[Elaboration and use of the methods of computer-assisted diagnosis of cerebrovascular pathology].

The purpose of the work was to analyze the experience with the trial of diagnostic algorithms to specify possibilities of differential diagnosis and forecasting the course or outcome of cerebrovascular diseases. Detailed study of 922 clinico-anatomic cases (547 with hemorrhagic brain stroke, 157 with ischemic, 218 with pseudostroke) enabled revealing differential-diagnostic criteria for the pathological patterns under study bearing in mind disease pathomorphism. Analysis was made of the tabular diagnosis of cerebrovascular pathology and of the experience gained with the trail of diverse diagnostic algorithms for staged diagnosis of brain stroke and pseudostroke, different varieties of cerebral hemorrhages (10 varieties). The results of the use of four algorithms for forecasting hemorrhagic brain stroke were studied and compared. The accuracy of the diagnosis and forecasting amounts from 55 to 82.3% depending on the problem to be solved and the algorithm used. The data show the necessity of the careful providing of the algorithm choice as dependent on the problem raised.