Deciphering the Genomic Landscape of Renal Cell Carcinoma Brain Metastases

Brain metastases from renal cell carcinoma (RCC) remain a major cause of morbidity and mortality, yet the genomic features associated with metastatic dissemination remain poorly understood. Whole-exome sequencing was performed on 72 RCC brain metastasis samples with matched normal. To identify candidate genomic alterations associated with brain metastasis, the genomic alterations detected in the brain metastases were compared against alterations in extracranial metastases from the MSK-ECM cohort (n=137) and primary RCC tumors from TCGA (n=432) by case-control analyses. Candidate alterations were also identified through matched-pair analyses comparing brain metastases with matched primary tumors or extracranial metastases from the same patient (n=25). A random survival forest model incorporating the candidate CNA events was developed to predict overall survival. The candidate CNAs were further evaluated using functional experimental data from MetMap and DepMap. Survival analyses were conducted to assess the prognostic relevance of these alterations. We identified recurrent CNAs enriched in RCC brain metastases, including 4q loss, 7p gain, 7q gain, 8p loss, 8q gain, 9p21.3 deletion, 12q15 amplification, and 14q loss. These alterations were associated with significantly poorer patient survival among RCC patients. A random survival forest model based on these CNA events stratified TCGA-KIRC patients into prognostically distinct risk groups (C-index = 0.64). Among the recurrent CNAs, 8p loss, 8q gain, 9p21.3 deletion were associated with increased incidence of brain metastases across multiple primary cancer types in xenograft mouse models. These alterations were also strongly associated with metastatic progression and poor prognosis across RCC, lung adenocarcinoma, breast cancer, and melanoma. These findings indicate a shared genomic basis for brain tropism and highlight the potential utility of copy-number alterations as biomarkers for risk stratification and clinical decision-making.

renal cell carcinoma, brain metastases, whole-exome sequencing, copy number alterations, cancer genomics

Brain metastases from renal cell carcinoma (RCC) cause substantial morbidity and remain a therapeutic challenge1,2. Historically reported in 2–15% of patients1, recent studies indicate that brain metastases now occur in 20–28% of RCC cases3,4, underscoring their rising clinical relevance. Prognoses for brain metastasis patients remain poor. In the JAVELIN Renal 101 trial, patients with asymptomatic brain metastases had markedly shorter progression-free survival (PFS) at 2.8–4.9 months, compared with 8.4–13.8 months in the overall trial population5. Limited treatment efficacy against RCC brain metastases stems from their biological complexity and tumour heterogeneity. Although targeted therapies and immunotherapies have transformed primary RCC management, their efficacies against brain metastases remain modest1,6,7, leaving radiation and surgery as the mainstays of care. This underscores the need to understand the genomic landscape of RCC brain metastasis, which can differ substantially from primary tumors and extracranial metastases8,9. The TRACERx study identified 9p and 14q loss as key drivers of RCC metastasis and poor prognosis9,10; however, it included only one patient with brain metastasis, leaving the genomic basis of CNS spread largely unexplored. To address this gap, we performed whole-exome sequencing (WES) of RCC brain metastases, primary tumors, and extracranial metastases to identify copy number alterations (CNAs) associated with brain metastasis, which may guide future development of therapeutic strategies.

Detailed methodology is provided in the Supplementary Methods. Study cohort Tissue specimens were collected under individual Institutional Review Board (IRB) approved protocols at the University of Texas – MD Anderson Cancer Center, Emory University and Hacettepe University, and all processing procedures performed on patient samples were in accordance with the ethical standards of the Helsinki Declaration and its later amendments. Consent was obtained, and annotation of specimens was performed via standard operating procedures of the respective Biobank framework and personnel, with the consent obtained through either in-person signatures or IRB-approved remote consent procedures. The study cohort is explained in Figure 1A. Histologic subtypes of samples are shown in Supplementary Table 1. Sample preparation and sequencing DNA was extracted from frozen and FFPE tumor tissue using commercial kits following institutional biobank protocols, requiring tumor purity ≥40%. Matched normal tissue was confirmed by genotyping. Whole-exome libraries were prepared with Twist library preparation and exome capture kits, followed by Illumina NovaSeq 6000 paired-end sequencing. Library quality, concentration, and target enrichment were assessed using fluorometric assays, TapeStation, and qPCR. Full sequencing procedures are provided in the Supplementary Methods. Copy-number and single-nucleotide variant analysis Copy-number analysis of WES data followed modified versions of established pipelines11. Read coverage across exome targets (±250 bp) was obtained using GATK4-CNV and normalized to diploid controls to generate total copy-ratio profiles, which were segmented using circular binary segmentation12. Allele-specific copy numbers were inferred by combining total copy ratios with B-allele frequencies from germline heterozygous SNPs11, and ASCAT was used to generate allele-specific CNA profiles; rescaled copy-number values were computed as previously