中华急诊医学杂志  2022, Vol. 31 Issue (8): 1105-1109   DOI: 10.3760/cma.j.issn.1671-0282.2022.08.001
精准分型指导ARDS临床精准救治
孙健1 , 罗悦1 , 邵勉1 , 宋振举1,2     
1. 复旦大学附属中山医院急诊科, 上海 200032;
2. 上海市重大传染病与生物安全研究院,上海 200032

急性呼吸窘迫综合征(acute respiratory distress syndrome, ARDS)是一种以肺泡-毛细血管屏障破坏和非心源性肺水肿致严重低氧血症为特征的急危重症。全球范围内ICU患者ARDS患病率为10%,在需要机械通气的患者中患病率为23%,住院病死率达35%~45%[1]。目前ARDS临床诊治存在五大痛点:高度异质性、分型不精准、协作研究少、疗效评估难、研究模型不匹配。ARDS的临床和生物学高度异质性,被认为是影响临床研究效果评价、同质化治疗程度低的重要因素。对ARDS精准分型,形成分型导向的精准治疗方案,是提高ARDS救治水平的关键[2]。近年来,针对ARDS异质性和精准分型开展了一系列的临床研究,并取得了较好的效果;精准分型也由单一变量发展为基于临床特征、生物学标志物和影像学资料的多变量、多模态模型。

1 单一变量的分型

通过病因、发病时间、严重程度等单一变量对ARDS患者进行简单分型具有一定的临床价值。ARDS患者病情危重且复杂,快速精准识别并作出相应治疗至关重要。对患者进行分型的一个重要目的是进行预后富集和预测富集。在没有掌握足够多信息的情况下,通过单一变量进行初步分型,有利于指导ARDS患者的早期精准识别及干预。

1.1 通气指标

柏林定义[3]根据氧合指数(P/F)把ARDS分为轻、中和重度。最新研究[4]显示使用150 mmHg(1 mmHg=0.133 kpa)的P/F阈值将患者分为两个亚型会使组内患者更为同质,解剖学和临床数据都支持这一观点,肺重(衡量肺水肿的指标)在P/F低于150 mmHg时显著增加。肺泡死腔分数(dead-space fraction)升高是早期ARDS的一个特征,与不良预后的风险增加相关[5-6]。但目前肺泡死腔分数的床旁快速测定技术尚未普及,限制了其在ARDS分型中的应用。在各个通气变量中,驱动压(潮气量/呼吸系统的静态顺应性)与患者生存率最为相关,并能很好地对不良预后的风险进行分层[7],但同样也存在无法常规获取的问题。通气比率(ventilation ratio= [分钟通气(mL/min) × PaCO2 (mmHg)]/(估算体重×100 × 37.5)与肺泡死腔分数相关,测量值与死亡风险呈正相关[8],并可床旁快速计算获取。

这些指标在病程中的快速变化是限制其临床使用的主要原因。例如,高呼气末正压(PEEP)的机械通气可以使氧合指数快速改善,也有患者在接受气管插管时出现严重低氧血症(满足ARDS诊断标准),但在气管插管后数小时后改善。快速改善的ARDS (rapidly improving ARDS,riARDS)在临床上并不少见,可实现早期拔管,其预后优于插管时间>1 d的ARDS,然而63%的riARDS患者存在中度或重度低氧血症[9],在临床研究中纳入riARDS患者会导致疾病预后的评估出现误差。生理指标难以捕捉到ARDS患者潜在的病理生理学差异,如脓毒症相关ARDS与创伤所致ARDS患者可能具有相同的氧合指数,然而预后截然不同。

