Spatial distribution and prognostic significance of hemorrhagic imaging features in brain metastases from metastatic renal cell carcinoma.

purposeBrain metastases (BM) occur in approximately 5-15% of patients with metastatic renal cell carcinoma (mRCC) and are associated with substantial morbidity and inferior survival. Although RCC brain metastases frequently demonstrate hemorrhagic features, the spatial distribution of BM and the prognostic significance of lesion-level imaging characteristics remains incompletely defined.

methodsWe conducted a retrospective study of adult patients with mRCC who underwent craniotomy for BM at Johns Hopkins Hospital between 2003 and 2021. Clinical characteristics and treatment data were collected at the time of BM diagnosis. We mapped each metastasis to a standardized brain space and used publicly available brain atlases for anatomic and vascular distribution analyses. We reviewed brain MRI and CT imaging at the time of BM diagnosis and assessed the presence of hemosiderin (in MRI) or hemorrhage (in CT) in each BM lesion. We further analyzed the association between BM features with survival outcomes including overall survival (OS) from diagnosis of brain metastasis and central nervous system progression-free survival (CNS-PFS) from date of craniotomy by Kaplan-Meier methods and multivariable Cox regression. Multivariable models were considered exploratory because of the limited sample size.

resultsForty-six patients with 67 brain metastases were included. The median age was 59 years, and 76.09% were male. 21.74% had renal vein thrombosis at primary diagnosis. International metastatic renal cell carcinoma database consortium (IMDC) risk classification at BM diagnosis was favorable in 19.57%, intermediate in 60.87%, and poor in 19.57%. By lobar distribution, 44.78% were frontal, 20.90% parietal, 17.91% occipital, 8.96% temporal, and 5.97% cerebellar. Lesions were significantly overrepresented in the PCA territory (34.3% observed vs. 16% expected; adjusted P = 0.003) and in the posterior circulation overall (41.8% observed vs. 30% expected; P = 0.045). Hemosiderin deposition was observed in 82.25% of evaluable lesions and hemorrhage on CT in 81.82%. On univariable analysis, renal vein thrombosis (HR 5.00, 95% CI 1.76-14.20, p = 0.003) and hemosiderin deposition (HR 4.91, 95% CI 1.08-22.37, p = 0.040) were associated with shorter CNS-PFS. Exploratory multivariable analyses demonstrated similar associations, with IMDC poor-risk classification (HR 31.55, 95% CI 2.47-403.65, p = 0.008) and renal vein thrombosis (HR 16.19, 95% CI 2.57-101.90, p = 0.003) remaining independently associated with inferior CNS-PFS. For OS, in exploratory multivariable analysis, systemic therapy before brain metastasis (HR 3.80, 95% CI 1.56-9.29, p = 0.003), IMDC poor-risk classification (HR 4.97, 95% CI 1.17-21.12, p = 0.030) and BM hemosiderin (HR 4.36, 95% CI 1.40-13.60, p = 0.011) were associated with shorter OS while immunotherapy post-craniotomy was associated with longer OS (HR 0.21, 95% CI 0.08-0.58, p = 0.003).

conclusionsIn this surgically managed RCC BM cohort, brain metastases from mRCC demonstrate nonrandom spatial distribution, with overrepresentation in posterior circulation and posterior cerebral artery territories. Hemorrhagic imaging feature, specifically hemosiderin deposition, was highly prevalent and showed potential prognostic significance for survival outcomes. These findings support further investigation of spatial and imaging biomarkers in larger prospective studies.