<p>Uvod: Stratifikacija rizika je postala integralna komponenta savremenog pristupa tretmanu u kliničkoj praksi. Danas se u dijagnostici i lečenju akutnog infarkta miokarda (AIM) koriste različiti skorovi rizika kao prognostički instrumenti za kratkoročan i dugoročan ishod bolesti. Nužni proceduralni procesi, u toku primarne perkutane koronarne intervencije (pPKI), kao i saznanja o distribuciji i vrstama lezija koronarnih arterija su od velikog značaja, te se preporučuje finalna evaluacija rizika neposredno nakon izvršene pPKI. Metode istraživanja podataka omogućavaju pronalaženje skrivenih obrazaca u podacima, otkrivanje njihovih uzročno-posledičnih veza I odnosa, te razvoj savremenih prediktivnih modela. Cilj: Kreiranje i testiranje jednostavnog, praktičnog i u svakodnevnoj praksi upotrebljivog prediktivnog modela za procenu intra-hospitalnog ishoda lečenja pacijenata obolelih od AIM sa ST-elevacijom (STEMI) lečenih pPKI. Metode: Istraživanje je unicentrična, retrospektivna, ali I prospektivna studija. U retrospektivnu studiju je uključeno 1495 pacijenta sa STEMI koji su lečeni na Klinici za kardiologiju Instituta za kardiovaskularne bolesti Vojvodine (IKVBV) kod kojih je u cilju rekanalizacije infarktne arterije izvršena pPKI, u periodu od decembra 2008. godine do decembra 2011. godine. Svaki pacijent je inicijalno predstavljen sa 629 obeležja sadržanih u postojećem IKVBV informacionom sistemu, koja čine demografske karakteristike, podaci iz anamneze i kliničkog nalaza, parametri biohemijskih analiza krvi priprijemu, parametri ehokardiografskog pregleda, angiografski i proceduralni detalji i šifre prijemnih dijagnoza. U svrhu istraživanja podataka korišćeno je programsko rešenje otvorenog koda Weka. Tokom evaluacije različitih algoritama izabran je algoritam koji daje najbolje rezultate po tačnosti predikcije i ROC parametru. U sklopu retrospektivnog dela izvršena je validacija prediktivnog modela desetostrukom unakrsnom validacijom na celom skupu podataka. Prospektivnom studijom je na uzorku od 400 pacijenata sa STEMI lečenih pPKI u toku 2015. godine izvršena dodatna validacija razvijenog prediktivnog modela. Za iste pacijente je izračunavat i GRACE skor rizika, te je upoređena njegova, i prediktivna moć razvijenog modela. Rezultati: Alternativno stablo odluke (ADTree) izdvojen je kao algoritam sa najboljim performansama u odnosu na ostale evaluirane algoritme. Cost sensitive klasifikacija je korišćena kao dodatna metodologija da bi se pojačala tačnost. ADTree stablo odluke izdvojilo je osam ključnih parametara koji najviše utiču na ishod intra-hospitalnog lečenja: sistolni krvni pritisak pri prijemu, ejekciona frakcija leve komore, udarni volumen leve komore, troponin, kreatinin fosfokinaza, ukupni bilirubin, T talas i<br />rezultat intervencije. Performanse razvijenog modela su: tačnost predikcije je 93.17%, ROC 0.94. Razvijeni model je na prospektivnoj validaciji zadržao performanse: tačnost predikcije 90.75%, ROC 0.93. Široko korišćeni GRACE skor je na prospektivnom skupu postigao ROC=0.86, što pokazuje da je razvijeni prediktivni model superiorniji u odnosu na njega. Zaključak: Razvijeni prediktivni model je jednostavan i pouzdan. Njegova implementacija u svakodnevnu kliničku praksu, omogućila bi kliničarima da izdvoje visokorizične pacijente, nakon reperfuzionog tretmana, a potom kod njih intenziviraju tretman i kliničko praćenje, a sa ciljem smanjenja incidence intra-hospitalnih komplikacija i povećanja njihovog preživljavanja.</p> / <p>Introduction: Risk stratification has become an integral component of modern treatment in clinical practice. Today, the diagnosis and treatment of acute myocardial infarction (AMI) use different risk scores as a prognostic instruments for short-term and long-term outcome of the disease. The necessary procedural processes during primary percutaneous coronary intervention (pPCI) as well as knowledge about the distribution and types of lesions in coronary arteries are of great importance, and a final risk evaluation is recommended directly after the pPCI. Methods of data mining allow finding hidden patterns in data, disclosure of their causal connections and relationships, and the development of modern predictive models. Aim: To create and test a simple, practical and usable predictive model in daily practice for the assessment of intrahospital treatment outcome of patients with AMI with STsegment elevation (STEMI) treated with pPCI. Methods: Presented research is unicentric, retrospective but also prospective study. Retrospective study included 1495 patients with STEMI who were admitted to the Clinics of cardiology of the Institute of Cardiovascular Diseases Vojvodina (IKVBV). For the purpose of recanalization of the infarct artery, pPCI has been performed to these patients during the period from December 2008 to December 2011. Each patient was initially described with 629 attributes from the existing information system of IKVBV. Those attributes consist of demographic characteristics, data from history and clinical findings, biochemical parameters of blood tests on admission, the echocardiographic parameters, angiographic and procedural details and admission diagnosis codes. For model development, an open source software solution Weka was used. During the evaluation of different algorithms, algorithm that gives the best results in terms of accuracy and ROC parameter was chosen. As part of the retrospective study, in order to assess the models performance, ten-fold cross-validation on the entire data set was used. A prospective study, on a sample of 400 patients with STEMI, treated with pPCI in 2015, performed additional validation of the developed predictive model. GRACE risk score was calculated for the prospective study patients and comparison with the developed model has been performed. Results: Alternative decision tree (ADTree) was isolated as an algorithm with the best performance in relation to other algorithms evaluated. Cost sensitive classification was used as an additional methodology to enhance accuracy. ADTree selected eight key parameters that most influence the outcome of intra-hospital treatment: systolic blood pressure on admission, left ventricular ejection fraction, stroke volume of the left ventricle, troponin, creatine phosphokinase, total bilirubin, T wave and the result of the intervention. The performance of the developed model are: the accuracy of the prediction is 93.17%, ROC 0.94. The developed model kept its performance in prospective validation: accuracy of prediction 90.75%, ROC 0.93. Widely used GRACE score achieved ROC = 0.86 in the prospective study patients, indicating that developed predictive model is superior to him. Conclusion: Developed predictive model is simple and reliable. Its implementation in everyday clinical practice, would allow clinicians to distinguish high-risk patients after reperfusion treatment, and then for them to intensify treatment and clinical follow-up, with an aim of reducing the incidence of intra-hospital complications and increase their survival.</p>
Identifer | oai:union.ndltd.org:uns.ac.rs/oai:CRISUNS:(BISIS)101088 |
Date | 28 September 2016 |
Creators | Sladojević Miroslava |
Contributors | Ćulibrk Dubravko, Jung Robert, Dejanović Jadranka, Lončar-Turukalo Tatjana, Čemerlić Ađić Nada, Petrović Milovan, Pavlović Katica |
Publisher | Univerzitet u Novom Sadu, Medicinski fakultet u Novom Sadu, University of Novi Sad, Faculty of Medicine at Novi Sad |
Source Sets | University of Novi Sad |
Language | Serbian |
Detected Language | English |
Type | PhD thesis |
Page generated in 0.2863 seconds