Investigation of 3D Imaging Support of the Aortic Arch Using Plain Chest CT Images in Patients before Angiography

We developed an image-based unsupervised survival prediction model, called pix2surv, based on a conditional generative adversarial network (cGAN), and evaluated its performance based on chest CT images of patients with the coronavirus disease 2019 (COVID-19). The architecture of the pix2surv model includes a time generator that consists of an encoding convolutional network and a fully connected prediction network, and a discriminator network. The time generator is trained to generate survival-time images from chest CT images of each patient. The discriminator is a patch-based convolutional network that is trained to differentiate between “fake pairs” of a chest CT image and a generated survival-time image from “true pairs” of the chest CT image and the corresponding observed survival-time image of the patient. For evaluation, we retrospectively collected high-resolution chest CT images of COVID-19 patients. The survival predictions of the pix2surv model on these patients were compared with those of existing clinical prognostic biomarkers by use of a two-sided t-test with bootstrapping. Concordance index (C-index) and relative absolute error (RAE) were used as measures of the prediction performance. The bootstrap evaluation yielded C-index and RAE values of 80.4% and 15.6% for the pix2surv model, whereas those for the extent of the well-aerated lung parenchyma were 49.8% and 33.6%, and for a combination of blood tests of lactic dehydrogenase, lymphocyte, and C-reactive protein were 69.8% and 25.5%, respectively. The increase in survival prediction by the pix2surv model was statistically significant (p < 0.0001), indicating high effectiveness of the pix2surv model as a prognostic biomarker for the survival of patients with COVID-19.

To investigate the CT findings of patients with different clinical types of coronavirus disease 2019 (COVID-19). A total of 67 patients diagnosed as COVID-19 by nucleic acid testing were collected and divided into 4 groups according to the clinical stages based on Diagnosis and treatment of novel coronavirus pneumonia (trial version 6). The CT imaging characteristics were analyzed among patients with different clinical types. Among 67 patients, 3(4.5%) were mild, 35 (52.2%) were moderate, 22 (32.8%) were severe, and 7(10.4%) were critical ill. No significant abnormality in chest CT imaging in mild patients. The 35 cases of moderate type included 3 (8.6%) single lesions, the 22 cases of severe cases included 1 (4.5%) single lesion and the rest cases were with multiple lesions. CT images of moderate patients were mainly manifested by solid plaque shadow and halo sign (18/35, 51.4%); while fibrous strip shadow with ground glass shadow was more frequent in severe cases (7/22, 31.8%). Consolidation shadow as the main lesion was observed in 7 cases, and all of them were severe or critical ill patients. CT images of patients with different clinical types of COVID-19 have characteristic manifestations, and solid shadow may predict severe and critical illness.

The purpose of this study was to compare the diagnostic performance of manually fused PET images obtained using 18F-FDG and CT images with that of CT alone, PET alone, and conventional side-by-side review of PET images and CT images (hereafter referred to as "PET + CT") in patients with suspected recurrent colorectal cancer. Ethics committee approval and informed consent were obtained. Sixty-three patients with suspected recurrent colorectal cancer underwent whole-body 18F-FDG PET followed by diagnostic CT. The acquired PET and CT images were merged on a workstation on a pixel-to-pixel basis. CT, PET, PET + CT, and fused images were evaluated separately in terms of the presence or absence of recurrence, new metastases, or both using a 5-point grading scale (0 = definitely negative, 1 = probably negative, 2 = equivocal, 3 = probably positive, and 4 = definitely positive). Lesions determined to be grade 3 or 4 were considered positive, and diagnostic accuracy and certainty were evaluated with statistical analysis using the chi-square test for independence. Of 119 pathologically or clinically confirmed lesions in 36 patients, evaluation of CT, PET, PET + CT, and fused images resulted in the detection of 75 (63%), 84 (71%), 91 (76%), and 111 (93%) lesions, respectively (p < 0.01) with the number of grade 4 lesions detected being 59 (50%), 72 (61%), 84 (71%), and 108 (91%), respectively (p < 0.01). Overall, the diagnostic accuracy of CT, PET, PET + CT, and fused images according to patient were 78%, 79%, 84%, and 92%, respectively (p = 0.13). Interpreting fused images provided more accurate diagnoses than interpreting CT, PET, or PET + CT images. This method of manually fusing separately obtained PET and CT images increased the diagnostic certainty for detecting colorectal cancer recurrence and decreased the number of equivocal cases.

