• Title/Summary/Keyword: weighted kappa

Search Result 89, Processing Time 0.031 seconds

Reproducibility of an Automatic Quantitation of Regional Myocardial Wall Motion and Systolic Thickening on Gated Tc-99m-MIBI Myocardial SPECT (게이트 Tc-99m-MIBI SPECT에서 국소 심근운동과 수축기 심근두꺼워짐 자동정량화법의 재현성)

  • Paeng, Jin-Chul;Lee, Dong-Soo;Cheon, Gi-Jeong;Kim, Yu-Kyeong;Chung, June-Key;Lee, Myung-Chul
    • The Korean Journal of Nuclear Medicine
    • /
    • v.34 no.6
    • /
    • pp.487-496
    • /
    • 2000
  • Purpose: The aim of this study is to investigate the reproducibility of the quantitative assessment of segmental wall motion and systolic thickening provided by an automatic quantitation algorithm. Materials and Methods: Tc-99m-MIBI gated myocardial SPECT with dipyridamole stress was performed in 31 patients with known or suspected coronary artery disease (4 with single, 6 with two, 11 with triple vessel disease; ejection fraction $51{\pm}14%$) twice consecutively in the same position. Myocardium was divided into 20 segments. Segmental wall motion and systolic thickening were calculated and expressed in mm and % increase respectively, using $AutoQUANT^{TM}$ software. The reproducibility of this quantitative measurement of wall motion and thickening was tested. Results: Correlations between repeated measurements on consecutive gated SPECT were excellent for wall motion (r=0.95) and systolic thickening (r=0.88). On Bland-Altman analysis, two standard deviation was 2 mm for repeated measurement of segmental wall motion, and 20% for that of systolic thickening. The weighted kappa values of repeated measurements were 0.807 for wall motion and 0.708 for systolic thickening. Sex, perfusion, or segmental location had no influence on reproducibility. Conclusion: Segmental wall motion and systolic thickening quantified using $AutoeUANT^{TM}$ software on gated myocardial SPECT offers good reproducibility and is significantly different when the change is more than 2 mm for wall motion and more than 20% for systolic thickening.

  • PDF

Assessment of Posterior Globe Flattening: Two-Dimensional versus Three-Dimensional T2-Weighted Imaging

  • Ann, Jun Hyung;Kim, Eung Yeop
    • Investigative Magnetic Resonance Imaging
    • /
    • v.19 no.3
    • /
    • pp.178-185
    • /
    • 2015
  • Purpose: To compare the frequency of posterior globe flattening between two-dimensional T2-weighted imaging (2D T2WI) and three-dimensional (3D T2WI). Materials and Methods: Sixty-nine patients (31 female; mean age, 44.4 years) who had undergone both 5-mm axial T2WI and sagittal 3D 1-mm isovoxel T2WI of the whole brain for evaluation of various diseases (headache [n = 30], large hemorrhage [n = 19], large tumor or leptomeningeal tumor spread [n = 15], large infarct [n = 3], and bacterial meningitis [n = 2]) were used in this study. Two radiologists independently reviewed both sets of images at separate sessions. Axial T2WI and multi-planar imaging of 3D T2WI were visually assessed for the presence of globe flattening. The optic nerve sheath diameter (ONSD) was measured at a location 4 mm posterior to each globe on oblique coronal imaging reformatted from 3D T2WI. Results: There were significantly more globes showing posterior flattening on 3D T2WI (105/138 [76.1%]) than on 2D T2WI (27/138 [19.6%], P = 0.001). Inter-observer agreement was excellent for both 2D T2WI and 3D T2WI (Cohen's kappa = 0.928 and 0.962, respectively). Intra-class correlation coefficient for the ONSD was almost perfect (Cohen's kappa = 0.839). The globes with posterior flattening had significantly larger ONSD than those without on both 2D and 3D T2WI (P < 0.001; $6.14mm{\pm}0.44$ vs. $5.74mm{\pm}0.44$ on 2D T2WI; $5.90mm{\pm}0.47$ vs. $5.56mm{\pm}0.34$ on 3D T2WI). Optic nerve protrusion was significantly more frequent on reformatted 1-mm 3D T2WI than on 5-mm 2D T2WI (8 out of 138 globes on 3D T2WI versus one on 2D T2WI; P = 0.018). Conclusion: Posterior globe flattening is more frequently observed on 3D T2WI than on 2D T2WI in patients suspected of having increased intracranial pressure. The globes with posterior flattening have significantly larger ONSD than those without.

