http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
( Ju Hyun Oh ),( Grace Hyun J. Kim ),( Jonathan G Goldin ),( Jooae Choe ),( Jin Woo Song ) 대한결핵 및 호흡기학회 2021 대한결핵 및 호흡기학회 추계학술대회 초록집 Vol.129 No.-
Background Interstitial lung abnormality (ILA) refers to incidental findings of parenchymal abnormalities suggestive of early interstitial lung disease (ILD) affecting >5% of lungs on CT scan. We aimed to determine the prevalence and progression rate of ILA evaluated by automated quantification system (AQS). Methods A total of 2890 subjects (mean age: 49 years, male: 79.4%), who participated in a health screening program, and had serial chest CT images (median interval: 6.5 years), were included. The quantitative lung fibrosis (QLF) and quantitative ILD (QILD, sum of QLF, honeycombing and ground glass opacity) were measured by AQS. ILA was defined as QILD ≥5 and QLF ≥3, and progression as an increase in QLF changes compared to baseline CT images. Results In the baseline scan, ILA was identified in 251 participants (8.6%). Those with ILA showed older age, higher body mass index (BMI), and low-density lipoprotein level than those without. The prevalence of ILA increased from 2.9% to 19.2% with age (Figure 1a). During follow-up, 21.1% (53/251) of participants initially identified with ILA progressed, while improvement or no change was noted in 78.9%. ILA was identified in 13.4% (387/2890) in follow-up CT images. Those who newly developed ILA were 11% (290/2639) of those who didn’t have ILA on initial CT images. Older age (HR: 1.026, 95% CI: 1.011-1.041) and higher BMI (HR: 1.056, 95%CI: 1.008-1.107) were independent risk factors for ILA development. When comparing the degree of QLF increase, the ILA group on the initial CT images showed a larger increase than the no-ILA group (Figure 1b). Conclusions ILA was not uncommon in the Korean population, with an increased prevalence in the group of subjects followed for 6.5-years. Progression was noted in 21.1% of the initially identified ILA cohort. Older age and higher BMI were risk factors for ILA development.
( Ju Hyun Oh ),( Grace Hyun J. Kim ),( David W Dai ),( S Sam Weigt ),( Jonathan G Goldin ),( Lila Pourzand ),( Jooae Choe ),( Fereidoun Abtin ),( Matthew S. Brown ),( Pangyu Teng ),( Jin Woo Song ) 대한결핵 및 호흡기학회 2021 대한결핵 및 호흡기학회 추계학술대회 초록집 Vol.129 No.-
Background Interstitial lung disease (ILD) includes a heterogeneous group of disease entities. Idiopathic pulmonary fibrosis (IPF) is ultimately fatal, and accurate diagnosis of IPF is critical to clinical decision making. Visual interpretation of chest high resolution CT (HRCT) is subjective and has limited reproducibility, especially with early disease. So we have previously developed attention-gated deep learning algorithm to diagnosis IPF and machine learning to predict IPF progression. The overall aim of IS-IPF is to collect the data from two centers of excellence and evaluate the robustness of the algorithm. We present the preliminary data of the patients studied following disease classification by the multidisciplinary review committees (MDCs) at UCLA and Asan Medical Center (AMC). Methods The IS-IPF study plans to include 234 IPF and 266 non-IPF cases from two large ILD centers (UCLA and AMC). Eligible patients were evaluated in ILD MDC, were >18 years old, had a HRCT, pulmonary function testing, and a committee diagnosis of IPF or non-IPF. Relevant demographic information was collected from the medical record. Results Total 185 IPF and 266 non-IPF patients’ HRCT images have been collected in the IS-IPF study. By center, 51 IPF and 133 non- IPF patients’ HRCT were collected from UCLA, and 134 IPF and 133 non-IPF patients’ HRCT were collected from AMC. On MDC diagnosis, non-IPF cohorts consisted of 33% hypersensitivity pneumonitis, and 67% other connective tissue disease-ILD. Mean age was 61 years (63 IPF and 58 non-IPF), and 63% were male (82% IPF and 57 % non-IPF). Up to date, the predicted FVC was 74.7% and the predicted DLco was 61.6 % in the IPF cohort. Data collection is on-going. Conclusions These well-characterized cohorts will be used to evaluate HRCT image signatures for distinguishing IPF from other ILD, and predicting patient-specific IPF progression within 2 years of diagnosis.