en
×

分享给微信好友或者朋友圈

使用微信“扫一扫”功能。
通讯作者:

盛梅笑,E-mail:yfy0075@njucm.edu.

中图分类号:R692.5

文献标识码:A

文章编号:1007-4368(2024)11-1605-07

DOI:10.7655/NYDXBNSN240470

参考文献 1
TEITELBAUM I.Peritoneal dialysis[J].N Engl J Med,2021,385(19):1786-1795
参考文献 2
NEPHROLOGY I S O.ISN-global kidney health atlas[EB/OL][.2024-03-27].https://www.theisn.org/initiatives/global-kidney-health-atlas/
参考文献 3
《中国腹膜透析管理现状白皮书》项目组.中国腹膜透析管理现状白皮书[J].中华肾脏病杂志,2022,38(12):1076-1104
参考文献 4
倪兆慧,金海姣.中国腹膜透析发展70年[J].中国血液净化,2019,18(10):661-663
参考文献 5
辛洋洋,阳晓.腹膜透析标准化临床结局的研究进展[J].中华肾脏病杂志,2023,39(8):630-634
参考文献 6
MANERA K E,JOHNSON D W,CRAIG J C,et al.Estab-lishing a core outcome set for peritoneal dialysis:report of the SONG-PD(standardized outcomes in nephrology-peri-toneal dialysis)consensus workshop[J].Am J Kidney Dis,2020,75(3):404-412
参考文献 7
曹雪莹,周建辉,蔡广研,等.腹膜透析患者腹膜转运功能的影响因素及其维护[J].中华肾病研究电子杂志,2014,3(3):160-165
参考文献 8
LIN C Y,CHEN W P,YANG L Y,et al.Persistent trans-forming growth factor-beta 1 expression may predict peri-toneal fibrosis in CAPD patients with frequent peritonitis occurrence[J].Am J Nephrol,1998,18(6):513-519
参考文献 9
王超超,林永强,陈恬恬,等.腹膜透析患者透析液中 SGLT1、TGF-β1、VEGF水平与腹膜纤维化的关系及应用价值[J].中华全科医学,2022,20(2):251-254
参考文献 10
HO-DAC-PANNEKEET M M,HIRALALL J K,STRUIJK D G,et al.Markers of peritoneal mesothelial cells during treatment with peritoneal dialysis[J].Adv Perit Dial,1997,13:17-22
参考文献 11
VISSER C E,BROUWER-STEENBERGEN J J,BETJES M G,et al.Cancer antigen 125:a bulk marker for the me-sothelial mass in stable peritoneal dialysis patients[J].Nephrol Dial Transplant,1995,10(1):64-69
参考文献 12
SAMPIMON D E,KORTE M R,BARRETO D L,et al.Early diagnostic markers for encapsulating peritoneal sclerosis:a case-control study[J].Perit Dial Int,2010,30(2):163-169
参考文献 13
HAO N,CHIOU T T,WU C H,et al.Longitudinal chang-es of PAI-1,MMP-2,and VEGF in peritoneal effluents and their associations with peritoneal small-solute trans-fer rate in new peritoneal dialysis patients[J].Biomed Res Int,2019,2019:2152584
参考文献 14
WANG L,LIU N,XIONG C X,et al.Inhibition of EGF re-ceptor blocks the development and progression of perito-neal fibrosis[J].J Am Soc Nephrol,2016,27(9):2631-2644
参考文献 15
SHI Y F,LI J Q,CHEN H,et al.Inhibition of EZH2 sup-presses peritoneal angiogenesis by targeting a VEGFR2/ERK1/2/HIF-1α-dependent signaling pathway[J].J Pathol,2022,258(2):164-178
参考文献 16
YANG X X,TONG Y J,YAN H,et al.High intraperitone-al interleukin-6 levels predict peritonitis in peritoneal di-alysis patients:a prospective cohort study[J].Am J Nephrol,2018,47(5):317-324
参考文献 17
RATNA A,LIM A,LI Z H,et al.Myeloid endoplasmic re-ticulum resident chaperone GP96 facilitates inflammation and steatosis in alcohol-associated liver disease[J].Hepa-tol Commun,2021,5(7):1165-1182
参考文献 18
FANG J Y,TONG Y,JI O Y,et al.Glycoprotein 96 in peritoneal dialysis effluent-derived extracellular vesicles:a tool for evaluating peritoneal transport properties and in-flammatory status[J].Front Immunol,2022,13:824278
参考文献 19
LI X R,YANG S K,ZENG B Y,et al.Relationship be-tween peritoneal solute transport and dialysate inflamma-tory markers in peritoneal dialysis patients:a cross-sec-tional study[J].Nefrologia,2023,43(3):335-343
参考文献 20
MATERA L,NENNA R,FRASSANITO A,et al.Low lym-phocyte count:a clinical severity marker in infants with bronchiolitis[J].Pediatr Pulmonol,2022,57(7):1770-1775
参考文献 21
CECCATO A,PANAGIOTARAKOU M,RANZANI O T,et al.Lymphocytopenia as a predictor of mortality in pa-tients with ICU-acquired pneumonia[J].J Clin Med,2019,8(6):843
参考文献 22
HE Y J,HUANG X Y,ZHANG J W,et al.Decreased peripheral blood lymphocyte count predicts poor treat-ment response in peritoneal dialysis-associated peritoni-tis[J].J Inflamm Res,2023,16:5327-5338
参考文献 23
YANG Y,XU Y Y,ZHANG P,et al.Predictive value of objective nutritional indexes in technique failure in perito-neal dialysis patients[J].J Ren Nutr,2022,32(5):605-612
参考文献 24
ELGHIATY A,KIM J,JANG W S,et al.Preoperative con-trolling nutritional status(CONUT)score as a novel im-mune-nutritional predictor of survival in non-metastatic clear cell renal cell carcinoma of ≤ 7cm on preoperative imaging[J].J Cancer Res Clin Oncol,2019,145(4):957-965
参考文献 25
ZHOU H,CHAO W Y,CUI L,et al.Controlling Nutritional Status(CONUT)score as immune-nutritional predictor of outcomes in patients undergoing peritoneal dialysis[J].Clin Nutr,2020,39(8):2564-2570
参考文献 26
吴爱华,沐晓蝶,巢文英,等.CONUT 评分和透析龄对腹膜透析相关性腹膜炎的预测价值[J].中华医学杂志,2023,103(10):720-726
参考文献 27
TIAN Z,YAO N J,WANG F,et al.Thymosin β4 suppresses LPS-induced murine lung fibrosis by attenuating oxida-tive injury and alleviating inflammation[J].Inflamma-tion,2022,45(1):59-73
参考文献 28
TIAN J K,ZHANG R,ZHU N,et al.Association of serum thymosin β4 with malnutrition-inflammation-atherosclero-sis syndrome in peritoneal dialysis patients:a cross-sec-tional study[J].Ren Fail,2023,45(1):2202761
参考文献 29
REIS F M,DA SILVA M Z C,REIS N S D C,et al.Asso-ciation between phase angle and coronary artery calcium score in patients on peritoneal dialysis[J].Front Nutr,2022,9:912642
参考文献 30
DO J Y,KIM A Y,KANG S H.Association between phase angle and sarcopenia in patients undergoing peritoneal dialysis[J].Front Nutr,2021,8:742081
参考文献 31
WEN Y Q,ZHAN X J,WANG N S,et al.Monocyte/lym-phocyte ratio and cardiovascular disease mortality in peri-toneal dialysis patients[J].Mediators Inflamm,2020,2020:9852507
参考文献 32
ROSALES C.Neutrophils at the crossroads of innate and adaptive immunity[J].J Leukoc Biol,2020,108(1):377-396
参考文献 33
HUANG N Y,LI H Y,FAN L,et al.Serum phosphorus and albumin in patients undergoing peritoneal dialysis:in-teraction and association with mortality[J].Front Med,2021,8:760394
参考文献 34
YU Y,ZHONG Z,YANG W Z,et al.Neutrophil percent-age-to-albumin ratio and risk of mortality in patients on peritoneal dialysis[J].J Inflamm Res,2023,16:6271-6281
参考文献 35
李洋,王亚芬,李相,等.TyG、CRP/Alb、25(OH)D 与维持性腹膜透析患者预后的关联性[J].中国医师杂志,2022,24(9):1335-1339
参考文献 36
罗亚维,冯胜,沈华英,等.C反应蛋白/白蛋白比值是腹膜透析患者死亡的独立影响因素[J].中华肾脏病杂志,2022,38(6):528-535
参考文献 37
MA S,CAI Y M,WANG Z,et al.Derivation and valida-tion of a risk score predicting risk of early-onset peritoni-tis among patients initializing peritoneal dialysis:a cohort study[J].Int J Infect Dis,2020,99:301-306
参考文献 38
WANG Y X,SHEN Q Q,WANG J N,et al.The risk fac-tors and predictive model for cardiac valve calcification in patients on maintenance peritoneal dialysis:a single-cen-ter retrospective study[J].Ren Fail,2023,45(2):2271069
参考文献 39
刘晨媛,吴娟洁,董金玲,等.腹膜透析患者感染的影响因素及其风险预测列线图模型[J].中华医院感染学杂志,2023,33(20):3185-3189
参考文献 40
陈婷,龙海波,黄千殷,等.预测腹膜透析患者预后模型的构建与验证[J].中华肾脏病杂志,2020,36(9):680-687
参考文献 41
KANG S H,DO J Y.Phase angle as a risk factor for mor-tality in patients undergoing peritoneal dialysis[J].Nutri-ents,2023,15(23):4991
参考文献 42
HUANG D D,LI Y Y,QI X M,et al.A nomogram pre-dicts cardiovascular events in patients with peritoneal dialysis-associated peritonitis[J].Ren Fail,2022,44(1):1558-1567
参考文献 43
LI J N,ELLIS D G,PEPE A,et al.Back to the roots:re-constructing large and complex cranial defects using an image-based statistical shape model[J].J Med Syst,2024,48(1):55
参考文献 44
CARANFIL E,LAMI K,UEGAMI W,et al.Artificial in-telligence and lung pathology[J/OL].Adv Anat Pathol,2024(2024-05-23)[2024-05-27].DOI:10.1097/PAP.0000000000000448
参考文献 45
VAMATHEVAN J,CLARK D,CZODROWSKI P,et al.Applications of machine learning in drug discovery and development[J].Nat Rev Drug Discov,2019,18(6):463-477
参考文献 46
DASHTBAN A,MIZANI M A,PASEA L,et al.Identify-ing subtypes of chronic kidney disease with machine learning:development,internal validation and prognostic validation using linked electronic health records in 350 067 individuals[J].EBio Medicine,2023,89:104489
参考文献 47
DACK E,CHRISTE A,FONTANELLAZ M,et al.Artifi-cial intelligence and interstitial lung disease:diagnosis and prognosis[J].Invest Radiol,2023,58(8):602-609
参考文献 48
SMITH L A,OAKDEN-RAYNER L,BIRD A,et al.Ma-chine learning and deep learning predictive models for long-term prognosis in patients with chronic obstructive pulmonary disease:a systematic review and meta-analy-sis[J].Lancet Digit Health,2023,5(12):e872-e881
参考文献 49
XU L P,CAO F,WANG L,et al.Machine learning model and nomogram to predict the risk of heart failure hospital-ization in peritoneal dialysis patients[J].Ren Fail,2024,46(1):2324071
参考文献 50
YANG J,WAN J F,FENG L,et al.Machine learning algo-rithms for the prediction of adverse prognosis in patients undergoing peritoneal dialysis[J].BMC Med Inform De-cis Mak,2024,24(1):8
参考文献 51
ZANG Z Y,XU Q J,ZHOU X L,et al.Random forest can accurately predict the technique failure of peritoneal dialy-sis associated peritonitis patients[J].Front Med,2023,10:1335232
参考文献 52
唐雯,高峻逸,马辛宇,等.循环神经网络模型在腹膜透析临床预后预测中的初步应用[J].北京大学学报(医学版),2019,51(3):602-608
参考文献 53
XU Z Y,XU X,ZHU X M,et al.Attention-based deep learning model for prediction of major adverse cardiovas-cular events in peritoneal dialysis patients[J].IEEE J Biomed Health Inform,2024,28(2):1101-1109
参考文献 54
MA L T,ZHANG C H,GAO J Y,et al.Mortality predic-tion with adaptive feature importance recalibration for peritoneal dialysis patients[J].Patterns,2023,4(12):100892
目录contents

