Roche unveiled the full scope of its failed TIGIT blockade program at ESMO 2025, presenting results from four pivotal studies of tiragolumab that collectively led to the drug's discontinuation in July. The comprehensive data release, including two late-breaking presentations, revealed consistent failures across multiple cancer types and represents one of the most significant setbacks in recent cancer immunotherapy development.
Late-Breaking Failures Seal TIGIT's Fate
The most anticipated presentations came from Imbrave-152, tiragolumab's final opportunity for success in first-line liver cancer, and Skyscraper-03 in stage III non-small cell lung cancer maintenance. Both studies failed their primary progression-free survival endpoints with hazard ratios near 1.0, indicating virtually no benefit from adding tiragolumab to standard care.
In Imbrave-152, the combination of tiragolumab plus Tecentriq and Avastin achieved a median PFS of 8.3 months compared to 8.2 months for Tecentriq/Avastin alone (HR=0.97, p=0.7464). The PFS curves for both treatment arms "lying virtually on top of each other" demonstrated the complete lack of efficacy. While overall survival data remains immature with a hazard ratio of 0.94, investigators stated the results are "not expected to reach statistical significance."
Skyscraper-03, testing tiragolumab plus Tecentriq versus Imfinzi in stage III NSCLC maintenance, similarly failed with a median PFS of 19.4 months versus 16.6 months in PD-L1≥1% patients (HR=0.96, p=0.7586). The study showed "absolutely no benefit across the entire study" with crossing survival curves that eliminated any meaningful clinical advantage.
Broader Pattern of Failure Across Indications
The ESMO presentations extended beyond the late-breakers to include results from Skyscraper-07 in esophageal squamous cell cancer and Skyscraper-09 in head and neck cancer. Skyscraper-07 delivered perhaps the most concerning results, where tiragolumab plus Tecentriq actually performed numerically worse than Tecentriq alone on both PFS and OS while adding toxicity.
The three-cohort design of Skyscraper-07 revealed that the tiragolumab combination achieved a median PFS of 20.8 months versus 16.6 months for placebo (HR=0.82, p=0.947) and median OS of 38.6 months versus 36.4 months (HR=0.91, p=0.4772). However, the Tecentriq monotherapy cohort demonstrated "clinically meaningful improvements" with a PFS of 29.1 months versus 16.6 months for placebo (HR=0.74) and overall survival that was not reached versus 36.4 months (HR=0.69).
Small Glimmers of Hope Amid Comprehensive Failure
Despite the overwhelming negative results, investigators identified limited areas of potential interest. The Skyscraper-09 study in PD-L1≥5% head and neck cancer showed numerical improvements in response rates (21.3% vs 15.4%) and median OS (16.2 vs 13.6 months). A post hoc analysis of patients with PD-L1≥20% expression appeared "especially promising," delivering median OS of 22.7 versus 10.0 months.
However, Skyscraper-09 was a small phase 2 study "not designed to compare efficacy across its two cohorts" and was discontinued in January. The real comparator in head and neck cancer remains Keytruda, which gained approval based on the Keynote-048 trial.
Lessons from a Costly Misstep
The tiragolumab program's failure stems from what appears to have been an over-interpretation of the mid-stage Cityscape study in first-line PD-L1-expressing NSCLC. That promising result "was shown to have been illusory" when the phase 3 Skyscraper-01 trial yielded only a non-significant 13% reduction in risk of death (p=0.22) at AACR earlier this year.
The comprehensive failure across multiple indications will have cost Roche "many hundreds of millions of dollars to run," representing a significant setback for the company's oncology pipeline. The presentations at ESMO serve primarily to "guide future science and pivotal study decision-making" rather than indicating any path forward for tiragolumab development.
TIGIT Mechanism Under Scrutiny
The tiragolumab failures add to growing concerns about the TIGIT mechanism across the industry. Other prominent TIGIT programs, including Merck & Co's vibostolimab and iTeos/GSK's belrestotug, did not feature in ESMO presentations, while Arcus/Gilead's domvanalimab "looks increasingly risky."
AstraZeneca remains the most prominent company still invested in TIGIT with its anti-TIGIT x PD-1 bispecific rilvegostomig, which featured in ESMO oral sessions through the Artemide-01 and Tropion-Pantumor03 studies.
Despite the comprehensive failure, Roche received recognition for its transparency in presenting the negative data rather than "quietly burying" the results. The company's willingness to share these failures at major medical conferences provides valuable insights for the broader oncology community about the challenges of targeting the TIGIT pathway in cancer immunotherapy.