described11. Segmented copy-number and gene-level profiles for TCGA and MSK cohorts were retrieved from PanCanAtlas13 and cBioPortal14. Arm-level cytogenetic alterations were defined using segment-length–weighted mean copy-number states15. For focal loci (2q35, 9p21.3, 12q15), gene identities were extracted using Organism.dplyr, and locus-level copy number was assigned by averaging gene-level values from cBioPortal14. Somatic SNVs were analyzed using dNdScv16 to identify genes mutated at frequencies exceeding background. Co-occurrence analysis To characterize genomic co-alteration patterns, we constructed a binary event matrix for recurrent arm-level and focal CNAs in RCC brain metastases. Pairwise co-occurrence was evaluated using a probabilistic framework, and significant associations were visualized as a network. These patterns informed downstream classification of RCC-BM samples into CNA-defined subgroups for oncoprint visualization. Comparison of mutation frequencies in the PI3K/AKT/mTOR pathway We compared the mutation frequencies of MTOR, PIK3CA, and PTEN between our RCC brain metastasis cohort and primary RCC tumors from the TCGA-KIRC cohort. Mutation data for TCGA-KIRC were retrieved from cBioPortal. CNA-based survival modeling using Random Survival Forests A random survival forest (RSF) model incorporating these CNAs was trained and validated following established protocols for genomic survival modeling, with performance summarized using the concordance index. Detailed statistical specifications, model parameters, and multiple-testing procedures are provided in the Supplementary Methods. MetMap To evaluate whether candidate CNAs associated with RCC brain metastasis generalize across other cancers, we integrated organ-specific metastatic potential scores from the MetMap500 project17 with corresponding copy-number profiles from DepMap. LUAD, BRCA, and melanoma cell lines were stratified by the presence of 8p loss, 8q gain, or 9p21.3 deletion, and brain, liver, lung, and bone metastatic-potential scores were compared using Wilcoxon rank-sum tests. Statistical analysis All analyses were performed in R (v3.2.3). Two-sided tests were used throughout, and P < 0.05 was considered statistically significant unless otherwise specified. CNA and mutation enrichment across cohorts was assessed using weighted logistic regression with matching weights, or Wilcoxon rank-sum tests for continuous metastatic-potential scores. For survival analysis, OS and PFI metrics were obtained from TCGA-KIRC and MSK cohorts. Patients lacking all candidate events (4q loss, 7p gain, 7q gain, 8p loss, 8q gain, 9p21.3 deletion, 12q15 gain, 14q loss) served as the reference group. Associations between CNAs and survival were assessed using Kaplan–Meier estimates, log-rank tests, and Cox-based hazard ratios. Details are provided in the Supplementary Methods.

Case-control analyses identified copy-number alterations enriched in brain metastases in RCC To identify genomic alterations that drive RCC-brain metastasis, we performed WES in 72 brain metastasis samples (Fig. 1A), comprehensively characterizing focal and broad copy number alterations, as well as single nucleotide variations (Supplementary Fig. 1A–B). We conducted this study in three phases. First, we constructed a discovery cohort consisting three groups: RCC brain metastases (RCC-BM; n=72), extracranial metastases from MSKCC (MKS-RCC-EM; n=137), and primary RCC tumors from TCGA (TCGA-RCC-P; n=432)18,19. We performed case-control analyses comparing brain metastases against extracranial metastases and primary tumors. We also performed matched-pair analyses on 25 brain metastases with matched primary or extracranial samples. Second, candidate genomic alterations associated with brain metastasis identified in the discovery cohort were evaluated using MetMap, which interrogated the propensity of genomically characterized cells lines to metastasize to different organs in xenograft mouse experiments20,21. Third, the candidate genomic alterations were validated in a non-overlapping pan-cancer cohort of brain metastasis (n=475) extracted from the MSKCC dataset. We first identified recurrent copy-number alterations (CNAs) that occur at significantly higher frequencies vs. background, as previously described15 (Fig. 1B–C). This analysis confirmed the cytogenetic hallmarks of RCC: 3p loss and 5q gain, as recurrent CNAs in all three groups (TCGA-RCC-P, MSK-RCC-EM, and RCC-BM) in the discovery cohort. Owing to positive selection during metastatic dissemination to the brain, we expect metastatic drivers to occur across brain metastases at significantly elevated frequencies compared against extracranial metastases and primary tumours (Supplementary Fig. 1C–D). Therefore, we compared the frequencies of recurrent CNAs across these three cohorts and identified CNAs that were enriched in brain metastases, including losses of 4p, 4q, 8p, 9p, 9q, 14q and 18q, as well as gains of 5p, 7p, 7q, 8q, and 20q (Fig. 1D). Notably, 8p loss was the most frequent BM CNA event, observed in ~50% of brain metastasis samples vs. 15% in primary and ECM samples (Fig. 1C–D). Furthermore, we identified distinct focal regions with significantly higher positive selection scores in brain metastasis samples vs. primary and ECM samples, using a previously validated Gaussian process model22. Homozygous deletions of 9p21.3 and 2q35, as well as amplification of 12q15 were under positive selection in RCC brain metastases compared extracranial metastases and primary tumors (Fig. 1E). We also looked for recurrent single nucleotide variants (SNVs) in RCC16. We identified VHL, PBRM1, TP53, MTOR, SETD2 genes as significantly