• Proceedings of the Korean Environmental Health Society Conference
• /
• 2005.06a
• /
• pp.105-139
• /
• 2005

Increased signal in T1-weighted images was observed in the experimental manganese (Mn) poisoning of the non-human primate and a patient with Mn neurointoxication. However, our study showed that the increased signals in magnetic resonance images (MRI) were highly prevalent (41.6%) in Mn-exposed workers. Blood Mn concentration correlated with pallidal index. These changes in MRI tend to disappear following the withdrawal from the source of Mn accumulation, despite permanent neurological damage. Thus increased signal intensities on a T1-weighted image reflect exposure to Mn, but not necessarily manganism. Our study also showed that the concentration of Mn required to produce increased signal intensities on MRI is much lower than the threshold necessary to result in overt clinical signs of manganism. Increased signal intensities in the globus pallidus were determined by manganese accumulation in the animal experiment. Reanalysis of the previous data with the structural equation model revealed that pallidal index (Pl) on MRI reflects target organ dose of occupational Mn exposure

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.7
• /
• pp.3578-3593
• /
• 2017

Given two collections of objects that carry both spatial and textual information in the form of tags, a $\text\underline{S}patio$-$\text\underline{T}extual$-based object $\text\underline{S}imilarity$ $\text\underline{JOIN}$ (ST-SJOIN) retrieves the pairs of objects that are textually similar and spatially close. In this paper, we have proposed a block index-based approach called BIST-JOIN to facilitate the efficient ST-SJOIN processing. In this approach, a dual-feature distance plane (DFDP) is first partitioned into some blocks based on four segmentation schemes, and the ST-SJOIN is then transformed into searching the object pairs falling in some affected blocks in the DFDP. Extensive experiments on real and synthetic datasets demonstrate that our proposed join method outperforms the state-of-the-art solutions.

• Journal of radiological science and technology
• /
• v.41 no.3
• /
• pp.209-214
• /
• 2018

The results of CBCT was obtained using image guided radiation therapy for radiation therapy in 5 prostate cancer patients. Using these results, we compared and evaluated the dose changes according to the treatment plan depending on the volume and position of bladder, rectum, and prostate. The 28 images of CBCT were acquired using On-Board Imaging device before radiotherapy. After the outline of bladder, rectum, and PTV, pCT images and CBCT images for radiotherapy were treated respectively. The volume of the bladder was increased by 105.6% and decreased by 45.2%. The volume of the rectum was increased by 30.5% and decreased by 20.3%. Prostate volume was increased by 6.3% and decreased by 12.3%. The mean dose of the rectum was higher in the CBCT than in the pCT, and V40 (equivalent to 40 Gy) of the bladder showed a reduction in all treatment regimens in the CBCT than in the pCT. Conformity treatment and homogeneity index of PTV showed better results in all treatment regimens using pCT than CBCT. It was found that the dose distribution of the pelvic internal organs varied greatly according to the patient 's condition and pretreatment.

• 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 radiological science and technology
• /
• v.40 no.1
• /
• pp.57-62
• /
• 2017

The purpose of this study is to reconstruct the treatment plan by applying CBCT and DIR to dose changes according to the change of the patient's motion and breast shape in the large breast cancer patients and to compare the doses using TWF, FIF and IMRT. CT and CBCT were performed with MIM6 to create DIRCT and each treatment plan was made. The patient underwent computed tomography simulation in both prone and supine position. The homogeneity index (HI), conformity index (CI), coverage index (CVI) to the left breast as planning target volume (PTV) were determined and the doses to the lung, heart, and right breast as organ at risk (OAR) were compared by using dose-volume histogram and the unique property of each organ. The value of HI of the PTV breast increased in all treatment planning methods using DIRCT, and CVI and CI were decreased in the treatment planning methods using DIRCT.

• Journal of radiological science and technology
• /
• v.40 no.3
• /
• pp.385-392
• /
• 2017

The purpose of this study is to investigate the image quality and exposure dose according to kVp and mAs in CT and to confirm improvement in image quality according to None IR and IR(Iterative Reconstruction) levels. Measurement results of image quality using Image J, HU(Hounsfield units) and BN(Background Noise) are decreased, while SNR(Signal to Noise Ratio) and $CTDI_{vol}$(CT dose index volume) are increased as the kVp increases and there was no change of BHU(Background Hounsfield units). BN was reduced due to increased kVp, while SNR and $CTDI_{vol}$ were increased. Also, the higher IR stage, the lower BN, SI(Signal Intensity) and HU while SNR was improved by about 10~60%. Based on this, when applying IR for clinical applications, it is necessary to finely adjust kVp and mA with a phased approach.

