Ji Kang Park

Jeju National University, Tse-tsiu, Jeju, South Korea

Are you Ji Kang Park?

Claim your profile

Publications (28)52.74 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: The aim of this study was to evaluate whether contrast enhanced fluid attenuated inversion recovery (CE-FLAIR) imaging can be used to predict the severity of meningitis based on leptomeningeal enhancement (LE) score and cerebrospinal fluid signal intensity (CSF-SI) on CE-FLAIR. We retrospectively analyzed data collected from 43 consecutive patients admitted to our hospital due to meningitis. Clinical factors including initial Glasgow Coma Scale (GCS) score, CSF glucose ratio, log CSF protein, log CSF WBC, and prognosis were evaluated. The LE score was semi-quantitatively scored, and we evaluated CSF-SI ratio at the interpeduncular or quadrigerminal cisterns on CE-FLAIR. We evaluated the differences in clinical variables, LE scores and CSF-SI ratios between the recovery and the complication group. We assessed the correlation between clinical variables, LE scores and CSF-SI ratios. The values of log CSF protein, CSF-SI ratio, and LE score were significantly higher in the complication group (p value <0.05). GCS score and CSF glucose ratio were significantly lower in the complication group (p value <0.01). The LE scores had significant negative correlation with GCS scores and CSF glucose ratios (p value <0.001). The LE score was significantly positively correlated with the value of log CSF protein and CSF-SI ratio (p value <0.01). The CSF-SI ratio was negatively correlated with GCS score and CSF glucose ratio (p value <0.01). The CSF-SI ratio was positively correlated with the value of log CSF protein (p value <0.05). Our results suggest that LE score and CSF-SI ratio are well correlated with clinical prognostic factors. We may predict the clinical severity of meningitis by using LE scores and CSF-SI ration on CE-FLAIR imaging.
    Journal of Neurology 02/2014; · 3.84 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To evaluate the clinical and CT findings in patients with small-bowel Anisakiasis. Nineteen patients with small-bowel Anisakiasis and who underwent abdominal CT between 2005 and 2012 were enrolled in our study. All of these patients were diagnosed using either a serologic test for Anisakiasis (n = 18) or by pathology (n = 1). Their medical records were reviewed in order to determine the clinical findings. CT images were retrospectively reviewed by two radiologists to evaluate the characteristics of the involved bowel wall and the ancillary findings. All patients had presented with the acute onset of severe abdominal pain as well as a history of having recently eaten raw fish. The mean time interval from eating the fish to the onset of abdominal pain was 1.7 days. Eighteen patients were treated conservatively and experienced resolution of their symptoms within seven days of hospitalization. One patient underwent surgical exploration for presumed small-bowel ischemia. The sites of involvement included the ileum (16/19, 84%) and jejunum (3/19, 16%). All patients had circumferential bowel-wall thickening (mean, 0.8 cm) with an intermediate length of involved bowel (mean, 7.9 cm). Small-bowel obstruction occurred in 16 patients (84%). The target sign was present in 17 patients (89%), ascites, particularly in the perilesional area, in 16 patients (84%), and mesenteric edema in 15 patients (79%). Small-bowel Anisakiasis should be considered in the differential diagnosis of acute abdomen in order to avoid unnecessary surgery when patients present with abdominal pain after having recently eaten raw fish, concentric bowel-wall thickening with the target sign in the ileum, perilesional ascites, as well as bowel obstruction seen on CT.
    Abdominal Imaging 02/2014; · 1.91 Impact Factor
  • Seong Hyung Kim, Ju Hyun Park, Ji Kang Park
    Journal of the Korean Society of Radiology. 01/2013; 69(1):23.