frequently mutated in all three groups. However, the mutation frequencies in the genes were not significantly different among RCC primary tumors, extracranial metastases, or brain metastases (Supplementary Fig. 1D). Matched-pair analyses of RCC brain metastases supported candidate CNAs To further evaluate the candidate genomic alterations associated with brain metastatis, we compared brain metastases samples against the matched primary or extracranial tumor samples (Fig 1A). Shared and compartment-specific genomic alterations were identified in these matched pairs (Fig 2A). Mutations in well-established RCC genes, including VHL, PBRM1, BAP1, MTOR, TP53, and KDM5C, were shared by brain metastasis and matched ECM/primary samples (Fig. 2A, 2D). Matched-pair analyses on the CNA events revealed that gains of 1q, 7q, 7p, and 8q, as well as losses of 4q, 10q, and 14q were significantly enriched in brain metastases compared against matched primaries or extracranial metastases (Fig. 2B–C). Indeed, gains of chromosome 7 and 8q, as well as losses in 4q and 14q were significantly enriched under both case-control and matched-pair analyses; these results highlight these events as potential CNAs associated with brain metastasis in RCC (Supplementary Table 2). Co-occurrence patterns of genomic alterations in RCC brain metastases We next characterized the co-occurrence pattern of the candidate CNAs associated with brain metastasis. Several events including 8p loss, 14q loss, 7p gain, and 8q gain demonstrated a strong co-occurrence with each other (Fig. 3A). Significant pairwise co-occurrences were observed between 8p loss and 14q loss, 8p loss and 8q gain, as well as 7p gain and 7q gain. From these associations, we constructed a co-occurrence network plot (Fig. 3B), which identifies 8p loss as a key hub, suggesting its pivotal role in the development of brain metastasis. To determine whether the co-occurrence patterns observed in brain metastases are already detectable in primary tumors, we examined pairwise CNA co-occurrence in TCGA-KIRC primary RCC (Supplementary Fig. 2). The analysis revealed several significant co-occurring events, consistent with findings from our RCC BM cohort. Notably, the loss of 9p21.3 frequently co-occurred with the loss of 14q (P < 0.0001), while the gain of 8q frequently co-occurred with loss of 8p. These co-occurrence patterns of events in primary tumors suggest that these genomic constellations are established early and may predispose to subsequent brain metastatic dissemination. Arm-level CNAs may cooperate with pathway-activating point mutations to drive brain metastasis. The PI3K/AKT/mTOR pathway has been recurrently implicated in brain metastasis across breast cancer, lung cancer, and melanoma, and several candidate CNA regions (e.g., PTEN on 10q) overlap with this pathway8,23–27. We therefore looked at whether PI3K/AKT/mTOR pathway mutations were enriched in RCC brain metastases compared to primary tumors. Analysis of our RCC brain metastases revealed higher combined mutation frequencies of MTOR, PIK3CA, and PTEN compared to the primary RCC tumors (P = 0.007; Fig. 3C–D). Recurrent CNAs associated with brain metastasis are linked with poor patient survival To assess the prognostic significance of the candidate CNAs enriched in brain metastasis in RCC (2q25 loss, 4q loss, 7p gain, 7q gain, 8p loss, 8q gain, 9p21.3 loss, 12q15 gain, and 14q loss), we conducted survival analyses using the TCGA-KIRC cohort. We first identified 147 patients whose tumors lacked all of the candidate CNAs, and we used this group as the reference in the survival analyses (Fig. 4A). Except for 2q35 loss, the candidate CNAs were associated with poor prognosis (Fig. 4B, Supplementary Fig. 3A). Collectively, patients who presented with one or more of the prognostically significant alterations had poorer overall survival (P=0.004) compared to the reference (Fig. 4D). We next looked at the impact of the candidate CNAs on progression-free interval (PFI). Similarly, all CNAsexcept for 2q35 loss were associated with shorter PFI (Fig. 4C). Patients with RCC tumors harboring one or more of the alterations exhibited significantly shorter PFIs (P < 0.0001; Fig. 4E). Given that most samples without candidate alterations were stage I tumors, we performed further survival analyses on stage I patients. This analysis showed a trend toward poorer overall survival among Stage I RCC patients with the candidate CNAs, although the difference did not reach statistical significance (P=0.06), likely due to the limited number of death events observed among Stage 1 patients (Supplementary Fig. 3A–B). Additionally, we developed a random survival forest model based on the CNV events (4q, 7p, 7q, 8p, 8q, 9p21.3, 12q15, and 14q), and we stratified TCGA-KIRC patients into prognostically distinct groups. Survival analysis revealed significantly poorer overall survival in the high-risk group than in the low-risk group (HR = 2.53, 95% CI: 1.85–3.47, p < 0.001; Fig. 4F). To assess the potential of model overfitting, we randomly split the patients into training and validation sets. Indeed, consistent survival differences were observed in the training (HR = 2.60, 95% CI: 1.78–3.81, p < 0.001) and validation sets (HR = 2.34, 95% CI: 1.34–4.11, p = 0.001) (Fig. 4G). The model yielded a mean concordance index (C-index) of 0.64 across the training and validation sets. Functional evidence for the role of 8p loss, 8q gain, and 9p21.3 deletion in promoting brain metastasis No RCC-specific experimental metastasis mapping dataset currently exists. We therefore leveraged MetMap500 data from lung adenocarcinoma (LUAD), breast cancer (BRCA), and melanoma