• Korean Journal of Radiology
• /
• v.21 no.7
• /
• pp.869-879
• /
• 2020

Objective: To evaluate the performance of a convolutional neural network (CNN) model that can automatically detect and classify rib fractures, and output structured reports from computed tomography (CT) images. Materials and Methods: This study included 1079 patients (median age, 55 years; men, 718) from three hospitals, between January 2011 and January 2019, who were divided into a monocentric training set (n = 876; median age, 55 years; men, 582), five multicenter/multiparameter validation sets (n = 173; median age, 59 years; men, 118) with different slice thicknesses and image pixels, and a normal control set (n = 30; median age, 53 years; men, 18). Three classifications (fresh, healing, and old fracture) combined with fracture location (corresponding CT layers) were detected automatically and delivered in a structured report. Precision, recall, and F1-score were selected as metrics to measure the optimum CNN model. Detection/diagnosis time, precision, and sensitivity were employed to compare the diagnostic efficiency of the structured report and that of experienced radiologists. Results: A total of 25054 annotations (fresh fracture, 10089; healing fracture, 10922; old fracture, 4043) were labelled for training (18584) and validation (6470). The detection efficiency was higher for fresh fractures and healing fractures than for old fractures (F1-scores, 0.849, 0.856, 0.770, respectively, p = 0.023 for each), and the robustness of the model was good in the five multicenter/multiparameter validation sets (all mean F1-scores > 0.8 except validation set 5 [512 x 512 pixels; F1-score = 0.757]). The precision of the five radiologists improved from 80.3% to 91.1%, and the sensitivity increased from 62.4% to 86.3% with artificial intelligence-assisted diagnosis. On average, the diagnosis time of the radiologists was reduced by 73.9 seconds. Conclusion: Our CNN model for automatic rib fracture detection could assist radiologists in improving diagnostic efficiency, reducing diagnosis time and radiologists' workload.

• Journal of the Korean Society of Radiology
• /
• v.84 no.1
• /
• pp.253-262
• /
• 2023

Purpose To assess the clinicoradiological features of pulmonary cryptococcosis in immunocompetent patients. Materials and Methods This retrospective study included immunocompetent patients who had been diagnosed with pulmonary cryptococcosis on the histopathologic exam and underwent chest CT between January 2008 and November 2019. Imaging features were divided into major imaging patterns, distributions, and ancillary imaging findings. Univariable analysis was performed to evaluate clinicoradiological features according to the presence of serum cryptococcal antigen. Results Thirty-one patients were evaluated (mean age: 60 years, range: 19-78 years). A single nodular lesion confined to a single lobe was the most common imaging pattern (14/31, 45.2%). Serum cryptococcal antigen tests were performed in 19 patients (19/31, 61.3%). The presence of serum cryptococcal antigen was observed in six patients (6/19, 31.6%), all of whom showed a consolidation-dominant pattern. The presence of serum cryptococcal antigen was significantly associated with the consolidationdominant pattern compared to those associated with a nodule-dominant pattern (p = 0.011). Conclusion A combination of CT findings of consolidation and a positive serum cryptococcal antigen test may be helpful for diagnosing pulmonary cryptococcosis in immunocompetent patients.

• Korean Journal of Radiology
• /
• v.22 no.3
• /
• pp.344-353
• /
• 2021

Objective: The mitotic count of gastrointestinal stromal tumors (GIST) is closely associated with the risk of planting and metastasis. The purpose of this study was to develop a predictive model for the mitotic index of local primary GIST, based on deep learning algorithm. Materials and Methods: Abdominal contrast-enhanced CT images of 148 pathologically confirmed GIST cases were retrospectively collected for the development of a deep learning classification algorithm. The areas of GIST masses on the CT images were retrospectively labelled by an experienced radiologist. The postoperative pathological mitotic count was considered as the gold standard (high mitotic count, > 5/50 high-power fields [HPFs]; low mitotic count, ≤ 5/50 HPFs). A binary classification model was trained on the basis of the VGG16 convolutional neural network, using the CT images with the training set (n = 108), validation set (n = 20), and the test set (n = 20). The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated at both, the image level and the patient level. The receiver operating characteristic curves were generated on the basis of the model prediction results and the area under curves (AUCs) were calculated. The risk categories of the tumors were predicted according to the Armed Forces Institute of Pathology criteria. Results: At the image level, the classification prediction results of the mitotic counts in the test cohort were as follows: sensitivity 85.7% (95% confidence interval [CI]: 0.834-0.877), specificity 67.5% (95% CI: 0.636-0.712), PPV 82.1% (95% CI: 0.797-0.843), NPV 73.0% (95% CI: 0.691-0.766), and AUC 0.771 (95% CI: 0.750-0.791). At the patient level, the classification prediction results in the test cohort were as follows: sensitivity 90.0% (95% CI: 0.541-0.995), specificity 70.0% (95% CI: 0.354-0.919), PPV 75.0% (95% CI: 0.428-0.933), NPV 87.5% (95% CI: 0.467-0.993), and AUC 0.800 (95% CI: 0.563-0.943). Conclusion: We developed and preliminarily verified the GIST mitotic count binary prediction model, based on the VGG convolutional neural network. The model displayed a good predictive performance.