Brain Magnetic Resolution Imaging to Diagnose Bing-Neel Syndrome

  • Kim, Ho-Jung;Suh, Sang-Il;Kim, Joo-Han;Kim, Byung-Jo
    • Journal of Korean Neurosurgical Society
    • /
    • v.46 no.6
    • /
    • pp.588-591
    • /
    • 2009
  • Radiologic findings of Bing-Neel syndrome, which is an extremely uncommon complication resulting from malignant lymphocyte infiltration into the central nervous system (CNS) in patients with Waldenstr$\ddot{o}$m's macroglobulinemia (WM), have been infrequently reported due to extreme rarity of the case. A 75-year-old man with WM presented at a neurology clinic with progressive gait and memory disturbances, and dysarthria of 2 months duration. Cerebrospinal fluid and serum protein electrophoresis and immunofixation electrophoresis showed IgM kappa-type monoclonal gammopathy. Brain magnetic resonance imaging revealed multifocal, hyperintense lesions on T2 weighted-images. Brain diffusion-weighted imaging (DWI) demonstrated hyperintensities in cerebral and cerebellar lesions that appeared isointense on apparent diffusion coefficient maps, which were compatible with vasogenic edema. Although histologic analysis is a confirmative study to prove direct cell infiltration into the brain, brain MRI with DWI may be a good supportive study to diagnose Bing-Neel syndrome.

Modeling of a Pulverized Coal Combustion With Applying WSGGM (희체가스 가중합산모델을 적용한 미분탄 연소의 해석)

  • Yu, Myoung-Jong;Baek, Seung-Wook
    • 한국연소학회:학술대회논문집
    • /
    • 1999.10a
    • /
    • pp.155-163
    • /
    • 1999
  • A numerical study for simulating a swirling pulverized coal combustion in axisymmetric geometry is done here by applying the weighted sum of gray gases model (WSGGM) approach with the discrete ordinate method (DOM) to model the radiative heat transfer equation. In the radiative transfer equation, the same polynomial equation and coefficients for weighting factors as those for gas are adopted for the coal/char particles as a function of partial pressure and particle temperature. The Eulerian balance equations for mass, momentum, energy, and species mass fractions are adopted with the standard ${\kappa}-{\varepsilon}$ turbulence model, whereas the Lagrangian approach is used for the particulate phase for soot. The eddydissipation model is employed for the reaction rate for gaseous mixture, and the single-step first-order reaction model for the devolatilization process for coal. By comparing the numerical results with experimental ones, the models used here are confirmed and found to be one of good alternatives for simulating the combustion as well as radiative characteristics.

  • PDF

Development and validation of a quantitative food frequency questionnaire to assess nutritional status in Korean adults

  • Na, Youn Ju;Lee, Seon Heui
    • Nutrition Research and Practice
    • /
    • v.6 no.5
    • /
    • pp.444-450
    • /
    • 2012
  • This study was performed to evaluate the validity of the food frequency questionnaire (FFQ), which is being used at the Samsung Medical Center. In total, 305 (190 males and 115 females) participants consented and completed the 3-day diet records and FFQ. Age, gender and energy-adjusted and de-attenuated correlations ranged from 0.317 (polyunsaturated fatty acid) to 0.748 (carbohydrate) with a median value of 0.550. The weighted kappa value ranged from 0.18 (vitamin A) to 0.57 (carbohydrate) with a median value of 0.36. More than 75% of the subjects were classified into the same or adjacent quartiles. The FFQ had reasonably good validity compared with that of another study. Therefore, our FFQ is considered a proper method to assess nutrient intake in healthy Korean adults.

