<<
>>

the Analysis of modern approaches to forecasting serdechnoyososudistyh complications

Offered alternatives to an estimation of extent of risk SSO include variety noninvasive and invazivnyh the methods, allowing to reveal patients with the maximum risk of death as a result of SSO. The modern medicine offers a number of methods of an early and dynamic estimation of risk SSO as at bed sick, and the subsequent out-patient observation.

There are various methods by definition the forecast of a life sick of ischemic sickness of heart (IBS), including transferred IT, for example, tests with physical activity: the tredmil-test, veloergometrija.

Ultrasonic research of heart or EhoKG - one of the basic methods of diagnostic in cardiology. Prognostichesky value ehokardiograficheskih data considerably surpasses many clinical parametres. However in recent research of the British cardiologists it has been shown, that any of quantitative ehokardiograficheskih parametres did not possess independent prognosticheskim value. Ehokardiografichesky research in a combination to the loading test are considered more
Perspectiv for definition of the forecast of patient IBS. Ehokardiogrammu display before test with loading and right after the test terminations. This method - more informative, but complexity of a technique consists in subjectivity at total interpretation of the test, and also in existing possibility of very sweeping disappearrance of signs of a passing ischemia after the loading termination.

Recently to studying of the remote forecast of patients IBS began to apply stress-test with dobutaminom. The technique dobutaminovogo the test allows, unlike a usual stress-echocardiography, to conduct continuous observation over a myocardium condition, to observe emersion of an ischemia and its disappearrance after the loading termination. However and in this variety of ultrasonic research at treatment of results it is difficult to refuse use of qualitative characteristics, that also conducts to subjectivity.

Radio isotope methods began to be used for risk-stratification of patients with IT from the end of 70th years of the last century. By means of radionuclides it is possible to investigate a condition sokratitelnoj functions of a myocardium and to gain visual representation about sizes and localisation of area of a passing or resistant to ischemia. The important advantage of radio isotope techniques is reshennost standardization problems. radionuklidnaja ventrikulografija - one of the most exact techniques of measurement of functions left zheludochka yields well reproduced results. Is thus possible to investigate both global, and local function. radionuklidnaja ventrikulografija gives reliable prognosticheskuju the information that has shown research SOLVD (1991) in which the forecast of a life of patients with warm insufficiency was studied. Deficiency of the given method is that the method is expensive, is not routine, not in all hospitals there is a possibility of conducting of similar research, and also specially trained personnel is necessary.

Besides noninvasive methods of diagnostic, exist also invazivnye. One of them is koronaroangiografija (KAG) - invazivnoe
The diagnostic research spent for an estimation of a condition of coronary arteries and a coronary blood-groove, detection of pathological changes of arteries, and also for the qualitative and quantitative characteristic of the knocked segments of arteries.

Last decade it is research became a routine diagnostic method in complex diagnostic study of patients with IBS, giving the chance to define the forecast of the patient and further medical tactics.

Last years in clinical practice along with rentgenokontrastnoj angiografiej began to apply widely intravascular ultrasonic research (VSUZI) coronary vessels. Became accessible not only studying of a clearance of an artery, but also structure of a vascular wall, composition AB. The exact quantitative estimation and detailed qualitative analysis of a condition of coronary arteries allows to size up extent of atherosclerotic defeat and to choose the optimal method of treatment. This method of visualisation in a real time gives the chance to carry out the thin structural analysis ateroskleroticheski the changed section of an artery, to define the haemodynamic significance of a stenosis, to reveal complicated and structurally astable plaques. The given methods, unconditionally, are very informative, allow to define the forecast and tactics of the further treatment, but these techniques rather expensive, not in all hospitals there is a possibility of conducting of similar research, and also specially trained personnel is necessary.

The disease forecast in cardiology is observed in various foreshortenings (the instrumental researches, invazivnye and noninvasive researches, biological labels etc.), however every decade, being based on existing diagnostic and medical possibilities, introduces the corrective amendments in used approaches, develops and deepens them. One of most complicated questions of a problem of forecasting by IT is forecasting of relapse and the remote outcome of disease which allows to size up adequately weight of a condition sick of IT and to optimise prospektivnuju correction of medical-rehabilitation provisions IT already in its early period.

