Oxford spinout RADiCAIT uses AI to make diagnostic imaging more affordable and accessible — catch it at TechCrunch Disrupt 2025

Undergoing a PET scan can be a daunting experience. While these scans are crucial for detecting cancer and monitoring its progression, the process is often fraught with challenges for patients. From mandatory fasting for several hours to the logistical difficulties of reaching hospitals that house these specialized machines, the ordeal can be overwhelming. Once at the facility, patients must endure a lengthy waiting period after being injected with radioactive tracers before lying still for a significant amount of time to obtain the required images. Furthermore, those undergoing a PET scan must keep their distance from vulnerable populations, such as young children and pregnant women, for hours afterward due to residual radioactivity. A significant barrier to accessing PET scans is their concentration in urban centers, where the necessary radioactive tracers can be produced nearby. This leaves those in rural areas at a disadvantage. However, a groundbreaking solution is emerging from RADiCAIT, a startup from Oxford University, which aims to leverage artificial intelligence to convert more readily available CT scans into PET scans. Recently unveiling itself and securing $1.7 million in pre-seed financing, RADiCAIT is a contender in the Startup Battlefield at TechCrunch Disrupt 2025 and is currently raising an additional $5 million to propel its clinical trials forward. Sean Walsh, the CEO of RADiCAIT, explained, "We have taken the most complex and costly medical imaging solution in radiology and replaced it with a more accessible and affordable alternative: the CT scan." The company’s innovation hinges on a generative deep neural network, developed in 2021 at the University of Oxford, which identifies patterns by comparing CT and PET images. This model is adept at translating anatomical data into functional insights, focusing on key characteristics like tissue types and anomalies. RADiCAIT’s method has drawn comparisons to Google DeepMind’s AlphaFold, which has transformed protein structure predictions. The team asserts that its AI-generated PET images can statistically match those from traditional chemical PET scans, thus providing a reliable alternative for diagnostic evaluations. While RADiCAIT does not aim to replace PET scans in every medical scenario, it does envision its technology as a potential game-changer for diagnostics, staging, and monitoring, addressing the existing supply-demand gap in medical imaging. The startup has embarked on clinical trials for lung cancer in collaboration with major health systems like Mass General Brigham and UCSF Health. Now, RADiCAIT is moving toward FDA clinical trials as part of its expansion strategy, which is driving the current funding round. Future plans include launching similar initiatives for colorectal and lymphoma cases, with aspirations to broaden the application of AI in medical imaging. Sina Shahandeh, RADiCAIT’s chief technologist, believes that their approach could revolutionize the field, stating, "We are looking to extend our insights across various domains in radiology and beyond, exploring the hidden relationships in materials science, biology, chemistry, and physics." To learn more about RADiCAIT and its innovative technology, attendees can join the discussions at Disrupt, taking place from October 27 to 29 in San Francisco.