• The Korean Journal of Nuclear Medicine Technology
• /
• 제13권1호
• /
• pp.82-85
• /
• 2009

Purpose: The purpose of this study was to evaluate the effect of extended CT field of view (FOV) on PET/CT of Standardized uptake value (SUV) when imaging extends beyond the CT FOV. Materials and Methods: CT images were reconstructed at different FOV sizes (500 and 700 mm). Two sets of CT images were reconstructed from the CT projection data by using two FOV sizes. Twenty patients were used in this study. PET images were reconstructed using attenuation maps with 500 mm CT FOV and 700 mm extended CT FOV images. Region of interests (ROIs) drawn on the PET images. In addition, twenty patients' PET images reconstructed by 500 mm CT FOV and 700 mm extended CT FOV were compared with $SUV_{max}$. Results: When using attenuation maps with 700 mm extended CT FOV, the $SUV_{max}$ analysis of liver (p=0.000), lung (p=0.007), mediastinum (p=0.001) were statistically significant. Conclusions: 700 mm extended CT FOV helps to recover the true activity distribution in the PET emission data. In addition, 700 mm extended CT FOV has affected SUV measurement of liver, lung, mediastinum.

• Journal of Korea Multimedia Society
• /
• 제25권5호
• /
• pp.757-768
• /
• 2022

In this paper, we proposed a system that visualizes a hologram device in 3D by utilizing the CT image segmentation function based on artificial intelligence deep learning. The input axial CT medical image is converted into Sagittal and Coronal, and the input image and the converted image are divided into 3D volumes using ResUNet, a deep learning model. In addition, the volume is created by segmenting the tumor region in the segmented liver image. Each result is integrated into one 3D volume, displayed in a medical image viewer, and converted into a video. When the converted video is transmitted to the hologram device and output from the device, a 3D image with a sense of space can be checked. As for the performance of the deep learning model, in Axial, the basic input image, DSC showed 95.0% performance in liver region segmentation and 67.5% in liver tumor region segmentation. If the system is applied to a real-world care environment, additional physical contact is not required, making it safer for patients to explain changes before and after surgery more easily. In addition, it will provide medical staff with information on liver and liver tumors necessary for treatment or surgery in a three-dimensional manner, and help patients manage them after surgery by comparing and observing the liver before and after liver resection.

• Clinical and Experimental Pediatrics
• /
• 제48권3호
• /
• pp.276-283
• /
• 2005

Purpose : Visceral fat accumulation plays a major role in metabolic complications of obesity. It is known that nonalcoholic fatty liver in obese adults is associated with visceral fat accumulation. Body mass index(BMI) is used as the index of obesity in children. The aim of this study is to evaluate the correlation of BMI and visceral adipose tissue(VAT), and the correlation of BMI, body fat distribution, aminotransferases, and severity of fatty liver. Methods : Twenty three obese children with fatty liver diagnosed by non-contrast abdominal computed tomography(CT) were included in this study. Data on BMI, aminotransferase levels were collected from clinical records. Visceral adipose area was evaluated with CT. Results : BMI had a singnificant correlation with VAT(r=0.51719, P=0.0115). The severity of fatty liver had no significant correlations with BMI(r=-0.11938, P=0.5876), VAT(r=-0.31234, P=0.1468), aspartate aminotransferase(AST)(r=0.12729, P=0.5628) or alanine aminotransferase(ALT)(r=-0.00179, P=0.9935). Conclusion : BMI in obese children was correlated with VAT. But the severity of fatty liver cannot be assessed by BMI, VAT or aminotransferase levels.

• Korean Journal of Radiology
• /
• 제24권4호
• /
• pp.294-304
• /
• 2023

Objective: We aimed to investigate whether image standardization using deep learning-based computed tomography (CT) image conversion would improve the performance of deep learning-based automated hepatic segmentation across various reconstruction methods. Materials and Methods: We collected contrast-enhanced dual-energy CT of the abdomen that was obtained using various reconstruction methods, including filtered back projection, iterative reconstruction, optimum contrast, and monoenergetic images with 40, 60, and 80 keV. A deep learning based image conversion algorithm was developed to standardize the CT images using 142 CT examinations (128 for training and 14 for tuning). A separate set of 43 CT examinations from 42 patients (mean age, 10.1 years) was used as the test data. A commercial software program (MEDIP PRO v2.0.0.0, MEDICALIP Co. Ltd.) based on 2D U-NET was used to create liver segmentation masks with liver volume. The original 80 keV images were used as the ground truth. We used the paired t-test to compare the segmentation performance in the Dice similarity coefficient (DSC) and difference ratio of the liver volume relative to the ground truth volume before and after image standardization. The concordance correlation coefficient (CCC) was used to assess the agreement between the segmented liver volume and ground-truth volume. Results: The original CT images showed variable and poor segmentation performances. The standardized images achieved significantly higher DSCs for liver segmentation than the original images (DSC [original, 5.40%-91.27%] vs. [standardized, 93.16%-96.74%], all P < 0.001). The difference ratio of liver volume also decreased significantly after image conversion (original, 9.84%-91.37% vs. standardized, 1.99%-4.41%). In all protocols, CCCs improved after image conversion (original, -0.006-0.964 vs. standardized, 0.990-0.998). Conclusion: Deep learning-based CT image standardization can improve the performance of automated hepatic segmentation using CT images reconstructed using various methods. Deep learning-based CT image conversion may have the potential to improve the generalizability of the segmentation network.