    摘要

    改善腹膜透析(peritoneal dialysis,PD)患者预后并制定标准化PD临床结局预测方法近年来受到广泛关注。目前公认的影响PD预后的核心结局有PD相关感染、心血管疾病、死亡、生活质量下降及技术失败。预测和识别PD临床结局预后不良的高风险患者,及早进行干预,对提高PD患者的生存率和生活质量、改善预后具有重要意义。文章从生物学标志物、风险预测模型、人工智能3个方面对近年来的相关研究进展进行综述。

    Abstract

    Improving the prognosis of patients with peritoneal dialysis(PD)and developing standardized method for predicting PD clinical outcomes have been received widespread attention in recent years. The currently recognized core outcomes affecting PD prognosis are PD-associated infections,cardiovascular disease,death,reduced quality of life,and technical failure. Predicting and identifying high-risk patients with poor prognosis for PD clinical outcomes and intervening early are of great importance for improving survival rate and quality of life as well as improving prognosis of PD patients. This paper reviews the relevant research progress in recent years from three aspects:biological markers,risk prediction models,and artificial intelligence.

  • 腹膜透析(peritoneal dialysis,PD)是终末期肾病 (end stage renal disease,ESRD)肾脏替代治疗方法之一。因可居家操作、成本较低、能够较好地保护残肾功能等优点在许多国家被推荐为肾衰一体化治疗的首选[1]。数据显示,全球PD中位患病率约为 21/100万[2]。近10年在中国内地,接受PD治疗的患者人数急剧上升,截至2022年12月底,中国内地PD 患者已达14万例,与1999 年相比,增长超30倍,虽然 PD患者生存率较前大幅提升,但仍有诸多因素影响 PD患者的预后[3-4]。改善PD患者预后并制定标准化 PD临床结局预测方法受到广泛关注[5],目前公认的影响PD预后的核心结局有PD相关感染、心血管疾病(cardiovascular disease,CVD)、死亡、生活质量下降及技术失败[6]。因此,预测和识别PD临床结局预后不良的高风险患者,及早进行干预,对提高PD患者的生存率和生活质量、改善预后具有重要意义。文章从PD生物学标志物、风险预测模型、人工智能 (artificial intelligence,AI)3 个方面对近年来相关研究进展进行综述,供研究者参考。

