Deep Learning Prediction of Survival in Patients with Chronic Obstructive Pulmonary Disease Using Chest Radiographs.

Background Preexisting indexes for predicting the prognosis of chronic obstructive pulmonary disease (COPD) do not use radiologic information and are impractical because they involve complex history assessments or exercise tests. Purpose To develop and to validate a deep learning-based survival prediction model in patients with COPD (DLSP) using chest radiographs, in addition to other clinical factors. Materials and Methods In this retrospective study, data from patients with COPD who underwent postbronchodilator spirometry and chest radiography from 2011-2015 were collected and split into training (n = 3475), validation (n = 435), and internal test (n = 315) data sets. The algorithm for predicting survival from chest radiographs was trained (hereafter, DLSPCXR), and then age, body mass index, and forced expiratory volume in 1 second (FEV1) were integrated within the model (hereafter, DLSPinteg). For external test, three independent cohorts were collected (n = 394, 416, and 337). The discrimination performance of DLSPCXR was evaluated by using time-dependent area under the receiver operating characteristic curves (TD AUCs) at 5-year survival. Goodness of fit was assessed by using the Hosmer-Lemeshow test. Using one external test data set, DLSPinteg was compared with four COPD-specific clinical indexes: BODE, ADO, COPD Assessment Test (CAT), and St George's Respiratory Questionnaire (SGRQ). Results DLSPCXR had a higher performance at predicting 5-year survival than FEV1 in two of the three external test cohorts (TD AUC: 0.73 vs 0.63 [P = .004]; 0.67 vs 0.60 [P = .01]; 0.76 vs 0.77 [P = .91]). DLSPCXR demonstrated good calibration in all cohorts. The DLSPinteg model showed no differences in TD AUC compared with BODE (0.87 vs 0.80; P = .34), ADO (0.86 vs 0.89; P = .51), and SGRQ (0.86 vs 0.70; P = .09), and showed higher TD AUC than CAT (0.93 vs 0.55; P < .001). Conclusion A deep learning model using chest radiographs was capable of predicting survival in patients with chronic obstructive pulmonary disease. © RSNA, 2022 Online supplemental material is available for this article.

Pulmonary sclerosing pneumocytoma and mortality risk.

BACKGROUND: Surgical resection is usually recommended for the treatment of pulmonary sclerosing pneumocytoma (PSP). However, no comparative study has demonstrated that surgical resection leads to improved outcomes. We aimed to compare all-cause mortality between patients with PSP who underwent surgery or did not and those without PSP. METHODS: Participants aged ≥18 years who had pathologically diagnosed PSP between 2001 to 2018, at 3 hospitals were included. Randomly selected (up to 1:5) age-, sex-, and smoking status-matched controls without PSP who were randomly selected from those who underwent health checkups including chest CT were included. Mortality was compared using Kaplan-Meier estimates and Cox proportional hazards regression models. Literature review of studies reporting PSP was also conducted. RESULTS: This study included 107 patients with PSP (surgery:non-surgery, 80:27) and 520 matched controls. There were no cases of lymph node or distant metastasis, recurrence, or mortality from PSP. No significant difference in all-cause mortality risk was observed between the PSP surgery, PSP non-surgery, and non-PSP groups (log rank test P = 0.78) (PSP surgery vs. non-PSP: adjusted hazards ratio [aHR], 1.80; 95% confidence interval [CI], 0.22-14.6; PSP non-surgery vs. non-PSP: aHR, 0.77; 95% CI, 0.15-3.86; PSP surgery vs. PSP non-surgery: aHR, 2.35; 95% CI, 0.20-28.2). In the literature review, we identified 3469 patients with PSP from 355 studies. Only 1.33% of these patients reported metastasis, recurrence, or death. CONCLUSIONS: All-cause mortality did not differ between patients with PSP and those without, irrespective of undergoing surgery. Our study and the literature review suggest that PSP has less impact on increased mortality risk.

Usefulness of Post-bronchoscopy Sputum Culture for Diagnosis of Nontuberculous Mycobacterial Pulmonary Disease.

