Back

Development and validation of a digital twin for the analog scoliometer

Suresh, S.; Stubbs, A.; Amiri, A.; Izatt, M. T.; Labrom, R. D.; Askin, G. N.; Little, J. P.

2023-12-03 pediatrics
10.1101/2023.11.30.23298978 medRxiv
Show abstract

Scoliosis is a complex 3D spine deformity characterised by an abnormal lateral curvature of the spine and associated rotation of the spine and ribcage. The rotational aspect of scoliosis is most commonly quantified in the Adams forward flexed position using an analog scoliometer. The scoliometer has a known user error of 5-8{degrees}, which is largely dependent on examiner experience, location of curve, patient positioning and BMI. The device is also limited by the 30{degrees} scale and parallax errors. Additionally, the scoliometer loses accuracy when the patients torso cannot be positioned parallel to the ground . This study describes the development of the first digital twin for the analog scoliometer to enable fast, gravity-independent reliable and accurate digital measurements of the Angle of Torso Rotation (ATR) from patient-specific 3D virtual models. A robust semi-automated algorithm of generative design which measures ATR from surface topography was developed. With an operating time of just a few seconds, it provides quick and reliable ATR measurements from simple parametric user inputs. 150 calibrated 3D virtual models of AIS patients treated at the Queensland Childrens Hospital Spine Clinic (QCHSC) obtained from our existing database of 3D surface scans (3DSS) and healthy non-scoliotic controls recruited for this study were used to validate the digital scoliometer tool. The tool showed excellent reliability in both intra-user (0.99) and inter-user (0.98) conditions. The digital values had a high positive correlation (0.897) and agreement (92.7%) with the analog ATR measurements made clinically. The tool also showed high sensitivity (95.83%) and specificity (76.76%). The development and validation of this virtual digital tool is significant for telehealth implementation in paediatric spine deformity management and is expected to enhance the remote health management of scoliosis.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.