• Korean Journal of Radiology
• /
• v.22 no.3
• /
• pp.476-488
• /
• 2021

Objective: We aimed to develop a deep neural network for segmenting lung parenchyma with extensive pathological conditions on non-contrast chest computed tomography (CT) images. Materials and Methods: Thin-section non-contrast chest CT images from 203 patients (115 males, 88 females; age range, 31-89 years) between January 2017 and May 2017 were included in the study, of which 150 cases had extensive lung parenchymal disease involving more than 40% of the parenchymal area. Parenchymal diseases included interstitial lung disease (ILD), emphysema, nontuberculous mycobacterial lung disease, tuberculous destroyed lung, pneumonia, lung cancer, and other diseases. Five experienced radiologists manually drew the margin of the lungs, slice by slice, on CT images. The dataset used to develop the network consisted of 157 cases for training, 20 cases for development, and 26 cases for internal validation. Two-dimensional (2D) U-Net and three-dimensional (3D) U-Net models were used for the task. The network was trained to segment the lung parenchyma as a whole and segment the right and left lung separately. The University Hospitals of Geneva ILD dataset, which contained high-resolution CT images of ILD, was used for external validation. Results: The Dice similarity coefficients for internal validation were 99.6 ± 0.3% (2D U-Net whole lung model), 99.5 ± 0.3% (2D U-Net separate lung model), 99.4 ± 0.5% (3D U-Net whole lung model), and 99.4 ± 0.5% (3D U-Net separate lung model). The Dice similarity coefficients for the external validation dataset were 98.4 ± 1.0% (2D U-Net whole lung model) and 98.4 ± 1.0% (2D U-Net separate lung model). In 31 cases, where the extent of ILD was larger than 75% of the lung parenchymal area, the Dice similarity coefficients were 97.9 ± 1.3% (2D U-Net whole lung model) and 98.0 ± 1.2% (2D U-Net separate lung model). Conclusion: The deep neural network achieved excellent performance in automatically delineating the boundaries of lung parenchyma with extensive pathological conditions on non-contrast chest CT images.

• Korean Journal of Radiology
• /
• v.22 no.6
• /
• pp.867-879
• /
• 2021

Objective: To compare the screening performance of diffusion-weighted (DW) MRI and combined mammography and ultrasound (US) in detecting clinically occult contralateral breast cancer in women with newly diagnosed breast cancer. Materials and Methods: Between January 2017 and July 2018, 1148 women (mean age ± standard deviation, 53.2 ± 10.8 years) with unilateral breast cancer and no clinical abnormalities in the contralateral breast underwent 3T MRI, digital mammography, and radiologist-performed whole-breast US. In this retrospective study, three radiologists independently and blindly reviewed all DW MR images (b = 1000 s/mm² and apparent diffusion coefficient map) of the contralateral breast and assigned a Breast Imaging Reporting and Data System category. For combined mammography and US evaluation, prospectively assessed results were used. Using histopathology or 1-year follow-up as the reference standard, cancer detection rate and the patient percentage with cancers detected among all women recommended for tissue diagnosis (positive predictive value; PPV2) were compared. Results: Of the 30 cases of clinically occult contralateral cancers (13 invasive and 17 ductal carcinoma in situ [DCIS]), DW MRI detected 23 (76.7%) cases (11 invasive and 12 DCIS), whereas combined mammography and US detected 12 (40.0%, five invasive and seven DCIS) cases. All cancers detected by combined mammography and US, except two DCIS cases, were detected by DW MRI. The cancer detection rate of DW MRI (2.0%; 95% confidence interval [CI]: 1.3%, 3.0%) was higher than that of combined mammography and US (1.0%; 95% CI: 0.5%, 1.8%; p = 0.009). DW MRI showed higher PPV2 (42.1%; 95% CI: 26.3%, 59.2%) than combined mammography and US (18.5%; 95% CI: 9.9%, 30.0%; p = 0.001). Conclusion: In women with newly diagnosed breast cancer, DW MRI detected significantly more contralateral breast cancers with fewer biopsy recommendations than combined mammography and US.