BACKGROUND: Bronchoscopy is recommended for patients with suspected nontuberculous mycobacterial pulmonary disease (NTM-PD) whose sputum culture results are consistently negative or from whom adequate sputum samples cannot be obtained. Post-bronchoscopy sputum (PBS) collection is recommended for patients with suspected tuberculosis who undergo bronchoscopy. However, it remains unclear whether PBS collection can increase the diagnostic yield of NTM-PD. METHODS: Patients with suspected NTM-PD who underwent diagnostic bronchoscopy from January 1, 2017 to June 30, 2020 at the Seoul National University Hospital were included in the study. They were divided into the sputum culture-negative and scanty sputum groups. The results of mycobacterial cultures from bronchial washing specimens and PBS were compared between these groups. RESULTS: In total, 141 patients were included in the study; there were 39 and 102 patients in the sputum culture-negative and scanty sputum groups, respectively. Nontuberculous mycobacteria were cultured from bronchial washing specimens collected from 38.3% (54/141) of all patients (30.7% [12/39] patients in the sputum culture-negative group and 41.2% [42/102] patients in the scanty sputum group; P = 0.345). Nontuberculous mycobacteria were exclusively cultured from PBS collected from 3.5% (5/141) of all patients (7.7% [3/39] patients in the sputum culture-negative group and 2.0% [2/102] patients in the scanty sputum group; P = 0.255). CONCLUSIONS: Additional PBS collection improved diagnostic yield marginally in patients with suspected NTM-PD who undergo bronchoscopy.

The Impact of Erosive Reflux Esophagitis on the Decline of Lung Function in the General Population.

BACKGROUND: The impact of reflux esophagitis on the decline of lung function has been rarely reported. This study was performed to evaluate the association between erosive reflux esophagitis and lung function changes. METHODS: We included patients with normal lung function who underwent esophagogastroduodenoscopy for health screening from a health screening center. Patients with persistent erosive reflux esophagitis on two discrete endoscopic examinations were designated as the erosive reflux esophagitis group. We also selected patients without erosive reflux esophagitis and matched them 1:4 with patients from the erosive reflux esophagitis group. We estimated annual forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC) changes from baseline and compared these estimates by the linear mixed regression model. We also estimated the biannual incidence of chronic obstructive pulmonary disease (COPD). RESULTS: In total, 1,050 patients (210 patients with erosive reflux esophagitis, and 840 matched controls) were included. The median follow-up duration for spirometry was six years. In patients with erosive reflux esophagitis, mild reflux esophagitis (A grade) was most common (165 patients, 78.6%). The adjusted annual FEV1 change in patients with erosive reflux esophagitis was -51.8 mL/yr, while it decreased by 46.8 mL/yr in controls (P = 0.270). The adjusted annual FVC decline was similar between the two groups (-55.8 vs. -50.5 mL/yr, P = 0.215). The estimated COPD incidence during the follow-up period was not different between the erosive reflux esophagitis and control groups. CONCLUSION: In patients with normal lung function, the presence of erosive reflux esophagitis did not affect the annual declines in FEV1 or FVC.

Low-dose chest computed tomographic screening and invasive diagnosis of pulmonary nodules for lung cancer in never-smokers.

