Thoracic Computed Tomography Images Research Articles

Myosteatosis is associated with coronary artery calcification in patients with type 2 diabetes.

Myosteatosis, rather than low muscle mass, is the primary etiologic factor of sarcopenia in patients with type 2 diabetes mellitus (T2DM). Myosteatosis may lead to a series of metabolic dysfunctions, such as insulin resistance, systematic inflammation, and oxidative stress, and all these dysfunctions are closely associated with the acceleration of T2DM and atherosclerosis. To investigate the association between myosteatosis and coronary artery calcification (CAC) in patients with T2DM. Patients with T2DM, who had not experienced major cardiovascular events and had undergone both abdominal and thoracic computed tomography (CT) scans, were included. The mean skeletal muscle attenuation was assessed using abdominal CT images at the L3 level. The CAC score was determined from thoracic CT images using the Agatston scoring method. Myosteatosis was diagnosed according to Martin's criteria. Severe CAC (SCAC) was defined when the CAC score exceeded 300. Logistic regression and decision tree analyses were performed. A total of 652 patients with T2DM were enrolled. Among them, 167 (25.6%) patients had SCAC. Logistic regression analysis demonstrated that myosteatosis, age, duration of diabetes, cigarette smoking, and alcohol consumption were independent risk factors of SCAC. Myosteatosis was significantly associated with an increased risk of SCAC (OR = 2.381, P = 0.003). The association between myosteatosis and SCAC was significant in the younger patients (OR = 2.672, 95%CI: 1.477-4.834, P = 0.002), but not the older patients (OR = 1.456, 95%CI: 0.863-2.455, P = 0.188), and was more prominent in the population with lower risks of atherosclerosis. The decision tree analyses prioritized older age as the primary variable for SCAC. In older patients, cigarette smoking was the main contributing factor for SCAC, while in younger patients, it was myosteatosis. Myosteatosis is a novel risk factor for atherosclerosis in patients with T2DM, especially in the population with younger ages and fewer traditional risk factors.

Automated size-specific dose estimates framework in thoracic CT using convolutional neural network based on U-Net model.

This study aimed to develop an automated method that uses a convolutional neural network (CNN) for calculating size-specific dose estimates (SSDEs) based on the corrected effective diameter (Deff corr ) in thoracic computed tomography (CT). Transaxial images obtained from 108 adult patients who underwent non-contrast thoracic CT scans were analyzed. To calculate the Deff corr according to Mihailidis etal., the average relative electron densities for lung, bone, and other tissues were used to correct the lateral and anterior-posterior dimensions. The CNN architecture based on the U-Net algorithm was used for automated segmentation of three classes of tissues and the background region to calculate dimensions and Deff corr values. Then, 108 thoracic CT images and generated segmentation masks were used for network training. The water-equivalent diameter (Dw ) was determined according to the American Association of Physicists in Medicine Task Group 220. Linear regression and Bland-Altman analysis were performed to determine the correlations between SSDEDeff corr(automated) , SSDEDeff corr(manual) , and SSDEDw . High agreement was obtained between the manual and automated methods for calculating the Deff corr SSDE. The mean values for the SSDEDeff corr(manual) , SSDEDw , and SSDEDeff corr(automated) were 14.3±2.1 mGy, 14.6±2.2 mGy, and 14.5±2.4 mGy, respectively. The U-Net model was successfully trained and used to accurately predict SSDEs, with results comparable to manual-labeling results. The proposed automated framework using a CNN offers a reliable and efficient solution for determining the Deff corr SSDE in thoracic CT.

Robust explanation supervision for false positive reduction in pulmonary nodule detection.