• Korean Journal of Radiology
• /
• v.22 no.5
• /
• pp.697-705
• /
• 2021

Objective: To investigate the feasibility and the accuracy of the coronary CT angiography (CCTA)-derived Registry of Crossboss and Hybrid procedures in France, the Netherlands, Belgium and United Kingdom (RECHARGE) score (RECHARGECCTA) for the prediction of procedural success and 30-minutes guidewire crossing in percutaneous coronary intervention (PCI) for chronic total occlusion (CTO). Materials and Methods: One hundred and twenty-four consecutive patients (mean age, 54 years; 79% male) with 131 CTO lesions who underwent CCTA before catheter angiography (CA) with CTO-PCI were retrospectively enrolled in this study. The RECHARGECCTA scores were calculated and compared with RECHARGECA and other CTA-based prediction scores, including Multicenter CTO Registry of Japan (J-CTO), CT Registry of CTO Revascularisation (CT-RECTOR), and Korean Multicenter CTO CT Registry (KCCT) scores. Results: The procedural success rate of the CTO-PCI procedures was 72%, and 61% of cases achieved the 30-minutes wire crossing. No significant difference was observed between the RECHARGECCTA score and the RECHARGECA score for procedural success (median 2 vs. median 2, p = 0.084). However, the RECHARGECCTA score was higher than the RECHARGECA score for the 30-minutes wire crossing (median 2 vs. median 1.5, p = 0.001). The areas under the curve (AUCs) of the RECHARGECCTA and RECHARGECA scores for predicting procedural success showed no statistical significance (0.718 vs. 0.757, p = 0.655). The sensitivity, specificity, positive predictive value, and the negative predictive value of the RECHARGECCTA scores of ≤ 2 for predictive procedural success were 78%, 60%, 43%, and 87%, respectively. The RECHARGECCTA score showed a discriminative performance that was comparable to those of the other CTA-based prediction scores (AUC = 0.718 vs. 0.665-0.717, all p > 0.05). Conclusion: The non-invasive RECHARGECCTA score performs better than the invasive determination for the prediction of the 30-minutes wire crossing of CTO-PCI. However, the RECHARGECCTA score may not replace other CTA-based prediction scores for predicting CTO-PCI success.

• Korean Journal of Radiology
• /
• v.22 no.5
• /
• pp.706-713
• /
• 2021

Objective: To evaluate the impact of surgical simulation training using a three-dimensional (3D)-printed model of tetralogy of Fallot (TOF) on surgical skill development. Materials and Methods: A life-size congenital heart disease model was printed using a Stratasys Object500 Connex2 printer from preoperative electrocardiography-gated CT scans of a 6-month-old patient with TOF with complex pulmonary stenosis. Eleven cardiothoracic surgeons independently evaluated the suitability of four 3D-printed models using composite Tango 27, 40, 50, and 60 in terms of palpation, resistance, extensibility, gap, cut-through ability, and reusability of. Among these, Tango 27 was selected as the final model. Six attendees (two junior cardiothoracic surgery residents, two senior residents, and two clinical fellows) independently performed simulation surgeries three times each. Surgical proficiency was evaluated by an experienced cardiothoracic surgeon on a 1-10 scale for each of the 10 surgical procedures. The times required for each surgical procedure were also measured. Results: In the simulation surgeries, six surgeons required a median of 34.4 (range 32.5-43.5) and 21.4 (17.9-192.7) minutes to apply the ventricular septal defect (VSD) and right ventricular outflow tract (RVOT) patches, respectively, on their first simulation surgery. These times had significantly reduced to 17.3 (16.2-29.5) and 13.6 (10.3-30.0) minutes, respectively, in the third simulation surgery (p = 0.03 and p = 0.01, respectively). The decreases in the median patch appliance time among the six surgeons were 16.2 (range 13.6-17.7) and 8.0 (1.8-170.3) minutes for the VSD and RVOT patches, respectively. Summing the scores for the 10 procedures showed that the attendees scored an average of 28.58 ± 7.89 points on the first simulation surgery and improved their average score to 67.33 ± 15.10 on the third simulation surgery (p = 0.008). Conclusion: Inexperienced cardiothoracic surgeons improved their performance in terms of surgical proficiency and operation time during the experience of three simulation surgeries using a 3D-printed TOF model using Tango 27 composite.

