• 한국방사선학회논문지
• /
• 제16권1호
• /
• pp.25-34
• /
• 2022

뇌 전산화단층촬영은 비침습성, 3차원 영상 제공, 저방사선량 등의 장점 때문에 뇌출혈과 같은 질병 진단을 위해 시행된다. 하지만 뇌 전산화단층영상 판독을 위한 전문의의 인력 공급 부족 및 막대한 업무량으로 인해 수많은 판독 오류 및 오진이 발생하고 있다. 이와 같은 문제를 해결하기 위해 객체 검출을 위한 다양한 인공지능 기술이 개발되고 있다. 본 연구에서는 뇌 전산화단층영상으로부터 뇌출혈 검출을 위한 딥러닝 기반 YOLOv5s 모델의 적용 가능성을 확인하였다. 또한 YOLOv5s 모델 학습 시 초매개변수를 변화시켜 학습된 모델의 성능을 평가하였다. YOLOv5s 모델은 backbone, neck 및 output 모듈로 구성하였고, 입력 CT 영상 내 뇌출혈로 의심되는 부위를 검출하여 출력할 수 있도록 하였다. YOLOv5s 모델 학습 시 활성화함수, 최적화함수, 손실함수 및 학습 횟수를 변화시켰고, 학습된 모델의 뇌출혈 검출 정확도 및 학습 시간을 측정하였다. 연구결과 학습된 YOLOv5s 모델은 뇌출혈로 의심되는 부위에 대한 경계 박스 및 해당 경계박스에 대한 정확도를 출력할 수 있음을 확인하였다. Mish 활성화함수, stochastic gradient descent 최적화함수 및 completed intersection over union 손실함수 적용 시 YOLOv5s 모델의 뇌출혈 검출 정확도 향상 및 학습 시간이 단축되는 결과를 확인하였다. 또한 YOLOv5s 모델의 뇌출혈 검출 정확도 및 학습 시간은 학습 횟수에 비례하여 증가하는 결과를 확인하였다. 따라서 YOLOv5s 모델은 뇌 전산화단층영상을 이용한 뇌출혈 검출을 위해 활용할 수 있으며, 최적의 초매개변수 적용을 통해 성능을 향상 시킬 수 있다.

• Journal of Radiation Protection and Research
• /
• 제38권4호
• /
• pp.214-227
• /
• 2013

MDCT의 등장과 함께 병적 증상이 없는 개인이 조기검진을 위해 CT검사를 하는 예가 꾸준히 증가하고 있다. 이러한 검사의 임상적 효용성과 방사선에 의한 암 발생위험에 대한 여러 가지 논란을 살펴보았다. 1. 저선량흉부CT(LDCT)를 이용한 폐암의 조기 검진 효용성을 보기 위해 설계된 무작위배정대조검사(randomized controlled trial)인 NLST(National Lung Screening Trial)의 분석 결과, CT선별검사가 선별검사를 하지 않은 그룹의 사망률 1.67%를 1.33%로, 20% (p=0.004) 감소시키는 것으로 나타났다. CT선별검사군에서 양성결과를 보인 사람의 96.4%는 위양성(false-positive)이었다. 그리고 방사선에 의한 폐암의 ERR (excess radiation risk)은 다른 고형암과는 달리 나이와 함께 감소하지 않으며, 더욱이 흡연 위험과 방사선 위험이 상승작용을 하여 폐암 검진 대상인 고위험군에게 더욱 위험할 수 있는 것으로 평가되었다. 따라서 방사선에 의한 위험이 NLST에서 관측한 이득을 상쇄할 수도 있다. 아직은 최적의 전략이 확립되지 않은 상태이므로 임상에서 광범위하게 적용하기에는 한계가 있다. 2. 대장CT검사(CT colonography)는 10 mm보다 큰 용종의 검출에는 정확도가 뛰어나지만 이보다 작은 경우에는 민감도와 특이도가 급격히 감소하고, 용종이 발견되어 제거하고자 하는 경우 다시 기존의 대장내시경을 이용해야 하는 불편함이 있다. 현재 대장CT검사 1회 시행의 평균유효선량은 약 8-10 $mS{\nu}$로 보고되어, 만약 이 검사가 많은 사람들을 대상으로 일상적으로 행해지면 공중보건에 영향을 미칠 수 있다. 따라서 집단검진 방법으로 정착하기 위해서 는 용종 크기 6-10 mm 범위에서 높은 민감도와 특이도를 확립하고, 하제를 사용하지 않는 방법을 개발하여야 하며, 선량을 더 감소시킬 수 있는 최적화된 전략을 세워 표준화하여야만 한다. 3. 전신CT검사는 그 효용성이 검증된 선행연구가 없는 상태이다. 이 검사는 약 90%의 검사자들에게서 최소한 하나 이상의 이상소견을 확인하였지만 그 중 약 2%만이 임상적으로 중요한 질환으로 판명되었다. 전신CT검사 1회 시행으로 인한 평균유효선량은 약 12 $mS{\nu}$이며 45세부터 75세까지 매년 검사를 받는 경우 LAR (lifetime attributable risk)이 1.9%로 보고되었다. 현재 의학계의 어떤 단체도 이 검사를 권고하고 있지 않다. 현재까지 CT선별검사는 정당화에 대한 충분한 논의를 위해 심도 있는 다양한 연구가 필요하며 최적화된 프로토콜이 확립된 후 이용하는 것이 바람직하다.

