Division of Paediatric Emergency Medicine, Department of Women, Child and Adolescent, Geneva University Hospitals, 47 Avenue de la Roseraie, 1205, Geneva, Switzerland.
Intelligent Global Health, Machine Learning and Optimization (MLO) Laboratory, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland.
BMC Pulm Med. 2021 Mar 24;21(1):103. doi: 10.1186/s12890-021-01467-w.
Lung auscultation is fundamental to the clinical diagnosis of respiratory disease. However, auscultation is a subjective practice and interpretations vary widely between users. The digitization of auscultation acquisition and interpretation is a particularly promising strategy for diagnosing and monitoring infectious diseases such as Coronavirus-19 disease (COVID-19) where automated analyses could help decentralise care and better inform decision-making in telemedicine. This protocol describes the standardised collection of lung auscultations in COVID-19 triage sites and a deep learning approach to diagnostic and prognostic modelling for future incorporation into an intelligent autonomous stethoscope benchmarked against human expert interpretation.
A total of 1000 consecutive, patients aged ≥ 16 years and meeting COVID-19 testing criteria will be recruited at screening sites and amongst inpatients of the internal medicine department at the Geneva University Hospitals, starting from October 2020. COVID-19 is diagnosed by RT-PCR on a nasopharyngeal swab and COVID-positive patients are followed up until outcome (i.e., discharge, hospitalisation, intubation and/or death). At inclusion, demographic and clinical data are collected, such as age, sex, medical history, and signs and symptoms of the current episode. Additionally, lung auscultation will be recorded with a digital stethoscope at 6 thoracic sites in each patient. A deep learning algorithm (DeepBreath) using a Convolutional Neural Network (CNN) and Support Vector Machine classifier will be trained on these audio recordings to derive an automated prediction of diagnostic (COVID positive vs negative) and risk stratification categories (mild to severe). The performance of this model will be compared to a human prediction baseline on a random subset of lung sounds, where blinded physicians are asked to classify the audios into the same categories.
This approach has broad potential to standardise the evaluation of lung auscultation in COVID-19 at various levels of healthcare, especially in the context of decentralised triage and monitoring.
PB_2016-00500, SwissEthics. Registered on 6 April 2020.
肺部听诊是呼吸疾病临床诊断的基础。然而，听诊是一种主观的实践，不同使用者之间的解释差异很大。听诊采集和解释的数字化是诊断和监测传染病（如冠状病毒病-19 疾病（COVID-19））的一种特别有前途的策略，在这种疾病中，自动化分析可以帮助分散护理，并在远程医疗中更好地为决策提供信息。本方案描述了在 COVID-19 分诊点对肺部听诊进行标准化采集，以及一种深度学习方法，用于对诊断和预后进行建模，以便将来纳入与人类专家解释相媲美的智能自主听诊器中进行基准测试。
从 2020 年 10 月开始，将在筛选点和日内瓦大学医院内科住院患者中招募总共 1000 名年龄≥16 岁且符合 COVID-19 检测标准的连续患者。COVID-19 通过鼻咽拭子的 RT-PCR 诊断，对 COVID-阳性患者进行随访，直到结局（即出院、住院、插管和/或死亡）。在纳入时，收集人口统计学和临床数据，如年龄、性别、病史以及当前发作的症状和体征。此外，将使用数字听诊器在每个患者的 6 个胸部部位记录肺部听诊。使用卷积神经网络（CNN）和支持向量机分类器的深度学习算法（DeepBreath）将对这些音频记录进行训练，以自动预测诊断（COVID 阳性与阴性）和风险分层类别（轻度至重度）。将该模型的性能与随机肺部声音子集上的人类预测基线进行比较，在该子集中，要求盲法医生将音频分类到相同的类别中。
这种方法具有广泛的潜力，可以在各级医疗保健中标准化 COVID-19 肺部听诊的评估，尤其是在去中心化分诊和监测的背景下。
PB_2016-00500，瑞士伦理。于 2020 年 4 月 6 日注册。
