Performance characteristics of a novel handheld web-based device for 3D topographical detection and assessment of scoliosis

Abstract

Introduction: Current topographical assessment devices for scoliosis are based on one-dimension measurement. We have developed a 3D novel handheld web-based device, SpineScan3D™ based on a built-in gyroscope, for topographic assessment of the spine. The data is captured when rolled on a patient’s back in standing and forward bent position, allowing posture and 3D assessments. This data is directly uploaded to the cloud storage and allows prospective analyses and longitudinal follow-up. Additional clinical information such as health quality of life scores can be captured by the associated Smartphone application. A prospective observational study was performed to evaluate the reliability and validity of this device compared with the scoliometer, and it correlation with radiographic Cobb angle measurements. Material and Methods: All patients referred to our spine specialist outpatient clinic were consecutively recruited. Measurements of the back topography were done three times in the standing and forward bending positions. Intra-rater reliability was assessed by Cronbach’s alpha, and Pearson’s correlation was used to analyse its relationship with the Scoliometer. Standing whole spine posteroanterior and lateral views were taken as part of their routine follow-up assessments. Multivariable regression was performed to analyse correlation between the maximal tilt as detected by SpineScan3D™ and radiographic Cobb angle on the posteroanterior view. Results: 397 patients had Scoliometer and 89 had radiographic data for analyses. Data output were expressed as maximal lateral tilt to right (-ve) and to the left (+ve), as well as by region of the spine based on the spinal length scanned. High intra-rater reliability for the 3 scans performed in standing and forward bending positions with Cronbach’s alpha > 0.9. Pearson’s correlation analyses with angle of trunk rotation (ATR) on Scoliometer showed a correlation coefficient r = 0.39 for thoracic humps, and r = 0.531 for loin humps. Maximal regional tilt at the thoracic (2.97, 95% CI 1,92, 4.01; p < 0.001) and lumbar (−2.24, 95% CI -3.19, -1.28; p < 0.001) regions had good correlation with radiographic Cobb angle. Conclusion: SpineScan3D™ is a reliable assessment tool for a patient’s back topography, and provides valid measurements for patients’ ATR. SpineScan3D™showed good correlation with radiographic Cobb measurements in patients with scoliosis, with screening parameters for scoliosis (sensitivity and specificity) to be defined. If the performance characteristics are proven to be consistent and specific in a large scale validation study, this device can not only be used as a non-invasive screening tool for scoliosis, but a surrogate for radiographs in patients with known scoliosis undergoing longitudinal follow-up.