Data Augmentation and Preprocessing to Improve Automated Essay Scoring Model (에세이 자동 평가 모델 성능 향상을 위한 데이터 증강과 전처리)

  • Kanghee Go;Doguk Kim
    • Annual Conference on Human and Language Technology
    • /
    • 2023.10a
    • /
    • pp.327-332
    • /
    • 2023
  • 데이터의 품질과 다양성은 모델 성능에 지대한 영향을 끼친다. 본 연구에서는 Topic을 활용한 데이터 전처리와 BERT 기반 MLM, T5, Random Masking을 이용한 증강으로 데이터의 품질과 다양성을 높이고자 했으며, 이를 KoBERT 기반 에세이 자동 평가 모델에 적용했다. 데이터 전처리만 진행했을 때, Quadratic Weighted Kappa Score(QWK)를 기준으로 모델이 에세이의 모든 평가 항목에 대해 베이스라인보다 더욱 높은 일치도를 보였으며 평가항목별 일치도의 평균을 기준으로 0.5368029에서 0.5483064(+0.0115035)로 상승했다. 여기에 제안하는 증강 방식을 추가 할 경우 MLM, T5, Random Masking 모두 성능 향상 효과를 보였다. 특히, MLM 데이터 증강 방식을 추가로 적용하였을 때 최종적으로 0.5483064에서 0.55151645(+0.00321005)으로 상승해 가장 높은 일치도를 보였으며, 에세이 총점으로 QWK를 기준으로 성능을 평가하면 베이스라인 대비 0.4110809에서 0.4380132(+0.0269323)로의 성능 개선이 있었다.

  • PDF

Usefulness of MRI 3D Image Reconstruction Techniques for the Diagnosis and Treatment of Femoral Acetabular Impingement Syndrome(Cam type) (대퇴 골두 충돌 증후군(Cam type)의 진단과 치료를 위한 자기공명 3D 영상 재구성 기법의 유용성)

  • Kwak, Yeong-Gon;Kim, Chong-Yeal;Cho, Yeong-Gi
    • The Journal of the Korea Contents Association
    • /
    • v.15 no.11
    • /
    • pp.313-321
    • /
    • 2015
  • To minimize CT examination for Hip FAI diagnosis and operation plan. also, whether the MRI 3D images can replace Hip Clock face image was evaluated when performing Hip FAI MRI by using additional 3D image. This study analyzed Hip MRI and 3D Hip CT images of 31 patients in this hospital. For the purpose of evaluating the images, one orthopedic surgeon and one radiology specialist reconstructed Clock face, at MR and CT modality, by superior 12 o'clock, labrum front 3 o'clock, and the other side 9 o'clock, centering on Hip joint articular transverse ligament 6 o'clock. Afterwards, by the Likert Scale 5 point scale (independent t-test p<0.005), this study evaluated the check-up of A. retinacular vessel, B. head neck junction at 11 o'clock, A. Epiphyseal line, B. Cam lesion at 12 o'clock, and Cam lesion, Posterior Cam lesion at 1,2,3 and 4 o'clock. As for the verification of reliability among observers, this study verified coincidence by Cohen's weighted Kappa verification. As a result of Likert scale for the purpose of qualitative evaluation about the image, 11 o'clock A. retinacular vessel MR average was $3.69{\pm}1.0$ and CT average was $2.8{\pm}0.78$. B. head neck juncton didn't have a difference between two observers (p <0.416). 12 o'clock A. Epiphyseal line MR average was $3.54{\pm}1.00$ and CT average was $4.5{\pm}0.62$(p<0.000). B. Cam lesion didn't have a difference between two observers (p <0.532). 1,2,3,4 Cam lesion and Posterior Cam lesion were not statistically significant (p <0.656, p <0.658). As a result of weighted Kappa verification, 11 o'clock A.retinacular vessel CT K value was 0.663 and the lowest conformity. As a result of coincidence evaluation on respective item, a very high result was drawn, and two observers showed high reliability.

Triage Accuracy of Pediatric Patients using the Korean Triage and Acuity Scale in Emergency Departments (한국형응급환자분류도구를 적용한 응급실에서 소아 환자의 중증도 분류 정확성)

  • Moon, Sun-Hee;Shim, Jae Lan
    • Journal of the Korea Academia-Industrial cooperation Society
    • /
    • v.19 no.11
    • /
    • pp.626-634
    • /
    • 2018
  • This retrospective study investigates the accuracy of triage procedures for pediatric patients in emergency departments (EDs) using the Korean Triage and Acuity Scale (KTAS). The study includes 250 randomly selected initial nursing records and clinical outcomes of pediatric patients who visited one regional ED or a local ED from October 2016 to September 2017. The collected data were analyzed by a qualified expert to determine the true triage score. The accuracy of triage was defined as the agreement between the triage score of the emergency nurses (ENs) and the true triage score as determined by the expert. Based on expert comments, the cause of the triage error was analyzed and the KTAS score was compared with the discharge, length of stay (LOS), and medical cost. The results showed that the degree of agreement in the triage score between the experts and the ENs was excellent (weighted kappa=0.77). Among the causes of triage discordance, the most frequent was the incorrect application of vital signs to the KTAS algorithm criteria (n=13). Patients with high severity KTAS levels 1 and 2 were discharged less often (${\chi}=43.25$, p<0.001). There were significant differences in the length of stay (F=12.39, p<0.001) and cost (F=11.78, p<0.001) between KTAS scores when adjusting for age. The results of this study indicate that KTAS is highly accurate in EDs. Hence, the newly developed triage tool is becoming well established in Korea.

