Artificial intelligence in radiology: Two new algorithms for spine images
The abstracts detail the validation of two novel, AI-based methods for the automatic analysis of spinal X-ray images. Future applications of the methods may include rapid analysis of large datasets for research and X-rays in clinical routine.
Development and validation of robust, clinically relevant, AI-based medical image analysis methods stands at the center of RAYLYTIC’s efforts. Therefore, leading spine surgeons are closely involved in this process and will be presenting the abstracts to the spine community come November.
On November 29, 2023, Dr. Jörg Franke, MD, PhD – director of orthopedic surgery at Clinical Magdeburg – will present “Automatic determination of lumbar segmental lordosis based on an AI-algorithm – a validation study on 170 preoperative full-body X-rays.” Segmental lumbar lordosis (LL), despite being a common parameter in the diagnosis and surgical treatment of spinal disorders, has not yet been the subject of automated analysis techniques.
Following Dr. Franke’s presentation, Dr. Clara Berlin, orthopedic and trauma surgeon at Schoen Clinic Group, will present “Validation of fully automatic measurement of pre- and postoperative coronal parameters of 100 adolescent idiopathic scoliosis patients using artificial intelligence” on November 30. This study builds on a previous publication in the Global Spine Journal through the inclusion of postoperative parameters – a notorious challenge for automatic analysis methods due to the presence of spinal instrumentation. Validation of the algorithm for preoperative X-ray analysis shows that the following parameters can be measured automatically in just 15 seconds:
- T1-Tilt
- Clavicle Angle
- Coronal Balance
- Cobb Angles (Proximal Thoracic, Thoracic, Thoracolumbar/Lumbar)
- Lumbar Modifier
This speed may significantly relieve surgeons, clinicians, and researchers of time-consuming manual measurements in clinical routine and in the analysis of large data sets.
Automatic segmentation (identification of relevant anatomy) of a patient radiograph displaying characteristics of scoliosis.
More on the German Spine Society (DWG)
The German Spine Society (DWG) was founded in 1987 in Hamburg. Today, over 1,250 members – primarily neurosurgeons, trauma surgeons, and orthopedics – participate in the DWG’s education and research activities, which aim to disseminate the latest technical and research developments in the field of spine surgery.
The DWG’s annual conference will take place this year in Stuttgart, Germany from November 29th to December 1st, 2023.