to test whether the same CNAs identified in RCC brain metastases are associated with brain-tropic metastatic potential across cancer types (Fig. 5A). In the MetMap study, investigators injected a pool of 500 barcoded cancer cell lines intracardially into mice to allow spontaneous metastasis to distant organs, and harvested organs were analyzed by RNA sequencing to quantify barcode abundance. By integrating MetMap with copy-number profiles of the cell lines from DepMap, we identified genomic alterations that are associated with the potential of the cell lines to metastasize to the brain and other organs17. We found that LUAD cell lines with higher brain metastatic potential predominantly exhibit gains in 8q and deletions in 9p21.3 (Fig. 5B–C). In contrast, compared against wildtype, these alterations were not associated increased metastatic potentials to the liver, bone, or lung (Supplementary Fig. 4A, D). Similarly, BRCA cell lines with 8p loss and 9p21.3 loss demonstrated significantly higher brain metastatic potential compared to WT cells. (Fig. 5D–E). There was also no significant differences in the lung, liver, bone metastasis potentials of 8p and 9p21.3 loss cell lines compared to wildtype (Supplementary Fig. 4B, E). In melanoma cell lines, we observed a similar pattern (Fig. 5F), where both loss of 8p and the deletion of 9p21.3 were significantly associated with increased metastatic potential to the brain, but not to the liver, bone, or lung (Fig. 5G, Supplementary Fig. 4E–F). Overall, our results suggest two sets of genomic alterations that are associated with brain metastases: 8q gain with 9p21.3 deletion, and 8p loss with 9p21.3 deletion. Interestingly, the curves for metastatic potential to the liver, lung, and bone generally follow a similar pattern across the three cancer types, while the metastatic potential to the brain seems to follow a different trend. Taken together, our results suggest that 8q gain, 8q loss, and 9p21.3 deletion can enhance metastatic potential to the brain across different cancer types. Pan-cancer analysis confirmed the roles of 8p loss, 8q gain, and 9p21.3 deletion in brain metastases across different cancers To externally validate the candidate CNAs (8p loss, 8q gain, 9p21 loss) associated with brain metastasis, we conducted a comprehensive analysis across cancer types from MSKCC cohorts19 (Supplementary Fig. 5A). Samples included in our discovery cohort for case-control analyses were removed from this external validation cohort. Significantly enriched cytogenetic alterations were identified using weighted logistic regression with matching weights, which were determined by coarsened exact matching to ensure exact balance of confounding covariates, including tumour type, as previously described11. Brain metastases (n=475) exhibited significantly higher frequencies of 8p loss, 9p21.3 deletion, and 8q gain, compared to primary tumors (n=15,030) and extracranial metastases (n=10,089) (Fig. 6A). These findings suggest a critical role of these alterations in promoting brain metastases. A distinct cluster of brain metastases is characterized by concurrent 9p21.3 deletion and chr8 alterations (Fig. 6B). We analyzed the survival of patients with brain metastases in BRCA, LUAD, and melanoma in MSK MetTropism dataset. Patients with brain metastasis had worse overall survival consistently across cancer types (Fig. 6C, Supplementary fig. 5B), highlighting the prognostic significance of brain metastases. We next examined the frequencies of the candidate CNAs in brain metastasis samples (BM), primary tumors from patients who would develop brain metastasis (Future BM), and primary tumors from patients who never developed brain metastasis (No BM) in breast cancer, lung cancer, and melanoma during the follow-up period (30 months)19. Indeed, the frequencies of 8p loss, 8q gain, and 9p21 deletion were consistently highest in the BM group (Fig. 6D). Furthermore, the Future BM group showed intermediate frequencies these alterations, suggesting that these alterations may serve as early indicators of potential brain metastasis. We observed similar findings in colorectal cancer (MSK-COAD), small-cell lung cancer (MSK-SCLC), and prostate cancer (MSK-PRAD) (Supplementary Fig. 5C). We next performed survival analyses to assess the prognostic significance of 8p loss, 8q gain, and 9p21.3 deletion across cancer types (Supplementary Fig. 6A–B). In breast cancer, 8p loss was significantly associated with worse OS, with (HR: 1.55, 95% CI: 1.23–1.95, P < 0.001), and 9p loss similarly showed a significantly worse prognosis (HR: 1.79, 95% CI: 1.40–2.29, P < 0.001). In lung cancer, 8q gain, 9p21.3 deletion, and 9p loss were significantly associated with worse OS, with hazard ratios of 1.57 (HR: 1.57, 95% CI: 1.33–1.87, P < 0.001), 1.83 (95% CI: 1.46–2.30, P < 0.001) and 1.63 (95% CI: 1.38–1.92, P < 0.001), respectively. In melanoma, both 9p21.3 deletion and 9p loss were significantly associated with worse OS with hazard ratios of 2.19 (95% CI: 1.39–3.46, P < 0.001) and 1.83 (95% CI: 1.21–2.76, p = 0.004), respectively. Although 8p loss showed a trend toward worse OS, it was not statistically significant (HR: 1.41, 95% CI: 0.85–2.37, p = 0.187). To further validate the candidate CNAs enriched in brain metastasis, we also conducted the same survival analyses in colorectal cancer (MSK-COAD), small-cell lung cancer (MSK-SCLC), and prostate cancer (MSK-PRAD). The results were largely consistent with our findings for the other cancer types above (Supplementary Fig. 7A-B). Above all, 8p loss, 8q gain, and 9p21.3 loss are consistently associated with brain metastasis and poor prognosis multiple cancer types.