BACKGROUND: Although lung cancer screening using low-dose computed tomography (LDCT) is now widely used in clinical practice, the characteristics and outcomes of diagnostic procedures related to screen-detected nodules in never-smokers remain unclear. We aimed to determine the incidence of nodules considered for invasive biopsy and evaluate the final diagnoses and procedure-related complications in never-smokers in comparison to ever-smokers who underwent LDCT screening. METHODS: We evaluated 37 436 asymptomatic adults (17 968 never-smokers and 19 468 ever-smokers) who underwent LDCT screening for lung cancer between January 2009 and December 2018 at a tertiary centre in South Korea. The rates of invasive diagnostic procedures for detected nodules and related complications, and the diagnostic outcomes were determined in the never-smoker and ever-smoker groups. RESULTS: Among the never-smokers, 2908 (16.2%) out of 17 968 had positive nodules. Overall, 139 (0.77%) out of 17 968 never-smokers and 194 (1.00%) out of 19 468 ever-smokers underwent invasive biopsy (p=0.022). Lung cancer was diagnosed in 84 (0.47%) out of 17 968 never-smokers and 123 (0.63%) out of 19 468 ever-smokers (p=0.032). The proportions of participants diagnosed with benign disease after invasive biopsy (false-positive) were 50 (0.28%) out of 17 968 and 69 (0.35%) out of 19 468 in the never-smoker and ever-smoker groups, respectively (p=0.191). Multivariate analyses revealed no significant associations of smoking with the risk of a false-positive diagnosis (OR 0.98, 95% CI 0.62-1.57) and complications (OR 1.33, 95% CI 0.65-3.73) after biopsy. Of the 84 never-smokers with lung cancer, 82 (97.6%) had adenocarcinoma, and 75 (89.3%) were in stage I with a favourable prognosis. CONCLUSIONS: LDCT screening in never-smokers resulted in a notable detection rate of lung nodules, which warranted invasive biopsy. The lung cancer detection rate was lower in never-smokers than in ever-smokers. However, no significant differences in the false-positive and complication rates were observed between the two groups. Accordingly, a more specifically tailored management strategy is needed for screen-detected nodules in Asian never-smokers.

The impact of incidental pulmonary cement embolism on mortality risk.

The prognosis of pulmonary cement embolism found incidentally on chest radiography after vertebroplasty has been rarely investigated. This study was performed to elucidate the impact of incidentally found pulmonary cement embolism on all-cause mortality. Patients with pulmonary cement embolism diagnosed using chest radiography between 2008 and 2014 at one tertiary referral hospital were included. Their mortality risk was compared to that of randomly selected, age-, sex-, and year-matched patients without pulmonary cement embolism (ratio, 1:10) by using Kaplan-Meier estimates and covariate-adjusted Cox proportional regression analysis. The study included 11 patients with pulmonary cement embolism and 110 patients without pulmonary cement embolism. The patients showed no significant intergroup differences in baseline characteristics, except comorbid heart failure. During a mean follow-up duration of 1.7 ± 1.6 years, five patients (45.5%) with pulmonary cement embolism and 60 (50.0%) without pulmonary cement embolism died, mostly because of underlying malignancy. Although the patients with pulmonary cement embolism were not treated, they did not show a higher mortality risk than did those without pulmonary cement embolism (adjusted hazard ratio, 1.10; 95% confidence interval, 0.43-2.85). Subgroup analyses showed similar results. Incidentally found pulmonary cement embolism had no significant impact on all-cause mortality.

Pulmonary Histoplasmosis Identified by Video-Assisted Thoracic Surgery (VATS) Biopsy: a Case Report.

Histoplasmosis is a common endemic mycosis in North, Central, and South America, but Korea is not known as an endemic area. We treated an immunocompetent Korean patient who had histoplasmosis. A 65-year-old Korean man presented with multiple pulmonary clumps of tiny nodules in the both lungs. He had been diagnosed 40 years earlier with pulmonary tuberculosis (TB) and a fungus ball had been diagnosed 4 years earlier. He denied any history of overseas travel. The patient visited our hospital with dyspnea, blood-tinged sputum, and weight loss, which had appeared 2 months earlier. The patient underwent video-assisted thoracic surgery (VATS) lung biopsy. The biopsy sample showed necrotizing granuloma and the presence of multiple small yeast-like fungi. Tissue culture confirmed Histoplasma capsulatum, and he was finally diagnosed with pulmonary histoplasmosis. Therapy was initiated with 200 mg itraconazole orally once per day. The symptoms disappeared 1 week after the start of treatment. After 4 months, low-dose chest computed tomography showed improvement in the ground glass opacity and size of the lung lesions. In conclusion, we report a case of an immunocompetent patient who developed histoplasmosis in Korea. When a patient shows unexplainable progressive infiltrative lung lesions, histoplasmosis should be considered as one of differential diagnoses although Korea is not an endemic area.

