Computed tomography-derived cardiac parameters versus biochemical inflammatory indices for discriminating disease severity and identifying cardiovascular comorbidity in obstructive sleep apnoea: A cross-sectional study.
Koç A, Ataş A, Yosunkaya &, Vatansev H • Sleep & breathing = Schlaf & Atmung • 2026
CT-derived cardiac parameters, particularly PA diameter, added modest discrimination over age, sex and BMI for severe OSAS, whereas haematological inflammatory indices did not; findings are exploratory and require prospective external validation.
Key Findings
Results
Adding CT-derived cardiac parameters to age, sex and BMI improved discrimination of severe OSAS, while adding haematological inflammatory indices did not.
Retrospective cross-sectional study of 401 participants (240 OSAS, 161 controls); nested logistic regression models were fitted in the OSAS cohort.
M2 (M0 plus CT parameters) vs M0: AUC 0.755 vs 0.716; likelihood ratio p = 0.005.
M1 (M0 plus CAR, NLR, PLR, SII) did not improve model fit (LR p = 0.62).
Multiple imputation (MICE, m = 20) was the primary approach for missing predictor data.
Results
The CT model produced significant reclassification improvement over the clinical baseline model for severe OSAS.
Continuous NRI +0.60 (95% CI 0.35-0.83).
IDI +0.055 (95% CI 0.027-0.083).
Comparison was M2 versus M0 (age, sex, BMI).
Results
Combining CT and biochemical indices added nothing beyond CT alone for severe OSAS.
M3 vs M2: ΔAUC = 0.007; LR p = 0.63.
Results
After internal validation, the incremental gain from CT parameters was modest.
Optimism-corrected AUCs were 0.705, 0.695, 0.721 and 0.716 for M0, M1, M2 and M3 respectively.
Internal validation used 500-iteration bootstrap optimism correction and repeated tenfold cross-validation.
Results
Nocturnal hypoxaemia indices discriminated severe OSAS much better than CT parameters, and CT added little once they were available.
ODI AUC 0.888 versus PA diameter AUC 0.699.
Once ODI and T90 were included, CT added little (ΔAUC < 0.01).
Results
For cardiovascular comorbidity, CT parameters gave modest reclassification while biochemical indices did not.
Outcomes included hypertension, coronary artery disease and a composite cardiovascular outcome.
Cardiovascular outcomes were defined from echocardiography to avoid predictor-outcome overlap.
Conclusions
The authors conclude that opportunistic PA measurement on existing thoracic CT may complement clinical risk stratification, but results are exploratory.
PA diameter was the CT parameter highlighted as adding discrimination.
Findings require prospective external validation.
What This Means
This study asked whether blood-test inflammation markers or measurements of the heart and lung arteries taken from CT scans help identify severe obstructive sleep apnoea (OSAS) and related heart problems, beyond simple information like age, sex and body mass index. It looked back at records from 401 people (240 with OSAS and 161 controls).
The CT measurements, especially the width of the pulmonary artery, modestly improved the ability to identify severe OSAS. The blood inflammation markers did not help, and combining them with CT added nothing. After statistical correction for over-optimism, the gain from CT was small. Overnight oxygen measurements, such as how often oxygen levels dropped, identified severe OSAS far better than CT did, and when these were available CT added very little.
This research suggests that when a chest CT has already been done for another reason, measuring the pulmonary artery might give some extra information about sleep apnoea severity. It does not suggest CT should replace sleep-study oxygen measurements. The study was retrospective and exploratory, so the findings need confirmation in prospective studies at other centres.
Koç A, Ataş A, Yosunkaya &, Vatansev H. (2026). Computed tomography-derived cardiac parameters versus biochemical inflammatory indices for discriminating disease severity and identifying cardiovascular comorbidity in obstructive sleep apnoea: A cross-sectional study.. Sleep & breathing = Schlaf & Atmung. https://doi.org/10.1007/s11325-026-03820-x