This study was designed to evaluate inter and intraobserver concordance in gross tumor volume (GTV) delineation on megavoltage CT (MVCT) images in patients with postoperative vault recurrences. Three observers delineated GTV on MVCT and CECT and two observers recontoured on MVCT images. Tumor volumes were calculated and correlated using Spearman correlation. The standard deviation of centre of mass was averaged on per patient basis. The ratio of common volume and encompassing volume was used to determine intra and interobserver spatial concordance. Lack of difference in spatial concordance ratio between MVCT and CECT images was used as an index of usability of MVCT images. Thirty six datasets were available for seven patients. High intraobserver GTV correlation was recorded for observer 1 and 2 (r = 0.93 and r = 0.98; P=0.03 and 0.0001). The average intraobserver spatial concordance ratio was 0.57 and 0.62 respectively. The mean GTV of observers 1, 2 and 3 were 31.6 (18.7-52.2); 28.2 (16.7-51.8) and 46.3 cc (29.1-90.5) respectively. Average standard deviation of centre of mass of all observers was less than 5 mm in either direction. Largest interobserver discordance was observed in anterior, inferior and lateral direction. The interobserver spatial concordance of GTV on MVCT and CECT images was 0.34 and 0.36 (P=0.24) respectively. Moderate to good inter and intraobserver GTV correlation was observed on MVCT images, however, was associated with low interobserver spatial concordance on both MVCT and CECT images. Strategies to improve contouring reproducibility on MVCT and KVCT images are desirable.

Radiotherapy requires the target area and the organs at risk to be contoured on the CT image of the patient. During the process of organs-at-Risk (OAR) of the chest and abdomen, the doctor needs to contour at each CT image. The delineations of large and varied shapes are time-consuming and laborious. This study aims to evaluate the results of two automatic contouring softwares on OARs definition of CT images of lung cancer and rectal cancer patients. The CT images of 15 patients with rectal cancer and 15 patients with lung cancer were selected separately, and the organs at risk were manually contoured by experienced physicians as reference structures. And then the same datasets were automatically contoured based on AiContour (version 3.1.8.0, Manufactured by Linking MED, Beijing, China) and Raystation (version 4.7.5.4, Manufactured by Raysearch, Stockholm, Sweden) respectively. Deep learning auto-segmentations and Atlas were respectively performed with AiContour and Raystation. Overlap index (OI), Dice similarity index (DSC) and Volume difference (Dv) were evaluated based on the auto-contours, and independent-sample t-test analysis is applied to the results. The results of deep learning auto-segmentations on OI and DSC were better than that of Atlas with statistical difference. There was no significant difference in Dv between the results of two software. With deep learning auto-segmentations, auto-contouring results of most organs in the chest and abdomen are good, and with slight modification, it can meet the clinical requirements for planning. With Atlas, auto-contouring results in most OAR is not as good as deep learning auto-segmentations, and only the auto-contouring results of some organs can be used clinically after modification.