  • 1 生物学标志物

  • 能够预测 PD 结局的生物学标志物,主要取自腹膜透析液或血液。但是目前尚未有任何生物学标志物被纳入预测PD风险预后的常规临床实践中。因此,有必要探究敏感、特异的评估PD患者临床结局和预后风险的生物学标志物,目前的研究指标主要聚焦在腹膜纤维化(peritoneal fibrosis,PF)、腹膜透析相关腹膜炎症(peritoneal dialysis-associated peritonitis, PDAP)、营养不良、全因死亡和CVD等临床结局。

  • 1.1 PF指标

  • 腹膜功能障碍是PD患者发生不良结局的核心因素[7],由此导致透析不充分使患者退出PD治疗。 PF是导致腹膜功能障碍的重要病理基础。临床诊断PF的金标准是腹膜活检,但其是有创操作,难以在PD期间实施。目前临床上评估腹膜功能主要借助 PD 充分性指标(尿素清除指数、肌酐清除率)和腹膜平衡试验,但缺乏特异性。腹膜透析液中的转化生长因子(transforming growth factor,TGF)-β1 是关键的致纤维化因子,腹膜透析液中的葡萄糖和葡萄糖降解产物可激活TGF-β1,通过Smad和非Smad 通路调控纤维化基因表达。Lin等[8] 对15例腹膜炎低发(low peritonitis occurrence,LPO,频次<1次/年)患者和5例腹膜炎高发(high peritonitis occurrence,HPO, 1次/年≤频次<5次/年)患者进行随访,发现PDAP发作前和发作期间,腹透液中TGF-β1 表达增加,LPO 组在 PDAP 早期可检出 TGF-β1,2 周后检测不到,而HPO组TGF-β1持续存在。通过原位杂交技术探究腹膜活检标本中TGF-β1 mRNA表达与PF之间的关系,在植入PD导管时采集第1份样本,发现间皮细胞形态正常,检测不到TGF-β1 mRNA,在第 4 次腹膜炎发作后进行拔管时采集第 2 份样本,显示腹膜表面有一层坏死组织碎片和纤维蛋白,间皮细胞呈立方形,TGF-β1 mRNA在整个腹膜间皮细胞中分散表达,表明腹膜中持续的 TGF-β1 表达可作为预测 HPO患者发生PF的有效指标。王超超等[9] 对60例 PD患者进行研究,分为研究组25例(腹膜活检确诊为PF),对照组35例(经活检排除PF),发现研究组腹膜透析液中TGF-β1水平高于对照组,且其与超滤量、腹膜转运特性指标D/Pcr(0、2、4 h的腹膜透析液肌酐浓度/血肌酐浓度)呈负相关(P <0.05),与D/D0(2、 4 h 腹膜透析液葡萄糖含量/0 h 腹膜透析液葡萄糖含量)呈正相关(P <0.05),说明 TGF-β1 可参与 PF 形成并影响腹膜功能。机制可能是TGF-β1介导组织纤维化,其水平升高会导致体内上皮细胞及纤维母细胞过表达,从而促进腹膜组织上皮-间质转化及纤维化形成,故TGF-β1可作为一种预测PF及评估腹膜功能的生物学标志物。

  • 癌抗原 125(cancer antigen 125,CA125)的高水平表达主要见于上皮卵巢癌组织及其患者血清中,但在慢性肾衰竭、肝硬化等疾病中也有不同程度升高。有研究指出腹膜透析液中的 CA125 是评估腹膜间皮细胞质量最特异性的标志物[10],其浓度与腹膜透析液中的间皮细胞数量呈正相关[11],而腹膜间皮细胞发生间皮-间充质转化是PF早期的可逆环节,故腹膜透析液中的 CA125 可用于评估 PD 患者的腹膜间皮细胞质量以预测PF。Sampimon等[12] 对 11例以PF为特征的包裹性腹膜硬化症(encapsulat-ing peritoneal sclerosis,EPS)患者和 31 例对照患者进行研究,采用自回归模型(autoregressive model, AR)纳入接受PD治疗时间、腹膜透析液中的CA125 及白细胞介素(interleukin,IL)-6 浓度,测算诊断 EPS的灵敏度及特异度,发现CA125<33 U/min同时 IL-6 >350 pg/min 共同诊断 EPS 具有 70%的灵敏度和 89%的特异度,提示 CA125 合并 IL-6 可用于评估未感染患者的腹膜功能以预测PF。

  • 腹膜透析液中的血管内皮生长因子(vascular endothelial growth factor,VEGF)可通过诱导腹膜内皮细胞增殖,促进新生血管形成进而导致PF。Hao 等[13] 发现VEGF水平随PD时间延长而升高。Wang 等[14] 研究表明腹膜间皮层血管的数量与腹膜透析液中的VEGF水平呈正相关,抑制其表达可抑制血管新生从而延缓 PF。故监测 VEGF 的表达可预测 PF的发生。Zeste基因同源蛋白2(enhancer of Zeste homolog2,EZH2)是一种组蛋白甲基转移酶,可调控基因表达。研究发现腹膜透析液中的EZH2在长期 PD患者腹膜和小鼠PF模型中的表达显著增加,其可通过激活VEGF受体2-细胞外信号调节激酶(ex-tracellular signal-regulated kinase,ERK)1/2-缺氧诱导因子 1α(hypoxia inducible factor 1α,HIF1α)信号轴(VEGFR2-ERK1/2-HIF1α),促进腹膜血管新生,导致腹膜超滤失败,可作为一种潜在的非侵入性预测诊断工具和治疗新靶点[15]

  • 1.2 PDAP指标

  • PDAP是PD过程中最常见的并发症,目前其临床诊断主要根据腹膜透析液中白细胞计数、细菌培养及症状体征。随着对PDAP 研究的深入,涌现了一些生物学诊断标志物。IL-6是反映机体感染的重要炎症因子,腹膜内IL-6表达增多提示腹膜内存在炎症状态。Yang等[16] 纳入149例PD患者,根据腹膜透析液中IL-6浓度高低进行分组统计分析,研究发现PDAP发生与腹膜透析液中IL-6水平升高显著相关,腹膜透析液中IL-6 水平是PD患者发生PDAP的潜在预测因子。糖蛋白96(glycoprotein 96,GP96)属于热休克蛋白90家族,可通过激活巨噬细胞和释放炎症细胞因子引发炎症反应。Ratna等[17] 研究证实 GP96抑制剂可显著减少大鼠PD模型中腹膜的局部炎症细胞浸润和促炎基因表达。Fang 等[18] 发现富含 GP96 的细胞外囊泡可刺激巨噬细胞分泌促炎细胞因子,且在高转运患者的PD滤出液中GP96表达水平较高,而高转运腹膜特征与炎症状态密切相关[19],因此腹膜透析液中的 GP96 可作为评估 PDAP 的潜在指标。淋巴细胞可识别特异性抗原、激活免疫反应,在免疫应答中发挥核心作用。血淋巴细胞计数 (lymphocyte count,LYC)减少与疾病的不良预后有关[20-21]。He等[22] 研究证实血清中LYC减少与PDAP 治疗失败风险增加有关,且较其他免疫炎症指标具有采样简便、成本低廉等优点,是预测PDAP不良预后较为实用的指标。