1.2 临床特征

相比于通气指标,临床特征更能反映患者的病理生理学改变。创伤相关的ARDS患者90 d病死率明显低于非创伤性ARDS患者[10],其较好的预后可能部分源于较轻的肺泡内皮和上皮损伤。此外,研究表明间接(非肺源性)ARDS首先影响肺泡毛细血管内皮通透性,导致弥漫性肺水肿,而直接(肺源性)ARDS则首先影响肺泡上皮[11]。相较于间接ARDS患者,直接ARDS患者病情程度较轻、更少的器官衰竭、更多的肺上皮损伤,而肺泡毛细血管内皮损伤的血浆生物标志物浓度更低[12]。这表明,直接和间接ARDS患者可能从分别针对肺泡上皮细胞损伤和血管内皮细胞损伤的疗法中获益。研究也发现尽管直接ARDS患者病情程度较低,器官功能紊乱程度较轻,但其病死率与间接ARDS患者相似[13]。此外,回顾性研究发现合并急性肾损伤(AKI)的ARDS患者的病死率明显高于无AKI的患者[14-15]。ARDS的发病时间也与预后有关,有研究者将ARDS分为社区获得性和医院(ICU)获得性,后者定义为ARDS发病于入院(ICU)48 h后,比前者有更高的病死率[16-18]

基于CT的肺部形态学特征也可以将ARDS患者分成更同质的亚组。研究发现影像学表现为非局灶性浸润的ARDS患者的病死率高于局灶性浸润的患者[19],根据肺部影像学形态采用个性化的通气策略可降低病死率[20]。用于量化胸片肺泡水肿程度的RALE (肺水肿影像学评估)评分,已被证明可用于评估ARDS的严重程度和优化治疗方案[21]

基于临床特征进行分型的难点在于对患者临床特征的准确识别。有研究发现,37%的患者无法确定是肺源性还是非肺源性ARDS[22]。在另一项评估ARDS个性化机械通气治疗效果的临床试验中,21%的患者被错误的归类为局灶性或非局灶性浸润ARDS,并导致了阴性结果[20]

2 基于多变量的精准分型

ARDS的高度异质性决定了精准分型的困难。结合生理指标、临床特征、生物标志物和影像学等多模态数据,构建基于多变量的精准分型,是理解ARDS潜在生物学规律,提高ARDS救治成功率的关键,也是未来研究的方向。

2.1 生物标志物

生物标志物一定程度上反映疾病的病理生理学改变,具有诊断和监测的价值,也是潜在的治疗靶点[23]。通过生物标志物确定亚型并给予针对性的靶向治疗,在其他异质性疾病如乳腺癌、黑色素瘤及哮喘中都实现了治疗突破[24-26]。尽管几乎没有靶向药物在大型ARDS临床试验被证实有效[27-29],但生物标志物已被证明在ARDS预后方面的价值。挖掘高度敏感和特异度的生物标志物,是构建精准分型的基石。

有研究证实表面活性蛋白D(SP-D)水平与ARDS不良预后呈正相关,低潮气量治疗策略有助于降低血浆SP-D水平[30]。血管性血友病因子(vWF)也与ARDS预后密切相关,但没有发现其血浆水平变化与通气策略的相关性[31]。血管生成素2(Ang-2)对ARDS的预后有评估价值,较低血浆Ang-2水平的患者更易从液体保守治疗中获益[32]。ARDS炎症级联反应诱导凝血纤溶系统激活,PAI-1水平降低与蛋白质C(PC)水平升高与ARDS不良预后相关[33]。可溶性晚期糖基化终末产物受体(sRAGE)参与介导肺泡炎症反应,在直接ARDS和CT显示为弥漫浸润性ARDS中水平较高,与病死率相关[34]。细胞间黏附分子-1(ICAM-1)介导细胞间黏附反应,高水平ICAM-1与ARDS患者不良预后相关[35]。此外IL-6、IL-8、TNF等也是ARDS生物标志物研究的热点。

有关ARDS生物标志物的研究数量多,异质性大,生物标志物在ARDS中扮演的角色仍需更多的研究去验证。目前尚不能证实生物标志物可以独立预测ARDS病死率[36],生物标志物和临床指标的组合,是探索ARDS精准分型的最佳方案。