We identified several candidate genomic alterations associated with brain metastases in RCC, including 8p loss, 8q gain, and 9p21.3 deletion. Beyond RCC, the candidate CNAs also occur frequently in brain metastases from lung adenocarcinoma, breast cancer, and melanoma, potentially indicating shared mechanisms of brain metastatic dissemination across cancer types. Importantly, these genomic events were associated with poor prognosis even in stage I RCC, highlighting their potential value for early risk stratification and personalized monitoring. Our results, along with findings in other studies, collectively suggest that 9p21.3 deletion contributes to the formation of brain metastases from RCC and other cancer types. This is important because 9p21.3 deletion occurs frequently in ~13% of cancers28. The 9p21.3 locus contains CDKN2A/B and MTAP, key regulators of cell cycle control and tumor metabolism29,30. Alterations in CDKN2A/B have been implicated in brain metastases across NSCLC31, LUAD11, breast cancer32, and melanoma33, and CDKN2A/B co-deletion in resected brain metastases correlates with increased local and distant CNS progression34. Recent work shows that 9p21.3 deletion promotes a “cold” immune microenvironment through type I IFN pathway suppression and loss of regulatory elements28,35, contributing to poor responses to immune checkpoint therapy28,35. Notably, RCC tumors with 9p21 deletions exhibit CD8+ T-cell infiltration yet still experience poor survival after PD-1 blockade36, potentially explaining the limited immunotherapy benefit in RCC brain metastasis37 and challenging the assumption that immune infiltration predicts improved outcomes38. Functionally, 9p21.3 loss appears to be an early metastasis-enabling event. In pancreatic cancer, loss of type I IFN genes in this region promotes immune evasion during metastatic seeding39, and 9p21 loss has also been shown to confer a chromosomal-instability–tolerant phenotype that supports aneuploid tumor growth and metastasis40. Together, these findings support 9p21.3 deletion as an early genomic trigger in metastatic progression, especially to the brain. Therapeutically, CDK4/6 inhibitors are being evaluated to target CDKN2A/B loss in brain metastases (NCT02896335; NCT03994796). Gain of 8q is also a promising driver of RCC brain metastasis. It is associated with metastasis and poor survival in ccRCC41. Multiple studies have shown focal amplification of 8q24 in ~15–23% of ccRCC, consistently involving MYC within the amplified region18,42,43. MYC amplification has previously been identified and functionally validated as a brain metastatic driver in lung adenocarcinoma11.. Mechanistically, MYC promotes brain colonization by enhancing invasiveness, recruiting macrophages, and inducing gap junction formation between cancer cells and astrocytes44. MYC also upregulates redox enzymes that mitigate ROS generated by activated microglia, potentially protecting metastatic cells within the brain microenvironment45. Several preclinical studies and early-phase clinical trials are currently investigating MYC-targeting approaches, but none are specifically designed for patients with brain metastases46. We also observed frequent 8p loss in RCC brain metastases, which has been associated with metastatic potential in multiple cancer types47,48. Evidence suggests that 8p loss promotes metastasis through deregulation of multiple tumor suppressors including MSRA, NAT1, PPP2CB, and DLC1, rather than a single driver48–50. In prostate cancer, loss of NKX3.1 (also on 8p) increases oxidative stress and disrupts mitochondrial function, thereby promoting progression51. Notably, the 8p-loss cluster was enriched in breast cancers that developed brain, but not lung or bone, metastases17. It alters fatty acid and ceramide metabolism, increasing invasiveness and autophagy49. Together, these findings suggest that 8p loss may contribute to biological processes relevant to brain metastasis, although causal mechanisms remain to be fully elucidated. We observed significant associations of these recurrent CNAs in brain metastasis across multiple cancer types, supporting the concept of shared genomic determinants of organ-specific metastasis. This is consistent with the pan-cancer brain metastasis genomic analysis by Brastianos et al.8, which identified recurrent PI3K/AKT pathway alterations across brain metastases from diverse primary sites. However, whether these alterations directly confer brain tropism or reflect broader features of aggressive disease remains to be determined. To translate these genomic insights into a clinically relevant framework, we developed a RSF model incorporating recurrent CNAs.). To our knowledge, this is the first CNA-based prognostic model in RCC, achieving a concordance index of 0.64, comparable to established clinical models such as IMDC, highlighting the additive prognostic value of genomic data52. Therefore, our CNA-based survival model is promising candidate for future validation in prospective cohorts. This study has some limitations. First, copy-number and SNV calls were derived from bulk WES, which cannot fully resolve intratumoral heterogeneity or spatially restricted clones within the brain microenvironment. Second, functional validation is indirect: MetMap analyses rely on cell-line systems and intracardiac injection models that may not recapitulate all aspects of human CNS colonization, including immune contexture, blood–brain barrier dynamics, and stromal–neural interactions. Third, RCC-specific experimental models of brain metastasis remain limited, restricting direct mechanistic validation of the observed associations. In summary, we identify recurrent CNAs that are enriched in RCC brain metastasis and associated with adverse clinical outcomes across multiple cancer types. These findings suggest a potential genomic basis underlying brain metastasis and support further investigation of can-based biomarkers for risk stratification. Our results uncover new areas of brain metastasis research, such as dissecting how 8p loss, 8q gain, and 9p21.3 deletion cooperate to promote brain tropism by using CNS-competent PDX systems and in vivo CRISPR perturbations53. Another future direction is explore how the CNA drivers remodel tumour microenvironment remodeling. Finally, potential avenues for therapeutic development for brain metastases include CDK4/6 inhibition in CDKN2A/B-deleted tumors, synthetic-lethality in MTAP co-deleted disease, and targeting lipid metabolism in RCC with 8p loss.