  • [Show abstract] [Hide abstract]
    ABSTRACT: The purpose of this study was to determine if there is a significant difference in the CT Hounsfield Unit (CTHU) of the carotid atherosclerotic plaque (CAP) on CTpre and CTartery depending on the presence of symptomatology. Quantitative analysis of the CTHUs for the CAP was performed in 43 arteries (11 symptomatic arteries, 32 asymptomatic arteries). The CTHUs were measured using the manual ROI method for each CAP. Group differences in the CTHU of the CAP depending on the symptomatology were evaluated by using the independent t-test. The CTHUs of CAP on CTpre were significantly higher in the symptomatic arteries than those of the asymptomatic arteries (P=0.036). The CTHUs of CAP on CTartery were not significantly different regardless of symptomatology. The CTHUs of the CAP on CT(pre) may be used as one of the biomarkers to distinguish the vulnerable CAP.
    European journal of radiology 07/2011; 81(9):2386-8. · 2.65 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We evaluated the feasibility of black-blood double inversion recovery magnetic resonance imaging (BBDIR) and CT imaging (CTI) for depiction of IAPs. We performed BBDIR on 20 control subjects and 13 patients with acute ischemic stroke. We measured the thickness of the normal vessel wall in control subjects and the maximal and minimal thickness of IAPs in patients on BBDIR. We evaluated signal intensity (SI) and the eccentricity of the IAP on BBDIR, and abnormal wall thickening and CT attenuation of IAPs on CTI. We correlated imaging features of BBDIR and CTI in the patients. The difference of wall thickness between control and patient group was statistically significant (control subjects; basilar artery 0.6 mm, MCA 0.51 mm, and patients; maximal 2.34 mm, minimal 1.3 mm, P value≤.001). The IAP showed eccentric remodeling and heterogeneous SI with the regions of high SI on BBDIR. CTI could not reveal abnormality in 10 patients. Suspicious intraplaque hemorrhage and calcification was demonstrated in 3 patients by CTI. BBDIR could reveal normal and abnormal wall of large intracranial arteries. CTI had limited role for detection of IAP, however, correlation of BBDIR and CTI could provide further characterization of the IAP's in terms of intraplaque calcification and hemorrhage.
    Journal of neuroimaging: official journal of the American Society of Neuroimaging 04/2011; 21(2):e64-8. · 1.82 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Herniation of coil loops into the parent artery is one of the complications of endovascular embolization with detachable coils. In this clinical setting, we cannot predict the consequence of the herniated coil loop. We report an unusual case of a superior cerebellar artery (SCA) aneurysm with delayed coil herniation into the basilar artery and spontaneous reposition into the SCA.
    Neurointervention. 02/2011; 6(1):31-3.
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this study, we evaluated the wall of saccular cerebral aneurysms (SCAs) using two-dimensional double inversion recovery black-blood sequence (BBDI). We examined 14 patients with an unruptured SCA (USCA). The BBDI was peripheral-pulse gated, and was acquired during the mid-diastolic period. We evaluated whether the aneurysmal wall could be visualized with BBDI, and the wall thickness in the neck and dome portion of the aneurysm was measured in cases with acceptable imaging quality. BBDI demonstrated the USCA walls in ten patients. In four patients, the USCA walls were poorly delineated from the adjacent brain parenchyma or cerebrospinal fluid. The mean aneurysm size was 8.0 mm. The mean thickness of the aneurysmal wall in the neck portion was 0.60 +/- 0.13 mm in 10 cases. The mean thickness at the dome portion was 0.46 +/- 0.05 mm in five cases. In this study, BBDI revealed some portion of the USCA wall, despite the limited spatial and contrast resolution for delineation of the entire USCA wall. In our opinion, this technique may be used as an additional imaging tool for the evaluation of the aneurysmal wall.