Lung cancer is the deadliest and second most common cancer in the United States due to the lack of symptoms for early diagnosis. Pulmonary nodules are small abnormal regions that can be potentially correlated to the occurrence of lung cancer. Early detection of these nodules is critical because it can significantly improve the patient's survival rates. Thoracic thin-sliced computed tomography (CT) scanning has emerged as a widely used method for diagnosing and prognosis lung abnormalities. The standard clinical workflow of detecting pulmonary nodules relies on radiologists to analyze CT images to assess the risk factors of cancerous nodules. However, this approach can be error-prone due to the various nodule formation causes, such as pollutants and infections. Deep learning (DL) algorithms have recently demonstrated remarkable success in medical image classification and segmentation. As an ever more important assistant to radiologists in nodule detection, it is imperative ensure the DL algorithm and radiologist to better understand the decisions from each other. This study aims to develop a framework integrating explainable AI methods to achieve accurate pulmonary nodule detection. A robust and explainable detection (RXD) framework is proposed, focusing on reducing false positives in pulmonary nodule detection. Its implementation is based on an explanation supervision method, which uses nodule contours of radiologists as supervision signals to force the model to learn nodule morphologies, enabling improved learning ability on small dataset, and enable small dataset learning ability. In addition, two imputation methods are applied to the nodule region annotations to reduce the noise within human annotations and allow the model to have robust attributions that meet human expectations. The 480, 265, and 265 CT image sets from the public Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI) dataset are used for training, validation, and testing. Using only 10, 30, 50, and 100 training samples sequentially, our method constantly improves the classification performance and explanation quality of baseline in terms of Area Under the Curve (AUC) and Intersection over Union (IoU). In particular, our framework with a learnable imputation kernel improves IoU from baseline by 24.0% to 80.0%. A pre-defined Gaussian imputation kernel achieves an even greater improvement, from 38.4% to 118.8% from baseline. Compared to the baseline trained on 100 samples, our method shows less drop in AUC when trained on fewer samples. A comprehensive comparison of interpretability shows that our method aligns better with expert opinions. A pulmonary nodule detection framework was demonstrated using public thoracic CT image datasets. The framework integrates the robust explanation supervision (RES) technique to ensure the performance of nodule classification and morphology. The method can reduce the workload of radiologists and enable them to focus on the diagnosis and prognosis of the potential cancerous pulmonary nodules at the early stage to improve the outcomes for lung cancer patients.

COVID 19 Hastalarında Arcus Aorta Açısı ve Arcus Aorta Morfometrisi: Radyoanatomik bir Çalışma

Aim: In endovascular surgery, knowing the morphometry of the aortic arch increases the success of surgery. The aim of this study was to examine the angle and morphometry of aortic arch in COVID 19 patients and to compare these with healthy individuals to find out the effect of the disease on the vessel.
 Material and Methods: A total of 120 individuals - 60 COVID 19 (30 females, 30 males) patients and 60 healthy (30 females, 30 males) individuals participated in the study. In the study, the parameters of aortic arch angle (AAA), aortic arch diameter (AAD), aortic arch (AA) branches of brachiocephalic trunk diameter (BTD), left common carotid artery diameter (LCCAD), left subclavian artery diameter (LSAD), transverse superior thoracic aperture length (TR-STAL) and anteroposterior superior thoracic aperture length (AP-STAL), transverse inferior thoracic aperture length (TR-ITAL) and anteroposterior inferior thoracic aperture length (AP-ITAL) were measured from thoracic computed tomography images.
 Results: As a result of the study, when female and male patients with COVID-19 were compared, LCCAD, LSAD, AP-ITAL, TR-ITAL values were found to be higher in favour of male patients. While Proximal AAD, BTD, LCCAD and LSAD values were higher in female patients with COVID 19 when compared with control group female patients, Proximal AAD, BTD, LCCAD, LSAD, AP-STAL, TR-STAL, AP-ITAL, TR-ITAL values were higher in male patients with COVID 19 when compared with control group male patients. When the measurements of COVID 19 and control group individuals were compared, Proximal AAD, BTD, ACCS, LSAD, TR-STAL, AP-ITAL and TR-ITAL values were found to be higher in favour of COVID 19 patients. 
 Conclusion: COVID 19 is an important disease that causes dilatation of the AA and its branches. We think that diseases that can change oxygen saturation such as COVID19 can change aortic morphology.

