A Novel Deep Learning Model to Distinguish Malignant Versus Benign Solid Lung Nodules.

BACKGROUND In this study we aimed to establish a new transfer learning model based on noncontrast and thin-layer computed tomography (CT) scans to distinguish between malignant and benign solid lung nodules. MATERIAL AND METHODS CT images from 202 patients with 210 lesions (malignant: 127, benign: 83) manifesting as solid lung nodules from January 2016 to December 2020 from 3 institutions were retrospectively collected, and each nodule was histopathologically confirmed. Two experienced thoracic radiologists reviewed all images and determined the regions of interest (ROIs) in the three-dimensional (3D) images layer-by-layer. We divided the lesions and images into training and testing sets at a ratio of 7: 3. The Inception V3 model was pretrained by the training dataset. Five-fold cross-validation was used to choose the optimal model. Receiver operator characteristic curves (ROC curves) for methods to evaluate the performance of the models were drafted. RESULTS In the validation set, the AUC, accuracy, sensitivity, and specificity of Inception V3 model (lesion-level) were 0.999, 0.989, 0.983, and 1.0, respectively, which is higher than the image-level (0.997, 0.933, 0.922, and 0.948, respectively). The Inception V3 model (lesion-level) performed better than the image-level but there was no significant difference between the models (P>0.05). The ResNet50 model based on image level achieved AUC, accuracy, sensitivity, and specificity of 0.963, 0.926, 0.916, and 0.944, respectively, which is lower than that of Inception V3. CONCLUSIONS Our study developed a novel deep learning model based on noncontrast and thin-layer CT scans to classify benign vs malignant lung nodules, and the Inception V3 model greatly improved the differentiation accuracy and specificity.

Magnetic Resonance Imaging has Better Accuracy in Detecting New-Onset Rib Fractures as Compared to Computed Tomography.

BACKGROUND The aim of this study was to explore the magnetic resonance imaging (MRI) manifestations of new-onset rib fractures and determine the utility of MRI through a comparative study of MRI and computed tomography (CT). MATERIAL AND METHODS Twenty-one patients with mild chest trauma who sought medical care from February 2019 to June 2020 were included in this study. The patients were subjected to CT and MRI scanning within 2 weeks after the trauma, and CT rescanning was scheduled 4-8 weeks later to classify rib fractures and determine the diagnostic accuracy of MRI and CT for new-onset rib fractures. RESULTS Seventy-seven rib fractures were confirmed by CT rescanning, of which 72 (93.51%) were type I fractures and 5 (6.49%) were type II. MRI identified 76 fractures, of which 3 were false positive, with the diagnostic accuracy rate of 91.25% and sensitivity rate of 94.81%. Among them, type I fractures (n=71, 3 were false positive) showed the MRI "sandwich" sign (heterogeneous high-signal shadow within bone marrow of the inner layer, low-signal bony cortex of the middle layer, and high-signal subperiosteal effusion of the outer layer) in T2-weighted fat-suppressed sequences; type II fractures (n=5) displayed intramedullary high-signal intensities and no subperiosteal effusion. Forty-four fractures (all type I) were discovered in the initial CT examination, and the corresponding diagnostic accuracy rate and sensitivity rate were 57.14%, which were lower than that of MRI. CONCLUSIONS MRI is highly sensitive for new-onset rib fractures, especially type I, so it is a preferred method for patients with mild chest trauma.

Precise anatomical localization and classification of rib fractures on CT using a convolutional neural network.

OBJECTIVE: To develop a convolutional neural network (CNN) model for the detection, precise anatomical localization (right 1-12th and left 1-12th) and classification (fresh, healing and old fractures) of rib fractures automatically, and to compare the performance with the experienced radiologists. MATERIALS AND METHODS: A total of 640 rib fracture patients with 340,501 annotations were retrospectively collected from three hospitals. They consisted of a classification training dataset (n = 482), a localization training dataset (n = 30), an internal testing dataset (n = 90) and an external testing dataset (n = 38). RetinaNet with rib localization postprocessing and the result merging technique were employed to structure the CNN model. ROC curve, free-response ROC curve, AUC, precision, recall, and F1-score were calculated to choose the better option between model I (training classification and localization data together) and model II (adding an additional classification model to model I). RESULTS: The detection and classification performance of rib fractures was better in model II than in model I. The sensitivity of localization reached 97.11% and 94.87% on the right and left ribs, respectively. In the external dataset with different CT scanner and slice thickness, model II showed better diagnostic performance. Moreover, the CNN model was superior in diagnosing fresh and healing fractures to 5 radiologists and consumed shorter diagnosis time. CONCLUSIONS: Our CNN model was capable of detection, precise anatomical localization, and classification of rib fractures automatically.