    Journal of Magnetic Resonance Imaging 11/2009; 30(5):1179-83. · 2.57 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We report herein a patient with aphasic status epilepticus (ASE) who underwent serial brain magnetic resonance imaging (MRI) including diffusion-weighted imaging (DWI), perfusion imaging (PI), and magnetic resonance angiography (MRA) in the peri-ictal phase. The initial DWI revealed a high signal intensity in the left medial temporal area and PI demonstrated hyperperfusion in the left temporoparietal area. MRA revealed an increased flow signal in the ipsilateral intracranial arteries. PI signal changes had normalized by the follow-up MRI. The DWI, PI, MRA, and clinical manifestation of ASE are described herein. J Korean Neurol Assoc 27(3):272-275, 2009
  • [Show abstract] [Hide abstract]
    ABSTRACT: PURPOSE Atherosclerotic plaque of the intracranial vessels has been evaluated by using high-resolution T2- and Proton density-weighted imaging. In this study, we used black-blood double inversion recovery imaging (BBDI) for the evaluation of the vessel wall of intracranial arteries. METHOD AND MATERIALS We studied 20 healthy control subjects (ten 5th decades, ten 8th decades) and 11 stroke patients (mean age 61 years). Imaging study included diffusion-weighted imaging, T2 or FLAIR, intracranial and carotid TOF MRA, and pulse-gated BBDI T1- and T2-weighted imaging (T1WI, T2WI) for intracranial arteries. Basilar and middle cerebral arteries (BA, MCA) were evaluated in control subjects, and symptomatic arteries were evaluated in patients. In-plane resolution for BBDI was 0.46x0.5mm and slice thickness was 3mm. The signal intensity ratio of the vessel wall and adjacent CSF (Vessel wall/CSF) in BBDI T1WI and T2WI imaging was evaluated in control subjects. We measured maximal thickness of the vessel wall and evaluated the nature of the vessel wall about homogeneity and evenness on the BBDI imaging. We evaluated degrees of stenosis on MRA. RESULTS BBDI T1WI had better contrast ratio between vessel wall and CSF in BA and MCA (T1WI: T2WI = 2.08:1.5 for BA, 1.68:1.4 for MCA). Mean thickness of BA and MCA was 0.61 and 0.5mm in control subjects, and 2.28 and 1.72mm in patients. We could not measure vessel wall thickness of MCA in two control subjects. In control subjects, vessel walls had always homogenous signal intensity and even thickness. In patients, vessel walls had inhomogeneity, eccentric thickening, and frequent T1 high signal (9 among 11 patients). In patients group, MRA revealed 7 severe stenotic lesions, but 4 arteries had normal luminal diameter. CONCLUSION BBDI T1WI had better contrast for the depiction of vessel wall of intracranial arteries. Vessel walls were very thin (one pixel size) in healthy control subjects. In stroke patients, vessel walls had 3~4 fold thickness and inhomogeneity and focal T1 high signal. CLINICAL RELEVANCE/APPLICATION MRA revealed only luminal pathology. Vessel wall state of the intracranial artery had been overlooked. BBDI vessel wall imaging aids us to understand intracranial atherosclerosis comprehensively.
    Radiological Society of North America 2008 Scientific Assembly and Annual Meeting; 12/2008
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We present here two patients that had an aberrant right subclavian artery and an anomalous origin of the right vertebral artery from the right common carotid artery. We review the previous literature and discuss herein the embryologic mechanism and clinical implications of this variation.
    Korean Journal of Radiology 07/2008; 9 Suppl:S39-42. · 1.81 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Despite denial as a disease entity, carotidynia had shown relatively concordant imaging finding as an inflammatory process surrounding the carotid bifurcation area. We report serial computed tomography (CT) images of a case with carotidynia that showed the layered thickening of the carotid wall with self-limiting disease course.
    Journal of neuroimaging: official journal of the American Society of Neuroimaging 06/2008; 19(1):84-5. · 1.82 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The purpose of this study was to compare three free-breathing T2-weighted MRI sequences in the evaluation of focal liver lesions. Forty-nine patients with 86 focal liver lesions (56 malignant, 30 benign) underwent liver MRI with free-breathing sequences: turbo spin-echo (TSE) with navigator-triggered prospective acquisition correction (PACE), respiration-triggered TSE, and HASTE with navigator-triggered PACE. The images were retrospectively reviewed by two independent observers. Diagnostic performance was evaluated with receiver operating characteristics and sensitivity. The images were assessed quantitatively by measurement of the liver signal-to-noise ratio (SNR) and the lesion-to-liver contrast-to-noise ratio (CNR). The PACE TSE sequence had better receiver operating characteristic curves for lesion detection and characterization than did the respiration-triggered TSE sequence, but the difference was not statistically significant (p > 0.05). The PACE TSE sequence had a significantly greater area under the curve for lesion detection (p < 0.01) and lesion characterization (p < 0.001) than did the PACE HASTE sequence. The composite sensitivity of the PACE TSE sequence for lesion detection was significantly higher than that of respiration-triggered TSE (p < 0.05) and PACE HASTE (p < 0.01). The mean signal-to-noise ratio for liver and the contrast-to-noise ratio for hepatic lesions were higher with the PACE HASTE than with the other sequences. The navigator-triggered PACE technique is a valid method for T2-weighted MRI of the liver and may replace conventional respiration-triggered techniques.