We wish to acknowledge the Pathology Core within the Brain Tumor Research Program at MD Anderson Cancer Center supported through the SPORE in Brain Cancer award (P50CA127001) and the generous philanthropic contributions to the MD Anderson Cancer Neuroscience Program. Funding E.H. was supported by the American Society of Clinical Oncology Conquer Cancer Foundation Young Investigator Award, the International Kidney Cancer Coalition Cecile and Ken Youner Scholarship, the Society for Immunotherapy of Cancer-NanoString Technologies Single Cell Biology Award, the Kidney Cancer Association Young Investigator Award, Gateway Cancer Research Foundation and the Pelotonia Institute for Immuno-Oncology Recruitment Startup Fund. Notably, the Society for Immunotherapy of Cancer-NanoString Technologies Single Cell Biology Award, the Kidney Cancer Association Young Investigator Award, and the Pelotonia Institute for Immuno-Oncology Startup Funds supported the correlative tissue sequencing and analyses conducted in this study. D.J.H.S. was supported by the HKU-100 Scholar Start-up Fund; the work described in this paper was partially supported by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (Project No. HKU 27102723). M.H. was supported by American Society of Clinical Oncology Conquer Cancer Foundation Young Investigator Award and start-up funds from The Ohio State University Comprehensive Cancer Center, which contributed to the analysis of this study. Competing interests E.H. reports receiving honoraria from Target Oncology, DAVA Oncology, participating on advisory boards for Telix Pharmaceuticals, Pfizer, and Eisai, and receiving research support to his institution from Exelixis. M.H. reports receiving honoraria from Curio Sciences, participating on advisory boards for Bristol Myers Squibb, and receiving research support to her institution from Exelixis. All other authors report no conflicts of interest. List of abbreviations - BM Brain metastasis - BRCA breast cancer - CNS central nervous system - ccRCC clear cell renal cell carcinoma - C-index concordance index - CNA copy-number alteration - ECM extracranial metastasis - FFPE formalin-fixed paraffin-embedded - GATK Genome Analysis Toolkit - HR hazard ratio - IFN interferon - IMDC International Metastatic Renal Cell Carcinoma Database Consortium - IRB institutional review board - KIRC kidney renal clear cell carcinoma - LUAD lung adenocarcinoma - MetMap Metastasis Map of Human Cancer Cell Lines - MSKCC Memorial Sloan Kettering (MSK) and Memorial Sloan Kettering Cancer Center - MTAP methylthioadenosine phosphorylase - NSCLC non–small cell lung cancer - OS overall survival - PD-1 programmed cell death protein 1 - PFI progression-free interval - PFS progression-free survival - PI3K phosphoinositide 3-kinase - PTEN phosphatase and tensin homolog - qPCR quantitative polymerase chain reaction - ROS reactive oxygen species - RCC renal cell carcinoma - RCC-BM renal cell carcinoma brain metastasis - RSF random survival forest - SNV single-nucleotide variant - SCLC small-cell lung cancer - TCGA-KIRC The Cancer Genome Atlas (TCGA), including the TCGA kidney renal clear cell carcinoma cohort - WT wild type - WES whole-exome sequencing Funding Statement E.H. was supported by the American Society of Clinical Oncology Conquer Cancer Foundation Young Investigator Award, the International Kidney Cancer Coalition Cecile and Ken Youner Scholarship, the Society for Immunotherapy of Cancer-NanoString Technologies Single Cell Biology Award, the Kidney Cancer Association Young Investigator Award, Gateway Cancer Research Foundation and the Pelotonia Institute for Immuno-Oncology Recruitment Startup Fund. Notably, the Society for Immunotherapy of Cancer-NanoString Technologies Single Cell Biology Award, the Kidney Cancer Association Young Investigator Award, and the Pelotonia Institute for Immuno-Oncology Startup Funds supported the correlative tissue sequencing and analyses conducted in this study. D.J.H.S. was supported by the HKU-100 Scholar Start-up Fund; the work described in this paper was partially supported by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (Project No. HKU 27102723). M.H. was supported by American Society of Clinical Oncology Conquer Cancer Foundation Young Investigator Award and start-up funds from The Ohio State University Comprehensive Cancer Center, which contributed to the analysis of this study. Footnotes Declaration of generative AI and AI-assisted technologies in the manuscript preparation process During the preparation of this work the author(s) used ChatGPT in order to improve the grammar and readability of the manuscript. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article. Data and Code Availability Processed DNA sequencing data, and supporting files will be made available through a project-specific request to the corresponding author. The scripts used for the analysis are available at the following GitHub repositories: https://github.com/hasanovlab (available upon publication).

