Characterisation and impact of intratumoural stroma in melanoma and carcinoma brain metastases

2020 Background: Lung cancer is one of the most common causes of brain metastases (BMs) and is always associated with poor prognosis. To evaluate the characteristics of the tumor immune microenvironment in brain metastases of non-small-cell lung cancer (NSCLC), we investigated the immunophenotype of primary NSCLC and paired brain metastases. Methods: Forty-three Chinese patients with NSCLC who had BMs at presentation or during the course of their disease were admitted to the Sun Yat-Sen University Cancer Center (Guangzhou, China) from 2000 to 2019. RNA sequencing (RNA-seq) of eighty-six formalin-fixed, paraffin embedded (FFPE) samples from primary lung tumors and paired brain metastases of 43 patients was conducted to comprehensively analyze the tumor immune microenvironment. Results: Our data revealed that brain metastases compared with primary lung tumors exhibited reduced tumor infiltrating lymphocytes (TILs) (all 28 immune cell subtypes P < 0.05), lower fraction of activated CD8 T cell and effector memory CD8 T cell in total TILs (P = 0.028, P < 0.001, respectively); higher fraction of macrophage and neutrophil in total TILs (P < 0.001, P < 0.01, respectively). Comparing with the primary lung tumors, the scores of some immune related signatures, including MHC non-class signature, IFN gamma signature and T-cell-inflamed gene-expression profile (GEP) signature, were significantly lower in brain metastases (P = 0.004, P = 0.009, P = 0.004, respectively), while the score of MHC class-II signature was higher in brain metastases (P = 0.045). We found the distributions of tumor microenvironment immune types (TMIT) in brain metastases and primary lung tumors were different. Brain metastases contained significantly lower proportion of TMIT I (high PD-L1/ high CD8A) (23%) than primary lung tumors (47%) (P < 0.05). Besides, we found three immune inhibitory checkpoint molecules, namely C10orf54 (VISTA), CTLA4 and CD274 (PD-L1) were downregulated in brain metastases than in primary lung tumors (P < 0.001, P < 0.001, P = 0.034, respectively). Moreover, there was poor correlation of PD-L1 expression between paired brain metastases and primary lung tumors (R = 0.28, P = 0.068). Unsupervised hierarchic cluster analysis revealed the primary lung tumors had two distinct patterns of immune gene signatures, namely Cluster A and Cluster B, and the tumors in Cluster B were immune rich, but associated with poor prognosis (log-rank P = 0.021). Conclusions: Our work illustrates the immune landscape of brain metastases from NSCLC, and suggests that the tumor immune microenvironment in brain metastases compared with primary lung tumors is further immunosuppressed, that may help guide immunotherapeutic strategies for NSCLC brain metastases.

Immunotherapy for NSCLC With Brain Metastases: What Can We Learn From Real-World Data?