  • 1.3 营养不良指标

  • PD患者普遍存在营养不良,原因有炎症、营养摄入不足、蛋白质丢失、透析不充分等。目前临床缺少理想标志物,大多依赖一些替代测量方法,如人体相关围度测量、人体成分分析、体重指数(body mass index,BMI)及血清白蛋白(albumin,Alb)等。控制性营养状况(controlling nutritional status,CONUT) 评分是评估营养状况的指标,包括血清LYC、Alb和总胆固醇3个参数,近年多项研究显示可用于预测 PD患者的不良预后[23-24]。Zhou等[25] 研究发现PD发生时 CONUT 评分>3 的人群更易退出治疗,与低 CONUT 评分组相比,CONUT 评分高的患者死亡风险比为 1.57,与单参数标志物相比,CONUT 评分能更好反映人体营养和免疫状况,增强预测能力,是 PD患者全因死亡、CVD和技术失败风险可靠的预后标志物。也有研究发现CONUT评分联合透析龄对 PDAP 有较高预测价值[26]。血清胸腺素β4(serum thymosin β4,sTβ4)是一种多功能蛋白,在损伤修复、抗炎、抗纤维化中发挥多种作用[27]。Tian等[28] 发现sTβ4水平与PD患者的营养状况呈正相关,其可通过抑制炎症,提高机体代谢改善 PD 患者营养状况,同时也发现sTβ4对预测PD患者发生营养不良-炎症-动脉粥样硬化综合征有一定作用;且由于sTβ4水平与BMI、血清Alb、生物电阻抗技术测量所得的相位角(phase angle,PhA)呈正相关,可考虑作为评估PD患者营养状态的新指标[29-30]

  • 1.4 全因死亡和CVD指标

  • CVD 是 PD 的严重并发症之一,与不良预后相关,是PD患者的主要致死原因。一项纳入1 753例 PD 患者的多中心回顾性研究显示,由 CVD 事件引起的 PD 患者的病死率为 54.3%[31]。中性粒细胞白蛋白比值(neutrophil percentage-to-albumin ratio, NPAR)是基于血清中的中性粒细胞和Alb两个评估参数的指标,近来被推荐为全身炎症反应的新标志物。中性粒细胞作为急性炎症的关键介质,在慢性炎症中也起着至关重要的作用[32]。Alb具有多种功能,如维持血浆渗透压、运输物质和抗氧化作用,低白蛋白血症是PD患者发生全因死亡的重要危险因素[33]。NPAR 能放大两个参数的变化,高 NPAR 水平是 PD 患者全因死亡和CVD的独立危险因素,与血清中的中性粒细胞/淋巴细胞比率(neutrophil-to-lymphocyte ratio,NLR)、血小板/淋巴细胞比率 (platelet-to-lymphocyte ratio,PLR)相比,NPAR对PD 死亡的预测价值更高[34]。李洋等[35] 利用受试者工作特征(receiver operating characteristic,ROC)曲线分析得出,血清中的甘油三酯葡萄糖乘积指数(tri-glyceride glucos,TyG)偏高、C 反应蛋白(C-reactive protein,CRP)与白蛋白比值(C-reaction protein/albu-min ratio,CAR)偏高、25-羟基维生素 D[25-hydroxy vitamin D,25(OH)D]偏低提示 PD 患者全因死亡风险更高,3 项指标联合的 ROC 曲线的曲线下面积 (area under the curve,AUC)高于单个指标,且各指标预测价值随透析时间增加而增大。罗亚维等[36] 发现当血清 CAR>0.19 mg/g 时,PD 患者的死亡风险升高,其准确性较 CRP、NLR 及 PLR 更高,可作为一种新型标志物预测 PD 患者的全因死亡和 CVD死亡风险。

  • 2 风险预测模型

  • 风险预测模型主要有Logistic 回归模型和比例风险回归模型(Cox回归模型)。前者主要用于预测某个事件是否发生或某个分类结果是否出现;后者可用于分析患者的生存时间、治疗方式、疾病进展等因素之间的关系。高性能的临床预测模型,有助于识别高危患者并进行风险分类,指导医生作出决策,还有助于行政部门进行更好的医疗质量管理,合理配置医疗资源。

  • 2.1 Logistic回归模型

  • Ma等[37] 纳入3 772例PD患者,开发并验证了初始 PD 患者发生 PDAP 的 Logistic 回归预测模型,结果显示PDAP发生的绝对风险为14.5%,年龄、心功能、血清电解质、血脂、肝功能、血尿素氮和白细胞计数是最终风险评分中的重要预测因子,该模型的一致性指数(concordance index,C-index)为 0.76,模型性能良好,适用于发展中国家的初始 PD 患者,可作为筛查 PDAP 高风险人群的工具。Wang 等[38] 利用Logistic回归模型和列线图构建了预测PD患者出现心脏瓣膜钙化的临床模型,纳入的危险因素包括年龄、性别、收缩压、钙磷乘积、透析龄和查尔森合并症指数(Charlson comorbidity index,CCI),模型的 C-index 为 0.845,有较高预测性能,具有良好的辨别力和准确性。刘晨媛等[39] 基于Logistic回归建立了预测 PD 患者发生医院感染的列线图,纳入的预测指标有Alb、CRP、血红蛋白、血磷、透析龄、营养状态、糖尿病病史,模型的C-index为0.962,预测性能较准。通常认为,构建Logistic回归分析模型对数据的完整度要求较高,预测效能与样本量、数据完整度密切相关,临床常用来筛选与 PD 相关的一些危险因素;不足的是其仅考虑PD随访结局,未考虑出现结局的时间长短,即无论结局变量发生在随访早期还是晚期,对其处理均相同。

  • 2.2 Cox回归模型

  • 陈婷等[40] 采用Cox回归模型筛选出高龄、血总胆固醇、卒中史、NLR 是 PD 死亡的独立危险因素, Alb是保护因素;用R语言构建模型,训练集和验证集的C-index分别为0.815和0.804,能较准确预测患者1年、3年生存率,具有较高的外部实用性。Kang 等[41] 为评估PhA在预测PD患者死亡率和技术失败方面的作用,利用Cox回归模型对PhA与其他危险因素(Alb、BMI、CRP、尿量)的可预测性进行比较,发现 PhA 较其他指标能更有效地预测患者死亡 (AUC为0.80),且患者生存率和技术生存率与PhA 呈正比,单因素和多因素Cox回归分析结果一致,所以PhA可能是预测PD患者生存率和技术生存率的有效指标,对临床治疗有指导意义。Huang等[42] 纳入150例PDAP患者,采用Cox回归和列线图构建预测PDAP 发生心血管不良事件的模型,危险因素有 Alb、碱性磷酸酶、65岁以上、有CVD史和腹膜透析液培养阳性,C-index为0.732,模型校准率一般。分析原因是研究样本量小,未进行外部验证,导致模型临床意义降低;回顾性研究不可避免地存在回忆和选择偏倚;未纳入脑钠肽、左心室功能、冠状动脉 CT等心脏功能评估指标。Cox回归模型可分析带有删失数据的资料并分析多个因素对生存时间的影响,但需注意删失比例过大会造成偏倚增大,故临床中需要更多大样本、多中心的数据,纳入更多因素去构建更可靠的Cox比例风险回归模型。

  • 3 AI

  • AI 是利用计算机学习模拟人类行为的一门综合性特殊学科,主要核心是机器学习,机器学习中的重要分支为深度学习。AI 利用非线性方法可更快更准地找到突破靶点,弥补传统方法的不足。随着大数据技术的蓬勃发展,AI 已广泛运用于与医疗相关的多个领域,如医学成像和诊断[43-44]、药物开发[45]、病历档案管理[46]、疾病诊断和预后[47-48] 等。近年来,深度学习技术飞速发展,对图像语音识别、自然语言理解等多个领域产生了颠覆性改变。在实际应用中,机器学习和深度学习也各有优缺点。