Detection and Semiquantitative Analysis of Cardiomegaly, Pneumothorax, and Pleural Effusion on Chest Radiographs.

PURPOSE: To develop and evaluate deep learning models for the detection and semiquantitative analysis of cardiomegaly, pneumothorax, and pleural effusion on chest radiographs. MATERIALS AND METHODS: In this retrospective study, models were trained for lesion detection or for lung segmentation. The first dataset for lesion detection consisted of 2838 chest radiographs from 2638 patients (obtained between November 2018 and January 2020) containing findings positive for cardiomegaly, pneumothorax, and pleural effusion that were used in developing Mask region-based convolutional neural networks plus Point-based Rendering models. Separate detection models were trained for each disease. The second dataset was from two public datasets, which included 704 chest radiographs for training and testing a U-Net for lung segmentation. Based on accurate detection and segmentation, semiquantitative indexes were calculated for cardiomegaly (cardiothoracic ratio), pneumothorax (lung compression degree), and pleural effusion (grade of pleural effusion). Detection performance was evaluated by average precision (AP) and free-response receiver operating characteristic (FROC) curve score with the intersection over union greater than 75% (AP75; FROC score75). Segmentation performance was evaluated by Dice similarity coefficient. RESULTS: The detection models achieved high accuracy for detecting cardiomegaly (AP75, 98.0%; FROC score75, 0.985), pneumothorax (AP75, 71.2%; FROC score75, 0.728), and pleural effusion (AP75, 78.2%; FROC score75, 0.802), and they also weakened boundary aliasing. The segmentation effect of the lung field (Dice, 0.960), cardiomegaly (Dice, 0.935), pneumothorax (Dice, 0.827), and pleural effusion (Dice, 0.826) was good, which provided important support for semiquantitative analysis. CONCLUSION: The developed models could detect cardiomegaly, pneumothorax, and pleural effusion, and semiquantitative indexes could be calculated from segmentations.Keywords: Computer-Aided Diagnosis (CAD), Thorax, CardiacSupplemental material is available for this article.© RSNA, 2021.

Infection Control and Management Strategy for COVID-19 in the Radiology Department: Focusing on Experiences from China.

Coronavirus disease 2019 (COVID-19) is a new infectious disease rapidly spreading around the world, raising global public health concerns. Radiological examinations play a crucial role in the early diagnosis and follow-up of COVID-19. Cross infection among patients and radiographers can occur in radiology departments due to the close and frequent contact of radiographers with confirmed or potentially infected patients in a relatively confined room during radiological workflow. This article outlines our experience in the emergency management procedure and infection control of the radiology department during the COVID-19 outbreak.

Intrapulmonary myelolipoma and its computed tomography features: A case report and literature review.

Intrapulmonary myelolipoma is a rare, benign tumor composed of mature adipose tissue and normal hematopoietic cells. To the best of our knowledge, 10 cases of intrapulmonary myelolipoma, including the present case, have been reported to date, and the majority have focused on the pathological diagnosis of the disease. The radiological features of intrapulmonary myelolipoma have not been studied. Therefore, the present study reports a case of primary myelolipoma in the lung, and examines its computed tomography features and pathology. Furthermore, other potential diagnoses are discussed in the context of the relevant literature. The present report describes the case of a 57-year-old female who experienced chills, but no coughing or expectoration, with an intermittent fever of 38.6°C that had been apparent for 13 days. Chest CT scan revealed a benign nodule and bronchiectasis in the lower lobe of the right lung. The patient then underwent a lobectomy of the lower right lung by thoracoscopy. The histological analysis of the excised specimen identifid a myelolipoma consisting of mature adipose tissue and hematopoietic cells. There was no recurrence after 513 days of follow-up, as shown by CT.

Lobular capillary hemangioma of the trachea.

As a polypoid form of capillary hemangioma, lobular capillary hemangioma (LCH) most commonly occurs on the cutaneous and mucosal surfaces and is frequently benign in children. Tracheal LCH is a rare benign tumor in adults, with hemoptysis being one of the most serious forms of presentation. A definite diagnosis of LCH depends on its histopathology. In clinical practice, however, radiological characteristics are an important component for making a suggestive diagnosis. We present a case of tracheal LCH and describe its features on computed tomography, pathology, and differential diagnosis. A review of the relevant literature is also provided.

The usefulness of combined diagnostic CT and (99m)Tc-octreotide somatostatin receptor SPECT/CT imaging on pulmonary nodule characterization in patients.