BACKGROUND: Brain metastases (BrM) arising from breast cancer (BC) are an increasing consequence of advanced disease, with up to half of patients (pts) with metastatic HER2+ or triple negative breast cancer experiencing central nervous system (CNS) recurrence. The genomic alterations driving CNS recurrence, along with contribution of the immune microenvironment, particularly by BC subtype remains unclear. METHODS: We characterized BrM from a cohort of n=42 BC pts by sequencing whole-exome DNA and total RNA libraries from frozen (n=31) and FFPE (n=34) BCBrM, FFPE extracranial (ECT, n=12) and blood DNA (n=26) tissues from the Duke Brain Tumor Biorepository. Analyses conducted and planned include inference of PAM50 intrinsic subtypes, somatic mutations, copy number alterations, immune cell type decompositions by CIBERSORTx, differential RNA expression, driver analysis, and associations with clinical outcomes. RESULTS: PAM50 subtypes across 31 frozen BrM were 23% Luminal (Lum) A, 13% LumB, 33% HER2-enriched, 27% Basal-like, 3% Normal-like. For 34 FFPE BrM, 15% were LumA, 24% LumB, 33% HER2-enriched, 24% Basal-like, 3% Normal-like. Across 26 paired FFPE and frozen BrM, subtype discrepancy was seen in 23% (6) cases of which frozen to FFPE classification differed: 3 LumA to LumB, 1 LumB to LumA, 1 Normal to LumB, and 1 Basal to LumB. Among 10 paired FFPE BrM and ECT, subtype concordance was observed in 70%. Discordance was seen in 1 case each for ECT to BrM: LumB to HER2, LumB to LumA, Basal to Normal-like. WES demonstrated frequent copy number alterations (CNA) in clinically-relevant genes including TGFB1, NOTCH1, CDK4, and ERBB3, in both BrM and ECT. For pts with matched blood, ERBB3 CNA (gain and loss) were more commonly found in FFPE BrM (7/19, 37%) compared to ECT (1/8, 13%). Among BrM, the most frequently altered clinically-relevant genes included TP53 (~75%), PTEN, EGFR, RB1, PIK3CA, NF1 and ESR1 (all ≤15%); in ECT, TP53 (67%), BRCA2, FBXW7, and ATM (all ≤33%). PI3K pathway genes (e.g. PTEN, PIK3CA) alterations were exclusive to BrM. One pt’s ECT showed an ATM mutation that was not observed in the paired BrM. Conversely, BRAF and CCND2 mutations were observed in BrM, but not ECT for 1 pt each. Inferring the relative abundance of immune populations in frozen BrM illustrated that 23% were CD4+ resting memory T cells, 25% M2 tumor-permissive macrophages, 13% M0 macrophages; M1 tumor-inhibiting macrophages were only 2%. When comparing immune cell populations between FFPE BrM and ECT, ECT had more M1 macrophages (Chi-sq 4.23, P = 0.04), while other immune cell populations were of similar relative abundance. Immune cell fractions did not vary by subtype in BrM or ECT with one exception: M2 macrophages were lower in Basal compared to LumB tissues (frozen BrM: Chi-sq = 9.28, P = 0.05; FFPE BrM: Chi-sq = 4.61, P = 0.33; FFPE EC: Chi-sq = 9.78, P = 0.04). In n=8 pt-matched FFPE BrM/ECT, hallmark pathways upregulated in BrM included MYC and E2F targets and oxidative phosphorylation, while those with lower expression in BrM included epithelial-mesenchymal transition, interferon gamma response, and JAK-STAT signaling. CONCLUSION: This analysis showed moderate discrepancy in subtype call of BrM by tissue preparation (frozen vs. FFPE), with LumB classification showing the highest discrepancy, and more commonly called in FFPE tissues. Subtype concordance between ECT and BrM was relatively high. Analysis of CNA illustrated deletions and amplifications in targetable genes, notably ERBB3 preferentially in BrM compared to ECT. Mutational analysis identified targetable alterations exclusive to BrM; this knowledge could lead to BrM-targeted treatments. Inferred immune cell populations illustrated a tumor permissive microenvironment in BrM. Therapeutic strategies repolarizing macrophages toward a tumor-inhibiting phenotype in BrM are warranted. Analyses of associations between genomic data and clinical outcomes, as well as driver analyses, are ongoing. Citation Format: Amanda Van Swearingen, Marissa Lee, Layne Rogers, Alexander Sibley, Pixu Shi, Xiaodi Qin, Michael Goodin, Kouros Owzar, Carey Anders. Genomic and immune profiling of breast cancer brain metastases [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO4-14-07.

Prognostic Factors Impacting Surgical Resection Outcomes in Elderly Patients With Brain Metastasis.

Metastasis is one of the important biological features of malignant tumors and one of the main factors responsible for poor prognosis. Although the widespread application of newer clinical technologies and their continuous development have significantly improved survival in patients with brain metastases, there is no uniform standard of care. More effective therapeutic measures are therefore needed to improve prognosis. Understanding the mechanisms of tumor cell colonization, growth, and invasion in the central nervous system is of particular importance for the prevention and treatment of brain metastases. This process can be plausibly explained by the "seed and soil" hypothesis, which essentially states that tumor cells can interact with various components of the central nervous system microenvironment to produce adaptive changes; it is this interaction that determines the development of brain metastases. As a novel form of intercellular communication, exosomes play a key role in the brain metastasis microenvironment and carry various bioactive molecules that regulate receptor cell activity. In this paper, we review the roles and prospects of brain metastatic tumor cells, the brain metastatic tumor microenvironment, and exosomes in the development and clinical management of brain metastases.