For the forecast IT the doctor should work overtime great volume of heterogeneous data and correctly them interprets. Process of installation of the forecast by such method is very labour-consuming while the forecast remains subjective enough. Hence, it is necessary to develop a uniform method of an estimation of weight IT and forecasting of its remote outcomes with use of the mathematical methods defining weight of each sign influencing the forecast and reducing the sum of burdening factors to a uniform parametre. Necessity of studying of structure and effective classification of variables FR from here outflows. With that end in view now apply machine methods of an intellectual data handling (Data Mining) on which basis form expert medical systems.

At the heart of intellectual systems (IS) the concept of an artificial intellect (AI) lays. AI problems are closely connected with the organisation of knowledge of world around in the form of mathematical structures, for example, assemblage, counts, algorithms which reflect real communications and relations between any installations in the nature (in particular, in a subject domain). The Artificial intellect is a scientific direction with which use of methods creative problems are solved, or ways of the solution (a deductive leading-out), inductive generalisation, frames) as similitude of system of reflexion are used by the person of a reality.

In creative problems the algorithm of the solution is not known in advance also a problem a priori assumes presence of sampling in the conditions of indeterminate form. It is possible to gate out a number of the reasons of existence of this indeterminate form:

1) the purpose ambiguity - often aim is specified in the course of the solution;

2) indeterminate form of all ways of achievement of the purpose, i.e. the imposed restrictions do not prohibit use of any methods which obviously in any way are not referring to to the given problem;

3) the ambiguity of statements of a problem originates because of unreliable data or the knowledge changing in a time;

4) ambiguity in a statement of problem;

5) impossibility of the account of all factors and the big space of search.

The programm system solving creative problems, or "conceiving" as the person, name intellectual, and the generalised ability of system to distinguish, be trained and remember - intellectuality. Generally under IS fathom the organizational-technical system which is consisting of an intellectual complex of means of a support of decision-making interconnected and co-operating with users and networks of the COMPUTER and carrying out the solutions of set problems [76]. Internal structure IS is formed under the influence of an environment of the images representing abstraction of installations of the real world.

One of the first ways of forecasting of a repeated heart attack of a myocardium (PIM), entered into clinical practice, is presented in-process JU.M.ball [70]. Forecasting was realised by means of inoculated method Shejnmana-Zavjalova, and weight were defined by construction of the moving apart metrics. Worth of the developed method decreases because the forecast each time is given only the proximal three days and repeated repeated definitions of the gated out factors of disease are necessary. In-process also the question on outcome PIM that would allow the doctor more differentsirovanno to prove and choose tactics of treatment [85] is not observed. Accuracy of the forecast for possibility of origination PIM on system JU.M. Ball makes 65 % [29] though the author overestimates this value to 80 %.

Other method of forecast PIM is presented in N.M.Ustinskovoj and A.L.Syrkina's works [101, 102]. Them IS was based on the discriminant function gained on method Bajesa, and at check of a solving rule the sliding control was used. At patients PIM which has survived in 72 hours after origination PIM, it is necessary to consider as essential deficiencies of a method of the forecast of the proximal outcome (the patient from a hospital will be written out or will die) multiplicity of the signs defining disease, subjectivity of some signs, and also that reliability of the selected most informative signs is non-uniform.

E.N.Amosov [5] in the research notes doubtless practical and theoretical interest of forecasting of development of relapses as this question actually is not studied, and considers, that bolshee it is necessary to give attention to definition of factors of origination of relapses.

In-process [104] the linear discriminant function presenting risk of development PIM in proximal three poda after research has been gained. Advantage of a method is accessibility in revealing of group of patients of high risk PIM and possibility of an estimation of risk in dynamics depending on a condition of a myocardium left zheludochka. However the percent of statement of the incorrect forecast remains high enough (14 %).

Besides the instrumental methods of diagnostic study there are various scales of an estimation of risk of development SSO [6]. In model of an estimation of risk PURSUIT such criteria, as age, CHSS, increased sistolicheskoe arterial pressure (slip-ring motor), depression of segment ST, signs of warm insufficiency, presence kardialnyh enzimov in blood whey are considered. They are associated with increment of death rate and risk IT (including repeated) during the 30-day period.

Also the model of an estimation of risk GRACE is known, which thicket is preferred by the European cardiologists. By means of this model it is possible to size up risk intrahospital letalnosti, death rates and developments by IT, and also death and developments by IT during half a year to define the most suitable to the given patient with an acute coronary syndrome (Construction Department) a way of treatment and its intensity.