Impact of Smoking and Chronic Obstructive Pulmonary Disease on All-Cause, Respiratory, and Cardio-Cerebrovascular Mortality.

Introduction: Mortality differences in chronic obstructive pulmonary disease (COPD) between nonsmokers and smokers remain unclear. We compared the risk of death associated with smoking and COPD on mortality. Methods: The study included participants aged ≥40 years who visited pulmonary clinics and were categorised into COPD or non-COPD and smoker or nonsmoker on the basis of spirometry results and cigarette consumption. Mortality rates were compared between groups using statistical analysis for all-cause mortality, respiratory disease-related mortality, and cardiocerebrovascular disease-related mortality. Results: Among 5811 participants, smokers with COPD had a higher risk of all-cause (adjusted hazard ratio (aHR), 1.69; 95% confidence interval (CI), 1.23-2.33) and respiratory disease-related mortality (aHR, 2.14; 95% CI, 1.20-3.79) than nonsmokers with COPD. Non-smokers with and without COPD had comparable risks of all-cause mortality (aHR, 1.39; 95% CI, 0.98-1.97) and respiratory disease-related mortality (aHR, 1.77; 95% CI, 0.85-3.68). However, nonsmokers with COPD had a higher risk of cardiocerebrovascular disease-related mortality than nonsmokers without COPD (aHR, 2.25; 95% CI, 1.15-4.40). Conclusion: The study found that smokers with COPD had higher risks of all-cause mortality and respiratory disease-related mortality compared to nonsmokers with and without COPD. Meanwhile, nonsmokers with COPD showed comparable risks of all-cause and respiratory mortality but had a higher risk of cardiocerebrovascular disease-related mortality compared to nonsmokers without COPD.

Impact of lung cancer screening with low-dose chest computed tomography on an older population: a retrospective cohort study.

Background: The older population is at high risk of lung cancer (LC). However, the importance of lung cancer screening (LCS) in this population is rarely investigated. Herein, we evaluated the effect of LCS with low-dose computed tomography (LDCT) in the older population. Methods: This retrospective cohort study was conducted in a single center and included patients aged 70-80 years who had undergone LCS with LDCT. They were categorized into the early 70s (70-74 years) and late 70s (75-80 years) groups based on their age. Using propensity score matching, the control group included patients with non-screening-detected LC from an LC cohort. LC detection, characteristics, and treatment were compared between the early and late 70s groups and between screening-detected LC and non-screening-detected LC. Results: The study included 1,281 participants who underwent LDCT for LCS, of whom 1,020 were in their early 70s and 261 in their late 70s. Among the screening groups, 87.7% of the patients were ever-smokers. The overall LC detection rate was 2.8%. Interestingly, the LC detection rate in the late 70s group was similar to that in the early 70s group (3.4% vs. 2.7%, P=0.485). Furthermore, the incidence of LC was 6.1 cases and 8.3 cases per 1,000 person-years in the early 70s and late 70s groups, respectively (P=0.428). When comparing LC characteristics, patients with screening-detected LC showed a higher proportion of stage I LC (52.8% vs. 30.6%, P=0.010) and a lower proportion of stage IV LC (19.4% vs. 42.2%, P=0.010) than those with non-screening-detected LC. Moreover, 80.6% of patients with screening-detected LC received appropriate tumor reduction treatment based on the cancer stage. Conclusions: In the older population, LCS using LDCT showed remarkable detection of LC, with a higher proportion of cases detected at an early stage.

Clinical outcomes of mild to moderate coronavirus disease 2019 patients treated with Regdanvimab in delta-variant outbreak: Retrospective cohort study.

