Title Lung ultrasound vs. CT scan for the assessment of post-ICU long COVID

Introduction: Long COVID includes signs and symptoms after acute COVID-19 [1]. Lung ultrasound (LUS) is increasingly used for lung assessment [2], but many patients still undergo traditional imaging (chest CT) for long COVID evaluation. Aims and objectives: To test the capability of LUS to identify post-ICU COVID-19 patients with significant alterations at chest CT scan. Method(s): Single-center retrospective study on post-ICU patients in recovery phase of long COVID. Patients were included if they had a complete LUS with computation of the LUS score and chest CT performed during the follow-up evaluation at least one month after hospital discharge. CT were classified by an expert radiologist as significant if focal/diffuse involvement was observed, as not significant if lung aeration was normal. Result(s): 40 patients were included so far (age 60.0 [51.0-66.0], males 73.8%, BMI 29.1 [27.7 - 29.4] kg/m2). Significant CTs were 15 (37.5%);LUS score was higher in these patients (7.0 [4.0-9.0] vs. 0.0 [0.0-4.0];p=0.0004). LUS score had an area under the ROC curve of 0.8347 [95% CI 0.7033-0.9662] for significant CT (Youden index cut off point >=3, sensitivity 86.7%, specificity of 70.8%). Conclusion(s): LUS score accurately identifies post-ICU long COVID patients with significant alterations at CT scan.

ALFABETO: Supporting COVID-19 hospital admissions with Bayesian Networks

The ongoing pandemics of coronavirus disease has accelerated the implementation of machine learning methods (ML) to support clinical decisions. Within this context, we present the ALFABETO project, whose aim is to aid clinicians during COVID-19 patients hospital admission through the application of ML approaches exploiting clinical and chest x-ray features. Yet, non linear ML classifiers are often perceived as not easily interpretable by users, thus hampering trust in ML predictions. Moreover, these ML models, such as Neural Networks or Random Forest, are not able to include pre-exisisting knowledge about a specific domain and are not designed to find causal relationships between variables. For these reasons, we wanted to investigate if Bayesian Networks were able to properly describe the hospital admission decision process. Bayesian Networks are probabilistic graphical models representing a set of variables and their conditional dependencies. The network structure was derived both from existing medical knowledge and from patients data collected during the first wave of the pandemic. While being explainable, we show that the Bayesian network has similar performance when compared to a less explainable ML model and that was able to generalize well across COVID-19 waves. © 2021 Copyright for this paper by its authors.