Scale GRACE (Global Registry of Acute Coronary Events) allows to size up risk letalnosti and developments by IT both at a hospital stage, and within the next half-year, and also to define an optimum way of treatment of the concrete patient. According to scale GRACE, gate out 3 classes of risk at patients from Construction Department with elevatsiej segment SU As scale GRACE is based on direct comparisons, it is recommended as the best at Construction Department for use at receipt of the patient and an extract in routine practice. At
The help of the given scale the risk of development of the proximal (in the course of hospital treatment) negative cardiovascular outcomes (death, by IT) under condition of sampling of conservative strategy of treatment is sized up.

Risk stratification in scale GRACE is based on:

• Clinical characteristics:

■ Age,

■ CHSS,

■ the Sistolichesky slip-ring motor,

■ Extent of warm insufficiency on classification Killip;

• Presence cardiac arrest at the moment of receipt of the patient;

• to the Estimation of changes on an electrocardiogram;

• Biochemical labels:

■ Kardiospetsifichesky enzymes,

■ Level kreatinina whey.

Estimated scale TIMI (Thrombolysis In Myocardial Infarction), offered E.M. Antman and soavt. In 2000, represents simple and popular enough tool for a risk estimation. For definition of extent of risk it is necessary to use 7 criteria. Presence of each of criteria is sized up in 1 point. Accordingly, the bolshee quantity of points the patient, the above for it gains risk of death and cardiovascular events.

The scale of risk TIMI and OIM without elevatsii segment ST is developed for an astable stenocardia in researches and tested on different patients for an estimation of efficiency of treatment of various groups of risk. It is less exact in a prediction of adverse events, but simplicity does its useful. For IT with lifting of segment ST scale TIMI represents the complicated integrated scale on which it is possible to size up easily a condition of the patient at receipt. This scale distributes patients on a basis anamnesticheskih data in addition to the fact of lifting of segment ST or BLNIG on an electrocardiogram, and also a time reperfuzii more than 4 ch. Among patients without reperfuzii a myocardium the risk scale underestimated probability of death and provided lower
prognosticheskuju worth. Despite it, the given scale - simple and reliable for the classifying of patients on different classes of risk.

Very often for forecasting PIM use artificial neural networks (INS) [54, 61, 75, 110]. INS, as well as live biological neural networks, are structure for machining of the information and are capable to be trained on the instances set by it and, thus, they to be adjusted on the solution of tasks in view. Besides, generalise the gained information, i.e. Give right answers at a presentation of "unfamiliar" data. Construction of the model including the most essential diagnostic signs which possess high informativnostju (sensitivity, specificity, accuracy) is as a result possible. Differently, machine IS at 1st stage processes a database, is trained to distinguish the set purpose (for example, the diagnosis), on 2nd makes of the solution on a fitting of symptoms and signs to this or that class-room and thus «makes the diagnosis».

The method of basic vectors refers to to family of linear qualifiers (installations dividing a set on groups by means of a line). He allows to classify signs by search of the hyperplane which are separating the variables magnitudes which belong to different class-rooms, with the maximum positive allowance in this space. Implementing this method, the mathematical program builds 2 hyperplanes parallel dividing, on boundary lines of class-rooms and solves a question, the new installation refers to to what class-room. The more distance between these hyperplanes, the less mean error of the qualifier.

The observed methods have a number of essential deficiencies:

- Definition of the individual forecast at bolshej remains to a part unsatisfactory.

- Practically there are no uniform systems of forecasting of origination and a disease outcome.

- Forecast formulation is labour-consuming work for the doctor.

1.4

<< | >>
Scientific source Kiselyov Alexey Viktorovich. HETEROGENEOUS QUALIFIERS With VIRTUAL STREAMS FOR INTELLECTUAL SYSTEMS of FORECASTING of the REPEATED HEART ATTACK of the MYOCARDIUM. The dissertation on competition of a scientific degree of a Cand.Tech.Sci. Kursk - 2019. 2019

Other medical related information the Analysis of modern approaches to forecasting serdechnoyososudistyh complications:

  1. Introduction
  2. Modern systems of computer non-invasive diagnostics of a functional condition of cardiovascular system and risk serdechnoyososudistyh accidents