- Singular and CSIRO collaboration completes successful automated segmentation of spinal vertebrae from CT scans with >95% accuracy.
- Target market spinal surgeries currently require pre-operative planning with manual segmentation consuming valuable time by surgeons and radiologists
- Artificial Intelligence (AI) model deployed within Singular’s MedVR & 3DicomViewer can automatically segment a spine in just 2 minutes.
- Segmented spine is editable by radiologists and surgeons to ensure 100% accuracy.
18 March 2021 – Singular Health Group Ltd announced today that a collaborative research project conducted with CSIRO, Australia’s national science agency, has successfully concluded the development of an artificial intelligence model capable of automatically segmenting spinal vertebrae from CT scans.
The automated instance segmentation and labelling of spinal vertebrae allows surgeons and radiologists to not only visualise the spine in 3D using the Company’s MedVR & 3DicomViewer software but to manipulate individual vertebrae, better plan their surgeries and even design custom surgical implants and guides.
The project is made possible through CSIRO Kick–Start, an initiative that provides funding and support for innovative Australian start-ups and small businesses to access CSIRO’s research and development (R&D) expertise and capabilities.
Dr Guan Tay, Executive Director of Innovation at Singular Health said: “The use of artificial intelligence in medical imaging, and more specifically radiology, has the ability to profoundly change the workflow for radiologists.
With around 45,000 Australians undergoing spinal surgery every year, the rapid segmentation will save thousands of hours for radiologists and surgeons who will only have to make small mark-ups and/or validate the segmentation as opposed to manual segmentation slice-by-slice.”
Taking just over two minutes to perform the analysis and segmentation, the AI model eliminates hours of tedious manual segmentation by medical practitioners and is being labelled as a semi-automated segmentation process as it will allow surgeons and radiologists to edit the output and ensure 100% compliance with their interpretation of the image whilst still saving vast amounts of processing time.
The CSIRO research lead, Dr Dadong Wang, Quantitative Imaging Research Team Leader at CSIRO’s Data61, said: “CSIRO has been able to apply data science and Artificial Intelligence to enable users to better visualise the spine and individual vertebrae. This semi-automated segmentation and labelling of the spinal vertebrae is both time saving and can be represented in 3D models, which can also be viewed in virtual reality, providing higher interactivity.”
The spine segmentation function will be further tested and validated before being considered for commercial release.