• Journal of Trauma and Injury
• /
• 제36권3호
• /
• pp.165-171
• /
• 2023

Trauma remains a significant healthcare burden, causing over five million yearly fatalities. Notably, the liver is a frequently injured solid organ in abdominal trauma, especially in patients under 40 years. It becomes even more critical given that uncontrolled hemorrhage linked to liver trauma can have mortality rates ranging from 10% to 50%. Liver injuries, mainly resulting from blunt trauma such as motor vehicle accidents, are traditionally classified using the American Association for the Surgery of Trauma grading scale. However, recent developments have introduced the World Society of Emergency Surgery classification, which considers the patient's physiological status. The diagnostic approach often involves multiphase computed tomography (CT). Still, newer methods like split-bolus single-pass CT and contrast-enhanced ultrasound (CEUS) aim to reduce radiation exposure. Concerning management, nonoperative strategies have emerged as the gold standard, especially for hemodynamically stable patients. Incorporating angiography with embolization has also been beneficial, with success rates reported between 80% and 97%. However, it is essential to identify the specific source of bleeding for effective embolization. Given the severity of liver trauma and its potential complications, innovations in diagnostic and therapeutic approaches have been pivotal. While CT remains a primary diagnostic tool, methods like CEUS offer safer alternatives. Moreover, nonoperative management, especially when combined with angiography and embolization, has demonstrated notable success. Still, the healthcare community must remain vigilant to complications and continuously seek improvements in trauma care.

• Smart Media Journal
• /
• 제9권3호
• /
• pp.59-70
• /
• 2020

This paper presents a fully automatic tool to recognize the liver region from CT images based on a deep learning model, namely Multiple Filter U-net, MFUnet. The advantages of both U-net and Multiple Filters were utilized to construct an autoencoder model, called MFUnet for segmenting the liver region from computed tomograph. The MFUnet architecture includes the autoencoding model which is used for regenerating the liver region, the backbone model for extracting features which is trained on ImageNet, and the predicting model used for liver segmentation. The LiTS dataset and Chaos dataset were used for the evaluation of our research. This result shows that the integration of Multiple Filter to U-net improves the performance of liver segmentation and it opens up many research directions in medical imaging processing field.

• The Korean Journal of Nuclear Medicine Technology
• /
• 제21권2호
• /
• pp.37-43
• /
• 2017

Purpose Recently PET/CT image's attenuation correction is used CTAC(Computed Tomgraphy Attenuation Correction). it can quantitative evaluation by SUV(Standard Uptake Value). This study's purpose is to evaluate SUV and to find proper CT kernel using CTAC with applied various CT kernel to PET/CT construction. Materials and Methods Biograph mCT 64 was used for the equipment. We were performed on 20 patients who had examed at our hospital from february through March 2017. Using NEMA IEC Body Phantom, The data was reconstructed PET/CT images with CTAC appiled various CT kernel. ANOVA was used to evaluated the significant difference in the result. Results The result of measuring the radioactivity concentration of Phantom was B45F 96% and B80F 6.58% against B08F CT kernel, each respectively. the SUVmax increased to B45F 0.86% and B80F 6.54% against B08F CT kernel, In case of patient's parts data, the Lung SUVmax increased to B45F 1.6% and B80F 6.6%, Liver SUVmax increased to B45F 0.7% and B80F 4.7%, and Bone SUVmax increased to B45F 1.3% and B80F 6.2%, respectively. As for parts of patient's about Standard Deviation(SD), the Lung SD increased to B45F 4.2% and B80F 15.4%, Liver SD increased to B45F 2.1% and B80F 11%, and Bone SD increased to B45F 2.3% and B80F 14.7%, respectively. There was no significant difference discovered in three CT kernel (P >.05). Conclusion When using increased noise CT kernel for PET/CT reconstruction, It tends to change both SUVmax and SD in ROI(region of interest), Due to the increase the CT kernel number, Sharp noise increased in ROI. so SUVmax and SD were highly measured, but there was no statistically significant difference. Therefore Using CT kernel of low variation of SD occur less variation of SUV.