UNLABELLED: The objective of this study was to evaluate the clinical value of combination of diagnostic computed tomography (CT) and somatostatin receptor imaging with (99m)Tc-octreotide acetate SPECT/CT in differentiation of benign pulmonary nodules from cancers. METHODS: This was a retrospective study, 29 patients with suspected pulmonary neoplasm underwent diagnostic CT and (99m)Tc-octreotide SPECT/CT scans, and the tumor-to-normal tissue tracer value (T/N) for (99m)Tc-octreotide was measured. Diagnosis was confirmed by histological analysis. RESULTS: Eighteen of the 29 patients included in this study had lung cancer: 2 with small cell lung cancer and 16 with nonsmall cell lung cancer. The other 11 patients had benign lung lesions: 5 with tuberculosis, 4 with nontuberculosis infection, 1 with hematoma, and 1 with fibroma. (99m)Tc-octreotide uptake (expressed as mean T/N±SD) was significantly higher in lung cancers (2.58±0.91) than benign lesions (1.38±0.79) (p=0.002). Specificity for pulmonary malignant nodule diagnosis was 63.6% for diagnostic CT, 72.7% for somatostatin receptor SPECT/CT imaging, and 81.8% for the combined use of diagnostic CT and somatostatin receptor SPECT/CT imaging. CONCLUSION: Somatostatin receptor imaging with (99m)Tc-octreotide SPECT/CT is useful for the differentiation of benign pulmonary nodules from lung cancers, the combination of diagnostic CT and (99m)Tc-octreotide SPECT/CT further increases the specificity of malignant pulmonary nodule detection.

[Associations of triple-phase enhancement multi-slice CT scan with histological differentiation and lymphovascular invasion in gastric cancer].

OBJECTIVE: To explore the associations of triple-phase enhancement multi-slice CT scan with histological differentiation and lymphovascular invasion in advanced gastric cancer. METHODS: Patients with gastric cancer undergoing surgical resection between January 2006 and December 2009 were included, all of whom underwent routine and two-phase contrast enhancement multi-slice CT examinations of the upper abdomen before surgery. Patients with incomplete data were excluded. Postoperative specimens were used for evaluation of histological differentiation, microscopic lymphovascular invasion, and CD34 and D2-40 expression. Associations between contrast enhancement ratio (CER) of triple-phase multi-slice CT scan and differentiation as well as lymphovascular invasion were analyzed. RESULTS: A total of 64 patients were included. There were significant associations between CER and tumor differentiation (P<0.05). Compared to those without microvascular invasion, CER of the arterial phase in gastric cancer with microvascular invasion was significantly higher (0.61±0.28 vs. 0.46±0.14, P<0.05), and that of the arterial-parenchymal phase was significantly lower(1.81±0.39 vs. 2.28±0.80, P<0.05). However, gastric cancers with lymphatic invasion were associated with significantly higher CER of the parenchymal phase than their counterparts(1.25±0.57 vs. 1.00±0.35, P<0.05). CONCLUSION: CER of triple-phase multi-slice CT scan can reveal the differentiation of gastric cancer and is associated with microvascular and lymphatic invasion.

Multi-slice CT contrast-enhanced presentations of advanced gastric cancer: associations with histo-differentiation and expression of p53 and P-glycoprotein.

BACKGROUND: This study aimed to investigate multi-slice CT contrast-enhanced presentation of gastric cancer and its correlation with histo-differentiation and p53 and P-glycoprotein (P-gp) expression. METHODS: Sixty-six patients with gastric cancer in the present study underwent a multi-slice CT preoperative routine and dual-phase contrast-enhanced examination of the upper abdomen; postoperative specimens were used to determine histo-differentiation and the expression of p53 and P-gp. The correlation of multi-slice CT contrast-enhanced presentation with histo-differentiation and expression of p53 and P-gp was analyzed. RESULTS: The dual-phase contrast-enhanced ratio (CER) was not correlated with the histo-differentiation of gastric cancer (P > 0.05). Positive expression of p53 and P-gp was significantly higher in the cases of layered or heterogeneous enhancement than in the cases of homogenous enhancement (P < 0.05). Positive expression of p53 was also correlated with the arterial phase CER, tumor size and lymph node metastasis (P < 0.05), but not with infiltration thickness of the gastric wall, nor was it correlated with the portal phase CER (P > 0.05). Positive expression of P-gp was only correlated with the portal phase CER (P = 0.005). CONCLUSIONS: Differently enhanced pattern and CER of the arterial and portal phase in gastric cancer correlate with its different histo-differentiation and expression of p53 and P-gp respectively. In addition, tumor size and lymph node metastasis of gastric cancer relate to the expression of p53.