G.I. Nazarenko (1,3), Yu.I. Zhuravlev (2), V.V. Ryazanov (2), M.V. Konstantinova (1), E.B. Kleymenova 1) 1 -Medical Center of Bank of Russia, Sevastopol'skiy pr., 66, Moscow, Russian Federation, 117593; 2 -A.A. Dorodnitsyn Computing Centre, Vavilova Str., 40, Moscow, Russian Federation, 119333; 3 -Institute for Systems Analysis, Pr. 60-letiya Oktyabrya, 9, Moscow, Russian Federation, 117312

Introduction. Ischemic stroke (IS) is a multifactorial disease, resulting from a complex interaction of genetic and environmental factors. Individual cause of the disease, multiple factors and symptoms require advanced mathematical methods of predicting IS risk, decision support for the diagnosis and treatment of a particular patient. The aim of the study. To evaluate the utility of pattern recognition methods for IS occurrence prediction on the base of risk factors family. Methods. The study included 257 patients aged 45 to 86 years, including 123 with IS history and 134 without cerebrovascular pathology or stroke history (control group). Total 18 traits were analyzed. Traditional risk factors included hypertension, hypercholesterolemia, diabetes, obesity, smoking, coronary heart disease, atrial fibrillation and a family history of heart disease. Besides, biochemical and genetic markers were also included: polymorphisms of MTHFR, ACE and ApoE genes. All patients were divided into 2 groups: control group was assigned to the first class, and patients with IS – to the second class. The data were analyzed with «RECOGNITION» system, which allows to construct on the basis of training set processing optimal algorithms for automatic classification (prediction) of a new patient’s state according to his individual factors. Results. First of all, the set of main prognostic factors and their informative values (weights) were determined with the help of «logical patterns» method. The most significant prognostic factors were: hypertension stage, diabetes mellitus, ischemic heart disease, atrial fibrillation, patient's age, smoking, as well as blood levels of cholesterol, triglycerides, high-density lipoprotein and homocysteine. There was a significant association between IS and combination of homozygous polymorphism of MTHFR gene (677TT) with either ApoE gene ε4 allele or ACE gene D-allele (p=0,04 and p=0,03, respectively). Logical patterns method yielded 88,3% prediction accuracy when using 16 factors. Further reduction of considered factors (9 and 11) resulted in a decrease of prediction accuracy up to 78,9%. Conclusion. Computer based algorithm accounting traditional IS risk factors, biochemical and genetic markers (polymorphisms of MTHFR, ACE and ApoE genes) proves to be relevant for the prediction of the risk of ischemic stroke.
ischemic stroke, risk factors, gene polymorphism, prognosis

Список литературы: 
  1. Skvorcova V.I. Snizhenie zabolevaemosti, smertnosti i invalidnosti ot insul`tov v Rossiyskoy Federacii. Zhurn. nevrologii i psihiatrii im. S.S. Korsakova. 2007. Spec. vyp. Insul`t. 25–7.
  2. Gusev E.I., Skvorcova V.I., Kilikovskiy V.V., Stahovskaya L.V. Problema insul`ta v Rossiyskoy Federacii. Kachestvo zhizni. Medicina. 2006; 2: 10–14.
  3. Bochkov V.B. Razrabotka modeley i algoritmov diagnostiki i prognozirovaniya ostryh narusheniy mozgovogo krovoobrashheniya: avtoref. … dis. kand. med. nauk. Kursk, 2000. 198 c.
  4. Misyuk N.S. Prognozirovanie i profilaktika ostryh narusheniy mozgovogo krovoobrashheniya. Minsk, 1987. 142 c.
  5. Dufouil C., Richard F., Fievet N. APOE genotype, cholesterol level, lipid-lowering treatment, and dementia: the three-city study. Neurology. 2005; 64: 1531–8.
  6. Raizada M.K., Paton J.F., Kasparov S. et al. Cardivascular genomics. Humana press, 2005. 362 p.
  7. Arnett D.K., Baird A.E., Barkley R.A. et al. Relevance of genetics and genomics for prevention and treatment of cardiovascular disease: a scientific statement from the American Heart Association Council on Epidemiology and Prevention, the Stroke Council, and the Functional Genomics and Translational Biology Interdisciplinary Working Group. Circulation. 2007; 115: 2878–901.
  8. Lusis A.J., Fogelman A.M., Fonarow G.C. Genetic basis of atherosclerosis, part II: clinical implications. Circulation. 2004; 110: 2066–71.
  9. Ryazanov V.V. Logicheskie zakonomernosti v zadachah raspoznovaniya (parametricheskiy podhod). Zhurn. vychislit. matematiki i matem. fiziki. 2007; 47 (10): 1793–1808.
  10. Zhuravlev Yu.I., Ryazanov V.V., Sen`ko O.V. «RASPOZNAVANIE». Matematicheskie metody. Programmnaya sistema. Prakticheskie primeneniya. M.: FAZIS, 2006. 176 s.
  11. Zhuravlev Yu.I. Izbrannye nauchnye trudy. M.: Magistr, 1998. 420 s.
  12. Nakata Y., Katsuya T. et al. Polymorphism of angiotensin converting enzyme, angiotensinogen and apolipoprotein E gene in Japanese population with cerebrovascular disease. Am. J. Hypertens. 1997; 10: 1391–5.
  13. Pfohl M., Fetter M., Koch M. Association between angiotensin converting enzyme genotypes. Extracranial artery stenosis and stroke. Atherosclerosis. 1998; 140 (1): 161–6.