Toraks bilgisayarlı tomografisinin yorumlanmasında acil tıp asistanlarının başarısı ve tanı koymadaki özgüvenleri

Purpose:
 Interpration of thoracic computed tomography (CT) scans by emergency department (ED) physicians is important especially in crowded EDs. The aim of this study is to assess the proficiency and confidence of ED physicians with varying levels of experience in interpreting thoracic CT.
 Materials and methods:
 A total of 25 pathological and 5 normal thoracic CT images were interpreted by 32 ED residents, initially without clinical information, then with. After each session the participants were asked to score their confidence on a scale of 1 to 10. At the end of the study, the results were compared between seniors and junior residents. 
 Results:
 The median age of the participants was 29 years (24–34). Twenty (62.5%) of the residents were junior residents. There were no significant differences between the two resident groups in terms of accurate diagnosis rates, regardless of the clinical information (p=0.307 and p=0.061). The physicians’ certainty of their own diagnosis mostly does not seem to be statistically different in these diagnoses. The seniors are more confident in the diagnosis of the CT scans they evaluated without clinical information (p=0.004), while when the clinical information is added, the confidence of the junior physicians also increase (p=0.087).
 Conclusion:
 Both senior and junior emergency medical residents are able to interpret thoracic CT images with a high degree of accuracy, both for COVID-19 and for other emergency diagnoses. Senior physicians could confidently diagnose CT images solely by visual inspection, whereas junior residents display similar confidence when clinical information is available.

Computed tomography reveals hypertrophic remodelling of the diaphragm in cystic fibrosis but not in COPD.

Computed tomography (CT) is increasingly used for assessing skeletal muscle characteristics. In cystic fibrosis (CF) and chronic obstructive pulmonary disease (COPD), reduced limb muscle mass predicts poor clinical outcomes. However, the degree to which quantity or quality of respiratory and nonrespiratory muscles is affected by these diseases remains controversial. Thoracic CT images of 29 CF, 21 COPD and 20 normal spirometry control subjects were analysed to measure indices of muscle quantity (volume or cross-sectional area) and quality (radiodensity) in respiratory (diaphragm, abdominal) and nonrespiratory (pectoralis, lumbar paraspinal) muscles. Multivariable linear regression assessed relationships of CT measurements with body mass index (BMI), forced expiratory volume in 1 s (FEV1) % pred, inflammation and infection biomarkers, nutritional status and CF genotype. Diaphragm volume in CF was significantly higher than in COPD (by 154%) or controls (by 140%). Abdominal muscle area in CF was also greater than in COPD (by 130%). Nonrespiratory muscles in COPD had more low radiodensity muscle (marker of lipid content) compared to CF and controls. In CF but not COPD, higher BMI and FEV1 % pred were independently associated with higher diaphragm and/or abdominal muscle quantity indices. Serum creatinine also predicted respiratory and nonrespiratory muscle quantity in CF, whereas other biomarkers including genotype correlated poorly with muscle CT parameters. Our data suggest that the CF diaphragm undergoes hypertrophic remodelling, whereas in COPD the nonrespiratory muscles show altered muscle quality consistent with greater lipid content. Thoracic CT can thus identify distinctive respiratory and nonrespiratory muscle remodelling signatures associated with different chronic lung diseases.

HCIU: Hybrid clustered inception‐based UNET for the automatic segmentation of organs at risk in thoracic computed tomography images

AbstractAutomatic segmentation is the most effective and fast method to plan radiotherapy treatment in thoracic computed tomography (CT), as manual segmentation takes time and produces variability in the results. This article proposes a hybrid HCIU method to segment the thoracic organs at risk (OAR) using prior localization techniques. The dataset of 40 Patients containing 7420 slices of thoracic CT has been used to perform the experiment on the Google Colab Pro+ framework. HCIU constitutes a hybrid deep learning model for features extraction, a K‐mean algorithm to form the clusters based on the extracted features, and the ensemble UNET InceptionV3 model trained on the data of each cluster for the segmentation of OAR. With HCIU, we achieved an average dice score of 92.05%, 97.39%, 95.75%, and 96.45% for the esophagus, heart, trachea, and aorta. The average sensitivity and specificity values achieved were 98.38% and 97.43% on the test dataset. Also, statistical tests have been applied to each organ's pixel values for analyzing the model's segmented results. A hybrid method has been proposed to improve the localization in thoracic OAR segmentation in CT. The proposed method achieved precise accuracy in segmenting the four organs compared with the other state‐of‐the‐art methods.

