Estimation of age in unidentified patients via chest radiography using convolutional neural network regression
- PDF / 378,723 Bytes
- 6 Pages / 595.276 x 790.866 pts Page_size
- 20 Downloads / 221 Views
ORIGINAL ARTICLE
Estimation of age in unidentified patients via chest radiography using convolutional neural network regression Carl F. Sabottke 1,2
&
Marc A. Breaux 1,2 & Bradley M. Spieler 2
Received: 29 February 2020 / Accepted: 16 April 2020 # American Society of Emergency Radiology 2020
Abstract Purpose Patient age has important clinical utility for refining a differential diagnosis in radiology. Here, we evaluate the potential for convolutional neural network models to predict patient age based on anterior-posterior chest radiographs for instances where patients may present for emergency services without the ability to provide this identifying information. Methods We used the CheXpert dataset of 224,316 chest radiographs from 65,240 patients to train CNN regression models with ResNet50 and DenseNet121 architectures for prediction of patient age based on anterior-posterior (AP) view chest radiographs. We evaluate these models on both the CheXpert validation dataset and a local hospital case in which a patient initially presented for emergency services intubated and without identification. Results Mean absolute error (MAE) for our ResNet50 model on the CheXpert dataset is 4.94 years for predicting patient age based on AP chest radiographs. MAE for our DenseNet121 model is 4.69 years. Both models have a correlation coefficient between true patient ages and predicted ages of 0.944. Wilcoxon rank-sum comparison between the two model architectures shows no significant difference (p = 0.33), but both show improvement over a baseline demographic-driven estimation (p < 0.001). Conclusions For circumstances in which patients present for healthcare services without readily accessible identification such as in the setting trauma or altered mental status, CNN regression models for age prediction have potential clinical utility for refining estimates related to this missing patient information. Keywords Convolutional neural networks . Chest radiography . Deep learning . Emergency medicine
Introduction Recent advances in convolutional neural networks (CNNs) have generated much excitement for applications to radiology [1–4]. One particular application of interest for radiology deep learning is using CNNs to estimate pediatric bone age based on hand radiographs [5, 6] in contrast to a more traditional method such as image atlas comparison pioneered by Greulich and Pyle [7, 8]. Indeed, in 2017, the Radiological Society of North America (RSNA) even hosted a competition, the
* Carl F. Sabottke [email protected]; [email protected] 1
Department of Internal Medicine, University Hospital and Clinics, Lafayette, LA, USA
2
Department of Radiology, Louisiana State University Health Sciences Center, New Orleans, LA, USA
Pediatric Bone Age Machine Learning Challenge, to focus attention on the utility of deep neural networks for performing this task [9–11]. More recently, knee MRI has been proposed for bone age estimation in patients 14–21 years old using a GoogLeNet CNN [12]. Multi-factorial MRI data (hand, clavicle, and tee
Data Loading...
