Department of Pharmacy Practice, Manipal College of Pharmaceutical Sciences, Manipal Academy of Higher Education, Manipal 576104, India.
School of Information Science, Manipal Academy of Higher Education, Manipal, Karnataka 576104, India.
Artif Intell Med. 2022 Sep;131:102361. doi: 10.1016/j.artmed.2022.102361. Epub 2022 Jul 19.
Acute respiratory distress syndrome (ARDS) is a life-threatening pulmonary disease with a high clinical and cost burden across the globe. Artificial intelligence (AI), an emerging area, has been used for various purposes in ARDS. We aim to summarize the currently available literature on various applications of AI in ARDS through a systematic review.
PubMed was searched from inception to February 2021 to collate all the studies. Additionally, a bibliographic search of included studies and a random search on Google, Google Scholar, and Research Gate were performed to identify relevant articles. Studies published in English language that employed data about developing and/or assessing the role of AI in the various aspects of ARDS were considered for this review. Three independent reviewers performed study selection and data extraction; any disagreements were settled through consensus or discussion with another member of the research team.
A total of 19 studies published between the year 2002 and 2020 were included. In these included studies, AI was used for various purposes in ARDS such as diagnosis (n = 10; 53 %), risk stratification (n = 1; 5 %), prediction of severity (n = 3; 17 %), management (n = 2; 10 %), prediction of mortality (n = 2; 10 %), and decision making (n = 1; 5 %). The area under the curve among the developed models in the included studies ranged between 0.8 and 1, which is considered to be very good to excellent.
AI is revolutionizing healthcare and has a wide range of applications in ARDS, such as minimizing cost and enhancing outcomes.
急性呼吸窘迫综合征（ARDS）是一种危及生命的肺部疾病，在全球范围内具有较高的临床和经济负担。人工智能（AI）作为一个新兴领域，已在 ARDS 中得到了各种应用。我们旨在通过系统评价总结目前关于 AI 在 ARDS 中各种应用的文献。
从创建到 2021 年 2 月，在 PubMed 上进行了检索，以整理所有研究。此外，还对纳入研究进行了文献追溯，并在 Google、Google Scholar 和 Research Gate 上进行了随机搜索，以确定相关文章。纳入的研究必须是发表在英语文献中，且使用了关于开发和/或评估 AI 在 ARDS 各个方面作用的数据。三位独立的评审员进行了研究选择和数据提取；任何分歧都通过达成共识或与研究团队的其他成员讨论来解决。
共纳入了 2002 年至 2020 年期间发表的 19 项研究。在这些纳入的研究中，AI 被用于 ARDS 的各种用途，如诊断（n=10；53%）、风险分层（n=1；5%）、严重程度预测（n=3；17%）、管理（n=2；10%）、死亡率预测（n=2；10%）和决策制定（n=1；5%）。纳入研究中开发的模型的曲线下面积在 0.8 到 1 之间，这被认为是非常好到优秀的。
AI 正在彻底改变医疗保健行业，在 ARDS 中有广泛的应用，如降低成本和提高治疗效果。
