Development of a daily predictive model for the exacerbation of chronic obstructive pulmonary disease.

Acute exacerbation (AE) of chronic obstructive pulmonary disease (COPD) compromises health status; it increases disease progression and the risk of future exacerbations. We aimed to develop a model to predict COPD exacerbation. We merged the Korean COPD subgroup study (KOCOSS) dataset with nationwide medical claims data, information regarding weather, air pollution, and epidemic respiratory virus data. The Korean National Health and Nutrition Examination Survey (KNHANES) dataset was used for validation. Several machine learning methods were employed to increase the predictive power. The development dataset consisted of 590 COPD patients enrolled in the KOCOSS cohort; these were randomly divided into training and internal validation subsets on the basis of the individual claims data. We selected demographic and spirometry data, medications for COPD and hospital visit for AE, air pollution data and meteorological data, and influenza virus data as contributing factors for the final model. Six machine learning and logistic regression tools were used to evaluate the performance of the model. A light gradient boosted machine (LGBM) afforded the best predictive power with an area under the curve (AUC) of 0.935 and an F1 score of 0.653. Similar favorable predictive performance was observed for the 2151 individuals in the external validation dataset. Daily prediction of the COPD exacerbation risk may help patients to rapidly assess their risk of exacerbation and will guide them to take appropriate intervention in advance. This might lead to reduction of the personal and socioeconomic burdens associated with exacerbation.

Prediction of Postoperative Lung Function in Lung Cancer Patients Using Machine Learning Models.

BACKGROUND: Surgical resection is the standard treatment for early-stage lung cancer. Since postoperative lung function is related to mortality, predicted postoperative lung function is used to determine the treatment modality. The aim of this study was to evaluate the predictive performance of linear regression and machine learning models. METHODS: We extracted data from the Clinical Data Warehouse and developed three sets: set I, the linear regression model; set II, machine learning models omitting the missing data: and set III, machine learning models imputing the missing data. Six machine learning models, the least absolute shrinkage and selection operator (LASSO), Ridge regression, ElasticNet, Random Forest, eXtreme gradient boosting (XGBoost), and the light gradient boosting machine (LightGBM) were implemented. The forced expiratory volume in 1 second measured 6 months after surgery was defined as the outcome. Five-fold cross-validation was performed for hyperparameter tuning of the machine learning models. The dataset was split into training and test datasets at a 70:30 ratio. Implementation was done after dataset splitting in set III. Predictive performance was evaluated by R2 and mean squared error (MSE) in the three sets. RESULTS: A total of 1,487 patients were included in sets I and III and 896 patients were included in set II. In set I, the R2 value was 0.27 and in set II, LightGBM was the best model with the highest R2 value of 0.5 and the lowest MSE of 154.95. In set III, LightGBM was the best model with the highest R2 value of 0.56 and the lowest MSE of 174.07. CONCLUSION: The LightGBM model showed the best performance in predicting postoperative lung function.

Association Between Quality of Life Questionnaire at Diagnosis and Survival in Patients With Lung Cancer.

BACKGROUND: Patients with lung cancer experience considerable symptom burden, which can decrease patients' QOL. Our aim was to investigate the association between QOL questionnaire at diagnosis and survival of lung cancer. PATIENTS AND METHODS: This was a multicenter study of lung cancer patients at 7 medical centers of the Catholic University of Korea that responded to a quality of life questionnaire between December 1, 2017 and December 31, 2020. We analyzed 5 functional (physical, role, emotional, cognitive, and social functioning) and nine symptom (fatigue, nausea and vomiting, pain, dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties) scales and examined their associations with survival. A Cox proportional hazards model was used to evaluate the prognostic value. RESULTS: In total, 1297 lung cancer patients were enrolled. The results of multivariable analysis showed that female, younger age, never smoker, stage I or II cancer, higher physical functioning, and emotional functioning were statistically significant favorable predictors for survival. On subgroup analysis according to early (stage I and II) or advanced (stage III or IV) stage, higher physical functioning and emotional functioning were each found to be favorable prognostic factors for survival. Meanwhile, fatigue, pain, insomnia, and financial difficulties were found to be associated with low scores on the emotional functioning scale; fatigue, pain, dyspnea, and financial difficulties were associated with low scores on the physical functioning scale. CONCLUSION: Assessing the physical functioning and emotional functioning scales of QOL questionnaire items at diagnosis can help clinicians predict the survival of patients with lung cancer.

The association between serum apolipoprotein B and fractional exhaled nitric oxide in bronchial asthma patients.

BACKGROUND: Asthma and lipid metabolism are associated with systemic inflammation. However, the studies about the relationship between lipid profile, fractional exhaled nitric acid (FeNO) and pulmonary function test (PFT) results are currently lacking. METHODS: We enrolled asthma patients who had serum lipid profiles including apolipoprotein levels from March 1, 2019 to December 31, 2019. We classified the asthma patients into two groups according to the diagnosis method: (I) patients who were diagnosed based on clinical symptoms/signs and PFT results and (II) patients diagnosed with clinical symptoms/signs. Clinical characteristics including age, underlying diseases, smoking status, allergy test results and treatment agents were compared between the two groups. The associations between blood cholesterol levels including apolipoprotein and pulmonary functions were analyzed. Moreover, patients were divided into two groups according to the median value of apolipoprotein B (Apo B), and lung function test results were compared between the patients who had high and low Apo B levels. RESULTS: Among the 167 patients, 93 (55.7%) were PFT-proven asthma patients. In PFT-proven asthma patients, the levels of total cholesterol (TC) (r =0.37, P=0.03), low-density lipoprotein (LDL) (r =0.46, P=0.01) and Apo B (r =0.38, P=0.02) showed a significant correlation with FeNO, which had no statistical significance in physician-diagnosed asthma group. In multivariate regression analysis, log (FeNO) showed a significant correlation with Apo B (P<0.01) after adjustment for presence of PFT-proven asthma (P=0.01) and current smoking (P=0.01). Patients with high Apo B levels had a lower post-bronchodilator (BD) forced expiratory volume in 1 second (FEV1)/forced vital capacity (FVC) ratio (69.8 vs. 74.9, P=0.02) and lower post-BD FEV1 (%) (77.5 vs. 85.0, P=0.04) compared with those showing low Apo B levels. CONCLUSIONS: The levels of Apo B and FeNO had positive correlations and high Apo B levels were associated with severe airflow obstruction and low FEV1 (%). Apo B could reflect the uncontrolled status of bronchial asthma and poor lung function.