• Journal of Microbiology and Biotechnology
• /
• 제31권10호
• /
• pp.1401-1408
• /
• 2021

This study examined whether the oral administration of low-molecular-weight collagen peptide (LMCP) containing 3% Gly-Pro-Hyp with >15% tripeptide (Gly-X-Y) content could ameliorate osteoarthritis (OA) progression using a rabbit anterior cruciate ligament transection (ACLT) model of induced OA and chondrocytes isolated from a patient with OA. Oral LMCP administration (100 or 200 mg/kg/day) for 12 weeks ameliorated cartilage damage and reduced the loss of proteoglycan compared to the findings in the ACLT control group, resulting in dose-dependent (p < 0.05) improvements of the OARSI score in hematoxylin & eosin (H&E) and Safranin O staining. In micro-computed tomography analysis, LMCP also significantly (p < 0.05) suppressed the deterioration of the microstructure in tibial subchondral bone during OA progression. The elevation of IL-1β and IL-6 concentrations in synovial fluid following OA induction was dose-dependently (p < 0.05) reduced by LMCP treatment. Furthermore, immunohistochemistry illustrated that LMCP significantly (p < 0.05) upregulated type II collagen and downregulated matrix metalloproteinase-13 in cartilage tissue. Consistent with the in vivo results, LMCP significantly (p < 0.05) increased the mRNA expression of COL2A1 and ACAN in chondrocytes isolated from a patient with OA regardless of the conditions for IL-1β induction. These findings suggest that LMCP has potential as a therapeutic treatment for OA that stimulates cartilage regeneration.

• 한국의학물리학회지:의학물리
• /
• 제24권4호
• /
• pp.278-283
• /
• 2013

방사선 치료를 진행한 오른편 유방암 환자에 비해 왼편 유방암 환자의 방사선 치료 시 왼편 유방의 인접 주요 장기인 심장에 전달되는 선량에 의한 심장 질환의 발병 및 기타 질환의 발병으로 인한 높은 치사율과 관련된 예후가 보고되고 있다. 방사선 치료에서 computed tomography (CT) 영상을 획득 하는 방법 중 deep inspiration breath hold (DIBH) 기법은 들숨 상태에서 일정 시간 동안 환자의 호흡을 정지시키고 영상을 획득 하는 방법으로 심장과 흉곽 사이의 거리가 최대가 되게 한다. 따라서 본 연구에서는 DIBH 영상 획득 기법을 활용하여 왼편 유방암 환자의 방사선 치료 시 DIBH 기법을 적용한 CT 영상을 토대로 심장과 왼편 유방까지의 거리 계산 및 심장에 전달되는 피해 선량을 정량화 함으로써 왼편 유방암 환자의 방사선 치료 시 DIBH 기법의 유용성을 평가하고자 하였다. Free breathing (FB)와 DIBH 기법을 적용한 여성 유방암 환자의 CT 영상을 각 10세트를 획득하고, 50 Gy를 28번으로 분할하여 처방하였으며, 쐐기 필터(wedge filter)를 이용한 대향 2문 접선 조사를 적용했다. 심장과 왼편 유방까지의 거리는 각 장기의 중심 좌표를 획득하고, 각 중심좌표 간의 거리를 계산하였다. DIBH 기법의 경우, 일반적인 FB 기법을 적용 했을 때보다 심장과 왼편 유방 사이의 거리가 평균 1.43 mm 증가하였으며, 통계적 유효성을 확인할 수 있었다. 심장에 전달된 피해 선량의 경우, 최대 선량 기준으로 크게는 3,555 cGy 가량 감소함을 확인할 수 있었다. 각 영상마다의 거리 및 심장의 피해 선량에 대한 정도의 차이는 있었지만, DIBH 기법을 적용하였을 경우, 심장과 왼편 유방까지의 거리의 증가 및 피해 선량 감소 등의 경향성을 확인할 수 있었다. 본 DIBH 기법은 기존 방사선 치료 과정 중 추가적인 시간 소모가 적고, 쉽게 적용할 수 있다는 장점이 있으며, 임상에서의 적용으로 여성 유방암 환자의 불필요한 심장 피해 선량 전달을 감소 시킬 수 있을 것으로 사료된다.

