Deep Learning Algorithms for Interpretation of Upper Extremity Radiographs: Laterality and Technologist Initial Labels as Confounding Factors

Convolutional neural networks (CNNs) trained to identify abnormalities on upper extremity radiographs achieved an AUC of 0.844 with a frequent emphasis on radiograph laterality and/or technologist labels for decision-making. Covering the labels increased the AUC to 0.857 (p = .02) and redirected CNN attention from the labels to the bones. Using images of radiograph labels alone, the AUC was 0.638, indicating that radiograph labels are associated with abnormal examinations. Potential radiographic confounding features should be considered when curating data for radiology CNN development.

Automated summary

Convolutional neural networks trained to identify abnormalities on upper extremity radiographs achieved an AUC of 0.844, with emphasis on radiograph laterality and/or technologist labels. Covering the labels increased the AUC to 0.857 (p = .02) and redirected CNN attention to bones. Using images of radiograph labels alone had an AUC of 0.638, indicating their association with abnormal examinations. Consideration should be given to potential radiographic confounding features when curating data for radiology CNN development.