Application of Medical Imaging Based on Deep Learning in the Treatment of Lumbar Degenerative Diseases and Osteoporosis with Bone Cement Screws.

Researchers

Journal

Modalities

Models

Abstract

To explore the application value of magnetic resonance spectroscopy (MRS) and GSI-energy spectrum electronic computed tomography (CT) medical imaging based on the deep convolutional neural network (CNN) in the treatment of lumbar degenerative disease and osteoporosis.There were 56 cases of suspected lumbar degenerative disease and osteoporosis. A group of 56 subjects were examined using 1.5 TMR spectrum (MRS) and dual-energy X-ray absorptiometry (DXA) to collect the lumbar L3 vertebral body fat ratio (FF) and L1~4 vertebral bone mineral density (BMD) value. We divided the subjects into 2 groups with T value -2.5 as the critical point. Set T value > -2.5 as the negative group and T value ≤ -2.5 as the positive group. Pearson’s method is used for FF-MRS and BMD correlation analyses. A group of all patients underwent GSI-energy spectrum CT scan, and X-ray bone mineral density (DXA) test results (bone density per unit area) were used as the gold standard to analyze the diagnosis of osteoporosis by the GSI-energy spectrum CT scan method value.The differences in FF and BMD between the negative group and the positive group were statistically significant (P < 0.01), and there was a highly negative correlation between the average value of FF and BMD. 30 cases were diagnosed as osteoporosis by DXA. The accuracy of GSI-energy spectrum CT medical imaging in diagnosing osteoporosis is 89.30%. The GSI-energy spectrum CT diagnosis of osteoporosis and DXA examination results have good consistency.Based on the deep convolutional neural network (CNN) MRS technology, GSI-energy spectrum CT medical imaging is used in the clinical diagnosis and treatment of lumbar degenerative lesions and osteoporosis. It has a good advantage in assessing bone quality and has good consistency with DXA examination and has better application value high.Copyright © 2021 Shengkai Mu et al.