Software for brain tumor diagnosis on magnetic resonance imaging

BACKGROUND: The main reason for the development and implementation of artificial intelligence (AI) technologies in neuro-oncology is high prevalence brain tumors reaching up to 200 cases per 100,000 population. incidence a primary focus 5%10%; however, 60%70% those who die from malignant neoplasms have metastases brain. Magnetic resonance imaging (MRI) most common method non-invasive diagnosis monitoring disease progression. One challenges classification tumor types determination clinical parameters (size volume) conduct, diagnosis, treatment procedures, including surgery.
 AIM: To develope software module differential on MRI images.
 METHODS: based developed Siberian Brain Tumor Dataset (SBT), which contains information over 1000 neurosurgical patients with fully verified (histologically immunohistochemically) postoperative diagnoses. data research was presented by Federal Neurosurgical Center (Novosibirsk). uses two- three-dimensional computer vision models pre-processed sequence included following packages: pre-contrast T1-weighted image (WI), post-contrast T1-WI, T2-WI, T2-WI fluid-attenuated inversion-recovery technique. allow detect recognize accuracy 4 neoplasms, such as meningioma, neurinoma, glioblastoma, astrocytoma, segment distinguish components sizes: ET (tumor core absorbing Gd-containing contrast), TC core) = + Necr (necrosis) NenTu, WT (whole tumor) Ed (peritumoral edema).
 RESULTS: shows segmentation results SBT Dice metric 0.846, 0.867, 0.9174, Sens 0.881, Spec 1.000 areas. testing validation were done at international BraTS Challenge 2021 competition. test dataset yielded DiceET 0.86588, DiceTC 0.86932, DiceWT 0.921 values, placing top ten. According classification, demonstrate rates 92% patient analysis (up 89% slice analysis), very potential, perspective future this area.
 CONCLUSIONS: may be used training specialists diagnostics.