RESUMEN
INTRODUCTION: Although thorax ultrasound has been used to diagnose pneumonia in recent years, the role of ultrasonic diaphragm evaluation in the prognosis of pneumonia is unknown. This study aimed to assess the impact of diaphragmatic excursion (Dex) measured by ultrasound on the prognosis of severe pneumonia in critical care patients. MATERIALS AND METHODS: We prospectively recruited patients with severe pneumonia who were admitted to the intensive care unit (ICU) between January 2019 and July 2021. Patients' Dex values, vital signs, clinical features, laboratory parameters, APACHE-II scores on the first admission day of ICU, mortality and respiratory support status at follow-up were recorded. RESULT: There were 39 patients enrolled in the study. Mean Dex of the study patients was 30.66 ± 12.17 mm. Mean Dex was significantly lower in deceased patients than survivors (18.37 ± 8.12 vs 34.90 ± 10.36 p< 0.001). Dex was lower in patients who required invasive mechanical ventilation than those not (24.90 ± 10.93 vs 34.26 ± 11.70, p= 0.017). The cut-off value of Dex was found 19.0 mm for significantly predicted (p≤ 0.001) survival with the sensitivity of 96.6% and specificity of 70%. Among the study group, diaphragm excursion was negatively correlated with APACHE-II score (r= -0.688, p≤ 0.001) and respiratory rate (r= -0.531, p= 0.001). CONCLUSIONS: Dex measured on the day of ICU admission can be used to evaluate the prognosis of patients with severe pneumonia.
Asunto(s)
Diafragma , Neumonía , APACHE , Diafragma/diagnóstico por imagen , Humanos , Unidades de Cuidados Intensivos , Neumonía/diagnóstico por imagen , Pronóstico , Curva ROC , Estudios Retrospectivos , UltrasonografíaRESUMEN
The generalizability of artificial intelligence (AI) models is a major issue in the field of AI applications. Therefore, we aimed to overcome the generalizability problem of an AI model developed for a particular center for pneumothorax detection using a small dataset for external validation. Chest radiographs of patients diagnosed with pneumothorax (n = 648) and those without pneumothorax (n = 650) who visited the Ankara University Faculty of Medicine (AUFM; center 1) were obtained. A deep learning-based pneumothorax detection algorithm (PDA-Alpha) was developed using the AUFM dataset. For implementation at the Health Sciences University (HSU; center 2), PDA-Beta was developed through external validation of PDA-Alpha using 50 radiographs with pneumothorax obtained from HSU. Both PDA algorithms were assessed using the HSU test dataset (n = 200) containing 50 pneumothorax and 150 non-pneumothorax radiographs. We compared the results generated by the algorithms with those of physicians to demonstrate the reliability of the results. The areas under the curve for PDA-Alpha and PDA-Beta were 0.993 (95% confidence interval (CI): 0.985-1.000) and 0.986 (95% CI: 0.962-1.000), respectively. Both algorithms successfully detected the presence of pneumothorax on 49/50 radiographs; however, PDA-Alpha had seven false-positive predictions, whereas PDA-Beta had one. The positive predictive value increased from 0.525 to 0.886 after external validation (p = 0.041). The physicians' sensitivity and specificity for detecting pneumothorax were 0.585 and 0.988, respectively. The performance scores of the algorithms were increased with a small dataset; however, further studies are required to determine the optimal amount of external validation data to fully address the generalizability issue.