2.2 LCA分型

结合患者多维度临床数据和生物标志物,构建ARDS临床精准分型,是ARDS研究的热点[37]。目前最广为接受的是Calfee等[38]2014年提出的潜类别分析(latent class analysis,LCA)分型。LCA是一种统计分析方法,通过对多个类别变量进行分析,将样本人群划分为几大类人群[39]。其本质上是一种潜变量测量模型,使用混合建模来为一组数据找到最佳拟合模型。Calfee等[38]对两个大型ARDS随机对照试验的临床和生物学数据进行潜类别分析,发现了两种不同的亚型:“低炎型”和“高炎型”,后者特征是更严重的炎症反应、休克和代谢性酸中毒,以及更差的预后。同时两种亚型对不同PEEP的通气策略表现出不同的治疗反应,高炎型更受益于高PEEP策略。Famous等[40]也证明两种亚型对液体复苏策略的不同治疗反应,低炎型更受益于限制性液体复苏治疗。除此之外,对一项辛伐他汀治疗ARDS的RCT数据二次分析再次验证了上述两种亚型的存在[41]。对患者分型后发现,在高炎症亚型中接受辛伐他汀治疗的患者28 d存活率明显高于安慰剂组[42]。对另一项大型RCT数据的二次分析显示,分型没有发现瑞舒伐他汀的治疗反应差异,但分型模型得到了验证[43]。截至目前,LCA分型在7个大型ARDS队列中得到了验证[38, 40-41, 43-44],在其中两个队列研究验证了LCA分型的稳定性[45]

LCA分型也在儿童ARDS中观察到[46],除此之外,单纯通过血浆生物标志物进行聚类分析[47],也识别出ARDS两种亚型“非炎症型(uninflamed phenotype)”和“反应型(reactive phenotype)”。与LCA分型类似,或许由LCA派生的“高炎症型”和“低炎症”已十分接近ARDS真实存在的亚型,但目前仅建立在回顾性数据分析的基础上,有待更多前瞻性研究验证。

2.3 临床分类模型

虽然LCA分型得到了大量试验数据的支持,但数十个变量使这一复杂模型在临床上的使用场景变得有限,同时血浆生物标志物如PAI-1、ICAM-1、TNFr1在临床上并非常规检测,因此研究者们尝试通过构建分类模型以实现对患者分型的快速床旁识别。在上述构建与验证LCA分型的研究中[38, 40-41, 43-44],研究者尝试仅通过几个常规测量的变量对ARDS患者亚型进行快速识别。以先前建立的LCA分型作为金标准。Famous等[40]选取了3个在LCA分型中占权重最大的变量——IL-8、血清碳酸氢盐、sTNFr1构建了识别模型,该模型的AUC在FACTT队列中为0.95。Sinha等[48]在6个大型队列中对临床分类模型进行了集中研究,通过机器学习算法选择重要的分类变量,然后用这些变量建立Logistic回归模型,发现3个变量(IL-8、碳酸氢盐和蛋白C)和4个变量(3个变量加上血管活性药的使用)效果最好。

Sinha等[49]使用现代机器学习算法开发出了高精度亚型分类模型,仅依赖现成的临床数据和常规实验室指标对LCA衍生亚型做出了快速精准识别。最新的研究[50]验证了两种机器学习临床分类模型,同时证明这些模型的应用能够提供有价值的预后信息,并指导个性化治疗。

新冠肺炎相关ARDS患者也观察到了不同的临床特征、炎症状态和呼吸力学表现,这提示该人群内部存在亚型的可能性。Gattinoni等[51]认为新冠肺炎相关ARDS可以分为H型(type high)和L型(type low),前者更类似典型ARDS,有较高的肺组织弹性、通气血流比值与肺重。两型表现出不同的治疗反应,H型受益于低潮气量和高PEEP,L型则相反[52]。而先前被广为接受的LCA衍生分型在新冠肺炎相关ARDS中同样被观察到[53],进一步证明了LCA衍生分型在ARDS患者人群中的普适性。糖皮质激素分别降低和提高了高炎型和低炎型的病死率[53],这为糖皮质激素在新冠肺炎中的个体化应用提供了很好的借鉴。