  • 3.1 机器学习

  • Xu等[49] 纳入1 006例PD患者,通过机器学习预测 PD 患者发生心力衰竭住院的风险,发现机器学习较Cox回归模型有更好的区分能力,且能准确预测 PD 患者出现心衰和全因死亡的风险,其中分布式梯度增强库模型(AUC为0.871)预测PD患者死亡率的效果最佳,并归纳出 PD 患者死亡的危险因素有年龄、CCI评分、肌酐、高密度脂蛋白胆固醇、总胆固醇、基线肾小球滤过率。Yang 等[50] 利用常见的 5 种机器学习算法构建预测 PD 预后的模型,发现 Cat Boost 模型预测性能最佳,该模型是 1 种基于梯度提升决策树的机器学习算法,基本原理是通过迭代添加新的决策树来改进现有预测模型的性能,每棵新的决策树都在负梯度方向上生长,从而使损失函数最小化。其还采用了1种“基于树的模型”的集成方法,可自动处理特征选择和特征缩放等任务,使模型更加高效,能有效处理分类特征,加快训练速度并提高模型性能。临床可用于预测患者是否适合PD治疗以及发生PDAP的可能性,从而为患者制定最佳透析方案。Zang 等[51] 通过多种机器学习开发并验证了预测PDAP 患者技术失败的模型,发现随机森林模型的预测性能最优异,同时发现N末端B型利钠肽原、纤维蛋白原和铁蛋白是 PDAP 患者技术失败的重要预测因子,借助预测模型,可准确识别PDAP 中易出现技术失败的高危患者,并提前进行干预。随机森林模型的原理主要是通过构建多个决策树来解决分类和回归问题,然后通过取平均值或取多数投票的方式来提高预测准确性、泛化能力和抗过拟合能力。随机森林模型与Cat Boost 模型相比,二者的区别主要在算法和实际应用效果方面。算法层面,前者通过对数据集进行随机采样来构建训练样本,其随机化有利于模型在测试集上取得更好的泛化性能,后者根据训练数据寻找所有决策树最优的线性组合,这种额外的调整可理解为两者的差异;应用层面,后者会根据观测值对预测结果进行调整,易受到噪声点影响,更易出现过拟合的情况,而前者抗过拟合能力比较强。总之,与传统风险预测模型相比,机器学习通常需要大量标记的数据来训练模型,因此能更好地处理具有大量变量的高维数据集,能自动从数据中学习和捕捉复杂的关系模式,提供更为灵活和强大的建模能力,但需要注意数据的质量和多样性会直接影响模型性能。

  • 3.2 深度学习

  • 唐雯等[52] 纳入 656 例 PD 患者,收集开始透析时、随访和预后数据,使用基于循环神经网络模型的死亡风险预测系统辅助医生进行临床决策和干预;发现模型对感染、胃肠道出血、营养不良和癌症等导致的死亡的预测效果较好,对心、脑血管疾病的预测效果不佳;同时相比传统Logistic回归模型,门控循环单元模型的性能更优,更适合预测 PD 患者的死亡风险。目前传统的预后模型难以处理具有高纬度、异质性和时间序列的电子病历数据。Xu 等[53] 基于7 539例PD患者数据开发并验证了一种神经网络CVD former模型,用于预测PD患者3个月内全因和心血管死亡风险,在测试集中,CVD former模型精度优于长短期记忆网络模型,模型的 AUC 为 0.88~0.90,利用低成本、广泛可用的自动化程序预测PD患者的临床预后风险,对于PD患者的管理具有重要意义。Ma等[54] 利用深度学习,整理656例PD 患者共13 091次门诊随访电子数据集,搭建个体化预后分析AI模型,智能预测不良结局的发生概率、个体化判别关键因素、重新判断指标参考值,提供详细的指标分析,如Alb、舒张压和血氯水平可作为 PD 患者的重要健康因素;血钠、血钾和体重是恶病质死亡的重要因素;Alb、尿素、体重、血钾、血压是 PDAP 死亡的重要因素。同时开发了 AI-医生交互预后预测系统,用于可视化患者疾病进展轨迹和关键因素,在临床中帮助医生尽早识别相关风险并进行干预。与机器学习相比,深度学习对大规模数据集的需求更为明显,其优点在于强大的数据处理分析能力,但是深度学习往往需要耗费大量的计算资源和训练时间,且对数据的依赖性较高,故在实际应用中还需根据具体问题和数据特点选择合适的模型。

  • 4 小结与展望

  • 综上,在 PD 患者临床结局预后风险预测研究中,生物学标志物研究进展相对缓慢,临床应用有限,目前尚未有任何生物学标志物被纳入 PD 的常规临床实践中,潜在的能够预测PF的指标有腹膜透析液中的 TGF-β1、CA125、VEGF 及 EZH2;预测 PDAP的指标有腹膜透析液中的IL-6、GP96及血清 LYC;预测营养不良的指标有CONUT评分、sTβ4;预测 CVD 的指标有血清 NPAR、TyG、CAR、25(OH)D。以上生物学标志物的优点是采样简单方便,可直观监测患者病情变化,但其临床准确性还需大量临床试验论证,研究证实多个生物学标志物联合运用较单项评估准确性更高,目前主要用于 PD 相关结局具体机制的探索研究。风险预测模型方面,常用的是Logistic回归和Cox回归模型,优点是可在一定程度上分析 PD 预后风险的危险因素并构建模型,但存在回归性数据偏倚和临床观察指标有限的缺点,模型的临床普适性参差不齐,准确度和可解释性方面仍待改进。目前主要是各个单中心构建个体化模型,确定影响 PD 预后的因素并运用于临床。AI 的优点是可利用非线性方法更快更准地找到突破点,弥补传统方法的不足,在多维度数据处理准确度、患者个性化服务和可解释性等方面较前两者效果更好,但临床上应结合患者的具体情况和客观检查结果综合考量,目前也只在少数研发单位进行应用,亟需进一步实践论证。未来在临床如何合理应用上述预测方法,在此基础上建立个性化的 PD 诊疗方案,提高患者生存率和生活质量、改善预后,值得探索。

  • 参考文献

    • [1] TEITELBAUM I.Peritoneal dialysis[J].N Engl J Med,2021,385(19):1786-1795

    • [2] NEPHROLOGY I S O.ISN-global kidney health atlas[EB/OL][.2024-03-27].https://www.theisn.org/initiatives/global-kidney-health-atlas/

    • [3] 《中国腹膜透析管理现状白皮书》项目组.中国腹膜透析管理现状白皮书[J].中华肾脏病杂志,2022,38(12):1076-1104

    • [4] 倪兆慧,金海姣.中国腹膜透析发展70年[J].中国血液净化,2019,18(10):661-663

    • [5] 辛洋洋,阳晓.腹膜透析标准化临床结局的研究进展[J].中华肾脏病杂志,2023,39(8):630-634

    • [6] MANERA K E,JOHNSON D W,CRAIG J C,et al.Estab-lishing a core outcome set for peritoneal dialysis:report of the SONG-PD(standardized outcomes in nephrology-peri-toneal dialysis)consensus workshop[J].Am J Kidney Dis,2020,75(3):404-412

    • [7] 曹雪莹,周建辉,蔡广研,等.腹膜透析患者腹膜转运功能的影响因素及其维护[J].中华肾病研究电子杂志,2014,3(3):160-165

    • [8] LIN C Y,CHEN W P,YANG L Y,et al.Persistent trans-forming growth factor-beta 1 expression may predict peri-toneal fibrosis in CAPD patients with frequent peritonitis occurrence[J].Am J Nephrol,1998,18(6):513-519

