A spin-out company from the University of Glasgow has secured £1.6M seed investment to launch AI technology that is learning ‘the language of cancer’ to improve diagnosis and treatment of the disease.

TileBio has developed a platform that allows AI systems to learn the language of diseased tissues directly from millions of unlabelled medical pathology images to increase the speed and accuracy of life-saving diagnosis.

The launch seed-funding round has been led by Twin Path Ventures with participation from Scottish Enterprise and GU Holdings Ltd, the University of Glasgow’s investment company, and will support AI model training at a vast scale, clinical validation studies and team expansion.

The investment will provide runway for the new company to achieve key technical and validation milestones required.

The company emerged from academic research led by Dr Ke Yuan, Dr Adalberto Claudio Quiros, Professor John Le Quesne, Professor David Chang, and Dr Christopher Walsh on computational pathology and representation learning applied to medical images.

A critical differentiator for TileBio is access to large-scale, real-world clinical data.

Its long-term aim is to train one of the largest pathology foundation models globally using more than two million whole slide images from NHS Greater Glasgow and Clyde, working closely with the organisation through formal data governance frameworks.

The self-learning nature of this model will enable cancer detection and triage in the clinic across all cancer types and human disease at the population scale. It will also enable novel biomarker discovery to accelerate drug development and improve therapeutic targeting, while maintaining model interpretability and providing biological insights.

As the company scales, recruitment will focus on deep learning engineers, data managers, regulatory specialists and commercial leads with experience in regulated medical software.

Dr Walsh, TileBio Chief Executive Officer, said: “We are delighted to launch and work with our partners and investors to grow our technology and scale up what we have worked so hard to achieve over many years.

“The core scientific insight underpinning TileBio is that histology contains an inherent structure that can be learned directly from raw images without requiring large volumes of human-annotated labels.

“Traditional medical AI has been constrained by the need for expensive, biased, and labour-intensive ground-truth data. We have developed a method that allows AI systems to interpret the language of tissue directly from millions of unlabelled images.”

Recommended reading

• Glasgow Spinouts Win £540K Boost to Turn Research Into Innovation
• UK Spinouts Look to US as Scale-Up Funding Stalls
• Report: UK Spinouts Are Powering a Deeptech Boom

Uzma Khan, Vice Principal, Economic Development and Innovation, University of Glasgow, said: “We are incredibly proud of the TileBio team in successfully achieving their first raise.  Their success is a powerful endorsement of the scientific excellence behind the venture combined with a leap forward in using AI models.

“It is also evidence of the level of entrepreneurial ambition emerging from the University of Glasgow. We have been delighted to support TileBio in their commercialisation journey through our de-risking funds and venture builder programmes such as the MedTech Innovation Fund and the Infinity G accelerator.

“The University looks forward to supporting their continued growth and success as they take this important next step.”