Search

This paper describes a method of characterizing the torso shape deformity associated with scoliosis by both its type and severity. This problem is challenging because regular human torsos show an astounding range of variations that is only compounded by scoliosis, and it is difficult to isolate natural shape variations from those caused by scoliosis. Torso shape characterization is important in the clinical management of scoliosis because torso aesthetics is a key concern that influences a patient's quality of life. Our method involves modeling 3-D torso range images into structured sequences of 3-D spline curves stacked along the spine. We obtain local shape measures from points of maximal curvature (dominant points) along each torso cross section by evaluating the relative symmetry of the spline curve at that cross section. This results in a scalable characterization scheme for torso deformity type and a measure of torso deformity severity. We assess the accuracy and precision of this shape characterization scheme, and its relationship to the actual deformities present in the underlying spine. Index Terms--Dominant points, object recognition, scoliosis, shape analysis, splines.

A support vector classifier (SVC) approach was employed in predicting the risk of progression of adolescent idiopathic scoliosis (AIS), a condition that causes visible trunk asymmetries. As the aetiology of AIS is unknown, its risk of progression can only be predicted from measured indicators. Previous studies suggest that individual indicators of AIS do not reliably predict its risk of progression. Complex indicators with better predictive values have been developed but are unsuitable for clinical use as obtaining their values is often onerous, involving much skill and repeated measurements taken over time. Based on the hypothesis that combining common indicators of AIS using an SVC approach would produce better prediction results more quickly, we conducted a study using three datasets comprising a total of 44 moderate AIS patients (30 observed, 14 treated with brace). Of the 44 patients, 13 progressed less than 5° and 31 progressed more than 5°. One dataset comprised all the patients. A second dataset comprised all the observed patients and a third comprised all the brace-treated patients. Twenty-one radiographic and clinical indicators were obtained for each patient. The result of testing on the three datasets showed that the system achieved 100% accuracy in training and 65%-80% accuracy in testing. It outperformed a "statistically equivalent" logistic regression model and a stepwise linear regression model on the said datasets. It took less than 20 min per patient to measure the indicators, input their values into the system, and produce the needed results, making the system viable for use in a clinical environment. [ABSTRACT FROM AUTHOR]