AI-SCoRE (artificial intelligence-SARS CoV2 risk evaluation): a fast, objective and fully automated platform to predict the outcome in COVID-19 patients.

PURPOSE: To develop and validate an effective and user-friendly AI platform based on a few unbiased clinical variables integrated with advanced CT automatic analysis for COVID-19 patients' risk stratification. MATERIAL AND METHODS: In total, 1575 consecutive COVID-19 adults admitted to 16 hospitals during wave 1 (February 16-April 29, 2020), submitted to chest CT within 72 h from admission, were retrospectively enrolled. In total, 107 variables were initially collected; 64 extracted from CT. The outcome was survival. A rigorous AI model selection framework was adopted for models selection and automatic CT data extraction. Model performances were compared in terms of AUC. A web-mobile interface was developed using Microsoft PowerApps environment. The platform was externally validated on 213 COVID-19 adults prospectively enrolled during wave 2 (October 14-December 31, 2020). RESULTS: The final cohort included 1125 patients (292 non-survivors, 26%) and 24 variables. Logistic showed the best performance on the complete set of variables (AUC = 0.839 ± 0.009) as in models including a limited set of 13 and 5 variables (AUC = 0.840 ± 0.0093 and AUC = 0.834 ± 0.007). For non-inferior performance, the 5 variables model (age, sex, saturation, well-aerated lung parenchyma and cardiothoracic vascular calcium) was selected as the final model and the extraction of CT-derived parameters was fully automatized. The fully automatic model showed AUC = 0.842 (95% CI: 0.816-0.867) on wave 1 and was used to build a 0-100 scale risk score (AI-SCoRE). The predictive performance was confirmed on wave 2 (AUC 0.808; 95% CI: 0.7402-0.8766). CONCLUSIONS: AI-SCoRE is an effective and reliable platform for automatic risk stratification of COVID-19 patients based on a few unbiased clinical data and CT automatic analysis.

Hybrid Convolutional Networks for End-to-End Event Detection in Concurrent PPG and PCG Signals Affected by Motion Artifacts.

The accurate detection of physiologically-related events in photopletismographic (PPG) and phonocardiographic (PCG) signals, recorded by wearable sensors, is mandatory to perform the estimation of relevant cardiovascular parameters like the heart rate and the blood pressure. However, the measurement performed in uncontrolled conditions without clinical supervision leaves the detection quality particularly susceptible to noise and motion artifacts. This work proposes a new fully-automatic computational framework, based on convolutional networks, to identify and localize fiducial points in time as the foot, maximum slope and peak in PPG signal and the S1 sound in the PCG signal, both acquired by a custom chest sensor, described recently in the literature by our group. The event detection problem was reframed as a single hybrid regression-classification problem entailing a custom neural architecture to process sequentially the PPG and PCG signals. Tests were performed analysing four different acquisition conditions (rest, cycling, rest recovery and walking). Cross-validation results for the three PPG fiducial points showed identification accuracy greater than 93 % and localization error (RMSE) less than 10 ms. As expected, cycling and walking conditions provided worse results than rest and recovery, however reaching an accuracy greater than 90 % and a localization error less than 15 ms. Likewise, the identification and localization error for S1 sound were greater than 90 % and less than 25 ms. Overall, this study showcased the ability of the proposed technique to detect events with high accuracy not only for steady acquisitions but also during subject movements. We also showed that the proposed network outperformed traditional Shannon-energy-envelope method in the detection of S1 sound, reaching detection performance comparable to state of the art algorithms. Therefore, we argue that coupling chest sensors and deep learning processing techniques may disclose wearable devices to unobtrusively acquire health information, being less affected by noise and motion artifacts.