Advanced Deep Learning Techniques Applied to Automated Femoral Neck Fracture Detection and Classification

Abstract

To use deep learning with advanced data augmentation to accurately diagnose and classify femoral neck fractures. A retrospective study of patients with femoral neck fractures was performed. One thousand sixty-three AP hip radiographs were obtained from 550 patients. Ground truth labels of Garden fracture classification were applied as follows: (1) 127 Garden I and II fracture radiographs, (2) 610 Garden III and IV fracture radiographs, and (3) 326 normal hip radiographs. After localization by an initial network, a second CNN classified the images as Garden I/II fracture, Garden III/IV fracture, or no fracture. Advanced data augmentation techniques expanded the training set: (1) generative adversarial network (GAN); (2) digitally reconstructed radiographs (DRRs) from preoperative hip CT scans. In all, 9063 images, real and generated, were available for training and testing. A deep neural network was designed and tuned based on a 20% validation group. A holdout test dataset consisted of 105 real images, 35 in each class. Two class prediction of fracture versus no fracture (AUC 0.92): accuracy 92.3%, sensitivity 0.91, specificity 0.93, PPV 0.96, NPV 0.86. Three class prediction of Garden I/II, Garden III/IV, or normal (AUC 0.96): accuracy 86.0%, sensitivity 0.79, specificity 0.90, PPV 0.80, NPV 0.90. Without any advanced augmentation, the AUC for two-class prediction was 0.80. With DRR as the only advanced augmentation, AUC was 0.91 and with GAN only AUC was 0.87. GANs and DRRs can be used to improve the accuracy of a tool to diagnose and classify femoral neck fractures.

Machine Learning Solutions for Osteoporosis—A Review

“…( 68,70–76,80,82,83,85–89,92,93,95,97 ) In some studies, class unbalancing was handled using sampling methods, ( 74,97 ) and overfitting was controlled with data augmentation. ( 80,83,84,95 ) Mutasa and colleagues demonstrated how data augmentation using digitally reconstructed and generated images was able to improve the test AUC from 0.80 (95% CI, not reported) to 0.92 (95% CI, not reported). ( 84 ) On the other hand, Adams and colleagues reported no significant improvement using data augmentation, whereas larger data sets had positive effects on the hip fracture detection accuracy.…”

“…Among the 32 studies that investigated fracture detection (Table 3), (66–97 ) 11 were on vertebral fractures, ( 66–76 ) 17 hip fractures, ( 74–90 ) and 10 other fracture sites such as humerus or wrist. ( 75,76,90–97 ) Nineteen studies developed CNN models for image analysis.…”

“…( 75,76,90–97 ) Nineteen studies developed CNN models for image analysis. ( 66,67,71,72,77–86,90,91,93–95 ) Others used features extracted from images or collected from non‐imaging data. ( 68–70,73–77,87–89,92,96,97 ) Studies reported average best AUC of 0.92 (range 0.63 to 1.00) and average best accuracy of 89.8% (range 78.4% to 99.1%].…”

“…( 78,80,81,85,90,91,94 ) Model decision‐making was highlighted using heat maps such as Grad‐CAM. ( 66,67,78,79,84,90,94 ) However, several studies did not use techniques that allowed for adequate understanding of the models' decision‐making (eg, data visualization). ( 68,70–76,80,82,83,85–89,92,93,95,97 ) In some studies, class unbalancing was handled using sampling methods, ( 74,97 ) and overfitting was controlled with data augmentation.…”

“…( 36,38 ) Several DL efforts in fracture detection visualized the findings of their models using activation maps (Grad‐Cam ( 20 ) ). ( 66,67,78,79,84,90,94 ) Other methods have also been applied such as t‐SNE (t‐distributed stochastic neighbor embedding ( 128 ) ) to visualize high‐dimensional data representations ( 77,81,99 ) or SHAP values (shapley additive explanations). ( 104 ) To assess the reliability of heat maps on images, Arun and colleagues explored their capability in localizing region of interest in medical imaging DL models.…”