Chest radiographs are among the most frequently acquired images in radiology and are often the subject of computer vision research. However, most of the models used to classify chest radiographs are derived from openly available deep neural networks, trained on large image datasets. These datasets differ from chest radiographs in that they are mostly color images and have substantially more labels. Therefore, very deep convolutional neural networks (CNN) designed for ImageNet and often representing more complex relationships, might not be required for the comparably simpler task of classifying medical image data. Sixteen different architectures of CNN were compared regarding the classification performance on two openly available datasets, the CheXpert and COVID-19 Image Data Collection. Areas under the receiver operating characteristics curves (AUROC) between 0.83 and 0.89 could be achieved on the CheXpert dataset. On the COVID-19 Image Data Collection, all models showed an excellent ability to detect COVID-19 and non-COVID pneumonia with AUROC values between 0.983 and 0.998. It could be observed, that more shallow networks may achieve results comparable to their deeper and more complex counterparts with shorter training times, enabling classification performances on medical image data close to the state-of-the-art methods even when using limited hardware. Chest radiographs are among the most frequently used imaging procedures in radiology. They have been widely employed in the field of computer vision, as chest radiographs are a standardized technique and, if compared to other radiological examinations such as computed tomography or magnetic resonance imaging, contain a smaller group of relevant pathologies. Although many artificial neural networks for the classification of chest radiographs have been developed, it is still subject to intensive research. Only a few groups design their own networks from scratch, while most use already established architectures, such as ResNet-50 or DenseNet-121 (with 50 and 121 representing the number of layers within the respective neural network) 1-6. These neural networks have often been trained on large, openly available datasets, such as ImageNet, and are therefore already able to recognize numerous image features. When training a model for a new task, such as the classification of chest radiographs, the use of pre-trained networks may improve the training speed and accuracy of the new model, since important image features that have already been learned can be transferred to the new task and do not have to be learned again. However, the feature space of freely available data sets such as ImageNet differs from chest radiographs as they contain color images and more categories. The ImageNet Challenge includes 1,000 possible categories per image, while CheXpert, a large freely available data set of chest radiographs, only distinguishes between 14 categories (or classes) 7 , and the COVID-19 Image Data Collection only differentiates between three classes 8. Although the ImageNet...

Supplemental digital content is available in the text.

No abstract

BACKGROUND AND PURPOSE: The Neck Imaging Reporting and Data System was introduced to assess the probability of recurrence in surveillance imaging after treatment of head and neck cancer. This study investigated inter-and intrareader agreement in interpreting contrast-enhanced CT after treatment of oral cavity and oropharyngeal squamous cell carcinoma. MATERIALS AND METHODS:This retrospective study analyzed CT datasets of 101 patients. Four radiologists provided the Neck Imaging Reporting and Data System reports for the primary site and neck (cervical lymph nodes). The Kendall's coefficient of concordance (W), Fleiss k (k F ), the Kendall's rank correlation coefficient (t B ), and weighted k statistics (k w ) were calculated to assess inter-and intrareader agreement.RESULTS: Overall, interreader agreement was strong or moderate for both the primary site (W ¼ 0.74, k F ¼ 0.48) and the neck (W ¼ 0.80, k F ¼ 0.50), depending on the statistics applied. Interreader agreement was higher in patients with proved recurrence at the primary site (W ¼ 0.96 versus 0.56, k F ¼ 0.65 versus 0.30) or in the neck (W ¼ 0.78 versus 0.56, k F ¼ 0.41 versus 0.29). Intrareader agreement was moderate to strong or almost perfect at the primary site (range t B ¼ 0.67-0.82, k w ¼ 0.85-0.96) and strong or almost perfect in the neck (range t B ¼ 0.76-0.86, k w ¼ 0.89-0.95).