse events | 16.1% | 9.9% |
| Required systemic corticosteroids for an immune-mediated adverse event | 24.3% | 12.5% |
| Most common adverse event: pruritus | 28.8% | 17.5% |
| Most common adverse event: rash | 25.5% | 13.3% |
Deaths due to an adverse event occurred in 29 patients (10.9%) with tiragolumab plus atezolizumab (4 considered treatment-related: pneumonitis, neurologic decompensation, autoimmune hemolytic anemia, and thrombophlebitis migrans — one case each) and 26 patients (9.9%) with placebo plus atezolizumab (2 considered treatment-related: immune-mediated encephalitis and myasthenia gravis — one case each). The most commonly reported immune-mediated adverse event in both arms was immune-mediated rash (42.3% vs. 19.0%), typically low-grade and managed with topical corticosteroids. The authors state the observed toxicity was consistent with the known safety profile of combined tiragolumab and atezolizumab, with no new safety signals identified.
Interpretation and Broader Context
SKYSCRAPER-01 did not meet either co-primary endpoint. The authors note that while numerical improvements in both INV-PFS and OS favored the tiragolumab-containing arm, neither difference reached statistical significance under the trial’s prespecified boundaries. An invited commentary accompanying the article offers a blunt assessment of what this negative result, in a PD-L1-selected population, means for the broader anti-TIGIT development landscape:
The negative results of this trial, in a PD-L1-selected patient population, discourages further investigation of anti-TIGIT agents in NSCLC.
— Thomas E. Stinchcombe, MD, JCO Associate Editor, invited commentary accompanying the article
The trial’s discussion section situates SKYSCRAPER-01 within a broader, and largely disappointing, pattern for anti-TIGIT combinations in NSCLC: a survival benefit for anti-PD-L1/PD-1 plus anti-TIGIT regimens over anti-PD-L1/PD-1 alone has not been demonstrated in other phase 3 randomized trials, despite earlier promise in phase 2. The authors point to mixed results elsewhere in the field — domvanalimab plus zimberelimab improved PFS and OS over zimberelimab alone in the phase 2 ARC-10 trial and is advancing into the phase 3 STAR-121 study, while the belrestotug program (dostarlimab plus belrestotug) was discontinued after the GALAXIES Lung-201 and GALAXIES Head & Neck-202 trials failed to show sufficient benefit.
Limitations
The trial enrolled only a PD-L1-high exploratory subgroup identified from the earlier CITYSCAPE study, so efficacy and safety in PD-L1-low or PD-L1-negative populations remain untested. Efficacy endpoints were assessed by investigators rather than an independent review facility, a design choice the authors acknowledge could introduce bias. Subgroup analyses, including by TIGIT expression level, were exploratory, with small sample sizes and wide confidence intervals that limit interpretation. The authors conclude that further research is needed to understand the underlying tumor biology driving inconsistent results across anti-TIGIT trials, and note that ongoing studies of rilvegostomig, a TIGIT/PD-1 bispecific antibody, may offer further mechanistic insight.
SKYSCRAPER-01, a global phase 3 trial in previously untreated PD-L1-high NSCLC, did not meet either co-primary endpoint: investigator-assessed progression-free survival and overall survival both numerically favored adding the anti-TIGIT antibody tiragolumab to atezolizumab, but neither difference reached the trial’s prespecified statistical significance boundary. Toxicity, including immune-mediated adverse events, was more frequent with the combination. The authors conclude the results discourage further investigation of anti-TIGIT agents in this PD-L1-selected population, though the trial does not address PD-L1-low or PD-L1-negative disease and used investigator-assessed rather than independently reviewed endpoints.