Chest CT-Derived Body Composition and Systemic Inflammation for Predicting One-Year Mortality in Hospitalised Acute Exacerbations of COPD: A Retrospective Cohort Study
- 1Chest Diseases
- 2Radiology
Abstract
Abdurrahman Koç,1 Abdullah Enes Ataş,2 Hasan Eryeşil3 1Department of Chest Diseases, Meram State Hospital, Konya, Türkiye; 2Department of Radiology, Necmettin Erbakan University, Konya, Türkiye; 3Department of Radiology, Tatvan State Hospital, Bitlis, TürkiyeCorrespondence: Abdurrahman Koç, Department of Chest Diseases, Meram State Hospital, Ayanbey Mh., Yeni Meram Cd, No: 97, 42090 Meram, Konya, Türkiye, Tel +905058512151, Email [email protected]: Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) carry substantial one-year mortality. Systemic inflammatory indices and computed tomography (CT)-derived body-composition measures each hold prognostic value but have not been combined. We examined whether integrating them improves prediction of one-year all-cause mortality in hospitalised AECOPD patients and quantified the incremental value of the radiological axis.Patients and Methods: In this single-centre retrospective cohort, 380 consecutive AECOPD patients admitted to a secondary-level hospital in Türkiye (May 2024–November 2025) were included; 336 with complete one-year follow-up formed the primary cohort. Inflammatory indices (C-reactive protein [CRP], albumin, LDH-to-albumin ratio [LAR], and others) and CT parameters (pectoralis muscle area [PMA], vertebral attenuation, pulmonary artery-to-aorta ratio) were evaluated. Predictors were selected by LASSO regression with stepwise refinement; nested logistic models were compared by the DeLong test on a single common analytic sample, and a nomogram was internally validated with 1000 bootstrap resamples.Results: Eighty-two patients (24.4%) died within one year. Cross-domain correlations between inflammatory indices and radiological parameters were uniformly weak (|r| < 0.21). In the common analytic sample (n = 272), a parsimonious model (age, male sex, ICU admission, albumin, CRP, LAR, PMA) reached an AUC of 0.849 (95% CI 0.793– 0.904), exceeding the radiological (0.810) and clinical (0.740) models; the increment of the full combined model over the inflammatory model was modest (ΔAUC +0.029; p = 0.029). In the model’s own complete-case cohort (n = 318) the apparent AUC was 0.864 (optimism-corrected C-index 0.846; Hosmer–Lemeshow p = 0.539). Nomogram tertiles stratified observed one-year survival from 98% to 42% (log-rank p < 0.001).Conclusion: Adding CT-derived body composition to a routine clinical–inflammatory panel produced a small but measurable gain in discrimination. Pending external validation, the tool may aid admission risk stratification, most usefully in patients whose serum albumin is already perturbed.Graphical Abstract: An infographic summarizing a seven-variable model and one-year risk groups after AECOPD admission.Top row summary: 336 patients hospitalised for an acute exacerbation of COPD; 24.4 percent one-year all-cause mortality; 7 predictors, all available at admission; 1000 bootstrap resamples internal validation. Routine tests at admission: Routine chest CT, Body-composition axis. Pectoralis muscle area (PMA). Vertebral attenuation. PA/A ratio. Note: largely independent axes (|r| less than 0.21). Admission bloods, Systemic-inflammation axis. Albumin. CRP. LAR. NLR. SII. MLR. CALLY. Seven-variable model: Parsimonious model. logistic regression, LASSO-selected. Predictor labels: Age. Sex. ICU. Albumin. PMA. CRP. LAR. Performance metrics: 0.846 optimism-corrected C-index. 0.119 Brier score (calibration). One-year risk groups: Observed one-year survival by model-based risk group (tertiles). Low 98 percent. Moderate 87 percent. High 42 percent. Text: log-rank p less than 0.001. Callout: Most informative when admission albumin is low (less than 35 g per L), external validation pending. Bottom statement: Adding routine chest CT body composition to systemic inflammatory markers gave a small but measurable gain in predicting one-year mortality in hospitalised AECOPD. Metric box: plus 0.029. AUC gain from adding the CT axis 0.836 to 0.865 (DeLong p equals 0.029).Keywords: prognostic model, pectoralis muscle area, systemic inflammation, sarcopenia, risk stratification, nomogram
Related
- Implementing an integrated inpatient–community rehabilitation pathway for adults hospitalised with acute exacerbations of COPD: a quality improvement study
- Machine-learning model for distinguishing cytomegalovirus infection status in hospitalised patients with systemic lupus erythematosus: a retrospective cohort study with temporal validation.