Automatic Benign and Malignant Classification of Pulmonary Nodules in Thoracic Computed Tomography Images based on CNN Algorithm

The majority of the time, cancer symptoms only appear in the final stages of the disease, but with the use of advanced technology in which computer-aided systems, we can identify cancer at an early stage, then we may be able to treat it. These automated detection systems use a variety of machine learning techniques to identify lung cancer in its early stages. This type of automatic diagnosis of lung cancer is detected in computed tomography image modalities using convolution neural network. Compared to MRI and X-ray, CT pictures have the advantage of having less noise interference. Median filtering is applied to de-noising the CT scans to enhance the image quality. These pre-processed images are then fed through the Re-Lu layer CNN architecture. This architecture's several layers handle the feature extraction and categorization work. Different low-level and distinct high-level characteristics are extracted during the feature extraction phase. The classification layer is in charge of determining whether the presented image contains a benign or malignant, and it can able to classify the severity of malignance based on the infected area. To carry out the proposed model, CT image datasets are taken from Imaging data base Resource Initiative (IDRI) and Lung image Database consortium (LIDC) datasets. According to the findings, the suggested proposed system works with optimal results. The comparison demonstrates that the suggested proposed system outperforms the current cutting-edge technologies. The suggested proposed system will be beneficial for health care systems and medical diagnosis research. The proposed methodology is verified on MATLAB software.

Thoracic CT Scan Findings in Patients with Confirmed Hematologic Malignancies Admitted to the Hospital with Acute Pulmonary Symptoms

Background: Hematologic malignancies are the fifth most common malignancies worldwide. Pulmonary symptoms are among the patient’s common complaints. Early diagnosis based on patient symptoms, laboratory testing, and imaging may lead to an appropriate therapeutic approach. Objectives: To evaluate thoracic computed tomography (CT) scan findings in known cases of hematologic malignancy with acute respiratory symptoms. Methods: A cross-sectional study was conducted on 200 patients with hematologic malignancies admitted to Khansari cancer center for acute pulmonary symptoms, from December 2019 to May 2020. Collected data included demographic data, physical examination, laboratory tests and findings of thoracic CT images. Descriptive analysis and chi-Square test were performed to analyze the data using SPSS version 20. Results: Non-Hodgkin lymphoma was the most common malignancy, accounting for 48% of total cases. The most common radiological findings included atelectasis (49%), followed by pleural effusion (42%), mediastinal lymphadenopathy (38.5%), consolidation (37%), and ground-glass opacities (33.5%). Mediastinal lymphadenopathy was more prevalent in patients diagnosed with lymphoma. Among pulmonary pathologies, pneumonia was the most prevalent disease (41.5%), followed by small airway disease (13.5%), mediastinal lymphadenopathy (12.5%), chronic pulmonary disease (6.5%), and pulmonary metastasis (6%). Normal thoracic CT scan was detected in 20% of participants. Overall assessment of clinical symptoms, laboratory tests, and imaging findings did not lead to the diagnosis of acute lung involvement in 5.5% of study participants, because complete clinical data were not available in some participants. Conclusions: Assessment of the most common thoracic imaging findings in patients diagnosed with hematologic malignancies, may help make better diagnosis and planning of a rapid therapeutic approach.

Assessment of Optimizers and their Performance in Autosegmenting Lung Tumors.

Optimizers are widely utilized across various domains to enhance desired outcomes by either maximizing or minimizing objective functions. In the context of deep learning, they help to minimize the loss function and improve model's performance. This study aims to evaluate the accuracy of different optimizers employed for autosegmentation of non-small cell lung cancer (NSCLC) target volumes on thoracic computed tomography images utilized in oncology. The study utilized 112 patients, comprising 92 patients from "The Cancer Imaging Archive" (TCIA) and 20 of our local clinical patients, to evaluate the efficacy of various optimizers. The gross tumor volume was selected as the foreground mask for training and testing the models. Of the 92 TCIA patients, 57 were used for training and validation, and the remaining 35 for testing using nnU-Net. The performance of the final model was further evaluated on the 20 local clinical patient datasets. Six different optimizers, namely AdaDelta, AdaGrad, Adam, NAdam, RMSprop, and stochastic gradient descent (SGD), were investigated. To assess the agreement between the predicted volume and the ground truth, several metrics including Dice similarity coefficient (DSC), Jaccard index, sensitivity, precision, Hausdorff distance (HD), 95th percentile Hausdorff distance (HD95), and average symmetric surface distance (ASSD) were utilized. The DSC values for AdaDelta, AdaGrad, Adam, NAdam, RMSprop, and SGD were 0.75, 0.84, 0.85, 0.84, 0.83, and 0.81, respectively, for the TCIA test data. However, when the model trained on TCIA datasets was applied to the clinical datasets, the DSC, HD, HD95, and ASSD metrics showed a statistically significant decrease in performance compared to the TCIA test datasets, indicating the presence of image and/or mask heterogeneity between the data sources. The choice of optimizer in deep learning is a critical factor that can significantly impact the performance of autosegmentation models. However, it is worth noting that the behavior of optimizers may vary when applied to new clinical datasets, which can lead to changes in models' performance. Therefore, selecting the appropriate optimizer for a specific task is essential to ensure optimal performance and generalizability of the model to different datasets.