3 展望

ARDS的高度异质性决定了其分型的困难。LCA亚型在目前ARDS分型研究中受到了最多关注,但其分型变量涉及维度仍然有限,数据分析方法还有探索的空间。目前基因组学、转录组学、蛋白组学和代谢组学的研究,大数据、人工智能等技术的兴起将会进一步推动ARDS临床精准分型研究。利用多组学技术平台,借助自然语言处理及数据融合、机器学习等新兴技术,进行影像数据、临床特征、生物标志物和呼吸力学参数处理及多模态数据融合,探索更接近ARDS内在生物学规律的精准分型。在此基础上,形成精准分型导向的集束化精准治疗方案,有望实现ARDS诊疗的同质化,提高ARDS的救治水平。

利益冲突  所有作者声明无利益冲突

参考文献
[1] Bellani G, Laffey JG, Pham T, et al. Epidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome in intensive care units in 50 countries[J]. JAMA, 2016, 315(8): 788-800. DOI:10.1001/jama.2016.0291
[2] Sinha P, Calfee CS. Phenotypes in acute respiratory distress syndrome: moving towards precision medicine[J]. Curr Opin Crit Care, 2019, 25(1): 12-20. DOI:10.1097/MCC.0000000000000571
[3] ARDS Definition Task Force, Ranieri VM, Rubenfeld GD, et al. Acute respiratory distress syndrome: the Berlin Definition[J]. JAMA, 2012, 307(23): 2526-2533. DOI:10.1001/jama.2012.5669
[4] Maiolo G, Collino F, Vasques F, et al. Reclassifying acute respiratory distress syndrome[J]. Am J Respir Crit Care Med, 2018, 197(12): 1586-1595. DOI:10.1164/rccm.201709-1804OC
[5] Morales-Quinteros L, Schultz MJ, Bringué J, et al. Estimated dead space fraction and the ventilatory ratio are associated with mortality in early ARDS[J]. Ann Intensive Care, 2019, 9(1): 128. DOI:10.1186/s13613-019-0601-0
[6] Nuckton TJ, Alonso JA, Kallet RH, et al. Pulmonary dead-space fraction as a risk factor for death in the acute respiratory distress syndrome[J]. N Engl J Med, 2002, 346(17): 1281-1286. DOI:10.1056/NEJMoa012835
[7] Amato MBP, Meade MO, Slutsky AS, et al. Driving pressure and survival in the acute respiratory distress syndrome[J]. N Engl J Med, 2015, 372(8): 747-755. DOI:10.1056/NEJMsa1410639
[8] Sinha P, Calfee CS, Beitler JR, et al. Physiologic analysis and clinical performance of the ventilatory ratio in acute respiratory distress syndrome[J]. Am J Respir Crit Care Med, 2019, 199(3): 333-341. DOI:10.1164/rccm.201804-0692OC
[9] Schenck EJ, Oromendia C, Torres LK, et al. Rapidly improving ARDS in therapeutic randomized controlled trials[J]. Chest, 2019, 155(3): 474-482. DOI:10.1016/j.chest.2018.09.031
[10] Calfee CS, Eisner MD, Ware LB, et al. Trauma-associated lung injury differs clinically and biologically from acute lung injury due to other clinical disorders[J]. Crit Care Med, 2007, 35(10): 2243-2250. DOI:10.1097/01.ccm.0000280434.33451.87
[11] Gattinoni L, Pelosi P, Suter PM, et al. Acute respiratory distress syndrome caused by pulmonary and extrapulmonary disease. Different syndromes?[J]. Am J Respir Crit Care Med, 1998, 158(1): 3-11. DOI:10.1164/ajrccm.158.1.9708031
[12] Calfee CS, Janz DR, Bernard GR, et al. Distinct molecular phenotypes of direct vs indirect ARDS in single-center and multicenter studies[J]. Chest, 2015, 147(6): 1539-1548. DOI:10.1378/chest.14-2454
[13] Luo L, Shaver CM, Zhao ZG, et al. Clinical predictors of hospital mortality differ between direct and indirect ARDS[J]. Chest, 2017, 151(4): 755-763. DOI:10.1016/j.chest.2016.09.004
[14] Liu KD, Glidden DV, Eisner MD, et al. Predictive and pathogenetic value of plasma biomarkers for acute kidney injury in patients with acute lung injury[J]. Crit Care Med, 2007, 35(12): 2755-2761.