The lungs of patients with COVID-19 exhibit distinctive lesion features in chest CT images. Fast and accurate segmentation of lesion sites from CT images of patients' lungs is significant for the diagnosis and monitoring of COVID-19 patients. To this end, we propose a progressive dense residual fusion network named PDRF-Net for COVID-19 lung CT segmentation. Dense skip connections are introduced to capture multi-level contextual information and compensate for the feature loss problem in network delivery. The efficient aggregated residual module is designed for the encoding-decoding structure, which combines a visual transformer and the residual block to enable the network to extract richer and minute-detail features from CT images. Furthermore, we introduce a bilateral channel pixel weighted module to progressively fuse the feature maps obtained from multiple branches. The proposed PDRF-Net obtains good segmentation results on two COVID-19 datasets. Its segmentation performance is superior to baseline by 11.6% and 11.1%, and outperforming other comparative mainstream methods. Thus, PDRF-Net serves as an easy-to-train, high-performance deep learning model that can realize effective segmentation of the COVID-19 lung CT images.

BackgroundPneumoconiosis has a significant impact on the quality of patient survival. This study aims to evaluate the performance and application value of improved Unet network technology in the recognition and segmentation of lesion areas of lung CT images in patients with pneumoconiosis.MethodsA total of 1212 lung CT images of patients with pneumoconiosis were retrospectively included. The improved Unet network was used to identify and segment the CT image regions of the patients’ lungs, and the image data of the granular regions of the lungs were processed by the watershed and region growing algorithms. After random sorting, 848 data were selected into the training set and 364 data into the validation set. The experimental dataset underwent data augmentation and were used for model training and validation to evaluate segmentation performance. The segmentation results were compared with FCN-8s, Unet network (Base), Unet (Squeeze-and-Excitation, SE + Rectified Linear Unit, ReLU), and Unet + + networks.ResultsIn the segmentation of lung CT granular region with the improved Unet network, the four evaluation indexes of Dice similarity coefficient, positive prediction value (PPV), sensitivity coefficient (SC) and mean intersection over union (MIoU) reached 0.848, 0.884, 0.895 and 0.885, respectively, increasing by 7.6%, 13.3%, 3.9% and 6.4%, respectively, compared with those of Unet network (Base), and increasing by 187.5%, 249.4%, 131.9% and 51.0%, respectively, compared with those of FCN-8s, and increasing by 14.0%, 31.2%, 4.7% and 9.7%, respectively, compared with those of Unet network (SE + ReLU), while the segmentation performance was also not inferior to that of the Unet + + network.ConclusionsThe improved Unet network proposed shows good performance in the recognition and segmentation of abnormal regions in lung CT images in patients with pneumoconiosis, showing potential application value for assisting clinical decision-making.

Pleural Fluid Analysis and Radiographic, Sonographic, and Echocardiographic Characteristics of Hepatic Hydrothorax

Purpose: Metal in patients creates streak artifacts in CT images. When used for radiation treatment planning, these artifacts make it difficult to identify internal structures and affects radiation dose calculations, which depend on HU numbers for inhomogeneity correction. This work quantitatively evaluates a new metal artifact reduction (MAR) CT image reconstruction algorithm (GE Healthcare CT-0521-04.13-EN-US DOC1381483) when metal is present. Methods: A Gammex Model 467 Tissue Characterization phantom was used. CT images were taken of this phantom on a GE Optima580RT CT scanner with and without steel and titanium plugs using both the standard and MAR reconstruction algorithms. HU values were compared pixel by pixel to determine if the MAR algorithm altered the HUs of normal tissues when no metal is present, and to evaluate the effect of using the MAR algorithm when metal is present. Also, CT images of patients with internal metal objects using standard and MAR reconstruction algorithms were compared. Results: Comparing the standard and MAR reconstructed images of the phantom without metal, 95.0% of pixels were within ±35 HU and 98.0% of pixels were within ±85 HU. Also, the MAR reconstruction algorithm showed significant improvement in maintaining HUs of non-metallic regions in the images taken ofmore » the phantom with metal. HU Gamma analysis (2%, 2mm) of metal vs. non-metal phantom imaging using standard reconstruction resulted in an 84.8% pass rate compared to 96.6% for the MAR reconstructed images. CT images of patients with metal show significant artifact reduction when reconstructed with the MAR algorithm. Conclusion: CT imaging using the MAR reconstruction algorithm provides improved visualization of internal anatomy and more accurate HUs when metal is present compared to the standard reconstruction algorithm. MAR reconstructed CT images provide qualitative and quantitative improvements over current reconstruction algorithms, thus improving radiation treatment planning accuracy.« less

