VheaRts is a virtual reality application built by the clinical cardiovascular engineering group at University College London in partnership with clinicians at Great Ormond Street Hospital for Children. It turns a patient's own CT and MRI scans into an interactive 3D anatomical model that a clinician can pick up, rotate, slice, highlight and measure inside a VR headset - spatial detail that is hard to reconstruct mentally from flat scan slices, particularly in complex congenital heart disease.
The application began as a research tool at UCL in 2018 and has been in use at GOSH ever since. A single-centre review published in 2024 reports 67 cases referred for VR assessment between late 2018 and May 2024, double outlet right ventricle being the most common lesion; by early 2023 the team reported the app had been used in planning more than 50 surgical procedures.
A networked version puts several people into the same virtual room around one model. GOSH used it in 2022 to plan the separation of craniopagus conjoined twins in Rio de Janeiro alongside the Brazilian surgical team, in months of shared VR sessions where both teams could see how the twins' skulls, brains and blood vessels were intertwined and work through the cuts that would be needed. The twins were successfully separated. The same shared-room feature is used to explain surgery to families: the parents of a child undergoing craniosynostosis correction entered a virtual room with their surgeon and were walked through the operation on their son's own skull model, which GOSH reports reduced procedural anxiety and improved the consent conversation.
On the education side, Professor Andrew Cook delivered his UCL MSc and iBSc congenital heart anatomy teaching in VheaRts through the COVID-19 lockdowns, sending headsets to students in place of the anatomy lab, and has run remote workshops with paediatric cardiologists and surgeons in Johannesburg. At the 2023 World Congress of Pediatric Cardiology and Cardiothoracic Surgery in Washington DC the team delivered four days of oversubscribed VR anatomy sessions - more than 30 people in the virtual lab at a time across eight workshops - with over 500 clinicians trying the application.
The app is built in Unity and runs on Oculus Rift and Quest headsets. Segmenting each patient's scans into a usable model is still done by hand by the engineering team, which the developers describe as the main barrier to routine, on-demand use.
