Hoffmann-La Roche has terminated its Phase 3 SKYSCRAPER-15 trial of tiragolumab plus atezolizumab in resected, PD-L1-positive non-small cell lung cancer, abandoning the primary efficacy objective and converting the study to a safety-only follow-up — the company’s fourth consecutive Phase 3 failure for its anti-TIGIT antibody and a further blow to the TIGIT checkpoint class that generated considerable industry optimism as recently as 2022.

The trial (NCT06267001) enrolled patients with completely resected stage IB–IIIA PD-L1-positive NSCLC and randomized them to receive tiragolumab, an investigational anti-TIGIT monoclonal antibody, in combination with atezolizumab (Tecentriq, an approved anti-PD-L1 antibody) or placebo plus atezolizumab as post-surgical adjuvant therapy. The trial enrolled just 56 participants — a fraction of its originally projected enrollment — before Roche suspended efficacy analysis and reclassified the study’s primary objective to safety follow-up only, as reflected in a ClinicalTrials.gov status update this week.

A pattern of failure

SKYSCRAPER-15 is the fourth successive Phase 3 setback for tiragolumab:

  • SKYSCRAPER-01 (first-line metastatic NSCLC, PD-L1 ≥50%): tiragolumab plus atezolizumab failed to improve progression-free or overall survival versus atezolizumab alone.
  • SKYSCRAPER-02 (extensive-stage small cell lung cancer): tiragolumab plus atezolizumab plus carboplatin/etoposide produced no benefit over placebo plus chemotherapy.
  • SKYSCRAPER-06 (advanced gastric or gastroesophageal junction cancer): no overall survival benefit observed.
  • SKYSCRAPER-15 (resected NSCLC, adjuvant): efficacy objective abandoned before analysis.

Broader TIGIT context

Roche’s experience with tiragolumab mirrors that of other developers. Merck’s vibostolimab-pembrolizumab combination failed to improve outcomes in the KEYVIBE-008 NSCLC trial. Results from domvanalimab-based programs (AstraZeneca/Sanofi) and ociperlimab-based programs (BeiGene/Novartis) have also been disappointing in lung cancer. The consecutive failures have renewed debate about whether TIGIT inhibition offers meaningful benefit as an add-on to PD-1/PD-L1 blockade, or whether the benefit, if any, is confined to specific tumor types, biomarker-defined subgroups, or TIGIT antibody formats that block different mechanisms of immune suppression.

Tiragolumab was designed to block TIGIT-mediated inhibition of T-cell and NK-cell activity in the tumor microenvironment. Its co-inhibitory relationship with PD-1/PD-L1 checkpoints was the theoretical basis for combination trials. Roche has not announced a broader program wind-down for tiragolumab, but the SKYSCRAPER-15 termination effectively forecloses the adjuvant lung cancer indication.