Evaluation of the Effectiveness of a Re-decontaminating Process with Bacterial Contaminated Specimens Showing a Positive MGIT Signal for the Detection of Mycobacteria (마이코박테리아 검출을 위하여 MGIT 양성 신호를 보인 세균에 오염된 검체 재처리 과정의 효율성 평가)

  • Jung, Haeyong;Bang, Hae In;Choi, Tae Youn
    • Korean Journal of Clinical Laboratory Science
    • /
    • v.51 no.2
    • /
    • pp.171-176
    • /
    • 2019
  • A comparative study between commercially available mycobacteria growth indicator tubes (MGIT) in the BACTEC MGIT 960 System and the conventional Ogawa media was carried out to assess the effectiveness of the re-decontaminating process for the recovery of mycobacteria. Processed specimens with 5% sodium hydroxide and 0.5% N-acetyl-L-cysteine were inoculated into MGIT and Ogawa media. The acid fast bacilli (AFB) recovered from the cultures were identified using a mycobacterium tuberculosis (TB) antigen kit. If contaminants were observed in the MGIT tubes within five days, a decontaminating process was repeated. A total of 1,190 out of 4,790 (24.8%) specimens showed positive results using the BACTEC MGIT 960 system. Among them, 278 specimens were reprocessed. When the MGIT and Ogawa results were compared, it showed discordant results (weighted kappa value: 0.283). One TB and 10 nontuberculous mycobacteria (NTM) were newly detected in MGIT only. The likely benefit of the re-decontaminating process is the detection of additional mycobacteria that could not be detected without a re-decontaminating process despite being small in number. In addition to the combination of MGIT and Ogawa, the re-decontaminating process is recommended in the case of contaminations to recover mycobacteria.

Rating criteria to evaluate student performance in digital wax-up training using multi-purpose software

  • Mino, Takuya;Kurosaki, Yoko;Tokumoto, Kana;Higuchi, Takaharu;Nakanoda, Shinichi;Numoto, Ken;Tosa, Ikue;Kimura-Ono, Aya;Maekawa, Kenji;Kim, Tae Hyung;Kuboki, Takuo
    • The Journal of Advanced Prosthodontics
    • /
    • v.14 no.4
    • /
    • pp.203-211
    • /
    • 2022
  • PURPOSE. The aim of this study was to introduce rating criteria to evaluate student performance in a newly developed, digital wax-up preclinical program for computer-aided design (CAD) of full-coverage crowns and preliminarily investigate the reliability and internal consistency of the rating system. MATERIALS AND METHODS. This study, conducted in 2017, enrolled 47 fifth-year dental students of Okayama University Dental School. Digital wax-up training included a fundamental practice using computer graphics (CG), multipurpose CAD software programs, and an advanced practice to execute a digital wax-up of the right mandibular second molar (#47). Each student's digital wax-up work (stereolithography data) was evaluated by two instructors using seven qualitative criteria. The total qualitative score (0-90) of the criteria was calculated. The total volumetric discrepancy between each student's digital wax-up work and a reference prepared by an instructor was automatically measured by the CAD software. The inter-rater reliability of each criterion was analyzed using a weighted kappa index. The relationship between the total volume discrepancy and the total qualitative score was analyzed using Spearman's correlation. RESULTS. The weighted kappa values for the seven qualitative criteria ranged from 0.62 - 0.93. The total qualitative score and the total volumetric discrepancy were negatively correlated (ρ = -0.27, P = .09, respectively); however, this was not statistically significant. CONCLUSION. The established qualitative criteria to evaluate students' work showed sufficiently high inter-rater reliability; however, the digitally measured volumetric discrepancy could not sufficiently predict the total qualitative score.