• Journal of the Korean Geosynthetics Society
• /
• v.23 no.2
• /
• pp.43-52
• /
• 2024

Recent abnormal climate conditions have increased the risk of slope collapses, which frequently result in significant loss of life and property due to the absence of early prediction and warning dissemination. In this paper, we develop a slope condition analysis system using IoT sensors and AI-based camera to assess the condition of slopes. To develop the system, we conducted hardware and firmware design for measurement sensors considering the ground conditions of slopes, designed AI-based image analysis algorithms, and developed prediction and warning solutions and systems. We aimed to minimize errors in sensor data through the integration of IoT sensor data and AI camera image analysis, ultimately enhancing the reliability of the data. Additionally, we evaluated the accuracy (reliability) by applying it to actual slopes. As a result, sensor measurement errors were maintained within 0.1°, and the data transmission rate exceeded 95%. Moreover, the AI-based image analysis system demonstrated nighttime partial recognition rates of over 99%, indicating excellent performance even in low-light conditions. Through this research, it is anticipated that the analysis of slope conditions and smart maintenance management in various fields of Social Overhead Capital (SOC) facilities can be applied.

• Korean Journal of Radiology
• /
• v.22 no.11
• /
• pp.1764-1776
• /
• 2021

Objective: This study aimed to validate a deep learning-based fully automatic calcium scoring (coronary artery calcium [CAC]_auto) system using previously published cardiac computed tomography (CT) cohort data with the manually segmented coronary calcium scoring (CAC_hand) system as the reference standard. Materials and Methods: We developed the CAC_auto system using 100 co-registered, non-enhanced and contrast-enhanced CT scans. For the validation of the CAC_auto system, three previously published CT cohorts (n = 2985) were chosen to represent different clinical scenarios (i.e., 2647 asymptomatic, 220 symptomatic, 118 valve disease) and four CT models. The performance of the CAC_auto system in detecting coronary calcium was determined. The reliability of the system in measuring the Agatston score as compared with CAC_hand was also evaluated per vessel and per patient using intraclass correlation coefficients (ICCs) and Bland-Altman analysis. The agreement between CAC_auto and CAC_hand based on the cardiovascular risk stratification categories (Agatston score: 0, 1-10, 11-100, 101-400, > 400) was evaluated. Results: In 2985 patients, 6218 coronary calcium lesions were identified using CAC_hand. The per-lesion sensitivity and false-positive rate of the CAC_auto system in detecting coronary calcium were 93.3% (5800 of 6218) and 0.11 false-positive lesions per patient, respectively. The CAC_auto system, in measuring the Agatston score, yielded ICCs of 0.99 for all the vessels (left main 0.91, left anterior descending 0.99, left circumflex 0.96, right coronary 0.99). The limits of agreement between CAC_auto and CAC_hand were 1.6 ± 52.2. The linearly weighted kappa value for the Agatston score categorization was 0.94. The main causes of false-positive results were image noise (29.1%, 97/333 lesions), aortic wall calcification (25.5%, 85/333 lesions), and pericardial calcification (24.3%, 81/333 lesions). Conclusion: The atlas-based CAC_auto empowered by deep learning provided accurate calcium score measurement as compared with manual method and risk category classification, which could potentially streamline CAC imaging workflows.