Regdanvimab is a novel neutralizing antibody agent used for the treatment of coronavirus disease 2019 (COVID-19). However, the effectiveness of regdanvimab in delta-variant patients has rarely been investigated. We examined the clinical outcomes and adverse events in COVID 19 patients treated with regdanvimab in the delta-variant era. Data were collected from laboratory-confirmed COVID-19 hospitalized patients who received regdanvimab in 2021 and categorized into pre-delta and delta variant groups. The primary outcome was the need for oxygen therapy. Rescue therapy, clinical improvement, and adverse events were analyzed. Among 101 patients treated with regdanvimab, 31 (30.7%) were delta patients and 49 (48.5) were pre-delta patients. 64.4% were male, the mean age was 60.3 years, and 70 patients (69%) had at least one underlying disease. The median interval from symptom onset to injection was 4 days. Twenty-three patients (23%) needed oxygen therapy, including 9 (29%) in the delta and 8 (16.3%) in the pre-delta group. (P = .176) The risk of early oxygen supplement was higher in the delta group (adjusted hazard ratio (aHR), 6.75; 95% confidence interval(CI), 1.53-29.8). The in-hospital survival rate was 100%, and no patients were admitted to the intensive care unit. Adverse events occurred in 43% of patients:13 (42%) delta patients and 23 (47%) pre-delta patients had any adverse events (P = .661). Patients treated with regdanvimab 4 days after symptom onset showed a favorable prognosis (aHR, 0.26; 95% CI, 0.26-0.91). We found that the high-risk mild to moderate COVID-19 patients treated with regdanvimab showed similar disease progression in delta-variant patients and pre-delta variants; however, we need to be more closely observed delta-variant patients than those in the pre-delta group despite regdanvimab treatment due to rapid disease aggravation.

Clinical Implications of Size of Cavities in Patients With Nontuberculous Mycobacterial Pulmonary Disease: A Single-Center Cohort Study.

BACKGROUND: The presence of cavities is associated with unfavorable prognosis in patients with nontuberculous mycobacterial pulmonary disease (NTM-PD). However, little is known about the characteristics of such cavities and their impact on clinical outcomes. The aim of this study was to investigate the size of cavities and their implications on treatment outcomes and mortality in patients with NTM-PD. METHODS: We included patients diagnosed with NTM-PD at Seoul National University Hospital between January 1, 2007, and December 31, 2018. We measured the size of cavities on chest computed tomography scans performed at the time of diagnosis and used multivariable logistic regression and Cox proportional hazards regression analysis to investigate the impact of these measurements on treatment outcomes and mortality. RESULTS: The study cohort comprised 421 patients (noncavitary, n = 329; cavitary, n = 92) with NTM-PD. During a median follow-up period of 49 months, 118 (35.9%) of the 329 patients with noncavitary and 64 (69.6%) of the 92 patients with cavitary NTM-PD received antibiotic treatment. Cavities >2 cm were associated with worse treatment outcomes (adjusted odds ratio, 0.41; 95% CI, 0.17-0.96) and higher mortality (adjusted hazard ratio, 2.52; 95% CI, 1.09-5.84), while there was no difference in treatment outcomes or mortality between patients with cavities ≤2 cm and patients with noncavitary NTM-PD. CONCLUSIONS: Clinical outcomes are different according to the size of cavities in patients with cavitary NTM-PD; thus, the measurement of the size of cavities could help in making clinical decisions.

The effect of Helicobacter pylori infection on the decline of lung function in a health screening population.

BACKGROUND: Although Helicobacter pylori (H. pylori) infection has long been associated with many extragastric diseases, its long-term effects on declining lung function and the development of chronic obstructive pulmonary disease (COPD) remains controversial. METHODS: We conducted a retrospective study in a health screening population (Gene-Environment of Interaction and Phenotype Cohort at Seoul National University Gangnam Health Center in Korea) recruited from 2004 to 2015. The annual decline rates of lung function and the risk for COPD were compared to the seropositivity for H. pylori. We also assessed the impact of the H. pylori eradication treatment on lung function. RESULTS: Among 3,619 subjects, 1,849 subjects (51.1%) tested positive for H. pylori-specific IgG. In the first year, 95 (2.6%) patients were diagnosed with COPD. however, there was no difference in seropositivity for H. pylori in patients with COPD compared to non-COPD subjects (P=0.756). There was no significant difference in the incidence of COPD between the seronegative (2.2%) and the seropositive group (2.0%; P=0.728) even during the follow-up period. The decline rates of the mean FVC and FEV1 were not different between the seronegative and seropositive group (35.38 vs. 34.34 mL/year for FVC, P=0.389; 39.23 vs. 37.49 mL/year for FEV1, P=0.086). The eradication treatment for H. pylori did not affect the decline rates of mean FVC and FEV1. CONCLUSIONS: Infection or eradication treatment of H. pylori was not associated with the decline of lung function or COPD progression in this general population health screening. As a result, H. pylori may not be a significant contributor to deteriorating lung function.

