A recent Nature Review Drug Discovery article caught our attention, a review of TIGIT as an IO target and the failed Phase 3 study of tiragolumab which is the TIGIT inhibitor from Roche.[1] The science writer interviewed a group of insiders that had a lead role in the development of the drug and deep understanding of its MoA. No doubt these researchers have insights and experiences to share about things gone right and things done wrong.
In corporate speak, this is called a ‘teaching moment’ or a ‘Lessons Learned’ exercise. In medicine, we have something similar, the post-mortem or M&M conference. The declared goal for either is pretty much the same: Learn from mistakes in order to avoid them in the future.
If you expected advice along these lines for TIGIT as a target in general and tiragolumab development in particular, you’d be disappointed, because that’s not what we get served in the article. This is not the interviewer’s fault but a systematic unwillingness to openly review a decision process that did not end with the desired outcome.
This article provides a rather superficial apology for the millions of dollars spent on the development of tiragolumab and the billions of dollars spent by companies trolling the TIGIT target. It tells the history of tiragolumab: its promising research beginnings, Phase 2 success, and the unexpected failure in the pivotal Phase 3 study. The key question “What exactly went wrong” is answered only in generic ways, with some platitudes mixed in for good measure.
This biological, brainchild of Roche/Genentech, made sense scientifically and biologically. Promising preclinical and excellent Phase 2 data suggested benefit in NSCLC. With seemingly compelling results in the CITYSCAPE Phase 2 trial [2], the project really took off at Roche. The competition did not want to miss out on the next blockbuster and was quick to jump on the TIGIT bandwagon.
Paired with a PD-1 inhibitor for synergistic efficacy in NSCLC , tiragolumab was seen as a potential blockbuster drug for Roche’s oncology portfolio.
Corporate risk-taking has become even more the norm since
the PD-1 race, when BMS’s nivolumab lost out to Merck’s pembrolizumab.
Following the Perlmutter / Merck example, tiragolumab was developed
like the second coming of pembrolizumab.
Alas, it was not to be…like Icaros, tiragolumab’s wings came too close to the sun. When projects fail late and unexpectedly, soul-searching starts, right? Well, not really.
Companies write off the losses of a failed program, lick their wounds, then move on quickly and reorganize. Frankly, they don’t spend much time on ‘Lessons Learned’; after all, it would be a painful exercise for the protagonists of the project and decision makers still around.
Corporate managers cannot divulge learnings; they are insiders and contractually held to confidentiality. Hence, the “Learnings” in the Nature article are generic, decrying “the heterogeneity of cancer and the complexity of the human immune system”. True, but we know that already!
Here we summarize the ‘TEACHINGS’ taken directly from the Nature article;
OUR COMMENTS and ‘LEARNINGS’ in BLUE:
It is our firm belief that the Phase 2 “High PD-1” analyses should have been treated as hypothesis-generating, to be confirmed in a new Phase 2 trial. This subset population should never have become the basis of a large pivotal Phase 3 trial like SKYSCRAPER.
SKYSCRAPER-01 enrolled 534 patients with “High PD-L1”- NSCLC in an attempt to confirm the subset data from Phase 2. We are not at all surprised that this trial failed; the entire program was subsequently stopped.[1],[2]
Actually, Roche’s decision to use data from an exploratory subset as the basis for a pivotal trial is a common mistake. We still remember tifacogin, the TFPI inhibitor program from Chiron/Novartis. In that program, a Phase 2 study in severe sepsis (OPTIMIST) showed little benefit, but a retrospective analysis of the trial identified a subset of septic patients with ‘CABP’ that seemed to benefit from tifacogin. This factoid clinched it for corporate decision-makers at Chiron. They embarked on a 2,100 patient Phase 3 program in septic CABP patients study called CAPTIVATE. It failed.[5],[6]
The tifacogin / TIFPI program was a LEARNING experience and a TEACHING moment for all involved (it was not reviewed in Nature at the time).
The Roche : Novartis score now stands at an even 1:1. We hope this game of blunders does not go into overtime!
REFERENCES
[1] Mullard A. TIGIT’s immuno-oncology teachings. Nature Reviews Drug Discovery 2026; 25: 157
[2] Cho B. Tiragolumab plus atezolizumab versus placebo plus atezolizumab as a first-line treatment for PD-L1-selected non-small-cell lung cancer (CITYSCAPE): primary and follow-up analyses of a randomised, double-blind, phase 2 study. Lancet Oncol 2022; 23: 781
[3] PR Roche Nov 25, 2024. Roche reports update on Phase III SKYSCRAPER-01 study results (accessed 4/18/2026)
[4] Peters S. Abstract CT051: SKYSCRAPER-01: A phase III, randomized trial of tiragolumab (tira) + atezolizumab (atezo) versus placebo (pbo) + atezo in patients (pts) with previously-untreated PD-L1-high, locally advanced unresectable/metastatic NSCLC. Cancer Res (2025) 85 (8_Supplement_2): CT051. https://doi.org/10.1158/1538-7445.AM2025-CT051
[5] Abraham E. Efficacy and Safety of Tifacogin (Recombinant Tissue Factor Pathway Inhibitor) in Severe Sepsis. A Randomized Controlled Trial. JAMA 2003;290:238
[6] Laterre P. A clinical evaluation committee assessment of recombinant human tissue factor pathway inhibitor (tifacogin) in patients with severe community-acquired pneumonia. Critical Care 2009, 13: R36 (doi:10.1186/cc7747). http://ccforum.com/content/13/2/R36
ABBREVIATIONS
IO immuno-oncology
ITIM Immunoreceptor Tyrosine-based Inhibitory Motif
M&M morbidity & mortality
NSCLC non-small cell lung cancer
TFPI Tissue Factor Pathway Inhibitor
TIGIT T-cell immunoreceptor with Ig and ITIM domain
