Currently, radiology departments are short-staffed, and Magnetic Resonance Imaging (MRI) machines are under-used. With improved AI technology to analyse MRI images, these expensive machines could be used for preventative medicine as well as diagnosis. This would allow radiologists to use their time more effectively, enable faster diagnosis, and ultimately save lives.
Twinn Health is solving a significant unmet need with its AI cloud-based platform designed to analyse MRI images and detect metabolic disease. Leveraging technology employed in oncology detection systems and a proprietary database, Twinn has developed segmentation and inference models trained on ethnically diverse, high- contrast MRI datasets where adipose tissue is seen as a clear white pixel. Creating a cloud-based platform that integrates into patient archive systems allows radiologists to process MRI images seamlessly. They can detect metabolic disease faster and more accurately than conventional blood tests, localising and assessing disease severity.
Building on a diverse dataset, Twinn Health has validated its algorithm accuracy using a representative dataset, recognising that visceral adiposity has different phenotypes across racial groups.
Based on initial feasibility studies, Twinn Health calculates that its platform saves five hours of radiologists’ time per patient assessment, valued at £720. The time saved by radiologists during the validation study will be monitored to confirm the cost value and determine the economic utility for adopters. Twinn Health predicts that its platform will save radiologists an average of five hours of analysis time, making MRI competitive.
Where did the idea originate?
The founder completed her PhD in Diagnostic Medical Imaging where she faced this problem on a daily basis. She analysed medical images for diagnosis manually for an average of 5 hours per scan going through every image to get a diagnosis. This process was time consuming, requiring high-skilled labour and a large budget. At Antler UK, the founder met with an AI image detection expert (Jamie) and a spinout consultant ( Chris) and together they started working on a technology platform for automating disease detection.
Other AI solutions in radiology focus on oncology, triage or diagnosing cardiovascular emergencies rather than identifying classically challenging adipose tissue and preventative medicine. BEIS has invested £50M through UKRI to establish five Digital Pathology and Imaging AI Centres of Excellence and drive NHS adoption (GOV.UK, 2021).
Twinn Health trialled a limited-functionality PoC with three radiologists in November 2021. Blind data showed a 97% sensitivity score and a specificity of 94%. Recommendations for algorithmic risk adjustment and clinical workflow improvement were fed into subsequent iterations to enhance UX and clinical compatibility.
How did the team meet?
The founder was part of a global early stage VC called Antler UK where she met the rest of the team.
Do you have any advisors?
We gathered top level advisors from the founder network and also some who were inspired by the company vision on live demos.
One of our advisors is Dr Rishi Ramaesh, UX/UI/regulatory lead, who is a certified consultant radiologist with a decade of experience in all stages of life science research and development. He is a regional AI-imaging lead and advises the Scottish government on digital healthcare and AI tools. Rishi has experience with startups, industry partners and academic institutions. He has helped turn innovative ideas into viable digital clinical products, validated products in clinical trials and advised on clinical aspects of digital healthcare.
Where are you now and where do you plan to be?
This industrial research project aims to deliver a cloud-based platform (TLR8) with a clinically validated AI system for diagnosing metabolic disease. We are raising £7-10 million to gain FDA approval of the technology and maximise commercial traction.
What support have you had from Imperial?
Twinn Health has participated in two advice programmes as the Enterprise Lab. They participated in the Experts-in-Residence programme in 2020, where the access to experts and mentorships particularly helped to move their idea forward. They also participated in the Idea Surgery programme in 2021, which identified main challenges as the team developed their ideas further.
What’s been your biggest success and challenge so far?
Securing our patent was our biggest success so far. Another success we are proud of is getting top notch advisors including the president of the European Society for Preventive medicine. Additionally, one of our clients has turned into an investor and shareholder. Nonetheless, the regulatory landscape especially with Brexit is one of our challenges.
What advice would you give an aspiring entrepreneur?
DO NOT QUIT.