Congruence of radiological scoring systems used in COVID-19 pneumonia and effect of comorbid diseases on radiological features

Background/Aim: Several scoring systems have been developed to standardize radiological findings in patients with COVID-19 pneumonia. The most commonly used scoring systems in the radiological examination of COVID-19 are those of the North American Radiology Association (RSNA), British Thoracic Society (BTS), and COVID-19 Reporting and Data System (CO-RADS). However, the compatibility between these radiological scoring systems has not been evaluated before. Therefore, this study evaluated the radiological features of COVID-19 pneumonia and congruence between radiological scoring systems and determined the effect of comorbidities and demographic characteristics on radiological features and thoracic computed tomography (TCT) findings in the context of COVID-19. Methods: A retrospective cohort study was performed on patients attending our unit with a suspicion of COVID-19 who also had a positive real-time transcriptase-polymerase chain reaction (RT-PCR) test. All TCT images were subjected to the RSNA, BTS, and CO-RADS scoring systems. Demographic data such as age and gender, and comorbid conditions were recorded. Results: TCT showed peripheral, posterior, and bilateral involvement in 97.7%, 97.7%, and 87.6% of the patients, respectively. The most common TCT finding was ground glass appearance, which was found in 95.5% of the patients. The Charlson Comorbidity Index (CCI) score was found to have an impact on RSNA and BTS criteria (P=0.011 and P=0.014), while age, gender, and the presence of comorbidities such as cardiovascular disease (CVD), diabetes mellitus (DM), and chronic pulmonary disease (CPD) did not have such an effect (P>0.05 for all). On the other hand, CCI scores and the presence of CPD had an association with CO-RADS, but there was no effect of age, gender, DM, and CVD (P=0.915 and P=0.730). Conclusion: TCT plays an important role in early management, isolation, and follow-up of patients with COVID-19 pneumonia. The radiological scoring systems were found to exhibit good compatibility, but comorbid conditions could have an impact on the assessment. Therefore, we conclude that these radiological assessment criteria are useful in the management and monitoring of such patients.

Automatic Segmentation of Organs-at-Risk in Thoracic Computed Tomography Images Using Ensembled U-Net InceptionV3 Model.

The objective of this article is to automatically segment organs at risk (OARs) for thoracic radiology in computed tomography (CT) scan images. The OARs in the thoracic anatomical region during the radiotherapy treatment are mainly the neighbouring organs such as the esophagus, heart, trachea, and aorta. The dataset of 40 patients was used in the proposed work by splitting it into three parts: training, validation, and test sets. The implementation was performed on the Google Colab Pro+ framework with 52 GB of RAM and 265 GB of storage space. An ensemble model was evolved for the automatic segmentation of four OARs in thoracic CT images. U-Net with InceptionV3 as the backbone was used, and different hyperparameters were used during the training of the model. The proposed model achieved precise accuracy for OARs segmentation with an average dice coefficient of 0.9413, Hausdorff value of 0.1838, sensitivity of 0.9783, and specificity of 0.9895 on the Test dataset. An ensembled U-Net InceptionV3 model has been proposed, improving the segmentation results compared with the state-of-the-art techniques such as U-Net, ResNet, Vgg16, etc. The results of the experiments revealed that the proposed model effectively improved the performance of the segmentation of the esophagus, heart, trachea, and aorta.