    • [9] 王超超,林永强,陈恬恬,等.腹膜透析患者透析液中 SGLT1、TGF-β1、VEGF水平与腹膜纤维化的关系及应用价值[J].中华全科医学,2022,20(2):251-254

    • [10] HO-DAC-PANNEKEET M M,HIRALALL J K,STRUIJK D G,et al.Markers of peritoneal mesothelial cells during treatment with peritoneal dialysis[J].Adv Perit Dial,1997,13:17-22

    • [11] VISSER C E,BROUWER-STEENBERGEN J J,BETJES M G,et al.Cancer antigen 125:a bulk marker for the me-sothelial mass in stable peritoneal dialysis patients[J].Nephrol Dial Transplant,1995,10(1):64-69

    • [12] SAMPIMON D E,KORTE M R,BARRETO D L,et al.Early diagnostic markers for encapsulating peritoneal sclerosis:a case-control study[J].Perit Dial Int,2010,30(2):163-169

    • [13] HAO N,CHIOU T T,WU C H,et al.Longitudinal chang-es of PAI-1,MMP-2,and VEGF in peritoneal effluents and their associations with peritoneal small-solute trans-fer rate in new peritoneal dialysis patients[J].Biomed Res Int,2019,2019:2152584

    • [14] WANG L,LIU N,XIONG C X,et al.Inhibition of EGF re-ceptor blocks the development and progression of perito-neal fibrosis[J].J Am Soc Nephrol,2016,27(9):2631-2644

    • [15] SHI Y F,LI J Q,CHEN H,et al.Inhibition of EZH2 sup-presses peritoneal angiogenesis by targeting a VEGFR2/ERK1/2/HIF-1α-dependent signaling pathway[J].J Pathol,2022,258(2):164-178

    • [16] YANG X X,TONG Y J,YAN H,et al.High intraperitone-al interleukin-6 levels predict peritonitis in peritoneal di-alysis patients:a prospective cohort study[J].Am J Nephrol,2018,47(5):317-324

    • [17] RATNA A,LIM A,LI Z H,et al.Myeloid endoplasmic re-ticulum resident chaperone GP96 facilitates inflammation and steatosis in alcohol-associated liver disease[J].Hepa-tol Commun,2021,5(7):1165-1182

    • [18] FANG J Y,TONG Y,JI O Y,et al.Glycoprotein 96 in peritoneal dialysis effluent-derived extracellular vesicles:a tool for evaluating peritoneal transport properties and in-flammatory status[J].Front Immunol,2022,13:824278

    • [19] LI X R,YANG S K,ZENG B Y,et al.Relationship be-tween peritoneal solute transport and dialysate inflamma-tory markers in peritoneal dialysis patients:a cross-sec-tional study[J].Nefrologia,2023,43(3):335-343

    • [20] MATERA L,NENNA R,FRASSANITO A,et al.Low lym-phocyte count:a clinical severity marker in infants with bronchiolitis[J].Pediatr Pulmonol,2022,57(7):1770-1775

    • [21] CECCATO A,PANAGIOTARAKOU M,RANZANI O T,et al.Lymphocytopenia as a predictor of mortality in pa-tients with ICU-acquired pneumonia[J].J Clin Med,2019,8(6):843

    • [22] HE Y J,HUANG X Y,ZHANG J W,et al.Decreased peripheral blood lymphocyte count predicts poor treat-ment response in peritoneal dialysis-associated peritoni-tis[J].J Inflamm Res,2023,16:5327-5338

    • [23] YANG Y,XU Y Y,ZHANG P,et al.Predictive value of objective nutritional indexes in technique failure in perito-neal dialysis patients[J].J Ren Nutr,2022,32(5):605-612

    • [24] ELGHIATY A,KIM J,JANG W S,et al.Preoperative con-trolling nutritional status(CONUT)score as a novel im-mune-nutritional predictor of survival in non-metastatic clear cell renal cell carcinoma of ≤ 7cm on preoperative imaging[J].J Cancer Res Clin Oncol,2019,145(4):957-965

    • [25] ZHOU H,CHAO W Y,CUI L,et al.Controlling Nutritional Status(CONUT)score as immune-nutritional predictor of outcomes in patients undergoing peritoneal dialysis[J].Clin Nutr,2020,39(8):2564-2570

    • [26] 吴爱华,沐晓蝶,巢文英,等.CONUT 评分和透析龄对腹膜透析相关性腹膜炎的预测价值[J].中华医学杂志,2023,103(10):720-726

    • [27] TIAN Z,YAO N J,WANG F,et al.Thymosin β4 suppresses LPS-induced murine lung fibrosis by attenuating oxida-tive injury and alleviating inflammation[J].Inflamma-tion,2022,45(1):59-73

    • [28] TIAN J K,ZHANG R,ZHU N,et al.Association of serum thymosin β4 with malnutrition-inflammation-atherosclero-sis syndrome in peritoneal dialysis patients:a cross-sec-tional study[J].Ren Fail,2023,45(1):2202761

    • [29] REIS F M,DA SILVA M Z C,REIS N S D C,et al.Asso-ciation between phase angle and coronary artery calcium score in patients on peritoneal dialysis[J].Front Nutr,2022,9:912642

    • [30] DO J Y,KIM A Y,KANG S H.Association between phase angle and sarcopenia in patients undergoing peritoneal dialysis[J].Front Nutr,2021,8:742081

    • [31] WEN Y Q,ZHAN X J,WANG N S,et al.Monocyte/lym-phocyte ratio and cardiovascular disease mortality in peri-toneal dialysis patients[J].Mediators Inflamm,2020,2020:9852507

    • [32] ROSALES C.Neutrophils at the crossroads of innate and adaptive immunity[J].J Leukoc Biol,2020,108(1):377-396

    • [33] HUANG N Y,LI H Y,FAN L,et al.Serum phosphorus and albumin in patients undergoing peritoneal dialysis:in-teraction and association with mortality[J].Front Med,2021,8:760394

    • [34] YU Y,ZHONG Z,YANG W Z,et al.Neutrophil percent-age-to-albumin ratio and risk of mortality in patients on peritoneal dialysis[J].J Inflamm Res,2023,16:6271-6281

    • [35] 李洋,王亚芬,李相,等.TyG、CRP/Alb、25(OH)D 与维持性腹膜透析患者预后的关联性[J].中国医师杂志,2022,24(9):1335-1339

    • [36] 罗亚维,冯胜,沈华英,等.C反应蛋白/白蛋白比值是腹膜透析患者死亡的独立影响因素[J].中华肾脏病杂志,2022,38(6):528-535

    • [37] MA S,CAI Y M,WANG Z,et al.Derivation and valida-tion of a risk score predicting risk of early-onset peritoni-tis among patients initializing peritoneal dialysis:a cohort study[J].Int J Infect Dis,2020,99:301-306

    • [38] WANG Y X,SHEN Q Q,WANG J N,et al.The risk fac-tors and predictive model for cardiac valve calcification in patients on maintenance peritoneal dialysis:a single-cen-ter retrospective study[J].Ren Fail,2023,45(2):2271069

    • [39] 刘晨媛,吴娟洁,董金玲,等.腹膜透析患者感染的影响因素及其风险预测列线图模型[J].中华医院感染学杂志,2023,33(20):3185-3189

    • [40] 陈婷,龙海波,黄千殷,等.预测腹膜透析患者预后模型的构建与验证[J].中华肾脏病杂志,2020,36(9):680-687

    • [41] KANG S H,DO J Y.Phase angle as a risk factor for mor-tality in patients undergoing peritoneal dialysis[J].Nutri-ents,2023,15(23):4991

