摘要 |
FIELD: medicine. SUBSTANCE: method involves recording several hundred to several thousand R-R intervals of electrocardiogram (N). Statistics are calculated using standard formulas: R-R interval duration means <EMI ID=0.22 HE=9 WI=24 TI=CHI> , mode variance Dr(M<SP>2</SP>-(Mr) , asymmetry <EMI ID=0.23 HE=12 WI=30 TI=CHI>, mode kurtosis <EMI ID=0.24 HE=12 WI=27 TI=CHI> and smoothness <EMI ID=0.25 HE=15 WI=30 TI=CHI>, n is calculated as the number of R-R intervals satisfying the condition <EMI ID=0.26 HE=3 WI=42 TI=CHI>, where ψ is calculated from formula <EMI ID=0.27 HE=9 WI=48 TI=CHI>. Probability p is calculated by applying Markov chains method to describe duration change dynamics of n(i) counted in this way in the direction of their growth or drop. Rhythmogram code is built as sequence of calculated values Mr, Dr, Ar, Er, Sr, p and the nearest patient code YZ and the nearest healthy person code IW are to be determined relative to Euclid's distance between the testee code and the rhythmogram codes available in the sample having groups of healthy and ill persons having oncological diseases. Discrimination function is calculated using formula f(x)= [dist(X,YW)-dist(X,IS)] / [dist(X,YW)-dist(X,IS)] 1000. If f(X)>10, conclusion is drawn malignant neoplasms are not detected in the examined person. . If f(X)<-10, conclusion is drawn cancer or tumor transformation process is the case. Otherwise if -10 <EMI ID=0.28 HE=3 WI=3 TI=CHI> f(X) <EMI ID=0.29 HE=3 WI=3 TI=CHI> 10, the examined person is to be included into the tumor growth risk group. EFFECT: noninvasive early stage diagnosis method; determined risk groups. 1 dwg
|