[15] McNicholas BA, Rezoagli E, Pham T, et al. Impact of early acute kidney injury on management and outcome in patients with acute respiratory distress syndrome: a secondary analysis of a multicenter observational study[J]. Crit Care Med, 2019, 47(9): 1216-1225. DOI:10.1097/CCM.0000000000003832
[16] Liao KM, Chen CW, Hsiue TR, et al. Timing of acute respiratory distress syndrome onset is related to patient outcome[J]. J Formos Med Assoc, 2009, 108(9): 694-703. DOI:10.1016/S0929-6646(09)60392-2
[17] Zhang RY, Wang ZX, Tejera P, et al. Late-onset moderate to severe acute respiratory distress syndrome is associated with shorter survival and higher mortality: a two-stage association study[J]. Intensive Care Med, 2017, 43(3): 399-407. DOI:10.1007/s00134-016-4638-3
[18] Kao KC, Hu HC, Hsieh MJ, et al. Comparison of community-acquired, hospital-acquired, and intensive care unit-acquired acute respiratory distress syndrome: a prospective observational cohort study[J]. Crit Care, 2015, 19: 384. DOI:10.1186/s13054-015-1096-1
[19] Mrozek S, Jabaudon M, Jaber S, et al. Elevated plasma levels of sRAGE are associated with nonfocal CT-based lung imaging in patients with ARDS: a prospective multicenter study[J]. Chest, 2016, 150(5): 998-1007. DOI:10.1016/j.chest.2016.03.016
[20] Constantin JM, Jabaudon M, Lefrant JY, et al. Personalised mechanical ventilation tailored to lung morphology versus low positive end-expiratory pressure for patients with acute respiratory distress syndrome in France (the LIVE study): a multicentre, single-blind, randomised controlled trial[J]. Lancet Respir Med, 2019, 7(10): 870-880. DOI:10.1016/S2213-2600(19)30138-9
[21] Warren MA, Zhao Z, Koyama T, et al. Severity scoring of lung oedema on the chest radiograph is associated with clinical outcomes in ARDS[J]. Thorax, 2018, 73(9): 840-846. DOI:10.1136/thoraxjnl-2017-211280
[22] Thille AW, Richard JC, Maggiore SM, et al. Alveolar recruitment in pulmonary and extrapulmonary acute respiratory distress syndrome: comparison using pressure-volume curve or static compliance[J]. Anesthesiology, 2007, 106(2): 212-217. DOI:10.1097/00000542-200702000-00007
[23] Prescott HC, Calfee CS, Thompson BT, et al. Toward smarter lumping and smarter splitting: rethinking strategies for Sepsis and acute respiratory distress syndrome clinical trial design[J]. Am J Respir Crit Care Med, 2016, 194(2): 147-155. DOI:10.1164/rccm.201512-2544CP
[24] FitzGerald JM, Bleecker ER, Nair P, et al. Benralizumab, an anti-interleukin-5 receptor α monoclonal antibody, as add-on treatment for patients with severe, uncontrolled, eosinophilic asthma (CALIMA): a randomised, double-blind, placebo-controlled phase 3 trial[J]. Lancet, 2016, 388(10056): 2128-2141. DOI:10.1016/S0140-6736(16)31322-8
[25] Partridge AH, Rumble RB, Carey LA, et al. Chemotherapy and targeted therapy for women with human epidermal growth factor receptor 2-negative (or unknown) advanced breast cancer: American Society of Clinical Oncology Clinical Practice Guideline[J]. J Clin Oncol, 2014, 32(29): 3307-3329. DOI:10.1200/JCO.2014.56.7479