A CT-based nomogram for differentiating invasive fungal disease of the lung from bacterial pneumonia

BackgroundThere is an annual increase in the incidence of invasive fungal disease (IFD) of the lung worldwide, but it is always a challenge for physicians to make an early diagnosis of IFD of the lung. Computed tomography (CT) may play a certain role in the diagnosis of IFD of the lung, however, there are no specific imaging signs for differentiating IFD of lung from bacterial pneumonia (BP).MethodsA total of 214 patients with IFD of the lung or clinically confirmed BP were retrospectively enrolled from two institutions (171 patients from one institution in the training set and 43 patients from another institution in the test set). The features of thoracic CT images of the 214 patients were analyzed on the picture archiving and communication system by two radiologists, and these CT images were imported into RadCloud to perform radiomics analysis. A clinical model from radiologic analysis, a radiomics model from radiomics analysis and a combined model from integrating radiologic and radiomics analysis were constructed in the training set, and a nomogram based on the combined model was further developed. The area under the ROC curve (AUC) of the receiver operating characteristic (ROC) curve was calculated to assess the diagnostic performance of the three models. Decision curve analysis (DCA) was conducted to evaluate the clinical utility of the three models by estimating the net benefit at a range of threshold probabilities.ResultsThe AUCs of the clinical model for differentiating IFD of lung from BP in the training set and test sets were 0.820 and 0.827. The AUCs of the radiomics model in the training set and test sets were 0.895 and 0.857. The AUCs of the combined model in the training set and test setswere 0.944 and 0.911. The combined model for differentiating IFD of lung from BP obtained the greatest net benefit among the three models by DCA.ConclusionOur proposed nomogram, based on a combined model integrating radiologic and radiomics analysis, has a powerful predictive capability for differentiating IFD from BP. A good clinical outcome could be obtained using our nomogram.

Correlation of Clinical Course with Computed Tomography Findings and Biochemical Parameters at the Time of Admission in COVID-19 Patients.

Objective:The primary objective of our study was to evaluate the predictive performance of serum inflammatory markers and the semi-quantitative computed tomography severity scoring system on diagnosing the Covid 2019 disease and its course.Methods:Our study is a single-center retrospective cohort study. The data of 213 adults who were confirmed to have coronavirus disease 2019 by polymerase chain reaction tests in the period between April 2020 and August 2020 were evaluated. One hundred eighty four of these patients whose C-reactive protein, d-dimer, and ferritin levels, lymphocyte counts, and thoracic computed tomography images were obtained at the time of admission were included in the study. The semi-quantitative computed tomography severity score was calculated for all patients.Results:The median age of the 184 patients included in the study was 51.5 (19-91) years. The incidence of intensive care need and mortality was 10.3% (n = 19) and 5.4% (n = 10), respectively. The intensive care need and mortality rate was significantly correlated with higher thoracic computed tomography involvement scores at admission. There was a statistically significant and positive correlation between the computed tomography scores and the C-reactive protein, d-dimer, and ferritin levels. Older age (>65 years-old) and thoracic computed tomography scores of 11 and higher were independent factors correlated with need for intensive care.Conclusion:Serum inflammatory markers and semi-quantitative computed tomography severity scoring system were predictive in diagnosing the Covid 2019 disease and its course.

A deep image-to-image network organ segmentation algorithm for radiation treatment planning: principles and evaluation

BackgroundWe describe and evaluate a deep network algorithm which automatically contours organs at risk in the thorax and pelvis on computed tomography (CT) images for radiation treatment planning.MethodsThe algorithm identifies the region of interest (ROI) automatically by detecting anatomical landmarks around the specific organs using a deep reinforcement learning technique. The segmentation is restricted to this ROI and performed by a deep image-to-image network (DI2IN) based on a convolutional encoder-decoder architecture combined with multi-level feature concatenation. The algorithm is commercially available in the medical products “syngo.via RT Image Suite VB50” and “AI-Rad Companion Organs RT VA20” (Siemens Healthineers). For evaluation, thoracic CT images of 237 patients and pelvic CT images of 102 patients were manually contoured following the Radiation Therapy Oncology Group (RTOG) guidelines and compared to the DI2IN results using metrics for volume, overlap and distance, e.g., Dice Similarity Coefficient (DSC) and Hausdorff Distance (HD95). The contours were also compared visually slice by slice.ResultsWe observed high correlations between automatic and manual contours. The best results were obtained for the lungs (DSC 0.97, HD95 2.7 mm/2.9 mm for left/right lung), followed by heart (DSC 0.92, HD95 4.4 mm), bladder (DSC 0.88, HD95 6.7 mm) and rectum (DSC 0.79, HD95 10.8 mm). Visual inspection showed excellent agreements with some exceptions for heart and rectum.ConclusionsThe DI2IN algorithm automatically generated contours for organs at risk close to those by a human expert, making the contouring step in radiation treatment planning simpler and faster. Few cases still required manual corrections, mainly for heart and rectum.

