Differentiation between Solitary Cerebral Metastasis and Astrocytoma on the Basis of Subventricular Zone Involvement on Magnetic Resonance Imaging

PURPOSE: To determine the relationship between the subventricular zone (SVZ) and astrocytoma based on magnetic resonance imaging (MRI) and whether SVZ involvement can be used to distinguish solitary cerebral metastases (SCMs) from astrocytomas. METHODS: This retrospective study involved 154 patients with solitary low-grade astrocytoma (LGA), high-grade astrocytoma (HGA), and SCM, who underwent T1-weighted imaging (T1WI), Gd-DTPA–enhanced T1WI, and T2-weighted imaging (T2WI) or fluid-attenuated inversion recovery (FLAIR) T2WI. The spatial relationship between the tumor and SVZ was classified as “involvement” or “segregation” on contrast-enhanced T1WI for enhanced tumors and T2WI/FLAIR T2WI for non-enhanced tumors. Patient-based SVZ-contact rates were compared between the LGA, HGA, and SCM groups. The frequencies of involvement of various lateral ventricle regions by astrocytoma were compared. The correlation between SVZ involvement and tumor necrosis was analyzed. RESULTS: Patient-based SVZ-contact rates in SCM, LGA, and HGA were 24.1%, 68.8%, and 85.4%, respectively. Univariate analysis showed that the SVZ-contact rate was significantly different between SCM and astrocytoma (24.1% vs. 75.2% P < 0.001), also between LGA and HGA (68.1% vs. 85.4% P=0.037). After the tumor volume was adjusted as a covariate, SVZ-contact rates still differed between SCMs and astrocytomas (Odds ratio [OR]: 4.58, 95% Confidence interval [CI]: 1.65 to 12.8, P=0.004). Tumor volume differed between LGA and HGA (P< 0.001), and influenced the association between SVZ involvement and astrocytoma grade (P = 0.05). Among the lateral ventricle regions, the frontal horn was the most frequently involved by astrocytomas. SVZ-contact rates were higher in necrosis group compared with non-necrosis groups (83.9% vs. 50.0%, P < 0.001) among astrocytoma patients. Necrosis positively correlated with SVZ involvement in astrocytomas (r(s) = 0.342, P < 0.001), but did not correlate with SVZ involvement in SCMs (P = 0.193). CONCLUSIONS: Compared to SCMs, solitary cerebral astrocytomas exhibited spatial proximity to the SVZ, which might distinguish the supratentorial astrocytomas from SCMs.