Development of a method for differential diagnostic of a recurrent glial tumor from radiation necrosis by clinical decision support system for MRI images

INTRODUCTION: Identification of brain glioma recurrence and necrosis — is actual neurosurgical problem, due to an absence of particular radiological signs on MRI image. Especially in the early stages. Identification of brain glioma recurrence and necrosis on MRI images — is a problem of object`s classification on medical images. Depending on this, an issue of chemotherapy is decided. OBJECTIVE: Development and practical approbation of a method for differential diagnostic of a recurrent glial tumor from radiation necrosis on MRI images by using mathematical model of medical image`s processing.MATERIALS AND METHODS: Analysis of existing methods for differential diagnostic of a recurrent glial tumor from radiation necrosis and description of principal deficiencies’ description. Authors proposed an algorithm for differential diagnostic of a recurrent glial tumor from radiation necrosis. Development of algorithmic, structural and mathematical models for proposed solution: prototype of an app and an interface. Testing of proposed method on the practice: dynamic examination for 98 patients with brain glioma Grade III–IV — brain MRI with contrast enhancement 1, 3, 6, 9, 12 months after surgical intervention and a chemotherapy.RESULTS: As a result of practical approbation, authors’ noticed that this method has a high level of brain glioma recurrence and radial necrosis detection (97,4%).DISCUSSION: The variety of scientific papers describing ways to differentiate recurrence of glial brain tumors and radiation necrosis based on MRI diagnostics confirms the relevance of the topic of the article in the medical scientific community. We have developed a method of differential diagnosis using MRI of the brain with contrast enhancement, due to the fact that this is the most accessible method of diagnosis in the postoperative period and a course of chemoradiotherapy.CONCLUSION: A high-precision method for differential diagnosis of radial necrosis and brain glioma recurrence using a clinical decision support system allows the neurosurgeon to timely adjust treatment tactics, thereby improving the survival rates of patients with Grade III–IV glial tumors.