Page 15 - Read Online

Page 221 Dababneh et al. Art Int Surg 2024;4:214-32 https://dx.doi.org/10.20517/ais.2024.50

Table 2. Compilation of published articles focusing on automated fracture detection in the hand or wrist using ML models

Study Fracture site Dataset size (view) ML model used AUC Sensitivity Specificity
Olczak et al. Wrist or hand or ankle 256,000 Multiple DL networks (BVLC Reference CaffeNet network, 0.83 - -
[25]
(2017) VGG CNN S network, VGG CNN, Network-in-network)
Lee et al. Distal radius, ulnar styloid or scaphoid 5,618 (AP, lat, and oblique) e-CRF, self-developed AI model Distal radius: Distal radius: 0.97 Distal radius: 0.83
[26]
(2023) 0.903 Ulnar styloid:0.98 Ulnar styloid:0.87
Ulnar styloid: Scaphoid: 0.87 Scaphoid: 0.74
0.925
Scaphoid: 0.808
Lindsey et al. Wrist 34,990 (PA, lat) Self-developed deep neural network 0.954 0.94 0.95
[27]
(2018)
Cohen et al. Distal radius or ulnar styloid or distal ulna 1,917 (AP, lat, oblique, BoneView (Gleamer) DCNN algorithm - 0.83 0.96
[28]
(2023) or metacarpal or scaphoid or carpal bone specific views of the
carpus)
Hardalac et al. Distal radius or distal ulna 542 WFD-C, deep-learning-based object detection model 0.864 - -
[29]
(2022)
Alammar et al. Humerus or wrist 10,558 TL adaption of ImageNet models Humerus: 0.879 Humerus: 0.87 Humerus: 0.87
[31]
(2023) Wrist: 0.856 Wrist: 0.89 Wrist: 0.93
Jacques et al. Distal radius or distal ulna or carpal bones 788 BoneView (Gleamer) 0.764 0.70 0.89
[32]
(2024) or scaphoid or finger
Kim and Distal radius or distal ulna 1,489 (lat) Inception v3 network, DCNNs 0.954 0.90 0.88
MacKinnon
[33]
(2018)
Gan et al. Distal radius 2340 wrist (AP) Inception-v4 0.96 0.90 0.96
[34]
(2019)
Oka et al. Distal radius, styloid process of ulna 1,464 (AP, lat) VGG16 Distal radius: 0.99 Distal radius: 0.99 Distal radius: 0.97
[36]
(2021) Ulna: 0.96 Ulna: 0.92 Ulna: 0.90
Russe et al. Distal radius 2,856 (AP, lat) Xception 0.97 0.95 0.95
[37]
(2024)
Zhang et al. Distal radius 6,536 (AP, lat) Ensemble model of RetinaNet, Faster RCNN and Cascade 0.97 0.96 0.98
[38]
(2023) RCNN
Anttila et al. Distal radius 3,785 (PA, lat) Self-developed DL algorithm developed as standalone 0.95 0.86 0.89
[39]
(2023) MATLAB application
Mert et al. Distal radius 150 (AP, lat) ChatGPT 4 0.93 0.88 0.98
[40]
(2024)
Ozkaya et al. Scaphoid 390 (AP) Pre-trained ResNet50 network 0.84 0.76 0.92
[41]
(2022)
Hendrix et al. Scaphoid 4,229 (AP, PA) Self-developed fracture detection and segmentation CNN 0.87 0.78 0.84
[42]
(2021)
Hendrix et al. Scaphoid 19,111 (AP, PA, ulnar- Self-developed fracture detection and segmentation CNN 0.88 0.72 0.93
[43]
(2023) deviated and oblique)