Relationship between Glioblastoma Heterogeneity and Survival Time: An MR Imaging Texture Analysis

• DOI: 10.3174/ajnr.a5279
• Publication Date: 2017-06-29T20:30:33Z
• Authors: Y. Liu; X. Xu; L. Yin; X. Zhang; L. Li; H. Lu
• Keywords: Adult; Aged, 80 and over; Male; Support Vector Machine; Adolescent; Brain Neoplasms; Contrast Media; Middle Aged; Prognosis; Magnetic Resonance Imaging; Survival Analysis; 3. Good health; Diagnosis, Differential; Young Adult; 03 medical and health sciences; 0302 clinical medicine; Image Processing, Computer-Assisted; Humans; Female; Glioblastoma; Aged; Retrospective Studies

<h3>BACKGROUND AND PURPOSE:</h3> The heterogeneity of glioblastoma contributes to the poor and variant prognosis. aim this retrospective study was assess with MR imaging textures evaluate its impact on survival time. <h3>MATERIALS METHODS:</h3> A total 133 patients primary who underwent postcontrast T1-weighted (acquired before treatment) whose data were filed times selected from Cancer Genome Atlas. On basis overall survival, divided into 2 groups: long-term (≥12 months, <i>n</i> = 67) short-term (&lt;12 66) survival. To measure heterogeneity, we extracted 3 types textures, co-occurrence matrix, run-length histogram, reflecting local, regional, global spatial variations, respectively. Then support vector machine classification used determine how different texture perform in differentiating groups, both alone combination. Finally, a recursive feature-elimination method find an optimal feature subset best differentiation performance. <h3>RESULTS:</h3> When alone, matrix performed best, while all features combined obtained stratification. According selection ranking, 43 top-ranked as subset. Among them, top 10 included 7 features, which 6 regional emphasizing high gray-levels ranked 7. <h3>CONCLUSIONS:</h3> results suggest that local may play important role stratification glioblastoma.