- 1.Hasanov E. et al. An interdisciplinary consensus on the management of brain metastases in patients with renal cell carcinoma. CA Cancer J. Clin. 72, 454–489 (2022). [DOI] [PubMed] [Google Scholar] - 2.Hasanov E. & Jonasch E. Management of brain metastases in metastatic renal cell carcinoma. Hematol. Oncol. Clin. North Am. 37, 1005–1014 (2023). [DOI] [PubMed] [Google Scholar] - 3.Tannir N. M. et al. Patterns of progression in patients treated with nivolumab plus ipilimumab (NIVO+IPI) versus sunitinib (SUN) for first-line treatment of advanced renal cell carcinoma (aRCC) in CheckMate 214. J. Clin. Oncol. 39, 313–313 (2021). [Google Scholar] - 4.Bowman I. A. et al. Improved survival outcomes for kidney cancer patients with brain metastases. Clin. Genitourin. Cancer 17, e263–e272 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar] - 5.Jonasch E. et al. Evaluation of brain metastasis in JAVELIN Renal 101: Efficacy of avelumab + axitinib (A+Ax) versus sunitinib (S). J. Clin. Oncol. 38, 687–687 (2020). [Google Scholar] - 6.Emamekhoo H. et al. Safety and efficacy of nivolumab plus ipilimumab in patients with advanced renal cell carcinoma with brain metastases: CheckMate 920. Cancer 128, 966–974 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar] - 7.Flippot R. et al. Safety and efficacy of nivolumab in brain metastases from renal cell carcinoma: Results of the GETUG-AFU 26 NIVOREN multicenter phase II study. J. Clin. Oncol. 37, 2008–2016 (2019). [DOI] [PubMed] [Google Scholar] - 8.Brastianos P. K. et al. Genomic Characterization of Brain Metastases Reveals Branched Evolution and Potential Therapeutic Targets. Cancer Discov. 5, 1164–1177 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar] - 9.Turajlic S. et al. Tracking Cancer Evolution Reveals Constrained Routes to Metastases: TRACERx Renal. Cell 173, 581–594.e12 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar] - 10.Monzon F. A. et al. Chromosome 14q loss defines a molecular subtype of clear-cell renal cell carcinoma associated with poor prognosis. Mod. Pathol. 24, 1470–1479 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar] - 11.Shih D. J. H. et al. Genomic characterization of human brain metastases identifies drivers of metastatic lung adenocarcinoma. Nat. Genet. 52, 371–377 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar] - 12.Olshen A. B., Venkatraman E. S., Lucito R. & Wigler M. Circular binary segmentation for the analysis of array-based DNA copy number data. Biostatistics 5, 557–572 (2004). [DOI] [PubMed] [Google Scholar] - 13.The Cancer Genome Atlas Research Network et al. The Cancer Genome Atlas Pan-Cancer analysis project. Nat Genet 45, 1113–1120 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar] - 14.Cerami E. et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2, 401–404 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar] - 15.Northcott P. A. et al. Subgroup-specific structural variation across 1,000 medulloblastoma genomes. Nature 488, 49–56 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar] - 16.Martincorena I. et al. Universal Patterns of Selection in Cancer and Somatic Tissues. Cell 173, 1823 (2018). [Google Scholar] - 17.Jin X. et al. A metastasis map of human cancer cell lines. Nature 588, 331–336 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar] - 18.Cancer Genome Atlas Research Network. Comprehensive molecular characterization of clear cell renal cell carcinoma. Nature 499, 43–49 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar] - 19.Nguyen B. et al. Genomic characterization of metastatic patterns from prospective clinical sequencing of 25,000 patients. Cell 185, 563–575.e11 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar] - 20.Jin X. et al. A metastasis map of human cancer cell lines. Nature 588, 331–336 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar] - 21.Ghandi M. et al. Next-generation characterization of the Cancer Cell Line Encyclopedia. Nature 569, 503–508 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar] - 22.Shih D. J. H. et al. CNGPLD: case-control copy-number analysis using Gaussian process latent difference. Bioinformatics 38, 2096–2101 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar] - 23.Fukumura K. et al. Genomic characterization of primary central nervous system lymphoma. Acta Neuropathol. 131, 865–875 (2016). [DOI] [PubMed] [Google Scholar] - 24.Fischer G. M. et al. Molecular profiling reveals unique immune and metabolic features of melanoma brain metastases. Cancer Discov. 9, 628–645 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar] - 25.Adamo B. et al. Phosphatidylinositol 3-kinase pathway activation in breast cancer brain metastases. Breast Cancer Res. 13, R125 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar] - 26.Chen G. et al. Molecular profiling of patient-matched brain and extracranial melanoma metastases implicates the PI3K pathway as a therapeutic target. Clin. Cancer Res. 20, 5537–5546 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar] - 27.Davies M. A. et al. Integrated molecular and clinical analysis of AKT activation in metastatic melanoma. Clin. Cancer Res. 15, 7538–7546 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar] - 28.Han G. et al. 9p21 loss confers a cold tumor immune microenvironment and primary resistance to immune checkpoint therapy. Nat. Commun. 12, 5606 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar] - 29.Alhalabi O. et al. MTAP deficiency creates an exploitable target for antifolate therapy in 9p21-loss cancers. Nat. Commun. 