An automated system for Perthes' disease prediction and treatment strategy selection

The authors have proposed an algorithm of modeling this disease development for different variants of its treatment based on the abovementioned discriminatory analysis in order to automate the selection of treatment in patients with Perthes disease. 13 clinical-andradiological patient parameters have been determined in the technique of Perthes disease outcome prediction: patients age at primary diagnosis, the upper subluxation of the femoral head, the angle of acetabulum opening, lateral displacement of the femoral head, change in the the acetabulum radius, the hip abduction, type of the treatment performed. The coordinate values of the canonical linear discriminatory functions have been calculated for clinical (RootF1k and RootF2k) and radiographic outcomes (Root F1r, Root F2r).