One-step algorithm for fast-track localization and multi-category classification of histological subtypes in lung cancer

BackgroundsAccumulated evidence has proven that computer-derived features from computed tomography (CT) through radiomics and deep learning technologies can identify extensive characteristics of pulmonary malignancies, such as nodules detection and malignant lesion discrimination. However, there are few studies on whether CT images can reflect histological subtypes of lung cancer through computer-derived features. MethodsContrast-enhanced CT images prior treatment from 417 patients diagnosed with small cell lung cancer (SCLC), lung adenocarcinoma (ADC), or lung squamous cell carcinoma (SCC) were collected. ITK-SNAP software was used by trained radiologists for the manual delineation of tumor volume. Patients of each category (SCLC, ADC, SCC) were then randomly split into training datasets and test datasets in an approximately ratio of 8:2. After image pre-processing and augmentation, 25,042 CT images from the training datasets were used to train our self-developed deep learning model for fast-tracking tumor lesions and classifying corresponding histological subtypes simultaneously. The performance of the network was evaluated by accuracy, F1-score and weighted F1-average using 1,921 testing images based on parameters generated during training. ResultsThe prediction accuracy of SCLC, ADC, and SCC were 0.83, 0.75 and 0.67, respectively. The weighted F1-average was 0.75. ADC obtained the best F1-score of 0.78, which was outperformed SCLC (0.77) and SCC (0.66). The corresponding AUC values of SCLC, ADC, and SCC were 0.87, 0.84, and 0.76, respectively. Only 0.24 s were required to simultaneously achieve functions of tumor localization and histological classification on a thoracic CT image slice. The heat map visualization illustrated the extracted tumor features to classify subtypes of lung cancer by the proposed model. ConclusionsThe newly developed multi-task algorithm provides a CNN-based DL approach in lung cancer for automatically fast-tracking tumor lesions and classifying corresponding histological subtypes in one-step.

FcTC-UNet: Fine-grained Combination of Transformer and CNN for Thoracic Organs Segmentation.

Precise segmentation of organs at risk (OARs) in computed tomography (CT) images is an essential step for lung cancer radiotherapy. However, the manual delineation of OARs is time-consuming and subject to inter-observer variation. Although U-like architecture has achieved great success in medical image segmentation recently, it exhibits the limitations in modeling long-range dependencies. As an alternative structure, Transformers have emerged due to the outstanding capability of capturing the global contextual information provided by Self-Attention(SA) mechanism. However, Transformers need more computational cost than CNNs for introducing the SA module. In this paper, we propose a novel module named fine-grained combination of Transformer and CNN(FcTC). FcTC module is composed of dual-path extractor and fusing unit to effectively extract local information and model long-distance dependency. Then we build FcTC-UNet to automatically segment the OARs in thoracic CT images. The experiments results demonstrate that the proposed method achieves better performance over other state-of-the-art methods.

Pandemi Hastanesine Başvuran Hastaların Bilgisayarlı Tomografi ve PCR Sonuçlarının COVID-19 Açısından Değerlendirilmesi

