The Establishment And Evaluation Of A New Three-Dimensional System For Intertrochanteric Fractures Based On Big Data And Machine Learning

BackgroundIntertrochanteric fractures are the most common fractures in elderly patients,often leading to a decline in limb function and even self-care ability,which brings heavy medical and economic burdens to society and families.A good reduction of the fracture end is the key to got successful surgical treatment.But the criteria for the treatment of closed reduction of the intertrochanteric fracture are still inconclusive.Fracture classification is crucial to guide the treatment of fractures and to expect the outcome of prognosis.However,the existing classifications have been based on the reference X-ray films,with combination of simple morphological features and summarization of the characteristics.This unclear description of the fracture details often leads to ambiguity in the clinical practice.At present,there is no fracture classification system that guides the closed reduction and forecast the stability after the intramedullary nail fixation.Purpose(1)To compare the clinical outcomes and complication rates of different reduction types-medial side of triangle structure in proximal femur using CT findings in 31-A2 intertrochanteric fractures treated with the intramedullary nail.(2)To determine the major trajectories of intertrochanteric fracture lines using computed tomography data from a series of surgically treated patients and lay the foundation for the three-dimensional classification of intertrochanteric fracture,which provide effective guidance for treatment.(3)To develop a systematic three-dimensional classification of intertrochanteric fractures by clustering the morphological features of fracture lines using the Hausdorff distance-based K-means approach and assess the usefulness of it in the clinical setting.(4)To verify the stability difference between the various types of new intertrochanteric fracture classification by finite element analysisMethod(1)Patients with selective inclusion of type A2 of intertrochanteric fractures were explored by CT reconstruction to explore how the quality of medial side reconstruction affects outcome after intramedullary nailing.(2)Included preoperative CT data of 504 intertrochanteric fractures and reconstructed the intertrochanteric fracture model with achieving the reduction of the intertrochanteric fracture block.Descriptive analysis was performed by depicting fracture lines on four sides of the proximal femur.(3)We retrospectively analyzed the data of 504 patients with intertrochanteric fractures who underwent closed reduction and intramedullary internal fixation.The morphological fracture lines of all patients extracted from computed tomography were transcribed freehand onto the template.All fracture lines were then clustered into five distinct types using the Hausdorff distance-based K-means clustering method.Five radiographic parameters and four functional parameters were used to evaluate the postoperative functional states and mobilization levels.Postoperative complications were also recorded.(4)We established fracture models of the new classification,reconstructed the intramedullary nail device,and performed finite element analysis to evaluate the effectiveness of fracture classification.Result(1)The incidence of complications after intramedullary nailing of intertrochanteric fractures with medial side reconstruction was significantly lower than that of unrealized cases(P=0.022),and the postoperative functional score was also higher than that of unsupported reduction(P<0.05).(2)The fracture line of the medial cortex of intertrochanteric fracture showed a variety of morphological features,of which 10.3%of the total amount of medial cortex structure was insufficient,and it would be difficult to achieve support reduction in this type of fracture during intraoperative reduction.Cases of complete rupture of the lateral wall account for approximately 10.7%of the total,and such fractures will lose the blocking effect of the lateral wall.(3)K-means cluster analysis divided 504 fracture models into five types.The incidence of complications was different in different types,and the incidence of complications gradually increased with the progressive classification of fractures.(4)The results of finite element analysis showed that in the new classification of intertrochanteric fractures,type ? and ? fractures were unstable,and the change of femoral shaft neck angle was significantly greater than that of other three types of fractures.And the bone volume exceeding the yield strain in the femoral head of type? and ? is significantly larger than the other three types of fracture.Conclusion(1)Reconstructing the triangular mechanical structure of the proximal femur and restoring physiological stress transfer is the key to get the successful treatment of intertrochanteric fractures.(2)The small trochanter plays a minor role in the physiological stress transfer process,but could carry some anterior bone mass by its avulsion.The size of the medial support cortex left behind has an important impact on the outcome of surgical treatment of intertrochanteric fractures.(3)The classification of intertrochanteric fracture based on artificial intelligence algorithm has certain regularity,high recognizability and clinical significance.(4)The medial side reconstruction of the intertrochanteric fracture is to restore the stress transmission channel to achieve uniform dispersion of stress in the skeletal system.If the channel is poorly restored or severely broken after fracture,the stress is difficult to diffusion,with the result of accumulating in the proximal bone block,causing stress overload of the bone block,and the physical properties of the bone block are extremely unstable,which is the main reason for the failure of fracture treatment.(5)Based on the theory of stress transmission channel,the research system uses deep learning to classify intertrochanteric fractures,which can guide the clinical treatment of intertrochanteric fractures and predict the risk of postoperative complications.