Artificial intelligence-guided distal radius fracture detection on plain radiographs in comparison with human raters

Table 2 Statistical analysis of AI performance compared to human raters

	AI on ap radiograph	Orthopaedic surgeon on ap radiograph	p-value	AI on lateral view radiograph	Orthopaedic surgeon on lateral view radiograph	p-value
Accuracy	95.90	94.95	0.239	94.81	96.10	0.646
Cohen’s Kappa	0.913	0.894	N/A	0.891	0.918	N/A
F1 score	0.947	0.935	N/A	0.934	0.950	N/A
Sensitivity	92.02	89.73	0.182	89.79	90.83	0.773
Specificity	98.45	98.48	1.000	98.25	99.71	0.131
Youden Index	90.47	88.22	N/A	88.04	90.55	N/A
True Positives	196	201	N/A	211	218	N/A
True Negatives	318	325	N/A	337	348	N/A
False Positives	5	5	N/A	6	1	N/A
False Negatives	17	23	N/A	24	22	N/A

ISSN: 1749-799X