[26] Silva IP, Long GV. Systemic therapy in advanced melanoma: integrating targeted therapy and immunotherapy into clinical practice[J]. Curr Opin Oncol, 2017, 29(6): 484-492. DOI:10.1097/CCO.0000000000000405
[27] Thompson BT, Chambers RC, Liu KD. Acute respiratory distress syndrome[J]. N Engl J Med, 2017, 377(6): 562-572. DOI:10.1056/NEJMra1608077
[28] Lewis SR, Pritchard MW, Thomas CM, et al. Pharmacological agents for adults with acute respiratory distress syndrome[J]. Cochrane Database Syst Rev, 2019, 7: CD004477. DOI:10.1002/14651858.CD004477.pub3
[29] Matthay MA, Arabi YM, Siegel ER, et al. Phenotypes and personalized medicine in the acute respiratory distress syndrome[J]. Intensive Care Med, 2020, 46(12): 2136-2152. DOI:10.1007/s00134-020-06296-9
[30] Eisner MD, Parsons P, Matthay MA, et al. Plasma surfactant protein levels and clinical outcomes in patients with acute lung injury[J]. Thorax, 2003, 58(11): 983-988. DOI:10.1136/thorax.58.11.983
[31] Ware LB, Eisner MD, Thompson BT, et al. Significance of von Willebrand factor in septic and nonseptic patients with acute lung injury[J]. Am J Respir Crit Care Med, 2004, 170(7): 766-772. DOI:10.1164/rccm.200310-1434OC
[32] Calfee CS, Gallagher D, Abbott J, et al. Plasma angiopoietin-2 in clinical acute lung injury: prognostic and pathogenetic significance[J]. Crit Care Med, 2012, 40(6): 1731-1737. DOI:10.1097/CCM.0b013e3182451c87
[33] Ware LB, Matthay MA, Parsons PE, et al. Pathogenetic and prognostic significance of altered coagulation and fibrinolysis in acute lung injury/acute respiratory distress syndrome[J]. Crit Care Med, 2007, 35(8): 1821-1828. DOI:10.1097/01.CCM.0000221922.08878.49
[34] Jabaudon M, Blondonnet R, Pereira B, et al. Plasma sRAGE is independently associated with increased mortality in ARDS: a meta-analysis of individual patient data[J]. Intensive Care Med, 2018, 44(9): 1388-1399. DOI:10.1007/s00134-018-5327-1
[35] Calfee CS, Eisner MD, Parsons PE, et al. Soluble intercellular adhesion molecule-1 and clinical outcomes in patients with acute lung injury[J]. Intensive Care Med, 2009, 35(2): 248-257. DOI:10.1007/s00134-008-1235-0
[36] van der Zee P, Rietdijk W, Somhorst P, et al. A systematic review of biomarkers multivariately associated with acute respiratory distress syndrome development and mortality[J]. Crit Care, 2020, 24(1): 243. DOI:10.1186/s13054-020-02913-7
[37] Ashbaugh D, Boyd Bigelow D, Petty T, et al. Acute respiratory distress in adults[J]. Lancet, 1967, 290(7511): 319-323. DOI:10.1016/S0140-6736(67)90168-7
[38] Calfee CS, Delucchi K, Parsons PE, et al. Subphenotypes in acute respiratory distress syndrome: latent class analysis of data from two randomised controlled trials[J]. Lancet Respir Med, 2014, 2(8): 611-620. DOI:10.1016/S2213-2600(14)70097-9
[39] Green BF Jr. A general solution for the latent class model of latent structure analysis[J]. Psychometrika, 1951, 16(2): 151-166. DOI:10.1007/BF02289112
[40] Famous KR, Delucchi K, Ware LB, et al. Acute respiratory distress syndrome subphenotypes respond differently to randomized fluid management strategy[J]. Am J Respir Crit Care Med, 2017, 195(3): 331-338. DOI:10.1164/rccm.201603-0645OC
