Mapping distribution of brain metastases in renal cell carcinoma (RCC) and association with clinical features.

Abstract

727 Background: Brain metastases (BM) are frequently observed in metastatic RCC (mRCC) with a tendency to present with hemorrhagic BM. Distribution patterns of BM in RCC and associations with clinical outcomes are not well understood. Here we present a detailed distribution analysis in a surgically resected BM RCC cohort with a focus on clinically relevant associations. Methods: A retrospective analysis was performed on patients (pts) with mRCC who underwent craniotomy for BM at our institution. We reviewed gadolinium-enhanced MRI brain and CT head (if available) at the time of BM diagnosis. We assessed the presence of hemosiderin on MRI or hemorrhage on CT in the BM lesion and marked each at their corresponding anatomic location. Public brain atlases were used for distribution analysis. Overall survival (OS) from the date of BM diagnosis and CNS progression free survival (CNS-PFS) from the date of craniotomy were calculated. Results: A total of 67 BM were analyzed from 46 pts. 23 with synchronous BM, 9 with two and 6 with 3 concurrent BM. Most patients were male (n=35, 76%) with a median age at diagnosis of 63 years. Primary histology was clear cell in 42 pts (91%) and 10 pts (22%) had renal vein thrombosis (RVT) at primary diagnosis. IMDC scores at time of BM diagnosis were favorable in 20%, intermediate in 60% and poor in 20%. Hemosiderin (MRI) or hemorrhage (CT) was seen in 82% and 80% of BM, respectively. 55% of lesions were left sided, 45% frontal, 21% parietal, 18% occipital, 7% temporal and 6% cerebellar. Posterior circulation and posterior cerebral artery territories contained 42% and 34% of BM despite supplying only 30% and 16% of the brain volume, respectively. Patients with solitary BM had higher Karnofsky performance status (p=0.006) and lower disease burden (p=0.004) than those with multiple BM. Shorter CNS-PFS was seen in association with hemosiderin on MRI (log-rank p=0.04) and RVT in primary specimen (log-rank p=0.002). IMDC risk scores (log-rank p<0.001) predicted OS. On multivariable OS analysis, BM lateralized to the right (HR 2.6, p=0.06), hemosiderin in BM (HR 6.5, p=0.02) and IMDC risk groups (intermediate HR 2.9, poor risk HR 25.6, p<0.04) were independent prognostic factors for shorter OS. Conclusions: The above-presented lesion mapping method shows clinically relevant findings in RCC BM. Tumoral hemosiderin deposits appear a potential parameter to predict outcomes in RCC BM which deserve further study and validation.