• Journal of the Korean Geotechnical Society
• /
• v.40 no.3
• /
• pp.93-100
• /
• 2024

Pore water freezes in low-temperature compaction, which leads to different compaction characteristics compared to room temperature conditions. In regions like Alberta, Canada, where organic soils are prevalent, compaction performance is influenced by the high water retention and compressibility of organic soils, as well as their sensitivity to freezing and thawing. Alberta's strict environmental regulations demand the reuse of excavated soil for backfill, and the long winter season creates challenging conditions for civil engineering projects. In this study, a laboratory compaction test was conducted to evaluate the low-temperature compaction characteristics of organic soils with varying organic content. The results indicate that the optimum moisture content increases as the organic content increases, and the maximum dry unit weight decreases by up to 21.9%. In addition, under temperature conditions below -4℃, no optimum moisture content was observed, and the dry unit weight decreased as the moisture content increased.

• Korean Journal of Radiology
• /
• v.22 no.2
• /
• pp.168-178
• /
• 2021

Objective: To provide an automatic method for segmentation and diameter measurement of type B aortic dissection (TBAD). Materials and Methods: Aortic computed tomography angiographic images from 139 patients with TBAD were consecutively collected. We implemented a deep learning method based on a three-dimensional (3D) deep convolutional neural (CNN) network, which realizes automatic segmentation and measurement of the entire aorta (EA), true lumen (TL), and false lumen (FL). The accuracy, stability, and measurement time were compared between deep learning and manual methods. The intra- and inter-observer reproducibility of the manual method was also evaluated. Results: The mean dice coefficient scores were 0.958, 0.961, and 0.932 for EA, TL, and FL, respectively. There was a linear relationship between the reference standard and measurement by the manual and deep learning method (r = 0.964 and 0.991, respectively). The average measurement error of the deep learning method was less than that of the manual method (EA, 1.64% vs. 4.13%; TL, 2.46% vs. 11.67%; FL, 2.50% vs. 8.02%). Bland-Altman plots revealed that the deviations of the diameters between the deep learning method and the reference standard were -0.042 mm (-3.412 to 3.330 mm), -0.376 mm (-3.328 to 2.577 mm), and 0.026 mm (-3.040 to 3.092 mm) for EA, TL, and FL, respectively. For the manual method, the corresponding deviations were -0.166 mm (-1.419 to 1.086 mm), -0.050 mm (-0.970 to 1.070 mm), and -0.085 mm (-1.010 to 0.084 mm). Intra- and inter-observer differences were found in measurements with the manual method, but not with the deep learning method. The measurement time with the deep learning method was markedly shorter than with the manual method (21.7 ± 1.1 vs. 82.5 ± 16.1 minutes, p < 0.001). Conclusion: The performance of efficient segmentation and diameter measurement of TBADs based on the 3D deep CNN was both accurate and stable. This method is promising for evaluating aortic morphology automatically and alleviating the workload of radiologists in the near future.

• 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.

• Advanced Industrial SCIence
• /
• v.3 no.2
• /
• pp.8-16
• /
• 2024

This paper shows the system of drug classification, the goal of this is to foretell the apt drug for the patients based on their demographic and physiological traits. The dataset consists of various attributes like Age, Sex, BP (Blood Pressure), Cholesterol Level, and Na_to_K (Sodium to Potassium ratio), with the objective to determine the kind of drug being given. The models used in this paper are K-Nearest Neighbors (KNN), Logistic Regression and Random Forest. Further to fine-tune hyper parameters using 5-fold cross-validation, GridSearchCV was used and each model was trained and tested on the dataset. To assess the performance of each model both with and without hyper parameter tuning evaluation metrics like accuracy, confusion matrices, and classification reports were used and the accuracy of the models without GridSearchCV was 0.7, 0.875, 0.975 and with GridSearchCV was 0.75, 1.0, 0.975. According to GridSearchCV Logistic Regression is the most suitable model for drug classification among the three-model used followed by the K-Nearest Neighbors. Also, Na_to_K is an essential feature in predicting the outcome.