• Journal of the Korean Society of Radiology
• /
• v.82 no.3
• /
• pp.670-681
• /
• 2021

Purpose This study aimed to investigate the optimal threshold value in Hounsfield units (HU) on CT to detect the solid components of pulmonary subsolid nodules using pathologic invasive foci as reference. Materials and Methods Thin-section non-enhanced chest CT scans of 25 patients with pathologically confirmed minimally invasive adenocarcinoma were retrospectively reviewed. On CT images, the solid portion was defined as the area with higher attenuation than various HU thresholds ranging from -600 to -100 HU in 50-HU intervals. The solid portion was measured as the largest diameter on axial images and as the maximum diameter on multiplanar reconstruction images. A linear mixed model was used to evaluate bias in each threshold by using the pathological size of invasive foci as reference. Results At a threshold of -400 HU, the biases were lowest between the largest/maximum diameter of the solid portion of subsolid nodule and the size of invasive foci of the pathological specimen, with 0.388 and -0.0176, respectively. They showed insignificant difference (p = 0.2682, p = 0.963, respectively) at a threshold of -400 HU. Conclusion For quantitative analysis, -400 HU may be the optimal threshold to define the solid portion of subsolid nodules as a surrogate marker of invasive foci.

• Journal of the Korean Society of Radiology
• /
• v.84 no.5
• /
• pp.1123-1133
• /
• 2023

Purpose Our study aimed to evaluate the association between automated quantified body composition on CT and pulmonary function or quantitative lung features in patients with chronic obstructive pulmonary disease (COPD). Materials and Methods A total of 290 patients with COPD were enrolled in this study. The volume of muscle and subcutaneous fat, area of muscle and subcutaneous fat at T12, and bone attenuation at T12 were obtained from chest CT using a deep learning-based body segmentation algorithm. Parametric response mapping-derived emphysema (PRM^emph), PRM-derived functional small airway disease (PRM^fSAD), and airway wall thickness (AWT)-Pi10 were quantitatively assessed. The association between body composition and outcomes was evaluated using Pearson's correlation analysis. Results The volume and area of muscle and subcutaneous fat were negatively associated with PRM^emph and PRM^fSAD (p < 0.05). Bone density at T12 was negatively associated with PRM^emph (r = -0.1828, p = 0.002). The volume and area of subcutaneous fat and bone density at T12 were positively correlated with AWT-Pi10 (r = 0.1287, p = 0.030; r = 0.1668, p = 0.005; r = 0.1279, p = 0.031). However, muscle volume was negatively correlated with the AWT-Pi10 (r = -0.1966, p = 0.001). Muscle volume was significantly associated with pulmonary function (p < 0.001). Conclusion Body composition, automatically assessed using chest CT, is associated with the phenotype and severity of COPD.

• Asian Pacific Journal of Cancer Prevention
• /
• v.17 no.8
• /
• pp.4025-4030
• /
• 2016

Background: Correlation between colorectal cancer (CRC) and abdominal obesity has been established, but there is a paucity of data on non-obese CRC patients. The aim of this study was to establish the characteristics of CRCs that occur in such patients. Materials and Methods: Consecutive CRC patients without cachexia were included. Unintended body weight loss, T4- or M1-staged CRCs, extensive lymph node involvement, or synchronous malignancy were classified as cachectic conditions. Abdominal fat volumes were measured using a multidetector CT unit with a software (Rapidia, INFINITT, Seoul, Korea). Results: Of the newly-diagnosed CRC patients, 258 non-cachectic and 88 cachectic patients were analyzed. The cancer size (p<0.001) and T stage (p<0.001) were inversely correlated with body mass index (BMI), visceral fat and subcutaneous fat volumes. Cancer size was the only independent factor related to BMI (p=0.016), visceral fat volume (p=0.002), and subcutaneous fat volume (p=0.027). In non-cachectic patients, a significant inverse correlation was found only between the cancer size and visceral fat volume (p=0.017). Conclusions: Non-obese CRC patients tend to have larger CRC lesions than their obese counterparts even under non-cachectic conditions. Such an inverse correlation between cancer size and visceral fat volume suggests that considerable CRCs are not correlated with abdominal obesity.