• The Korean Journal of Nuclear Medicine Technology
• /
• 제17권1호
• /
• pp.11-17
• /
• 2013

Purpose: The use of SUV which should be normalized by lean body mass (LBM) is recommended for PET response criteria in solid tumors. LBM which was determined by whole body CT was used for SUV normalization (SUL) in this study. The purpose of the present study was to assess interobserver and intraobserver reproducibility of SUL measurements in reference organs. Materials and Methods: F-18 FDG PET/CT was conducted on 52 subjects and LBMs were directly determine by whole body CT for normalization of SUV. The 3 cm diameter spherical VOI, $1\times2$ cm cylindrical VOI, 2 cm diameter spherical VOI were placed in the liver, descending aorta and spleen, respectively. Experienced two observers measured SULmax and SULmean in each organ. Repeated measurements were conducted two weeks apart by observer 1 blind to previous results. Similarly, measurements were conducted on the same patients by observer 2. For assessing reproducibility(or repeatability), the paired t-test, Pearson's correlation coefficients (CC), and technical error of measurement (TEM) were calculated. Results: For interobserver reproducibility in liver SULmax and SULmean, no significant differences were found between observers(paired t-test, P=0.536, 0.293, respectively). CC and TEM for liver SULmean were 0.909 (P=0.000) and 0.067 SUL unit, respectively. Corresponding figures for liver SULmax were 0.882 (P=0.000) and 0.117 SUL unit, respectively. For intraobserver reproducibility in liver SULmax and SULmean, no significant differences were observed within observer1 (paired t-test, P=0.374, 0.268, respectively). CC and TEM for liver SULmean were 0.924 (P=0.000) and 0.061 SUL, respectively. Corresponding figures for liver SULmax were 0.908 (P=0.000) and 0.104 SUL, respectively. Similarly, no significant differences were found in SULmax and SULmean of the spleen and aorta between observers. Conclusion: The current study demonstrated that both SULmean and SULmax measurements in normal reference organs are highly reproducible. Reproducibility of SULmean in reference organs were slightly better than SULmax. Interobsever technical error of measurement was less than 0.10 SUL unit for liver SULmean, and 0.12 SUL unit for liver SULmax. Intraobsever technical error of measurement was less than 0.07 SUL unit for liver SULmean, and 0.11 SUL unit for liver SULmax.

• Journal of Internet Computing and Services
• /
• 제21권3호
• /
• pp.103-111
• /
• 2020

Liver cancer is the most fatal cancer that occurs worldwide. In order to diagnose liver cancer, the patient's physical condition was checked by using a CT technique using radiation. Segmentation was needed to diagnose the liver on the patient's abdominal CT scan, which the radiologists had to do manually, which caused tremendous time and human mistakes. In order to automate, researchers attempted segmentation using image segmentation algorithms in computer vision field, but it was still time-consuming because of the interactive based and the setting value. To reduce time and to get more accurate segmentation, researchers have begun to attempt to segment the liver in CT images using CNNs, which show significant performance in various computer vision fields. The pixel value, or numerical value, of the CT image is called the Hounsfield Unit (HU) value, which is a relative representation of the transmittance of radiation, and usually ranges from about -2000 to 2000. In general, deep learning researchers reduce or limit this range and use it for training to remove noise and focus on the target organ. Here, we observed that the range of HU values was limited in many studies but different in various liver segmentation studies, and assumed that performance could vary depending on the HU range. In this paper, we propose the possibility of considering HU value range as a hyper parameter. U-Net and ResUNet were used to compare and experiment with different HU range limit preprocessing of CHAOS dataset under limited conditions. As a result, it was confirmed that the results are different depending on the HU range. This proves that the range limiting the HU value itself can be a hyper parameter, which means that there are HU ranges that can provide optimal performance for various models.

• The Journal of the Korea Contents Association
• /
• 제15권9호
• /
• pp.411-417
• /
• 2015

To distinguish between HCC and HH is one of the important test methods in determining the treatment method by determining the treatment method by distinguishing malignant growth and benign tumors in liver CT scan. Currently, the specialist is reading CT images by their subjective judgment. So, the purpose of this study is to treat reading the CT images even more objective way. The test times after injection contrast medium in this study are the before injection phase(Pre.), artery phase(35sec), portal phase(70sec) and delay phase(180sec). The general pattern change of HCC in change of contrast enhancement pattern shows 26.6% matching. And the case of HH shows 16.6% matching. In order to observe the change of HU value between HCC and HH, each average values and standard deviation was confirm and as a result, it shows the lagre difference between artery and portal phase in lesion.(HCC$19.76{\pm}23.52$, HH$60.23{\pm}29.43$). And it shows the 76.6% matching in HCC and 80.0% matching in HH. Thorough this study, to suggest a HU value as objective analysis method and if the anlaysis method was used in clinical will assist in the diagnosis.