Aim: COVID-19 is an infectious disease that primarily affects the respiratory system and spreads rapidly. Diagnostic methods as computed Tomography (CT) and reverse transcription-polymerase chain reaction (RT-PCR) have gained importance in the COVID-19 pandemic. The aim of this study is to evaluate the appearance and distribution of abnormal parenchymal findings with Thorax CT in patients diagnosed with COVID-19 by RT-PCR method. And to evaluate the relationship between the severity of lung infection and the clinical course of the disease in these patients. Materials and Methods: Patients (n:613) with a preliminary diagnosis of COVID-19 who applied to Konya Beyhekim Training and Research Hospital were evaluated retrospectively between December 2020 and February 2021. Nasopharyngeal samples were studied for COVID-19 with RT-PCR by Selcuk University Medical Faculty Microbiology Laboratory. Thoracic CT images of 361 patients with positive COVID-19 PCR tests were examined for the presence of COVID-19 pneumonia. The clinical course of patients with Covid-19 pneumonia was evaluated. Results: RT-PCR results was positive in 361 (58%) of 613 patients. While 243 (64.8%) of the PCR positive patients had signs of pneumonia, 118 of them had normal lung parenchyma (GGO). There was consolidation in 22% of the patients, and ground glass with consolidation in 20%. 34 patients (13.9%) had crazy-paving pattern and 19 (7.8%) patients had reverse halo appearance. A significant relationship was found between the severity of lung infection involvement and the clinical course of the disease. Conclusion: A comprehensive understanding of diagnostic imaging features is essential for effective patient management and treatment.

Ambient Air Pollution and Dysanapsis: Associations with Lung Function and Chronic Obstructive Pulmonary Disease in the Canadian Cohort Obstructive Lung Disease Study.

Rationale: Outdoor air pollution is a potential risk factor for lower lung function and chronic obstructive pulmonary disease (COPD). Little is known about how airway abnormalities and lung growth might modify this relationship. Objectives: To evaluate the associations of ambient air pollution exposure with lung function and COPD and examine possible interactions with dysanapsis. Methods: We made use of cross-sectional postbronchodilator spirometry data from 1,452 individuals enrolled in the CanCOLD (Canadian Cohort Obstructive Lung Disease) study with linked ambient fine particulate matter (PM2.5) and nitrogen dioxide (NO2) air pollution estimates. Dysanapsis, or the ratio of the airway-to-lung volume calculated from thoracic computed tomography images, was used to examine possible interactions. Measurements and Main Results: In adjusted models, 101.7 ml (95% confidence interval [CI], -166.2 to -37.2) and 115.0 ml (95% CI, -196.5 to -33.4) lower FEV1 were demonstrated per increase of 2.4 ug/m3 PM2.5 and 9.2 ppb NO2, respectively. Interaction between air pollution and dysanapsis was not statistically significant when modeling the airway-to-lung ratio as a continuous variable. However, a 109.8 ml (95% CI, -209.0 to -10.5] lower FEV1 and an 87% (95% CI, 12% to 213%) higher odds of COPD were observed among individuals in the lowest, relative to highest, airway-to-lung ratio, per 2.4 μg/m3 increment of PM2.5. Conclusions: Ambient air pollution exposure was associated with lower lung function, even at relatively low concentrations. Individuals with dysanaptic lung growth might be particularly susceptible to inhaled ambient air pollutants, especially those at the extremes of dysanapsis.

Cross-sectional area of erector spinae muscles is associated with activities of daily living at discharge in middle- to older-aged patients with coronavirus disease 2019.

ObjectivesMeasurement of skeletal muscle wasting using computed tomography (CT) is widely known to be useful in predicting prognosis. Although some reports have been found in patients with coronavirus disease 2019 (COVID-19), few reports have focused on the ability to perform activities of daily living (ADLs). This study retrospectively investigated the relationship between the erector spinae muscle area measured from CT images and ADL at the time of hospital discharge in patients with COVID-19.MethodsAmong patients aged 40 years or older, 271 patients (median age, 65 years; 180/271 male patients) who had CT cross-sectional images of the 12th thoracic vertebral level on admission were included. The Katz index was used to assess ADLs, and patients who were not completely independent were defined as dependent. Multivariable logistic and Poisson regression analyses were applied to examine the relationship between the cross-sectional area of the erector spinae muscles and the onset of ADL dependence at discharge.ResultsA total of 75 (27.7%) patients became dependent on ADL at the time of hospital discharge. Decreased erector spinae muscle area was significantly related to dependent ADL at discharge (adjusted odds ratio: 0.886, 95% confidence interval: 0.805–0.975). In addition, the erector spinae muscle area was significantly related to the number of ADL items for which independence was not achieved (adjusted incidence rate ratio: 0.959, P < 0.001).ConclusionsThe cross-sectional area of the erector spinae muscles from the thoracic CT image was associated with the ability to perform basic ADL at hospital discharge.