杂志文章Kanwar, Manreet, Gomberg-Maitland, Mardi Hoeper, Marius Pausch, Christine Pittow, David Strange, Geoff Anderson, James Zhao, Carol Scott, Jacqueline, druzzel, Marek Kraisangka, Jidapa Lohmueller, Lisa Antaki, James Benza，雷蒙德·L % T在肺动脉高血压危险分层使用贝叶斯分析% D J 2020% R 10.1183/13993003.00008 -2020%欧洲呼吸杂志% P 2000008 % X背景电流在肺动脉高血压危险分层工具(PAH)是歧视性的能力有限,部分原因是假设预后临床变量有一个独立的和临床结果的线性关系。我们试图证明基于贝叶斯网络(BN)的机器学习在增强现有的最先进的风险分层工具揭示2.0的预测能力方面的效用。方法使用与揭示2.0相同的变量和切点，我们获得了一种树增径朴素贝叶斯模型(标题为PHORA)来预测包括在揭示注册表中的PAH患者的一年生存期。PHORA模型在内部(在REVEAL注册表中)和外部(在COMPERA和PHSANZ注册表中)进行验证。根据2015年ESC/ERS指南，将患者分为低、中、高风险(分别小于5%、5-20%和>10个月死亡率)。结果PHORA预测1年生存率的AUC为0.80，优于REVEAL 2.0 (AUC为0.76)。在COMPERA和PHSANZ注册中验证后，PHORA显示AUC分别为0.74和0.80。根据PHORA预测的1年生存率在风险评分较低的患者中更高，而在风险评分较高的患者中更低(P<.001)，在所有三个登记中，低、中、高风险组之间有很好的分离。结论我们的BN衍生的风险预测模型，PHORA，显示了一个改进的区分现有模型。这反映了基于BN的模型对结果的临床变量之间的相互关系以及在计算预测时对缺失数据元素的容忍度的能力。这篇手稿最近已被欧洲呼吸杂志接受出版。它是在这里出版在其接受的形式之前，由我们的生产团队编辑和排版。 After these production processes are complete and the authors have approved the resulting proofs, the article will move to the latest issue of the ERJ online. Please open or download the PDF to view this article.Conflict of interest: Dr. Kanwar reports grants from NIH/NHBLI, during the conduct of the study.Conflict of interest: Dr. Gomberg-Maitland reports consultancy/Steering Committee, DSMB from Acceleron, Actelion, Complexa, Gossamer Bio, Reata, and Neuroderm. George Washington has received grants for research from Altavant and United Therapeutics. She is a member of the Scientific Advisory Board for United Therapeutics, outside the submitted work.Conflict of interest: Dr. Hoeper reports personal fees from Actelion, personal fees from Bayer, personal fees from MSD, personal fees from Pfizer, outside the submitted work.Conflict of interest: Dr. Pausch has nothing to disclose.David PittowConflict of interest: Dr. Strange reports grants from Actelion Pharmaceuticals, grants from GlasoSmithKline , grants from Bayer Pharmaceuticals, during the conduct of the study.Conflict of interest: Dr. Anderson reports grants from GlaxoSmithKline, non-financial support from Actelion, non-financial support from Bayer, personal fees from AstraZeneca, outside the submitted work.Conflict of interest: Carol Zhao reports personal fees from Actelion Pharmaceuticals US, Inc., A Janssen Pharmaceutical Company of Johnson & Johnson, during the conduct of the study; personal fees from Actelion Pharmaceuticals US, Inc., A Janssen Pharmaceutical Company of Johnson & Johnson, outside the submitted work.Conflict of interest: Dr. Scott has nothing to disclose.Conflict of interest: Dr. Druzdzel is a partner at BayesFusion, LLC.Conflict of interest: Dr. Kraisangka has nothing to disclose.Lisa LohmuellerConflict of interest: Dr. Antaki reports grants from NIH/NHLBI, during the conduct of the study.Conflict of interest: Dr. Benza reports grants from NIH/NHLBI, grants from Actelion, grants from United Therapeutics, grants from Bayer, during the conduct of the study.Conflict of interest: Dr. Lohmueller has nothing to disclose.Conflict of interest: Dr. Pittrow reports personal fees from Actelion, personal fees from Bayer, personal fees from Amgen, personal fees from Boehringer Ingelheim, personal fees from Sanofi, personal fees from MSD, personal fees from Biogen, outside the submitted work. %U //www.qdcxjkg.com/content/erj/early/2020/04/29/13993003.00008-2020.full.pdf