AI辅助HRCT构建COVID-19智能诊断和分级预后平台及不明原因肺炎早期预警模型研究

目前超过200个国家确诊COVID-19患者160万余人，死亡超10万人，疫情依然严峻。HRCT检测肺内病变敏感性和特异性高，并能定量分析，在疾病分级和预后中也具有一定优势，是国家卫健委有关新冠肺炎诊疗方案推荐的重要检测技术。但HRCT含海量数据，人工阅片工作量大、主观性强、量化分析缺乏标准。课题组在前期总结大量新冠肺炎患者CT表现与临床资料的基础上，组织AI专家联合影像医生的科研团队，拟构建COVID-19肺炎大数据库、完善其他类别肺炎数据库，开发COVID-19智能诊断平台，进一步针对患者进行分级预后研究，以期完善HRCT在COVID-19肺炎诊断、疗效评价和预后中的研究。预防重于治疗，不明原因肺炎暴发前的早期预警尤为重要，CT在此方面的价值有待于开发。在COVID-19智能诊断平台上研发基于HRCT和AI的预警模型，是传染性疾病预警体系的有益补充，具有重大社会效益和经济效益。

By far, more than 1.6 million people have been diagnosed with COVID-19 in more than 200 countries, and over 100,000 of them died. HRCT has the advantages of high sensitivity, high specificity for detecting lung lesions, and can be quantitatively analyzed. It also has an advantage in disease staging and prognosis prediction. HRCT has been recommended by the National Health Commission for the diagnosis and treatment of COVID-19 pneumonia. However, due to lots of information in HRCT images, manual reading has shortcomings of heavy workload, subjectivity, and lacks of standard for quantitative analysis. Based on earlier summary in CT and clinical manifestation of COVID-19 pneumonia, the research group organized a scientific research team of artificial intelligence (AI) experts and radiologists to build databases of COVID-19 and some other types of pneumonia, and then to develop a COVID-19 intelligent diagnostic platform. The platform will be further improved to conduct a system of staging and prognosis study, with a view to perfecting the role of HRCT in the diagnosis, curative effect evaluation and prognosis prediction of COVID-19 pneumonia. Prevention is more important than treatment, so early warning before an unexplained pneumonia outbreak is particularly vital. The value of HRCT in this field needs remains to be further explored. A warning model based on the COVID-19 intelligent diagnostic platform will be a useful supplement to the warning system for infectious diseases, with significant social and economic benefits.