Characteristics of female lung cancer in Korea: analysis of Korean National Lung Cancer Registry.

BACKGROUNDS: The present study evaluated Korean women with lung cancer and compared the clinical characteristics of ever-smoker and never-smoker groups using the National Lung Cancer Registry. METHODS: In affiliation with the Korean Central Cancer Registry, the Korean Association for Lung Cancer constructed a registry into which 10% of the lung cancer cases in Korea were registered. Female lung cancer patients with valid smoking history were evaluated. RESULTS: Among 735 female lung cancer patients, 643 (87.5%) were never-smokers and 92 (12.5%) were smokers. The median survival was significantly longer in the never-smoker group (28 vs. 14 months; P<0.001). Among 683 patients with non-small cell lung cancer (NSCLC), the never-smoker group showed significantly longer median survival (29 vs. 14 months; P=0.002) and a higher proportion of stage I cancer (40.3% vs. 25.7%; P<0.001). Survival analysis of the NSCLC patients showed that smoking status, receiving only supportive care, EGFR mutation status, lung cancer stage, and forced vital capacity (FVC) (%) were significantly associated with mortality in the multivariate analysis (P=0.025, HR 2.39, 95% CI: 1.12-5.11; P=0.017, HR 3.14, 95% CI: 1.22-8.06; P=0.033, HR 0.63, 95% CI: 0.41-0.96; P<0.001, HR 11.88, 95% CI: 5.79-24.38; P=0.002, HR 0.98, 95% CI: 0.96-0.99, respectively). CONCLUSIONS: In Korean women with NSCLC, smoking status, not receiving active anticancer treatment, EGFR mutation status, lung cancer stage, and pulmonary function were significantly associated with mortality.

Impact of smoking amount on clinicopathological features and survival in non-small cell lung cancer.

BACKGROUND: Screening for early detection of lung cancer has been performed in high-risk individuals with smoking history. However, researches on the distribution, clinical characteristics, and prognosis of these high-risk individuals in an actual cohort are lacking. Thus, the objective of this study was to retrospectively review characteristics and prognosis of patients with smoking history in an actual lung cancer cohort. METHODS: The present study used the lung cancer cohort of the Catholic Medical Centers at the Catholic University of Korea from 2014 to 2017. Patients with non-small cell lung cancer were enrolled. They were categorized into high and low-risk groups based on their smoking history using the national lung screening trial guideline. Distribution, clinical characteristics, and survival data of each group were estimated. RESULTS: Of 439 patients, 223 (50.8%) patients were in the high-risk group. Patients in the high-risk group had unfavorable clinical characteristics and tumor biologic features. Overall survival of the high-risk group was significantly shorter than that of the low-risk group with both early (I, II) and advanced stages (III, IV). In multivariate analysis, heavy smoking remained one of the most important poor clinical prognostic factors in patients with lung cancer. It showed a dose-dependent relationship with patients' survival. CONCLUSIONS: High-risk individuals had poor clinical outcomes. Patients' prognosis seemed to be deteriorated as smoking amount increased. Therefore, active screening and clinical attention are needed for high-risk individuals.

Metabolic health is more closely associated with decrease in lung function than obesity.

OBJECTIVE: Previous studies have evaluated the link between metabolic syndrome and obesity with impaired lung function, however findings have been controversial. We aimed to compare lung function among subjects with different metabolic health and obesity status. METHODS: Total 10,071 participants were evaluated at the Health Promotion Center in Seoul St. Mary's Hospital between January 2012 and December 2014. Being metabolically healthy was defined as having fewer than three of the following risk factors: high blood pressure, high fasting blood glucose, high triglyceride, low high-density lipoprotein cholesterol and abdominal obesity. Obesity status was defined as body mass index (BMI) higher than 25 kg/m2. Analyses of pulmonary function were performed in four groups divided according to metabolic health and obesity: metabolically healthy non-obese (MHNO), metabolically health obese (MHO), metabolically unhealthy non-obese (MUHNO), and metabolically unhealthy obese (MUHO). RESULTS: Metabolically unhealthy subjects were more prone to decreased lung function compared with their metabolically healthy counterparts, regardless of obesity status. When multinomial logistic regression analysis was performed according to quartiles of forced vital capacity (FVC) or forced expiratory volume in 1 second (FEV1) (% pred), after adjusting for age, sex, and smoking status, odds ratio (OR) for the lowest FVC and FEV1 (% pred) quartiles were significantly higher in MUHO subjects (1.788 [95% CI, 1.531-2.089] and 1.603 [95% CI, 1.367-1.881]) and lower in MHO subjects (0.768 [95% CI, 0.654-0.902] and 0.826 [95% CI, 0.700-0.976]) with MHNO group as the reference, when OR for highest FVC and FEV1 quartiles were considered as 1.0. CONCLUSION: Metabolic health is more closely associated with impaired lung function than obesity.