    • [42] HUANG D D,LI Y Y,QI X M,et al.A nomogram pre-dicts cardiovascular events in patients with peritoneal dialysis-associated peritonitis[J].Ren Fail,2022,44(1):1558-1567

    • [43] LI J N,ELLIS D G,PEPE A,et al.Back to the roots:re-constructing large and complex cranial defects using an image-based statistical shape model[J].J Med Syst,2024,48(1):55

    • [44] CARANFIL E,LAMI K,UEGAMI W,et al.Artificial in-telligence and lung pathology[J/OL].Adv Anat Pathol,2024(2024-05-23)[2024-05-27].DOI:10.1097/PAP.0000000000000448

    • [45] VAMATHEVAN J,CLARK D,CZODROWSKI P,et al.Applications of machine learning in drug discovery and development[J].Nat Rev Drug Discov,2019,18(6):463-477

    • [46] DASHTBAN A,MIZANI M A,PASEA L,et al.Identify-ing subtypes of chronic kidney disease with machine learning:development,internal validation and prognostic validation using linked electronic health records in 350 067 individuals[J].EBio Medicine,2023,89:104489

    • [47] DACK E,CHRISTE A,FONTANELLAZ M,et al.Artifi-cial intelligence and interstitial lung disease:diagnosis and prognosis[J].Invest Radiol,2023,58(8):602-609

    • [48] SMITH L A,OAKDEN-RAYNER L,BIRD A,et al.Ma-chine learning and deep learning predictive models for long-term prognosis in patients with chronic obstructive pulmonary disease:a systematic review and meta-analy-sis[J].Lancet Digit Health,2023,5(12):e872-e881

    • [49] XU L P,CAO F,WANG L,et al.Machine learning model and nomogram to predict the risk of heart failure hospital-ization in peritoneal dialysis patients[J].Ren Fail,2024,46(1):2324071

    • [50] YANG J,WAN J F,FENG L,et al.Machine learning algo-rithms for the prediction of adverse prognosis in patients undergoing peritoneal dialysis[J].BMC Med Inform De-cis Mak,2024,24(1):8

    • [51] ZANG Z Y,XU Q J,ZHOU X L,et al.Random forest can accurately predict the technique failure of peritoneal dialy-sis associated peritonitis patients[J].Front Med,2023,10:1335232

    • [52] 唐雯,高峻逸,马辛宇,等.循环神经网络模型在腹膜透析临床预后预测中的初步应用[J].北京大学学报(医学版),2019,51(3):602-608

    • [53] XU Z Y,XU X,ZHU X M,et al.Attention-based deep learning model for prediction of major adverse cardiovas-cular events in peritoneal dialysis patients[J].IEEE J Biomed Health Inform,2024,28(2):1101-1109

    • [54] MA L T,ZHANG C H,GAO J Y,et al.Mortality predic-tion with adaptive feature importance recalibration for peritoneal dialysis patients[J].Patterns,2023,4(12):100892

  • 参考文献

    • [1] TEITELBAUM I.Peritoneal dialysis[J].N Engl J Med,2021,385(19):1786-1795

    • [2] NEPHROLOGY I S O.ISN-global kidney health atlas[EB/OL][.2024-03-27].https://www.theisn.org/initiatives/global-kidney-health-atlas/

    • [3] 《中国腹膜透析管理现状白皮书》项目组.中国腹膜透析管理现状白皮书[J].中华肾脏病杂志,2022,38(12):1076-1104

    • [4] 倪兆慧,金海姣.中国腹膜透析发展70年[J].中国血液净化,2019,18(10):661-663

    • [5] 辛洋洋,阳晓.腹膜透析标准化临床结局的研究进展[J].中华肾脏病杂志,2023,39(8):630-634

    • [6] MANERA K E,JOHNSON D W,CRAIG J C,et al.Estab-lishing a core outcome set for peritoneal dialysis:report of the SONG-PD(standardized outcomes in nephrology-peri-toneal dialysis)consensus workshop[J].Am J Kidney Dis,2020,75(3):404-412

    • [7] 曹雪莹,周建辉,蔡广研,等.腹膜透析患者腹膜转运功能的影响因素及其维护[J].中华肾病研究电子杂志,2014,3(3):160-165

    • [8] LIN C Y,CHEN W P,YANG L Y,et al.Persistent trans-forming growth factor-beta 1 expression may predict peri-toneal fibrosis in CAPD patients with frequent peritonitis occurrence[J].Am J Nephrol,1998,18(6):513-519

    • [9] 王超超,林永强,陈恬恬,等.腹膜透析患者透析液中 SGLT1、TGF-β1、VEGF水平与腹膜纤维化的关系及应用价值[J].中华全科医学,2022,20(2):251-254

    • [10] HO-DAC-PANNEKEET M M,HIRALALL J K,STRUIJK D G,et al.Markers of peritoneal mesothelial cells during treatment with peritoneal dialysis[J].Adv Perit Dial,1997,13:17-22

    • [11] VISSER C E,BROUWER-STEENBERGEN J J,BETJES M G,et al.Cancer antigen 125:a bulk marker for the me-sothelial mass in stable peritoneal dialysis patients[J].Nephrol Dial Transplant,1995,10(1):64-69

    • [12] SAMPIMON D E,KORTE M R,BARRETO D L,et al.Early diagnostic markers for encapsulating peritoneal sclerosis:a case-control study[J].Perit Dial Int,2010,30(2):163-169

    • [13] HAO N,CHIOU T T,WU C H,et al.Longitudinal chang-es of PAI-1,MMP-2,and VEGF in peritoneal effluents and their associations with peritoneal small-solute trans-fer rate in new peritoneal dialysis patients[J].Biomed Res Int,2019,2019:2152584

    • [14] WANG L,LIU N,XIONG C X,et al.Inhibition of EGF re-ceptor blocks the development and progression of perito-neal fibrosis[J].J Am Soc Nephrol,2016,27(9):2631-2644

    • [15] SHI Y F,LI J Q,CHEN H,et al.Inhibition of EZH2 sup-presses peritoneal angiogenesis by targeting a VEGFR2/ERK1/2/HIF-1α-dependent signaling pathway[J].J Pathol,2022,258(2):164-178

    • [16] YANG X X,TONG Y J,YAN H,et al.High intraperitone-al interleukin-6 levels predict peritonitis in peritoneal di-alysis patients:a prospective cohort study[J].Am J Nephrol,2018,47(5):317-324

    • [17] RATNA A,LIM A,LI Z H,et al.Myeloid endoplasmic re-ticulum resident chaperone GP96 facilitates inflammation and steatosis in alcohol-associated liver disease[J].Hepa-tol Commun,2021,5(7):1165-1182

    • [18] FANG J Y,TONG Y,JI O Y,et al.Glycoprotein 96 in peritoneal dialysis effluent-derived extracellular vesicles:a tool for evaluating peritoneal transport properties and in-flammatory status[J].Front Immunol,2022,13:824278

    • [19] LI X R,YANG S K,ZENG B Y,et al.Relationship be-tween peritoneal solute transport and dialysate inflamma-tory markers in peritoneal dialysis patients:a cross-sec-tional study[J].Nefrologia,2023,43(3):335-343

    • [20] MATERA L,NENNA R,FRASSANITO A,et al.Low lym-phocyte count:a clinical severity marker in infants with bronchiolitis[J].Pediatr Pulmonol,2022,57(7):1770-1775