• Journal of the Korean Society of Radiology
• /
• v.81 no.1
• /
• pp.213-218
• /
• 2020

Coronary artery aneurysm is a rare disease. It occurs in the epicardial space, mostly along the course of major coronary arteries. Here, we report a case of a giant incidental aneurysm embedded in the basal posterior wall of the left ventricle. A 43-year-old woman was referred to our institution for the evaluation of cardiac palpitations that had been present from the previous 2 months. She reported no medical history (such as Kawasaki's disease or hypertension) or previous operative history. Echocardiogram and subsequent cardiac CT revealed a giant aneurysm in the left ventricle, with a direct fistulous connection to a dilated and tortuous left circumflex artery, which showed direct communication with the straight right coronary artery.

• Journal of the Korean Society of Radiology
• /
• v.84 no.6
• /
• pp.1266-1289
• /
• 2023

Malignant lymphoma typically presents with homogeneous enhancement of enlarged lymph nodes without internal necrotic or cystic changes on multiphasic CT, which can be suspected without invasive diagnostic methods. However, some subtypes of malignant lymphoma show atypical imaging features, which makes diagnosis challenging for radiologists. Moreover, there are several lymphoma-mimicking diseases in current clinical practice, including leukemia, viral infections in immunocompromised patients, and primary or metastatic cancer. The ability of diagnostic processes to distinguish malignant lymphoma from mimicking diseases is necessary to establish effective management strategies for initial radiological examinations. Therefore, this study aimed to discuss the typical and atypical imaging features of malignant lymphoma as well as mimicking diseases and discuss important diagnostic clues that can help narrow down the differential diagnosis.

• Imaging Science in Dentistry
• /
• v.46 no.1
• /
• pp.25-31
• /
• 2016

Purpose: This study was performed to evaluate condylar surface changes and remodeling after orthognathic surgery using three-dimensional computed tomography (3D CT) imaging, including comparisons between the right and left sides and between the sexes. Materials and Methods: Forty patients (20 males and 20 females) who underwent multi-detector CT examinations before and after surgery were selected. Three-dimensional images comprising thousands of points on the condylar surface were obtained before and after surgery. For the quantitative assessment of condylar surface changes, point-to-point (preoperative-to-postoperative) distances were calculated using 3D processing software. These point-to-point distances were converted to a color map. In order to evaluate the types of condylar remodeling, the condylar head was divided into six areas (anteromedial, anteromiddle, anterolateral, posteromedial, posteromiddle, and posterolateral areas) and each area was classified into three types of condylar remodeling (bone formation, no change, and bone resorption) based on the color map. Additionally, comparative analyses were performed between the right and left sides and according to sex. Results: The mean of the average point-to-point distances on condylar surface was $0.11{\pm}0.03mm$. Bone resorption occurred more frequently than other types of condylar remodeling, especially in the lateral areas. However, bone formation in the anteromedial area was particularly prominent. No significant difference was found between the right and left condyles, but condylar surface changes in males were significantly larger than in females. Conclusion: This study revealed that condylar remodeling exhibited a tendency towards bone resorption, especially in the lateral areas. Condylar surface changes occurred, but were small.

• Journal of the Korean Society of Radiology
• /
• v.84 no.4
• /
• pp.911-922
• /
• 2023

Purpose To compare small bowel distension and side effects between a diluted polyethylene glycol (PEG) solution and a low-density (0.1% w/v) barium sulfate suspension (LDBSS) for CT enterography (CTE) preparation. Materials and Methods Total 173 consecutive patients who underwent CTE were enrolled in this study. The LDBSS (1 L) was used in 50 patients, and the diluted iso-osmotic PEG solution (1 L) was used in 123 patients. Two blinded radiologists independently scored jejunal and ileal distensions on a 5-point scale. To compare side effects between the two groups, the patients reported whether they had immediate complications after the administration of the oral contrast media. Results For ileal and jejunal distension, the diluted PEG solution showed no difference from the LDBSS for either reader (ileum: reader 1, median, 4; 4, interquartile range, 3-4; 3-4, p = 0.997; reader 2, median, 4; 4, interquartile range, 3.3-4.0; 3-4, p = 0.064; jejunum: reader 1, median, 2; 2, interquartile range, 2-3; 2-3, p = 0.560; reader 2, median, 3; 2, interquartile range, 2-3; 2-3, p = 0.192). None of the patients complained of immediate complications following administration of either of the oral contrast media. Conclusion The diluted PEG solution showed comparable bowel distension compared to LDBSS and no immediate side effects; thus, it can be a useful alternative.