    American Journal of Roentgenology 01/2008; 190(1):W19-27. · 2.74 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We have used diffusion tensor tractography (DTT) for the evaluation of the somatotopic organization of corticospinal tracts (CSTs) in the posterior limb of the internal capsule (PLIC) and cerebral peduncle (CP). We imaged the brains of nine healthy right-handed subjects. We used a spin-echo echo-planar imaging (EPI) sequence with 12 diffusion-sensitized directions. DTT was calculated with an angular threshold of 35 degrees and a fractional anistropy (FA) threshold of 0.25. We determined the location of the CSTs by using two regions of interest (ROI) at expected areas of the pons and expected areas of the lateral half of the PLIC, in the left hemisphere of the brain. Fiber tracts crossing these two ROIs and the precentral gyrus (PCG) were defined as CSTs. Four new ROIs were then defined for the PCG, from the medial to lateral direction, as ROI 1 (medial) to ROI 4 (lateral). Finally, we defined each fiber tract of the CSTs between the pons and each ROI in the PCG by using two ROIs methods. In all subjects, the CSTs were organized along the long axis of the PLIC, and the hand fibers were located anterior to the foot fibers. The CSTs showed transverse orientation in the CP, and the hand fibers were located usually medial to the foot fibers. Corticospinal tracts are organized along the long axis of the PLIC and the horizontal direction of the CP.
    Korean Journal of Radiology 01/2008; 9(3):191-5. · 1.81 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: It is difficult to distinguish reactive joint fluid from infectious joint fluid by conventional MR images. In this presented case, we successfully distinguished these two conditions using diffusion-weighted imaging.
    Journal of Magnetic Resonance Imaging 05/2007; 25(4):859-61. · 2.79 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: To compare the efficacy of low-dose and standard-dose computed tomography (CT) for the diagnosis of ureteral stones. Unenhanced helical CT was performed with both a standard dose (260 mAs, pitch 1.5) and a low dose (50 mAs, pitch 1.5) in 121 patients suspected of having acute renal colic. The two studies were prospectively and independently interpreted for the presence and location of ureteral stones, abnormalities unrelated to stone disease, identification of secondary signs, i.e. hydronephrosis and perinephric stranding, and tissue rim sign. The standard-dose CT images were interpreted by one reviewer and the low-dose CT images independently by two reviewers unaware of the standard-dose CT findings. The findings of the standard and low-dose CT scans were compared with the exact McNemar test. Interobserver agreements were assessed with kappa analysis. The effective radiation doses resulting from two different protocols were calculated by means of commercially available software to which the Monte-Carlo phantom model was given. The sensitivity, specificity, and accuracy of standard-dose CT for detecting ureteral stones were 99%, 93%, and 98%, respectively, whereas for the two reviewers the sensitivity of low-dose CT was 93% and 95%, specificity 86%, and accuracy 92% and 94%. We found no significant differences between standard-dose and low-dose CT in the sensitivity and specificity for diagnosing ureter stones (P >0.05 for both). However, the sensitivity of low-dose CT for detection of 19 stones less than or equal to 2 mm in diameter was 79% and 68%, respectively, for the two reviewers. Low-dose CT was comparable to standard-dose CT in visualizing hydronephrosis and the tissue rim sign. Perinephric stranding was far less clear on low-dose CT. Low-dose CT had the same diagnostic performance as standard-dose CT in diagnosing alternative diseases. Interobserver agreement between the two low-dose CT reviewers in the diagnosis of ureter stones and alternative diseases, the identification of secondary signs, and tissue rim sign were high, with kappa values ranging from 0.769 to 0.968. On standard-dose CT scans, the calculated mean effective radiation dose was 7.30 mSv for males and 10.00 mSv for females. On low-dose CT scans, the calculated mean effective radiation dose was 1.40 mSv for males and 1.97 mSv for females. Compared with standard scans using 260 mAs, low-dose unenhanced helical CT using a reduced tube current of 50 mAs results in a concomitant decrease in the radiation dose of 81%. Although low-dose CT was limited in its ability to depict small-sized calculi less than or equal to 2 mm, it is still comparable to standard-dose CT for the diagnosis of ureter stones and alternative disease.