    • [21] CECCATO A,PANAGIOTARAKOU M,RANZANI O T,et al.Lymphocytopenia as a predictor of mortality in pa-tients with ICU-acquired pneumonia[J].J Clin Med,2019,8(6):843

    • [22] HE Y J,HUANG X Y,ZHANG J W,et al.Decreased peripheral blood lymphocyte count predicts poor treat-ment response in peritoneal dialysis-associated peritoni-tis[J].J Inflamm Res,2023,16:5327-5338

    • [23] YANG Y,XU Y Y,ZHANG P,et al.Predictive value of objective nutritional indexes in technique failure in perito-neal dialysis patients[J].J Ren Nutr,2022,32(5):605-612

    • [24] ELGHIATY A,KIM J,JANG W S,et al.Preoperative con-trolling nutritional status(CONUT)score as a novel im-mune-nutritional predictor of survival in non-metastatic clear cell renal cell carcinoma of ≤ 7cm on preoperative imaging[J].J Cancer Res Clin Oncol,2019,145(4):957-965

    • [25] ZHOU H,CHAO W Y,CUI L,et al.Controlling Nutritional Status(CONUT)score as immune-nutritional predictor of outcomes in patients undergoing peritoneal dialysis[J].Clin Nutr,2020,39(8):2564-2570

    • [26] 吴爱华,沐晓蝶,巢文英,等.CONUT 评分和透析龄对腹膜透析相关性腹膜炎的预测价值[J].中华医学杂志,2023,103(10):720-726

    • [27] TIAN Z,YAO N J,WANG F,et al.Thymosin β4 suppresses LPS-induced murine lung fibrosis by attenuating oxida-tive injury and alleviating inflammation[J].Inflamma-tion,2022,45(1):59-73

    • [28] TIAN J K,ZHANG R,ZHU N,et al.Association of serum thymosin β4 with malnutrition-inflammation-atherosclero-sis syndrome in peritoneal dialysis patients:a cross-sec-tional study[J].Ren Fail,2023,45(1):2202761

    • [29] REIS F M,DA SILVA M Z C,REIS N S D C,et al.Asso-ciation between phase angle and coronary artery calcium score in patients on peritoneal dialysis[J].Front Nutr,2022,9:912642

    • [30] DO J Y,KIM A Y,KANG S H.Association between phase angle and sarcopenia in patients undergoing peritoneal dialysis[J].Front Nutr,2021,8:742081

    • [31] WEN Y Q,ZHAN X J,WANG N S,et al.Monocyte/lym-phocyte ratio and cardiovascular disease mortality in peri-toneal dialysis patients[J].Mediators Inflamm,2020,2020:9852507

    • [32] ROSALES C.Neutrophils at the crossroads of innate and adaptive immunity[J].J Leukoc Biol,2020,108(1):377-396

    • [33] HUANG N Y,LI H Y,FAN L,et al.Serum phosphorus and albumin in patients undergoing peritoneal dialysis:in-teraction and association with mortality[J].Front Med,2021,8:760394

    • [34] YU Y,ZHONG Z,YANG W Z,et al.Neutrophil percent-age-to-albumin ratio and risk of mortality in patients on peritoneal dialysis[J].J Inflamm Res,2023,16:6271-6281

    • [35] 李洋,王亚芬,李相,等.TyG、CRP/Alb、25(OH)D 与维持性腹膜透析患者预后的关联性[J].中国医师杂志,2022,24(9):1335-1339

    • [36] 罗亚维,冯胜,沈华英,等.C反应蛋白/白蛋白比值是腹膜透析患者死亡的独立影响因素[J].中华肾脏病杂志,2022,38(6):528-535

    • [37] MA S,CAI Y M,WANG Z,et al.Derivation and valida-tion of a risk score predicting risk of early-onset peritoni-tis among patients initializing peritoneal dialysis:a cohort study[J].Int J Infect Dis,2020,99:301-306

    • [38] WANG Y X,SHEN Q Q,WANG J N,et al.The risk fac-tors and predictive model for cardiac valve calcification in patients on maintenance peritoneal dialysis:a single-cen-ter retrospective study[J].Ren Fail,2023,45(2):2271069

    • [39] 刘晨媛,吴娟洁,董金玲,等.腹膜透析患者感染的影响因素及其风险预测列线图模型[J].中华医院感染学杂志,2023,33(20):3185-3189

    • [40] 陈婷,龙海波,黄千殷,等.预测腹膜透析患者预后模型的构建与验证[J].中华肾脏病杂志,2020,36(9):680-687

    • [41] KANG S H,DO J Y.Phase angle as a risk factor for mor-tality in patients undergoing peritoneal dialysis[J].Nutri-ents,2023,15(23):4991

    • [42] HUANG D D,LI Y Y,QI X M,et al.A nomogram pre-dicts cardiovascular events in patients with peritoneal dialysis-associated peritonitis[J].Ren Fail,2022,44(1):1558-1567

    • [43] LI J N,ELLIS D G,PEPE A,et al.Back to the roots:re-constructing large and complex cranial defects using an image-based statistical shape model[J].J Med Syst,2024,48(1):55

    • [44] CARANFIL E,LAMI K,UEGAMI W,et al.Artificial in-telligence and lung pathology[J/OL].Adv Anat Pathol,2024(2024-05-23)[2024-05-27].DOI:10.1097/PAP.0000000000000448

    • [45] VAMATHEVAN J,CLARK D,CZODROWSKI P,et al.Applications of machine learning in drug discovery and development[J].Nat Rev Drug Discov,2019,18(6):463-477

    • [46] DASHTBAN A,MIZANI M A,PASEA L,et al.Identify-ing subtypes of chronic kidney disease with machine learning:development,internal validation and prognostic validation using linked electronic health records in 350 067 individuals[J].EBio Medicine,2023,89:104489

    • [47] DACK E,CHRISTE A,FONTANELLAZ M,et al.Artifi-cial intelligence and interstitial lung disease:diagnosis and prognosis[J].Invest Radiol,2023,58(8):602-609

    • [48] SMITH L A,OAKDEN-RAYNER L,BIRD A,et al.Ma-chine learning and deep learning predictive models for long-term prognosis in patients with chronic obstructive pulmonary disease:a systematic review and meta-analy-sis[J].Lancet Digit Health,2023,5(12):e872-e881

    • [49] XU L P,CAO F,WANG L,et al.Machine learning model and nomogram to predict the risk of heart failure hospital-ization in peritoneal dialysis patients[J].Ren Fail,2024,46(1):2324071

    • [50] YANG J,WAN J F,FENG L,et al.Machine learning algo-rithms for the prediction of adverse prognosis in patients undergoing peritoneal dialysis[J].BMC Med Inform De-cis Mak,2024,24(1):8

    • [51] ZANG Z Y,XU Q J,ZHOU X L,et al.Random forest can accurately predict the technique failure of peritoneal dialy-sis associated peritonitis patients[J].Front Med,2023,10:1335232

    • [52] 唐雯,高峻逸,马辛宇,等.循环神经网络模型在腹膜透析临床预后预测中的初步应用[J].北京大学学报(医学版),2019,51(3):602-608

    • [53] XU Z Y,XU X,ZHU X M,et al.Attention-based deep learning model for prediction of major adverse cardiovas-cular events in peritoneal dialysis patients[J].IEEE J Biomed Health Inform,2024,28(2):1101-1109

    • [54] MA L T,ZHANG C H,GAO J Y,et al.Mortality predic-tion with adaptive feature importance recalibration for peritoneal dialysis patients[J].Patterns,2023,4(12):100892

  • 关闭