We aim to analyze the ability to detect epithelial growth factor receptor (EGFR) mutations on chest CT images of patients with lung adenocarcinoma using radiomics and/or multi-level residual convolutionary neural networks (MCNNs). We retrospectively collected 1,010 consecutive patients in Shanghai Chest Hospital from 2013 to 2017, among which 510 patients were EGFR-mutated and 500 patients were wild-type. The patients were randomly divided into a training set (810 patients) and a validation set (200 patients) according to a balanced distribution of clinical features. The CT images and the corresponding EGFR status measured by Amplification Refractory Mutation System (ARMS) method of the patients in the training set were utilized to construct both a radiomics-based model (MRadiomics) and MCNNs-based model (MMCNNs). The MRadiomics and MMCNNs were combined to build the ModelRadiomics+MCNNs (MRadiomics+MCNNs). Clinical data of gender and smoking history constructed the clinical features-based model (MClinical). MClinical was then added into MRadiomics, MMCNNs, and MRadiomics+MCNNs to establish the ModelRadiomics+Clinical (MRadiomics+Clinical), the ModelMCNNs+Clinical (MMCNNs+Clinical) and the ModelRadiomics+MCNNs+Clinical (MRadiomics+MCNNs+Clinical). All the seven models were tested in the validation set to ascertain whether they were competent to detect EGFR mutations. The detection efficiency of each model was also compared in terms of area under the curve (AUC), sensitivity and specificity. The AUC of the MRadiomics, MMCNNs and MRadiomics+MCNNs to predict EGFR mutations was 0.740, 0.810 and 0.811 respectively. The performance of MMCNNs was better than that of MRadiomics (P=0.0225). The addition of clinical features did not improve the AUC of the MRadiomics (P=0.623), the MMCNNs (P=0.114) and the MRadiomics+MCNNs (P=0.058). The MRadiomics+MCNNs+Clinical demonstrated the highest AUC value of 0.834. The MMCNNs did not demonstrate any inferiority when compared with the MRadiomics+MCNNs (P=0.742) and the MRadiomics+MCNNs+Clinical (P=0.056). Both of the MRadiomics and the MCNNs could predict EGFR mutations on CT images of patients with lung adenocarcinoma. The MMCNNs outperformed the MRadiomics in the detection of EGFR mutations. The combination of these two models, even added with clinical features, is not significantly more efficient than MMCNNs alone.

Purpose:To assess image quality and radiation dose reduction in abdominal CT imaging, physical phantoms having realistic background textures and lesions are highly desirable. The purpose of this work was to construct a liver phantom with realistic background and lesions using patient CT images and a 3D printer.Methods:Patient CT images containing liver lesions were segmented into liver tissue, contrast‐enhanced vessels, and liver lesions using commercial software (Mimics, Materialise, Belgium). Stereolithography (STL) files of each segmented object were created and imported to a 3D printer (Object350 Connex, Stratasys, MN). After test scans were performed to map the eight available printing materials into CT numbers, printing materials were assigned to each object and a physical liver phantom printed. The printed phantom was scanned on a clinical CT scanner and resulting images were compared with the original patient CT images.Results:The eight available materials used to print the liver phantom had CT number ranging from 62 to 117 HU. In scans of the liver phantom, the liver lesions and veins represented in the STL files were all visible. Although the absolute value of the CT number in the background liver material (approx. 85 HU) was higher than in patients (approx. 40 HU), the difference in CT numbers between lesions and background were representative of the low contrast values needed for optimization tasks. Future work will investigate materials with contrast sufficient to emulate contrast‐enhanced arteries.Conclusion:Realistic liver phantoms can be constructed from patient CT images using a commercial 3D printer. This technique may provide phantoms able to determine the effect of radiation dose reduction and noise reduction techniques on the ability to detect subtle liver lesions in the context of realistic background textures.

