Insilico's generative AI-designed TEAD inhibitor shrinks mesothelioma tumours in Phase 1

Insilico Medicine reported first-in-human Phase 1 data for ISM6331, an oral pan-TEAD inhibitor its Chemistry42 generative AI platform helped design, on 26 October 2026, with the results presented as a Rapid Oral at the ESMO Congress 2026 in Madrid¹. Melina E. Marmarelis of the Abramson Cancer Center at the University of Pennsylvania presented the data¹.

What the trial found

The study (NCT06566079) had enrolled 36 participants across doses from 15 mg to 130 mg by the data cutoff of 14 August 2026, and dose escalation continues¹. Participants had a median of 2 prior lines of treatment, with a range of 1 to 12¹.

At 130 mg once daily, two of five response-evaluable mesothelioma participants responded, an objective response rate of 40.0%, with target-lesion reductions of 58% and 36%¹. Across the 30 mg to 130 mg cohorts the mesothelioma response rate was 18.2% (four of 22 evaluable participants) and the disease control rate 68.2%¹. Three of the four responders had already responded at the first Week-8 scan, partial responses appeared in patients after five and six prior lines, and three of the four partial responses were still ongoing at cutoff, one of them for roughly a year¹.

Insilico also reported target-lesion shrinkage in heavily pretreated non-small cell lung cancer and head and neck squamous cell carcinoma, and stable disease beyond four months in epithelioid hemangioendothelioma and colorectal cancer¹.

On safety, no dose-limiting toxicity occurred between 15 mg and 130 mg, and no treatment-related adverse event led a participant to stop¹. Grade 3 treatment-related events affected 8.3% of participants, among them asymptomatic amylase and lipase increases, fatigue and drug dermatitis¹. Treatment-related proteinuria, an adverse event of special interest, was seen in 11.1% and was low-grade and reversible¹. Falling levels of circulating CCN2 point to engagement of the TEAD pathway, and a terminal half-life of about 18 to 32 hours fits once-daily dosing¹.

What it means for the market

The readout matters less for its headline percentage than for what it says about generative chemistry reaching patients. TEAD has long been a hard target, and Insilico chief executive Alex Zhavoronkov described it as having "sat for years in the industry's 'too hard' category"¹. He called the result "generative chemistry becoming a clinical signal"¹. For pharma R&D teams weighing AI design platforms, an early efficacy signal against a difficult target is the kind of evidence that has so far been scarce.

The molecule already carries a US FDA Orphan Drug Designation, granted in June 2024, and a Fast Track Designation from July 2026, which is the company's first Fast Track programme¹. Insilico, listed in Hong Kong since 30 December 2025, also frames the Hippo-TEAD pathway as relevant to both cancer and ageing biology, a wider pitch than mesothelioma alone¹. Whether that broader opportunity holds up is not something these data address.

Caveats

The 40.0% figure comes from five evaluable patients at a single dose, and the broader 18.2% rate from 22¹. The trial is an open, early-phase dose-escalation study, the data are company-reported from a congress presentation, and dose escalation was still under way at cutoff¹.

Insilico plans to finish Phase 1 dose selection and then, depending on the emerging data, start a randomised Phase 2a dose-optimisation study in mesothelioma¹.