    Acta Radiologica 12/2005; 46(7):756-63. · 1.35 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We examined whole blood (MnB), red blood cell (MnRBC), plasma (MnP) and urinary Mn (MnU) concentrations in 22 liver cirrhotics and 10 healthy controls to evaluate Mn concentration in which a fraction of biological samples best reflects pallidal signal intensities (pallidal index; PI) on T1-weighted magnetic resonance images (MRI) in liver cirrhotics. Increased signal intensity in the globus pallidus was observed in 18 (81.8%) of the 22 patients with liver cirrhosis. In a transplanted patient, increased pallidal signals also resolved as his liver function tests normalized after liver transplantation. There were significant correlations between MnB/MnRBC and PI (rho=0.529, rho=0.573, respectively) in liver cirrhotics, although no significant correlation was observed between MnP/MnU and PI. According to a multiple linear regression, MnB and MnRBC reflected the signal intensities of T1-weighted MRI better than MnP or MnU.
    Journal of Occupational Health 02/2005; 47(1):68-73. · 1.10 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The objective of our prospective study was to evaluate the value of various operator-dependent techniques that allow graded compression sonography to detect normal or abnormal vermiform appendix. A total of 877 subjects were included in this study. This sample population consisted of two groups: 202 control subjects and 675 patients who were suspected of having acute appendicitis. If detection of the appendix failed after a sufficient number of trials using graded compression scanning, appropriate operator-dependent techniques were used to help graded compression scanning to increase the detectability of the appendix further. The detection rate for the appendix in both groups and the diagnostic accuracy for acute appendicitis were obtained. The initial graded compression sonography examination depicted the appendix in 170 (84%) of 202 subjects in the control group and 601 (89%) of the 675 patients in the patient group. We then added operator-dependent techniques to graded compression sonography for the remaining patients in whom the appendix could not be detected. The additional use of the posterior manual compression technique, low-frequency convex transducer, upward graded compression technique, or left oblique lateral decubitus change of body position allowed graded compression sonography to depict the appendix in an additional 10, eight, six, and four patients in the control group, respectively, and in an additional 27, 23, 11, and seven patients in the patient group. The number of identified appendixes was increased to 198 (98%) of the 202 patients in the control group and to 669 (99%) of the 675 patients in the patient group. Graded compression sonography with operator-dependent techniques in the patient group yielded a sensitivity of 99% (319/321 patients), specificity of 99% (350/354), and an accuracy of 99% (669/675) for acute appendicitis. The addition of various operator-dependent techniques to graded compression sonography is useful for allowing improved visualization of both normal and abnormal appendixes.
    American Journal of Roentgenology 02/2005; 184(1):91-7. · 2.74 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: To determine whether blood manganese (Mn) concentration is elevated in patients with iron deficiency anemia (IDA), and whether this affects signal intensities in the globus pallidus. Twenty-seven patients with IDA and 10 control subjects were tested for blood Mn, and brain magnetic resonance images (MRI) were also examined. Seventeen of the 27 patients were followed-up after iron therapy. IDA patients had a mean blood Mn concentration of 2.05 +/- 0.44 microg/dl, which was higher than controls. The mean pallidal index (PI) of anemic patients was not different from that of controls. There was a correlation between log blood Mn and PI (rho = 0.384, P = 0.048; n = 27) in IDA patients. None of the patients showed increased signals in the globus pallidus in T1-weighted MRI. Blood Mn levels decreased and hemoglobin levels increased after iron therapy (P < 0.05). Although blood Mn is elevated in IDA patients, there is no increase in globus pallidus MRI signal intensity. These findings stand in contrast to those of our other studies showing patients with chronic liver disease or occupational Mn exposure have elevated signal intensities remarkably.