Dense GAN and multi-layer attention based lesion segmentation method for COVID-19 CT images

Background Identifying the presence and extent of infarcted brain tissue at baseline plays a crucial role in the treatment of patients with acute ischemic stroke (AIS). Patients with extensive infarction are unlikely to benefit from thrombolysis or thrombectomy procedures. Purpose To develop an automated approach to detect and quantitate infarction by using non-contrast-enhanced CT scans in patients with AIS. Materials and Methods Non-contrast-enhanced CT images in patients with AIS (<6 hours from symptom onset to CT) who also underwent diffusion-weighted (DW) MRI within 1 hour after AIS were obtained from May 2004 to July 2009 and were included in this retrospective study. Ischemic lesions manually contoured on DW MRI scans were used as the reference standard. An automatic segmentation approach involving machine learning (ML) was developed to detect infarction. Randomly selected nonenhanced CT images from 157 patients with the lesion labels manually contoured on DW MRI scans were used to train and validate the ML model; the remaining 100 patients independent of the derivation cohort were used for testing. The ML algorithm was quantitatively compared with the reference standard (DW MRI) by using Bland-Altman plots and Pearson correlation. Results In 100 patients in the testing data set (median age, 69 years; interquartile range [IQR]: 59-76 years; 59 men), baseline non-contrast-enhanced CT was performed within a median time of 48 minutes from symptom onset (IQR, 27-93 minutes); baseline MRI was performed a median of 38 minutes (IQR, 24-48 minutes) later. The algorithm-detected lesion volume correlated with the reference standard of expert-contoured lesion volume in acute DW MRI scans (r = 0.76, P < .001). The mean difference between the algorithm-segmented volume (median, 15 mL; IQR, 9-38 mL) and the DW MRI volume (median, 19 mL; IQR, 5-43 mL) was 11 mL (P = .89). Conclusion A machine learning approach for segmentation of infarction on non-contrast-enhanced CT images in patients with acute ischemic stroke showed good agreement with stroke volume on diffusion-weighted MRI scans. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Nael in this issue.

Background and objectiveNovel coronavirus 2019 (COVID-19) outbreak was first reported in Wuhan, Hubei Province in China in December 2019; it has then spread quickly and exponentially beyond the Chinese borders and is now regarded as a global pandemic. We aimed to evaluate the chest CT radiological characteristics and lesion distribution patterns in patients of COVID-19 pneumonia in London, UK.MethodsWe performed a retrospective study and reviewed data of patients with clinically suspected COVID-19 who underwent chest CT between February 1 and May 5, 2020. All patients underwent the reverse transcription-polymerase chain reaction (RT-PCR) test. Lung lesion characteristics and distribution patterns were evaluated by two radiologists. Fisher’s exact test was used for statistical analysis, and a p-value of <0.05 was considered statistically significant.ResultsA total of 18 patients (nine men and nine women) were analyzed. All of them had bilateral patchy lesions in the chest CT images. There was no correlation between the severity score and mortality (p=0.790). The distinctive CT features included ground-glass opacity (GGO) and consolidative patchy amorphous lesions, bilateral posterior and peripheral multi-lobar lung involvement, pleural effusions, subpleural fibrotic lines, subpleural sparing, vascular engorgement, occasional crazy paving, occasional mediastinal lymphadenopathy, pleural thickening, lack of cavitation, and absence of reverse halo (atoll) signs.ConclusionCT can facilitate the diagnosis of COVID-19 pneumonia. Our UK cohort showed slight variations compared with previously reported Asian and continental European cases with respect to chest CT images.

Are We Choosing Wisely in Lymphoma? Excessive Use of Surveillance CT Imaging in Patients with Diffuse Large B-Cell Lymphoma (DLBCL) in Long-Term Remission