| Literature DB >> 31928116 |
Pishtiwan H S Kalmet1, Sebastian Sanduleanu2, Sergey Primakov2, Guangyao Wu2, Arthur Jochems2, Turkey Refaee2, Abdalla Ibrahim2, Luca V Hulst1, Philippe Lambin2, Martijn Poeze1,3.
Abstract
Artificial intelligence (AI) is a general term that implies the use of a computer to model intelligent behavior with minimal human intervention. AI, particularly deep learning, has recently made substantial strides in perception tasks allowing machines to better represent and interpret complex data. Deep learning is a subset of AI represented by the combination of artificial neuron layers. In the last years, deep learning has gained great momentum. In the field of orthopaedics and traumatology, some studies have been done using deep learning to detect fractures in radiographs. Deep learning studies to detect and classify fractures on computed tomography (CT) scans are even more limited. In this narrative review, we provide a brief overview of deep learning technology: we (1) describe the ways in which deep learning until now has been applied to fracture detection on radiographs and CT examinations; (2) discuss what value deep learning offers to this field; and finally (3) comment on future directions of this technology.Entities:
Mesh:
Year: 2020 PMID: 31928116 PMCID: PMC7144272 DOI: 10.1080/17453674.2019.1711323
Source DB: PubMed Journal: Acta Orthop ISSN: 1745-3674 Impact factor: 3.717
Figure 1.Visualization of Artificial Intelligence sub-family.
Figure 2.Visualization of artificial neuron model. Where A1–AN are the inputs, W1–WN are the weights for the input connections to neuron, b is the bias value, z is the output from the neuron.
Figure 3.Deep learning aided workflow in fracture detection
Summary of clinical studies involving computer-aided fracture detection
| Reference | Region of interest | Modality | Conclusion | Performance (metric) |
|---|---|---|---|---|
| Olczak et al. | Wrist/Hand/Ankle | Radiographs | This study supports the use of orthopaedic radiographs of artificial intelligence, which can perform at a human level | 0.83 (accuracy) |
| Kim et al. 2018 | Wrist | Radiographs | The AUC scores for this test were comparable tostate-of-the-art providing proof of concept for transfer learning from CNNs in fracture detection on plain radiographs | 0.95 (AUC) 0.90 (sensitivity) 0.88 (specificity) |
| Chung et al. | Proximal humerus | Radiographs | The use of artificial intelligence can accurately detect and classify proximal humerus fractures on plain shoulder AP radiographs | Detection: 0.96 (accuracy) 1 (AUC) 0.99 (sensitivity) 0.97 (specificity) Classification: 0.65–0.86 (accuracy) 0.90–0.98 (AUC) 0.88–0.97 (sensitivity) 0.83–0.94 (specificity) |
| Heimer et al. | Skull | CT | Classification based on the existence of skull fractures on CMIPs with deep learning is feasible | 0.97 (AUC) 0.91 (sensitivity) 0.88 (specificity) |
| Lindsey et al. | Wrist | Radiographs | Deep learning methods are a mechanism by which senior medical specialists can deliver their expertise to generalists on the front lines of medicine, thereby providing substantial improvements to patient care | 0.97 (AUC) on Test set1 0.98 (AUC) on Test set2 |
| Tomita et al. | Pelvis | CT | The proposed system will assist and improve OVF diagnosis in clinical settings by pre-screening routine CT examinations and flagging suspicious cases prior to review by radiologists | 0.89 (accuracy) 0.91 (F1 score) |
| Pranata et al. | Calcaneus | CT | The feasibility using deep CNN and SURF for computer-aided classification and detection of the location of calcaneus fractures in CT images | 0.98 (accuracy) |
| Adams et al. | Pelvis | Radiographs | As impressive as recognising fractures is for a DCNN, similar learning can be achieved by top-performing medically naпve humans with less than 1 hour of perceptual training | 0.91 (accuracy) 0.98 (AUC) |
Abbreviations: CT = computed tomography; AUC = area under curve; CNN = convolutional neural network; AP = plain anteroposterior;
CMIP = curved maximum intensity projections; OVF = Osteoporotic vertebral fractures; SURF = speeded-up robust features;
DCNN = deep convolutional neural networks.