Background Chest CT is used to assess the severity of lung involvement in COVID-19 pneumonia. Purpose To determine the change in chest CT findings associated with COVID-19 pneumonia from initial di...

One year after COVID-19 diagnosis, chest CT scans showed persistent abnormalities in 53 of 209 adult study participants (25%). Background The chest CT manifestations of COVID-19 from hospitalization to convalescence after 1 year are unknown. Purpose To assess chest CT manifestations of COVID-19 up to 1 year after symptom o...

The current artificial intelligence (AI) models are still insufficient in multi-disease diagnosis for real-world data, which always present a long-tail distribution. To tackle this issue, a long-tail public dataset, “ChestX-ray14,” which involved fourteen (14) disease labels, was randomly divided into the train, validation, and test sets with ratios of 0.7, 0.1, and 0.2. Two pretrained state-of-the-art networks, EfficientNet-b5 and CoAtNet-0-rw, were chosen as the backbones. After the fully-connected layer, a final layer of 14 sigmoid activation units was added to output each disease’s diagnosis. To achieve better adaptive learning, a novel loss (L_ours) was designed, which coalesced reweighting and tail sample focus. For comparison, a pretrained ResNet50 network with weighted binary cross-entropy loss (L_WBCE) was used as a baseline, which showed the best performance in a previous study. The overall and individual areas under the receiver operating curve (AUROC) for each disease label were evaluated and compared among different models. Group-score-weighted class activation mapping (Group-CAM) is applied for visual interpretations. As a result, the pretrained CoAtNet-0-rw + L_ours showed the best overall AUROC of 0.842, significantly higher than ResNet50 + L_WBCE (AUROC: 0.811, p = 0.037). Group-CAM presented that the model could pay the proper attention to lesions for most disease labels (e.g., atelectasis, edema, effusion) but wrong attention for the other labels, such as pneumothorax; meanwhile, mislabeling of the dataset was found. Overall, this study presented an advanced AI diagnostic model achieving a significant improvement in the multi-disease diagnosis of chest X-rays, particularly in real-world data with challenging long-tail distributions.

Detrimental to individuals and society, online hateful messages have recently become a major social issue. Among them, one new type of hateful message, “hateful meme”, has emerged and brought difficulties in traditional deep learning-based detection. Because hateful memes were formatted with both text captions and images to express users’ intents, they cannot be accurately identified by singularly analyzing embedded text captions or images. In order to effectively detect a hateful meme, the algorithm must possess strong vision and language fusion capability. In this study, we move closer to this goal by feeding a triplet by stacking the visual features, object tags, and text features of memes generated by the object detection model named Visual features in Vision-Language (VinVl) and the optical character recognition (OCR) technology into a Transformer-based Vision-Language Pre-Training Model (VL-PTM) OSCAR+ to perform the cross-modal learning of memes. After fine-tuning and connecting to a random forest (RF) classifier, our model (OSCAR+RF) achieved an average accuracy and AUROC of 0.684 and 0.768, respectively, on the hateful meme detection task in a public test set, which was higher than the other eleven (11) published baselines. In conclusion, this study has demonstrated that VL-PTMs with the addition of anchor points can improve the performance of deep learning-based detection of hateful memes by involving a more substantial alignment between the text caption and visual information.

Purpose Report of experience from a single institution in treating postoperative lymphatic leakage (PLL) applying conventional transpedal lymphangiography (TL). Materials and Methods 453 patients with the initial diagnosis of PLL receiving TL between 03/1993 and 09/2018 were identified in the database. Only patients with confirmed PLL were included in the study. The technical success, safety, and treatment success of TL were evaluated. Independent predictors of TL treatment failure were examined using univariate and multivariate logistic regression analysis. Results 355 consecutive patients (218 men, 137 women; median age of 62 years) who underwent TL for PLL (e. g., chylothorax) after ineffective conservative treatment were included. The median time between causal surgery and TL was 27 days. The median technical success rate of TL was 88.5 %, with a median volume of Lipiodol of 10.0 ml. No complication of TL was recorded. Three groups were defined according to the different clinical courses: group A (41/355, 11.5 %) – TL with technical failure; group B (258/355, 72.7 %) – “therapeutic” TL alone with technical success; and group C (56/355, 15.8 %) – “diagnostic” TL with simultaneously invasive treatment (incl. surgical revision and percutaneous sclerotherapy). Treatment success rate and median time to treatment success were higher in group C than in group B, but without significant differences (64.3 % vs. 61.6 %, p = 0.710; six vs. five days, p = 0.065). Univariate and multivariate logistic regression analyses for group B confirmed drainage volume (> 500 ml/d) and Lipiodol extravasation as independent predictors of TL clinical failure (odds ratios [ORs] of 2.128 and 2.372 [p = 0.005 and p = 0.003, respectively]). Conclusion TL is technically reliable, safe, and effective in treating PLL. When conservative treatment fails, TL can be regarded as the next treatment option. Key Points: Citation Format

This report describes a patient with COVID-19 who developed spontaneous pneumothorax and subpleural bullae during the course of the infection. Consecutive chest computed tomography images indicated that COVID-19-associated pneumonia had damaged the subpleural alveoli and distal bronchus. Coughing might have induced a sudden increase in intra-alveolar pressure, leading to the rupture of the subpleural alveoli and distal bronchus and resulting in spontaneous pneumothorax and subpleural bullae. At the 92-day follow-up, the pneumothorax and subpleural bullae had completely resolved, which indicated that these complications had self-limiting features.