RESUMO
PURPOSE: To compare the performance of a convolutional neural network (CNN) to that of 11 radiologists in detecting scaphoid bone fractures on conventional radiographs of the hand, wrist, and scaphoid. MATERIALS AND METHODS: At two hospitals (hospitals A and B), three datasets consisting of conventional hand, wrist, and scaphoid radiographs were retrospectively retrieved: a dataset of 1039 radiographs (775 patients [mean age, 48 years ± 23 {standard deviation}; 505 female patients], period: 2017-2019, hospitals A and B) for developing a scaphoid segmentation CNN, a dataset of 3000 radiographs (1846 patients [mean age, 42 years ± 22; 937 female patients], period: 2003-2019, hospital B) for developing a scaphoid fracture detection CNN, and a dataset of 190 radiographs (190 patients [mean age, 43 years ± 20; 77 female patients], period: 2011-2020, hospital A) for testing the complete fracture detection system. Both CNNs were applied consecutively: The segmentation CNN localized the scaphoid and then passed the relevant region to the detection CNN for fracture detection. In an observer study, the performance of the system was compared with that of 11 radiologists. Evaluation metrics included the Dice similarity coefficient (DSC), Hausdorff distance (HD), sensitivity, specificity, positive predictive value (PPV), and area under the receiver operating characteristic curve (AUC). RESULTS: The segmentation CNN achieved a DSC of 97.4% ± 1.4 with an HD of 1.31 mm ± 1.03. The detection CNN had sensitivity of 78% (95% CI: 70, 86), specificity of 84% (95% CI: 77, 92), PPV of 83% (95% CI: 77, 90), and AUC of 0.87 (95% CI: 0.81, 0.91). There was no difference between the AUC of the CNN and that of the radiologists (0.87 [95% CI: 0.81, 0.91] vs 0.83 [radiologist range: 0.79-0.85]; P = .09). CONCLUSION: The developed CNN achieved radiologist-level performance in detecting scaphoid bone fractures on conventional radiographs of the hand, wrist, and scaphoid.Keywords: Convolutional Neural Network (CNN), Deep Learning Algorithms, Machine Learning Algorithms, Feature Detection-Vision-Application Domain, Computer-Aided DiagnosisSee also the commentary by Li and Torriani in this issue.Supplemental material is available for this article.©RSNA, 2021.
RESUMO
UNLABELLED: Timely recognition of deterioration of hospitalised children is important to improve mortality. We developed a modified Paediatric Early Warning Score (PEWS) and studied the effects by performing three different cohort studies using different end points. Taking unplanned Paediatric Intensive Care Unit admission as end point and only using data until 2 h prior to end point, we found a sensitivity of 0.67 and specificity of 0.88 to timely recognise patients. This proves that earlier identification is possible without a loss of sensitivity compared to other PEWS systems. When determining the corresponding clinical condition in patients with an elevated PEWS dichotomously as 'sick' or 'well', this resulted in a total of 27 % false-positive scores. This can cause motivational problems for caregivers to use the system but is a consequence of PEWS design to minimise false-negative rates because of high mortality associated with paediatric resuscitation. Using the need for emergency medical interventions as end point, sensitivity of PEWS is high and it seems, therefore, that it is also fit to alert health-care professionals that urgent interventions may be needed. CONCLUSION: These data show the effectiveness of a modified PEWS in identifying critically ill patients in an early phase making early interventions possible and hopefully reduce mortality.