    NeuroToxicology 01/2005; 26(1):107-11. · 3.05 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Characteristic high signal intensities confined to the globus pallidus on T1-weighted magnetic resonance image (MRI) can be observed in manganese (Mn)-exposed workers, however, these high signals should be differentiated from those due to other causes such as fat, hemoglobin breakdown products, melanoma, neurofibromatosis, and calcification. A 39-year-old woman was admitted with mutism and involuntary movements which had developed the day before. She had ingested two packs of liquid herbal medicine containing 0.53 mg of Mn daily for 4 months prior to visiting our hospital. Her MRI showed high signals, confined mainly to the globus pallidus on T1-weighted images. Follow-up brain MRI at an interval of 11 months showed no interval change. Brain computed tomography (CT) at the time of the second MRI showed symmetric calcification on both globus pallidus. Blood levels of liver function tests, calcium, phosphorus, and parathyroid hormone were within normal ranges. The increased signals, which were first presumed to be induced by Mn, were concluded to be due to calcification based on the following reasons. First, follow-up brain MRI at an interval of 11 months did not show any interval change. Second, the ingested amount of 1.06 mg Mn daily for 4 months is even less than that added to mineral supplements for adults. Third, Mn-induced high signals in T1-weighted MRI do not show any abnormal findings in brain CT. The present case report suggests that brain CT should be performed to rule out symmetric calcification on basal ganglia in patients showing increased signals in T1-weighted MRI, but who do not have a significant exposure history to Mn. The present report also showed that the amount of 1.06 mg Mn daily ingested for 4 months did not cause the high signal in brain MRI.
    NeuroToxicology 01/2004; 24(6):835-8. · 3.05 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To evaluate whether the results of cerebrospinal fluid (CSF) flow quantification differ according to the anatomical location of the cerebral aqueduct that is used and the background baseline region that is selected. The CSF hydrodynamics of eleven healthy volunteers (mean age = 29.6 years) were investigated on a 1.5 T MRI system. Velocity maps were acquired perpendicular to the cerebral aqueduct at three different anatomical levels: the inlet, ampulla and pars posterior. The pulse sequence was a prospectively triggered cardiac-gated flow compensated gradient-echo technique. Region-of-interest (ROI) analysis was performed for the CSF hydrodynamics, including the peak systolic velocity and mean flow on the phase images. The selection of the background baseline regions was done based on measurements made in two different areas, namely the anterior midbrain and temporal lobe, for 10 subjects. The mean peak systolic velocities showed a tendency to increase from the superior to the inferior aqueduct, irrespective of the background baseline region, with the range being from 3.30 cm/sec to 4.08 cm/sec. However, these differences were not statistically significant. In the case of the mean flow, the highest mean value was observed at the mid-portion of the ampulla (0.03 cm(3)/sec) in conjunction with the baseline ROI at the anterior midbrain. However, no other differences were observed among the mean flows according to the location of the cerebral aqueduct or the baseline ROI. We obtained a set of reference data of the CSF peak velocity and mean flow through the cerebral aqueduct in young healthy volunteers. Although the peak systolic velocity and mean flow of the CSF differed somewhat according to the level of the cerebral aqueduct at which the measurement was made, this difference was not statistically significant.
    Korean Journal of Radiology 01/2004; 5(2):81-6. · 1.81 Impact Factor

Publication Stats

310 Citations
52.74 Total Impact Points


  • 2007–2014
    • Jeju National University
      Tse-tsiu, Jeju, South Korea
  • 2005
    • Daejeon University
      Sŏul, Seoul, South Korea
  • 2003–2005
    • Ulsan University Hospital
      Urusan, Ulsan, South Korea
  • 2004
    • University of Ulsan
      • College of Medicine
      Ulsan, Ulsan, South Korea