Role of Low-Dose Computerized Tomography in Lung Cancer Screening among Never-Smokers.

INTRODUCTION: The incidence of lung cancer among never-smokers has been increasing rapidly. The U. S. National Lung Screening Trial and the NELSON trial showed that screening using low-dose computerized tomography (LDCT) effectively reduced lung cancer mortality among heavy smokers. However, its effectiveness in never-smokers has not been well investigated. This study investigated the role of LDCT in lung cancer screening among never-smokers. METHODS: The study was designed as a single-center, retrospective cohort study. We analyzed the data on patients who underwent LDCT screening between May 2003 and June 2016. Nodules detected by computerized tomography were classified according to the Lung Imaging Reporting and Data System criteria. The detection rate and lung cancer outcomes (type of cancer, staging of lung cancer, and mortality) according to smoking history were determined. RESULTS: Of the 28,807 enrolled patients, 12,176 were never-smokers; of these patients, 7744 (63.6%) were women and 1218 (10.0%) were found to have lung nodules. Overall, lung cancer was diagnosed in 55 never-smokers (0.45%). In contrast, lung cancer was diagnosed in 143 (0.86%) of the 16,631 ever-smokers. Of the never-smokers with lung cancer, 51 (92.7%) presented with stage I disease, and all patients had adenocarcinomas. CONCLUSIONS: In the never-smoker population, LDCT screening helped to detect a significant number of lung cancers. Most of these lung cancers were detected at a very early stage. The positive results of the National Lung Screening Trial in the United States and the NELSON trial may have established the value of LDCT screening for heavy smokers, but future research should consider the value of using LDCT screening in the never-smoker population.

A Comparative Analysis of Survival Prediction Using PRESERVE and RESP Scores.

BACKGROUND: Venovenous (VV) extracorporeal membrane oxygenation (ECMO) can be a life-saving therapy for patients with severe acute lung failure refractory to conventional therapy. The respiratory ECMO survival prediction (RESP) score and the predicting death for severe acute respiratory distress syndrome on VV-ECMO (PRESERVE) score were created to predict survival at the time of initiation of ECMO. This study aimed to validate both of these scores externally and to compare their predictive accuracies in patients with non-Western acute respiratory distress syndrome (ARDS). METHODS: In this retrospective cohort study, we reviewed and extracted data from electronic medical records of consecutive adult ARDS patients who were treated with VV-ECMO from 2007 to 2015. The PRESERVE and RESP scores were calculated for each patient. The outcomes of interest were inhospital and 6-month survival. RESULTS: In all, 99 patients were included. The mean age of the patients was 54 years, and male patients constituted 70% of the cohort. The inhospital and 6-month survival rates were 23% and 22%, respectively. Receiver-operating characteristics curve analysis of the PRESERVE and RESP scores showed area under the curve values of 0.64 and 0.69, respectively (p = 0.53), for inhospital survival. The receiver-operating characteristics areas under the curve for 6-month survival were 0.66 and 0.69, respectively (p = 0.68). The prognostic accuracies of the PRESERVE and RESP scores were thus similar. CONCLUSIONS: Both PRESERVE and RESP scores are useful for predicting survival in Asian ARDS patients, and both scores had similar prognostic accuracies in our Korean ARDS cohort.