13, 1797 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar] - 30.Tang B., Lee H.-O., An S. S., Cai K. Q. & Kruger W. D. Specific targeting of MTAP-deleted tumors with a combination of 2’-fluoroadenine and 5’-methylthioadenosine. Cancer Res. 78, 4386–4395 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar] - 31.Huang R. S. P. et al. Clinicopathologic and genomic landscape of non-small cell lung cancer brain metastases. Oncologist 27, 839–848 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar] - 32.Furet E. et al. Increased risk of brain metastases in women with breast cancer and p16 expression in metastatic lymph-nodes. Oncotarget 8, 37332–37341 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar] - 33.Arnoff T. E. & El-Deiry W. S. MDM2/MDM4 amplification and CDKN2A deletion in metastatic melanoma and glioblastoma multiforme may have implications for targeted therapeutics and immunotherapy. Am. J. Cancer Res. 12, 2102–2117 (2022). [PMC free article] [PubMed] [Google Scholar] - 34.Morshed R. A. et al. CDKN2A/B co-deletion is associated with increased risk of local and distant intracranial recurrence after surgical resection of brain metastases. Neurooncol. Adv. 5, vdad007 (2023). [Google Scholar] - 35.Spiliopoulou P. et al. All is not lost: learning from 9p21 loss in cancer. Trends Immunol. 43, 379–390 (2022). [DOI] [PubMed] [Google Scholar] - 36.Braun D. A. et al. Interplay of somatic alterations and immune infiltration modulates response to PD-1 blockade in advanced clear cell renal cell carcinoma. Nat. Med. 26, 909–918 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar] - 37.Reed J. P., Posadas E. M. & Figlin R. A. Brain metastases in renal cell carcinoma: Immunotherapy responsiveness is multifactorial and heterogeneous. J. Clin. Oncol. 37, 1987–1989 (2019). [DOI] [PubMed] [Google Scholar] - 38.Fridman W. H., Zitvogel L., Sautès-Fridman C. & Kroemer G. The immune contexture in cancer prognosis and treatment. Nat. Rev. Clin. Oncol. 14, 717–734 (2017). [DOI] [PubMed] [Google Scholar] - 39.Barriga F. M. et al. MACHETE identifies interferon-encompassing chromosome 9p21.3 deletions as mediators of immune evasion and metastasis. Nat. Cancer 3, 1367–1385 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar] - 40.Perelli L. et al. Interferon signaling promotes tolerance to chromosomal instability during metastatic evolution in renal cancer. Nat. Cancer 4, 984–1000 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar] - 41.Klatte T. et al. Gain of chromosome 8q is associated with metastases and poor survival of patients with clear cell renal cell carcinoma. Cancer 118, 5777–5782 (2012). [DOI] [PubMed] [Google Scholar] - 42.Sato Y. et al. Integrated molecular analysis of clear-cell renal cell carcinoma. Nat. Genet. 45, 860–867 (2013). [DOI] [PubMed] [Google Scholar] - 43.Beroukhim R. et al. Patterns of gene expression and copy-number alterations in von-hippel lindau disease-associated and sporadic clear cell carcinoma of the kidney. Cancer Res. 69, 4674–4681 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar] - 44.Lee H. Y. et al. C-MYC drives breast cancer metastasis to the brain, but promotes synthetic lethality with TRAIL. Mol. Cancer Res. 17, 544–554 (2019). [DOI] [PubMed] [Google Scholar] - 45.Klotz R. et al. Circulating tumor cells exhibit metastatic tropism and reveal brain metastasis drivers. Cancer Discov. 10, 86–103 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar] - 46.Llombart V. & Mansour M. R. Therapeutic targeting of ‘undruggable’ MYC. EBioMedicine 75, 103756 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar] - 47.Macartney-Coxson D. P. et al. Metastatic susceptibility locus, an 8p hot-spot for tumour progression disrupted in colorectal liver metastases: 13 candidate genes examined at the DNA, mRNA and protein level. BMC Cancer 8, 187 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar] - 48.Huth T. et al. Chromosome 8p engineering reveals increased metastatic potential targetable by patient-specific synthetic lethality in liver cancer. Sci Adv 9, eadh1442 (2023). [Google Scholar] - 49.Cai Y. et al. Loss of Chromosome 8p Governs Tumor Progression and Drug Response by Altering Lipid Metabolism. Cancer Cell 29, 751–766 (2016). [DOI] [PubMed] [Google Scholar] - 50.Xue W. et al. A cluster of cooperating tumor-suppressor gene candidates in chromosomal deletions. Proc. Natl. Acad. Sci. U. S. A. 109, 8212–8217 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar] - 51.Papachristodoulou A. et al. NKX3.1 localization to mitochondria suppresses prostate cancer initiation. Cancer Discov. 11, 2316–2333 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar] - 52.Heng D. Y. et al. External validation and comparison with other models of the International Metastatic Renal-Cell Carcinoma Database Consortium prognostic model: a population-based study. The Lancet Oncology 14, 141–148 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar] - 53.Feng H. et al. Genome-wide Chromosome-specific Aneuploidy Engineering and Phenotypic Characterization with CRISPR-Taiji. Preprint at 10.1101/2025.04.25.650684 (2025). [DOI] [Google Scholar] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Data Availability Statement Processed DNA sequencing data, and supporting files will be made available through a project-specific request to the corresponding author. The scripts used for the analysis are available at the following GitHub repositories: https://github.com/hasanovlab (available upon publication).