The U.S. Food and Drug Administration has granted Fast Track Designation to ISM6331, an experimental cancer drug that Hong Kong-based Insilico Medicine says its generative AI software designed from scratch, marking the first time any program in the company’s AI-driven pipeline has won the status. The decision, confirmed July 29, 2026, applies to adults with unresectable malignant pleural mesothelioma whose disease has progressed after anti-PD-1 immunotherapy, with or without anti-CTLA-4 therapy, and platinum-based chemotherapy — a narrow, hard-to-treat population with few remaining options.
From Algorithm to Molecule
ISM6331 is a small-molecule pan-TEAD inhibitor, meaning it blocks a family of transcription factors that drive tumor growth when the Hippo signaling pathway malfunctions — a defect found in the large majority of malignant mesotheliomas, which are most often linked to past asbestos exposure and are diagnosed in roughly 3,000 Americans a year, according to figures the American Cancer Society has long cited. Insilico says the compound emerged from Chemistry42, its proprietary generative chemistry engine, which proposes and scores candidate molecules computationally before a single one is synthesized in a lab, compressing a discovery process that traditionally takes years into a fraction of the time. Insilico has used the same AI-first approach across its broader pipeline, pairing Chemistry42 with a separate AI system for identifying disease targets and another for helping design trial protocols, an end-to-end bet that software can compress drug development timelines that have historically run a decade or longer from first idea to approval.
A First for Insilico’s Pipeline
ISM6331 already held FDA Orphan Drug Designation for mesothelioma, granted in June 2024, and is now being tested in a global, multicenter Phase 1 trial (NCT06566079) that dosed its first patient in January 2025. The study is tracking dose-limiting toxicities and side effects to find a recommended Phase 2 dose, while also watching for early signs of tumor shrinkage, tracking secondary measures like objective response rate and duration of response. “Receiving fast track designation validates the strong clinical potential of ISM6331,” said Halle Zhang, Insilico’s vice president of clinical development for oncology, adding that the company hopes to work more closely with the FDA as it speeds up clinical evaluation. It is the first time any asset from Insilico’s AI-driven pipeline has reached Fast Track status, a marker the company is likely to point to as it seeks partners and investment for the rest of its roughly 30-program portfolio.
What Fast Track Actually Buys
Fast Track status is not an approval and carries no claim of proven benefit — the drug is still in Phase 1, the earliest and riskiest stage of human testing. What it does provide is more frequent meetings with FDA reviewers, written feedback on trial design and biomarker strategy, and, if the drug continues to show promise, eligibility for rolling review, priority review and accelerated approval later in development. Those mechanisms can shave months or years off a drug’s path to market, assuming the underlying data holds up.
Skepticism Persists
That caveat matters because the broader field of AI-designed drugs has yet to prove it beats the industry’s historically brutal odds. An ASCO 2026 analysis counted 117 AI-enabled therapeutic candidates from 63 companies in human trials, of which 60 had completed Phase 1 but only eight had finished Phase 2. A Boston Consulting Group analysis cited by the outlet ProMarket found that while Phase 1 safety success rates for AI-designed drugs have climbed to 80–90%, Phase 2 efficacy success remains stuck near the industry’s long-standing average of about 40%. In other words, AI appears to be getting molecules through the front door of clinical testing faster and more safely, but it has not yet shown it can out-predict biology on whether a drug actually works. Veteran drug developers who are skeptical of the generative-AI hype cycle note that mesothelioma itself has frustrated conventional and targeted therapies alike for decades, with chemotherapy remaining the backbone of first-line treatment for most patients since the early 2000s, so a Phase 1 safety signal, while encouraging, is a long way from changing that picture.
What’s Next
Insilico plans to present first-in-human data from the Phase 1 trial as a rapid oral presentation at the European Society for Medical Oncology Congress, running October 23–27 in Madrid — the first real public look at how ISM6331 is performing in patients. The company also points to preclinical signals suggesting the drug could work in combination with other therapies and help overcome drug resistance. If the Phase 1 results hold up, the Fast Track status could accelerate a path toward a pivotal trial; if they don’t, ISM6331 will join the long list of promising AI-designed compounds that stalled well short of a pharmacy shelf. Either way, mesothelioma patients and their doctors are watching a test case for whether generative chemistry can translate algorithmic promise into a genuine new treatment option for a cancer that has seen little innovation in decades.
Photo: Sunriseforever / PIXABAY via Pixabay