• Korean Journal of Radiology
• /
• 제20권5호
• /
• pp.709-718
• /
• 2019

Objective: To investigate the association of myocardial blood flow (MBF) quantified by dynamic computed tomography (CT) myocardial perfusion imaging (MPI) with troponin level and left ventricle (LV) function in patients with ST-segment elevated myocardial infarction (STEMI). Materials and Methods: Thirty-five STEMI patients who successfully had undergone reperfusion treatment within 1 week of their infarction were consecutively enrolled. All patients were referred for dynamic CT-MPI. Serial high-sensitivity troponin T (hs-TnT) levels and left ventricular ejection fraction (LVEF) measured by echocardiography were recorded. Twenty-six patients with 427 segments were included for analysis. Various quantitative parameters derived from dynamic CT-MPI were analyzed to determine if there was a correlation between hs-TnT levels and LVEF on admission and again at the 6-month mark. Results: The mean radiation dose for dynamic CT-MPI was 3.2 ± 1.1 mSv. Infarcted territories had significantly lower MBF (30.5 ± 7.4 mL/min/100 mL versus 73.4 ± 8.1 mL/min/100 mL, p < 0.001) and myocardial blood volume (MBV) (2.8 ± 0.9 mL/100 mL versus 4.2 ± 1.1 mL/100 mL, p = 0.044) compared with those of reference territories. MBF showed the best correlation with the level of peak hs-TnT (r = -0.682, p < 0.001), and MBV showed a moderate correlation with the level of peak hs-TnT (r = -0.437, p = 0.026); however, the other parameters did not show any significant correlation with hs-TnT levels. As for the association with LV function, only MBF was significantly correlated with LVEF at the time of admission (r = 0.469, p = 0.016) and at 6 months (r = 0.585, p = 0.001). Conclusion: MBF quantified by dynamic CT-MPI is significantly inversely correlated with the level of peak hs-TnT. In addition, patients with lower MBF tended to have impaired LV function at the time of their admission and at 6 months.

• Journal of Radiation Protection and Research
• /
• 제43권1호
• /
• pp.1-9
• /
• 2018

Background: In this study, we investigate the image quality of virtual monochromatic images synthesized from dual-energy computed tomography (DECT) at voltages of 80/140 kV and 100/140 kV. Materials and Methods: Virtual monochromatic images of a phantom are synthesized from DECT scans from 40 to 70 keV in steps of 1 keV under the two combinations of tube voltages. The dose allocation of dual-energy (DE) scan is 50% for both low- and high-energy tubes. The virtual monochromatic images are compared to single-energy (SE) images at the same radiation dose. In the DE images, noise is reduced using the 100/140 kV scan at the optimal monochromatic energy. Virtual monochromatic images are reconstructed from 40 to 70 keV in 1-keV increments and analyzed using two quality indexes: noise and contrast-to-noise ratio (CNR). Results and Discussion: The DE scan mode with the 100/140 kV protocol achieved a better maximum CNR compared to the 80/140 kV protocol for various materials, except for adipose and brain. Image noise is reduced with the 100/140 kV protocol. The CNR values of DE with the 100/140 kV protocol is similar to or higher than that of SE at 120 kV at the same radiation dose. Furthermore, the maximum CNR with the 100/140 kV protocol is similar to or higher than that of the SE scan at 120 kV. Conclusion: It was found that the CNR achieved with the 100/140 kV protocol was better than that with the 80/140 kV protocol at optimal monochromatic energies. Virtual monochromatic imaging using the 100/140 kV protocol could be considered for application in breast, brain, lung, liver, and bone CT in accordance with the CNR results.

• Korean Journal of Radiology
• /
• 제21권8호
• /
• pp.967-977
• /
• 2020

Objective: To evaluate the effects of tube voltage on image quality in coronary CT angiography (CCTA), the estimated radiation dose, and DNA double-strand breaks (DSBs) in peripheral blood lymphocytes to optimize the use of CCTA in the era of low radiation doses. Materials and Methods: This study included 240 patients who were divided into 2 groups according to the DNA DSB analysis methods, i.e., immunofluorescence microscopy and flow cytometry. Each group was subdivided into 4 subgroups: those receiving CCTA only with different tube voltages of 120, 100, 80, or 70 kVp. Objective and subjective image quality was evaluated by analysis of variance. Radiation dosages were also recorded and compared. Results: There was no significant difference in demographic characteristics between the 2 groups and 4 subgroups in each group (all p > 0.05). As tube voltage decreased, both image quality and radiation dose decreased gradually and significantly. After CCTA, γ-H2AX foci and mean fluorescence intensity in the 120-, 100-, 80-, and 70-kVp groups increased by 0.14, 0.09, 0.07, and 0.06 foci per cell and 21.26, 9.13, 8.10, and 7.13 (all p < 0.05), respectively. The increase in the DNA DSB level in the 120-kVp group was higher than those in the other 3 groups (all p < 0.05), while there was no significant difference in the DSBs levels among these latter groups (all p > 0.05). Conclusion: The 100-kVp tube voltage may be optimal for CCTA when weighing DNA DSBs against the estimated radiation dose and image quality, with further reductions in tube voltage being unnecessary for CCTA.