- Psychiatric Characteristics and Mental Health Service Use Among Hospitalised Youth in Foster Care, Adoption, and Child Welfare Supervision: A Retrospective Comparative Study
- Dysphagia in infective and non-infective exacerbations of chronic obstructive pulmonary disease: a retrospective cohort study
- Nosocomial Infections in Hospitalised Haematological Patients: A Retrospective Observational Record‐Based Study
BibTeX
@article{koc2026chest,
title = {{Chest CT-Derived Body Composition and Systemic Inflammation for Predicting One-Year Mortality in Hospitalised Acute Exacerbations of COPD: A Retrospective Cohort Study}},
author = {Koç A and Ataş AE and Eryeşil H},
journal = {International Journal of COPD},
year = {2026},
month = nov,
url = {https://www.dovepress.com/chest-ct-derived-body-composition-and-systemic-inflammation-for-predic-peer-reviewed-fulltext-article-COPD},
note = {CC BY},
}
RIS
TY - JOUR
TI - Chest CT-Derived Body Composition and Systemic Inflammation for Predicting One-Year Mortality in Hospitalised Acute Exacerbations of COPD: A Retrospective Cohort Study
AU - Koç A
AU - Ataş AE
AU - Eryeşil H
PY - 2026
DA - 2026/11/15
JO - International Journal of COPD
UR - https://www.dovepress.com/chest-ct-derived-body-composition-and-systemic-inflammation-for-predic-peer-reviewed-fulltext-article-COPD
AB - Abdurrahman Koç,1 Abdullah Enes Ataş,2 Hasan Eryeşil3 1Department of Chest Diseases, Meram State Hospital, Konya, Türkiye; 2Department of Radiology, Necmettin Erbakan University, Konya, Türkiye; 3Department of Radiology, Tatvan State Hospital, Bitlis, TürkiyeCorrespondence: Abdurrahman Koç, Department of Chest Diseases, Meram State Hospital, Ayanbey Mh., Yeni Meram Cd, No: 97, 42090 Meram, Konya, Türkiye, Tel +905058512151, Email [email protected]: Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) carry substantial one-year mortality. Systemic inflammatory indices and computed tomography (CT)-derived body-composition measures each hold prognostic value but have not been combined. We examined whether integrating them improves prediction of one-year all-cause mortality in hospitalised AECOPD patients and quantified the incremental value of the radiological axis.Patients and Methods: In this single-centre retrospective cohort, 380 consecutive AECOPD patients admitted to a secondary-level hospital in Türkiye (May 2024–November 2025) were included; 336 with complete one-year follow-up formed the primary cohort. Inflammatory indices (C-reactive protein [CRP], albumin, LDH-to-albumin ratio [LAR], and others) and CT parameters (pectoralis muscle area [PMA], vertebral attenuation, pulmonary artery-to-aorta ratio) were evaluated. Predictors were selected by LASSO regression with stepwise refinement; nested logistic models were compared by the DeLong test on a single common analytic sample, and a nomogram was internally validated with 1000 bootstrap resamples.Results: Eighty-two patients (24.4%) died within one year. Cross-domain correlations between inflammatory indices and radiological parameters were uniformly weak (|r| < 0.21). In the common analytic sample (n = 272), a parsimonious model (age, male sex, ICU admission, albumin, CRP, LAR, PMA) reached an AUC of 0.849 (95% CI 0.793– 0.904), exceeding the radiological (0.810) and clinical (0.740) models; the increment of the full combined model over the inflammatory model was modest (ΔAUC +0.029; p = 0.029). In the model’s own complete-case cohort (n = 318) the apparent AUC was 0.864 (optimism-corrected C-index 0.846; Hosmer–Lemeshow p = 0.539). Nomogram tertiles stratified observed one-year survival from 98% to 42% (log-rank p < 0.001).Conclusion: Adding CT-derived body composition to a routine clinical–inflammatory panel produced a small but measurable gain in discrimination. Pending external validation, the tool may aid admission risk stratification, most usefully in patients whose serum albumin is already perturbed.Graphical Abstract: An infographic summarizing a seven-variable model and one-year risk groups after AECOPD admission.Top row summary: 336 patients hospitalised for an acute exacerbation of COPD; 24.4 percent one-year all-cause mortality; 7 predictors, all available at admission; 1000 bootstrap resamples internal validation. Routine tests at admission: Routine chest CT, Body-composition axis. Pectoralis muscle area (PMA). Vertebral attenuation. PA/A ratio. Note: largely independent axes (|r| less than 0.21). Admission bloods, Systemic-inflammation axis. Albumin. CRP. LAR. NLR. SII. MLR. CALLY. Seven-variable model: Parsimonious model. logistic regression, LASSO-selected. Predictor labels: Age. Sex. ICU. Albumin. PMA. CRP. LAR. Performance metrics: 0.846 optimism-corrected C-index. 0.119 Brier score (calibration). One-year risk groups: Observed one-year survival by model-based risk group (tertiles). Low 98 percent. Moderate 87 percent. High 42 percent. Text: log-rank p less than 0.001. Callout: Most informative when admission albumin is low (less than 35 g per L), external validation pending. Bottom statement: Adding routine chest CT body composition to systemic inflammatory markers gave a small but measurable gain in predicting one-year mortality in hospitalised AECOPD. Metric box: plus 0.029. AUC gain from adding the CT axis 0.836 to 0.865 (DeLong p equals 0.029).Keywords: prognostic model, pectoralis muscle area, systemic inflammation, sarcopenia, risk stratification, nomogram
SN - 1178-2005
SN - 1176-9106
ER -