Background and Objectives: Acute respiratory distress syndrome (ARDS) carries high mortality, with cardiovascular complications frequently contributing to adverse outcomes. This study investigated the relationship between cardiac auscultation using electronic stethoscopy and echocardiographic findings and evaluated their prognostic significance in mechanically ventilated ARDS patients. Materials and Methods: This prospective observational study enrolled 173 consecutive adults with ARDS requiring mechanical ventilation (June 2020–June 2021). Cardiac auscultation was performed using an electronic stethoscope at four standard valvular positions. Bedside echocardiography assessed ventricular function, valvular regurgitation, right ventricular systolic pressure (RVSP), and inferior vena cava dimensions. Primary outcomes were ICU and 90-day mortality; the secondary outcome was ICU length of stay. Results: ICU mortality was 42.2% and 90-day mortality 46.8%. Auscultation findings correlated significantly with echocardiographic parameters: aortic stenosis murmur with an elevated aortic valve velocity (p = 0.009), and mitral/tricuspid regurgitation murmurs with corresponding color Doppler findings (p < 0.001). In multivariate analysis, the mean daily SOFA score (OR 2.39, 95% CI 1.57–3.64, p < 0.001) and RVSP (OR 1.07, 95% CI 1.02–1.11, p = 0.006) independently predicted ICU mortality. For 90-day mortality, the APACHE II score (OR 1.25, p = 0.006), mean daily SOFA score (OR 1.54, p = 0.039), RVSP (OR 1.07, p = 0.020), and mitral regurgitation severity (OR 2.98, p = 0.031) were independent predictors. ICU length of stay was predicted by the mean daily SOFA score (r = 0.35, p < 0.001) and tricuspid regurgitation severity (r = 0.25, p = 0.012). Conclusions: Electronic stethoscope auscultation correlates with the echocardiographic findings in ARDS patients. The RVSP and SOFA scores independently predict mortality, while valvular regurgitation severity provides additional prognostic information for long-term survival and ICU resource utilization.
Background: Chest radiography remains central to post-procedural assessment of central venous catheter (CVC) placement in intensive care units. Multimodal large language models (MLLMs) can process medical images, but their reliability for practical radiography tasks remains uncertain. This study assessed the diagnostic performance of MLLMs and intensivists for CVC access classification, CVC tip assessment, and pneumothorax-related radiographic findings. Methods: In this retrospective diagnostic performance study, consecutive portable anteroposterior chest radiographs obtained after CVC placement in adult critically ill patients were independently evaluated by four intensivists and five MLLMs. A radiologist consensus served as the reference standard. Interobserver agreement and diagnostic performance were assessed using Fleiss’ kappa, Gwet AC1, Cohen’s kappa, accuracy, sensitivity, specificity, precision, F1 score, balanced accuracy, and Matthews correlation coefficient. Results: The final cohort included 183 unique radiographs. Intensivist reviewers showed high performance for CVC access classification but lower and more heterogeneous performance for CVC tip-position assessment. Among MLLMs, CVC access accuracy ranged from 0.339 to 0.874, whereas CVC tip assessment was dominated by almost universal classification of tips as appropriate, with near-zero specificity and chance-level balanced accuracy. For pneumothorax-related findings, all MLLMs classified every case as negative. Intensivist reviewers had higher balanced accuracy than MLLMs for CVC access classification (difference, 0.420; 95% CI, 0.349–0.490; p < 0.001) and CVC tip assessment (difference, 0.247; 95% CI, 0.205–0.290; p < 0.001). Pneumothorax analyses were exploratory because only five positive cases were present. Conclusions: The evaluated MLLMs showed unreliable diagnostic performance compared with experienced intensivists. Apparent performance was influenced by class imbalance and dominant-response behavior, supporting cautious task-specific validation and complete diagnostic performance reporting.