• 한국멀티미디어학회논문지
• /
• 제23권6호
• /
• pp.729-737
• /
• 2020

Screening with low-dose spiral computed tomography (LDCT) has been shown to reduce lung cancer mortality by about 20% when compared to standard chest radiography. One of the problems arising from screening programs is that large amounts of CT image data must be interpreted by radiologists. To solve this problem, automated detection of pulmonary nodules is necessary; however, this is a challenging task because of the high number of false positive results. Here we demonstrate detection of pulmonary nodules using six off-the-shelf convolutional neural network (CNN) models after modification of the input/output layers and end-to-end training based on publicly databases for comparative evaluation. We used the well-known CNN models, LeNet-5, VGG-16, GoogLeNet Inception V3, ResNet-152, DensNet-201, and NASNet. Most of the CNN models provided superior results to those of obtained using customized CNN models. It is more desirable to modify the proven off-the-shelf network model than to customize the network model to detect the pulmonary nodules.

• Tuberculosis and Respiratory Diseases
• /
• 제84권2호
• /
• pp.89-95
• /
• 2021

Lung cancer is one of the leading causes of cancer-related deaths in Korea. Although the smoking rate has decreased over time, the prevalence of lung cancer still remains high. In this study, we reviewed recent trends on the incidence, epidemiology, screening, diagnosis, and treatment of lung cancer in Korea by analyzing data from the national lung cancer registry and recently-published studies. Although approximately 40% of patients with non-small cell lung cancer (NSCLC) were diagnosed as stage IV, the 5-year relative survival rate improved from 11.3% (1993-1995) to 30.2% (2013-2017), possibly due to advances in methods of diagnosis and therapy. In addition, the 2019 implementation of the national lung cancer screening program with low-dose computed tomography may have also contributed to these improvements in survival rates. Recently, molecular diagnosis has become more widely used in the identification of genetic mutations in tissue specimens. Target therapy and immune checkpoint inhibitors have also been successfully used, particularly in cases of advanced NSCLC. In the future, further research on the optimal management of lung cancer remains necessary.

• Korean Journal of Radiology
• /
• 제22권11호
• /
• pp.1850-1857
• /
• 2021

Objective: The purpose of this study was to assess whether a deep learning (DL) algorithm could enable simultaneous noise reduction and edge sharpening in low-dose lumbar spine CT. Materials and Methods: This retrospective study included 52 patients (26 male and 26 female; median age, 60.5 years) who had undergone CT-guided lumbar bone biopsy between October 2015 and April 2020. Initial 100-mAs survey images and 50-mAs intraprocedural images were reconstructed by filtered back projection. Denoising was performed using a vendor-agnostic DL model (ClariCT.AI^TM, ClariPI) for the 50-mAS images, and the 50-mAs, denoised 50-mAs, and 100-mAs CT images were compared. Noise, signal-to-noise ratio (SNR), and edge rise distance (ERD) for image sharpness were measured. The data were summarized as the mean ± standard deviation for these parameters. Two musculoskeletal radiologists assessed the visibility of the normal anatomical structures. Results: Noise was lower in the denoised 50-mAs images (36.38 ± 7.03 Hounsfield unit [HU]) than the 50-mAs (93.33 ± 25.36 HU) and 100-mAs (63.33 ± 16.09 HU) images (p < 0.001). The SNRs for the images in descending order were as follows: denoised 50-mAs (1.46 ± 0.54), 100-mAs (0.99 ± 0.34), and 50-mAs (0.58 ± 0.18) images (p < 0.001). The denoised 50-mAs images had better edge sharpness than the 100-mAs images at the vertebral body (ERD; 0.94 ± 0.2 mm vs. 1.05 ± 0.24 mm, p = 0.036) and the psoas (ERD; 0.42 ± 0.09 mm vs. 0.50 ± 0.12 mm, p = 0.002). The denoised 50-mAs images significantly improved the visualization of the normal anatomical structures (p < 0.001). Conclusion: DL-based reconstruction may enable simultaneous noise reduction and improvement in image quality with the preservation of edge sharpness on low-dose lumbar spine CT. Investigations on further radiation dose reduction and the clinical applicability of this technique are warranted.