• Journal of Gastric Cancer
• /
• v.12 no.2
• /
• pp.99-107
• /
• 2012

Purpose: Gastric mucosal neoplastic lesions should have characteristic endoscopic features for successful endoscopic submucosal dissection. Materials and Methods: Out of the 1,010 endoscopic submucosal dissection, we enrolled 62 patients that had the procedure cancelled. Retrospectively, whether the reasons for cancelling the endoscopic submucosal dissection were consistent with the indications for an endoscopic submucosal dissection were assessed by analyzing the clinical outcomes of the patients that had the surgery. Results: The cases were divided into two groups; the under-diagnosed group (30 cases; unable to perform an endoscopic submucosal dissection) and the over-diagnosed group (32 cases; unnecessary to perform an endoscopic submucosal dissection), according to the second endoscopic findings, compared with the index conventional white light image. There were six cases in the under-diagnosed group with advanced gastric cancer on the second conventional white light image endoscopy, 17 cases with submucosal invasion on endoscopic ultrasonography findings, 5 cases with a size greater than 3 cm and ulcer, 1 case with diffuse infiltrative endoscopic features, and 1 case with lymph node involvement on computed tomography. A total of 25 patients underwent a gastrectomy to remove a gastric adenocarcinoma. The overall accuracy of the decision to cancel the endoscopic submucosal dissection was 40% (10/25) in the subgroup that had the surgery. Conclusions: The accuracy of the decision to cancel the endoscopic submucosal dissection, after conventional white light image and endoscopic ultrasonography, was low in this study. Other diagnostic options are needed to arrive at an accurate decision on whether to perform a gastric endoscopic submucosal dissection.

• Journal of the Korean Society of Radiology
• /
• v.9 no.6
• /
• pp.343-348
• /
• 2015

The purpose of this study is a comparison of forward planning(FP) and inverse planning(IP) of a radiosurgery procedure. 10 patients of acoustic schwannoma MR image were used for treatment plan. FP-1,2 and IP were established under the same condition. FP and IP were compared by number of shot, conformity index(CI), paddic conformity index(PCI), gradiant index(GI) and treatment time. On average the treatment plan produced by IP tool provided an improved or similar CI, PCI, GI and reduced treatment time as compared to the FP (CI;FP-1:0.85, FP-2:0.86, IP:0.94, PCI;FP-1:0.79, FP-2:0.81, IP:0.78, GI;FP-1:2.94, FP-2:2.94, IP:3.01). The inverse planning system provides a clinically useful plan while reducing the planning time and treatment time.

• Korean Journal of Radiology
• /
• v.21 no.7
• /
• pp.891-899
• /
• 2020

Objective: To assess the diagnostic performance of a deep learning-based algorithm for automated detection of acute and chronic rib fractures on whole-body trauma CT. Materials and Methods: We retrospectively identified all whole-body trauma CT scans referred from the emergency department of our hospital from January to December 2018 (n = 511). Scans were categorized as positive (n = 159) or negative (n = 352) for rib fractures according to the clinically approved written CT reports, which served as the index test. The bone kernel series (1.5-mm slice thickness) served as an input for a detection prototype algorithm trained to detect both acute and chronic rib fractures based on a deep convolutional neural network. It had previously been trained on an independent sample from eight other institutions (n = 11455). Results: All CTs except one were successfully processed (510/511). The algorithm achieved a sensitivity of 87.4% and specificity of 91.5% on a per-examination level [per CT scan: rib fracture(s): yes/no]. There were 0.16 false-positives per examination (= 81/510). On a per-finding level, there were 587 true-positive findings (sensitivity: 65.7%) and 307 false-negatives. Furthermore, 97 true rib fractures were detected that were not mentioned in the written CT reports. A major factor associated with correct detection was displacement. Conclusion: We found good performance of a deep learning-based prototype algorithm detecting rib fractures on trauma CT on a per-examination level at a low rate of false-positives per case. A potential area for clinical application is its use as a screening tool to avoid false-negative radiology reports.

• Investigative Magnetic Resonance Imaging
• /
• v.18 no.1
• /
• pp.43-51
• /
• 2014

Purpose : To review MR imaging finding of papillary lesion identified as additional suspicious lesion on MR image in women with biopsy-proven breast cancer and to evaluate upgrading rates after subsequent surgical histopathological diagnosis. Materials and Methods: Among 1729 preoperative MR image of women with biopsy proven breast cancer, US-guided CNB-proven 22 papillary lesions from 21 patients, which showed additional suspicious contrast enhancement other than index cancer on MR image, were subjected to the study. Some of these lesions underwent surgery, thus the comparisons between the histopathologic results were able to be compared to the results of US-guided CNB. Also retrospective analysis was done for MR findings of these lesions by BI-RADS MRI lexicon. Results: On MR imaging, 8 mass lesions, 7 non-mass lesions, 7 focus lesions were detected. All of the focus lesion (100%, 7/7) was diagnosed as benign lesion and showed plateau and washout pattern in dynamic MR image. After excisional biopsy, one of 9 benign papilloma (11.1%), 3 of 3 papillary neoplasm with atypia component (100%), 3 of 5 papillary neoplasm (60%) were upgraded to malignancy such as ductal carcinoma in situ (DCIS), invasive ductal carcinoma (IDC), invasive lobular carcinoma (ILC). Conclusion: The MR images of papillary lesions diagnosed by US-guided CNB exhibit no significant differences between malignancy and benign lesion. Also 41.2% of the lesion (7/17) was upgraded after subsequent surgery. Thus all of the papillary lesions require excisional biopsy for definite diagnosis and the MR imaging, it's just not enough by itself.