[41] Calfee CS, Delucchi KL, Sinha P, et al. Acute respiratory distress syndrome subphenotypes and differential response to simvastatin: secondary analysis of a randomised controlled trial[J]. Lancet Respir Med, 2018, 6(9): 691-698. DOI:10.1016/S2213-2600(18)30177-2
[42] McAuley DF, Laffey JG, O'Kane CM, et al. Simvastatin in the acute respiratory distress syndrome[J]. N Engl J Med, 2014, 371(18): 1695-1703. DOI:10.1056/NEJMoa1403285
[43] Sinha P, Delucchi KL, Thompson BT, et al. Latent class analysis of ARDS subphenotypes: a secondary analysis of the statins for acutely injured lungs from sepsis (SAILS) study[J]. Intensive Care Med, 2018, 44(11): 1859-1869. DOI:10.1007/s00134-018-5378-3
[44] Sinha P, Delucchi KL, Chen Y, et al. Latent class analysis-derived subphenotypes are generalisable to observational cohorts of acute respiratory distress syndrome: a prospective study[J]. Thorax, 2022, 77(1): 13-21. DOI:10.1136/thoraxjnl-2021-217158
[45] Delucchi K, Famous KR, Ware LB, et al. Stability of ARDS subphenotypes over time in two randomised controlled trials[J]. Thorax, 2018, 73(5): 439-445. DOI:10.1136/thoraxjnl-2017-211090
[46] Dahmer MK, Yang GY, Zhang M, et al. Identification of phenotypes in paediatric patients with acute respiratory distress syndrome: a latent class analysis[J]. Lancet Respir Med, 2022, 10(3): 289-297. DOI:10.1016/S2213-2600(21)00382-9
[47] Bos LD, Schouten LR, van Vught LA, et al. Identification and validation of distinct biological phenotypes in patients with acute respiratory distress syndrome by cluster analysis[J]. Thorax, 2017, 72(10): 876-883. DOI:10.1136/thoraxjnl-2016-209719
[48] Sinha P, Delucchi KL, McAuley DF, et al. Development and validation of parsimonious algorithms to classify acute respiratory distress syndrome phenotypes: a secondary analysis of randomised controlled trials[J]. Lancet Respir Med, 2020, 8(3): 247-257. DOI:10.1016/S2213-2600(19)30369-8
[49] Sinha P, Churpek MM, Calfee CS. Machine learning classifier models can identify acute respiratory distress syndrome phenotypes using readily available clinical data[J]. Am J Respir Crit Care Med, 2020, 202(7): 996-1004. DOI:10.1164/rccm.202002-0347OC
[50] Maddali MV, Churpek M, Pham T, et al. Validation and utility of ARDS subphenotypes identified by machine-learning models using clinical data: an observational, multicohort, retrospective analysis[J]. Lancet Respir Med, 2022, 10(4): 367-377. DOI:10.1016/S2213-2600(21)00461-6
[51] Gattinoni L, Chiumello D, Caironi P, et al. COVID-19 pneumonia: different respiratory treatments for different phenotypes?[J]. Intensive Care Med, 2020, 46(6): 1099-1102. DOI:10.1007/s00134-020-06033-2
[52] Marini JJ, Gattinoni L. Management of COVID-19 respiratory distress[J]. JAMA, 2020, 323(22): 2329-2330. DOI:10.1001/jama.2020.6825
[53] Sinha P, Furfaro D, Cummings MJ, et al. Latent class analysis reveals COVID-19-related acute respiratory distress syndrome subgroups with differential responses to corticosteroids[J]. Am J Respir Crit Care Med, 2021, 204(11): 1274-1285. DOI:10.1164/rccm.202105-1302OC