Brain metastases are a very common and serious complication of oncological diseases. Despite the vast progress in multimodality treatment, brain metastases significantly decrease the quality of life and prognosis of patients. Therefore, identifying new targets in the microenvironment of brain metastases is desirable. Fibroblast activation protein (FAP) is a transmembrane serine protease typically expressed in tumour-associated stromal cells. Due to its characteristic presence in the tumour microenvironment, FAP represents an attractive theranostic target in oncology. However, there is little information on FAP expression in brain metastases. In this study, we quantified FAP expression in samples of brain metastases of various primary origin and characterised FAP-expressing cells. We have shown that FAP expression is significantly higher in brain metastases in comparison to non-tumorous brain tissues, both at the protein and enzymatic activity levels. FAP immunopositivity was localised in regions rich in collagen and containing blood vessels. We have further shown that FAP is predominantly confined to stromal cells expressing markers typical of cancer-associated fibroblasts (CAFs). We have also observed FAP immunopositivity on tumour cells in a portion of brain metastases, mainly originating from melanoma, lung, breast, and renal cancer, and sarcoma. There were no significant differences in the quantity of FAP protein, enzymatic activity, and FAP+ stromal cells among brain metastasis samples of various origins, suggesting that there is no association of FAP expression and/or presence of FAP+ stromal cells with the histological type of brain metastases. In summary, we are the first to establish the expression of FAP and characterise FAP-expressing cells in the microenvironment of brain metastases. The frequent upregulation of FAP and its presence on both stromal and tumour cells support the use of FAP as a promising theranostic target in brain metastases.

Differences in the immune microenvironment and responses to immunotherapy may exist between primary non-small cell lung cancer (NSCLC) and brain metastases (BMs). This study aimed to investigate discrepancies in programmed death-ligand 1 (PD-L1) expression, tumor-infiltrating lymphocytes (TILs), tertiary lymphoid structures (TLS), and tumor mutational burden (TMB) between matched BMs and primary tumors (PTs) in NSCLC. Twenty-six pairs of surgically resected BMs and corresponding PTs from NSCLC patients were collected. PD-L1 expression and TILs, including CD8, CD3, CD4, CD20, CD68, and CD21, were analyzed using immunohistochemistry (IHC) and quantitatively assessed through digital image analysis. Whole-exome sequencing (WES) was performed to investigate genomic discrepancies and variations in TMB. The density of PD-L1+ cells did not differ significantly between matched PTs and BMs (P>0.99). However, BMs exhibited a higher tumor proportion score (TPS) compared to PTs (mean TPS: 31.92% vs. 25.96%, P=0.049), with moderate agreement in categorized TPS (κ=0.653). Analysis of TILs revealed a significant reduction in CD3+ T cells (P<0.001), CD8+ cytotoxic T cells (P<0.001), CD20+ B cells (P<0.001), and CD68+ macrophages (P=0.02) in BMs compared to PTs. BMs also exhibited a loss of TLS, with no presence of mature TLS marked by CD21 expression. The number of non-synonymous mutations was generally higher in BMs than in PTs, with only 34.69% of mutations shared between paired PTs and BMs. TMB was slightly increased in BMs (mean TMB: 34.2 mutations/Mb in BMs vs. 26.8 mutations/Mb in PTs; P=0.30). Additionally, the log-rank test indicated that a higher density of CD20+ B cells in BMs was significantly associated with better overall survival (P=0.007). Compared to primary NSCLC tumors, matched BMs show an increase in TPS of PD-L1 expression and TMB, but a significant reduction in TILs and loss of mature TLS, suggesting an immune-suppressive microenvironment in BMs. The infiltration of CD20+ B cells may serve as a potential prognostic biomarker in NSCLC with BMs.

Background: Despite advances in the treatment of breast cancer, brain metastases (BM) remain a significant clinical problem. Detailed biological information is increasing on extracranial metastases through for example the AURORA program, however the biology of BM is less studied. Immune checkpoint inhibitors (ICI) are efficient in treatment of early and disseminated TNBC but the effect on BM seems to be modest. Aims: To compare gene expression (GE) signatures between BM and the primary breast tumours (PT) by use of NanoString nCounter 360™ Panel with a focus on signatures involved in the immune system and microenvironment (ME), as well as levels of CD4+, CD8+, CD68+ and Treg cells by immunohistochemistry. We also wanted to compared GE signatures between BM according to mutational status i.e. wildtype (wt) or somatic mutations in TP53 and PIK3CA respectively. Methods: Patients with surgery of BM between 1994-2014 were identified. Clinical data and presence of TP53 and PIK3CA mutations have been published previously. The Breast Cancer 360TM panel version 2 (nCounter® platform) was used according to the manufacturer’s specifications (NanoString Technologies, Inc., Seattle, WA, USA). IHC was performed according to the manufacturer’s specifications with the following antibodies: CD4(1:100; IR649), CD8 (1:100; M7103), CD68 (1:200; IR613) (DAKO, Glostrup, Denmark) and FOXP3 (1:200; AB20034) (Abcam, Cambridge, United Kingdom). Stained slides were scanned on Hamamatsu Photonics NanoZoomer s210. Differences between GE signatures in BM and PT were calculated with Anova and Bonferroni post hoc analysis and for IHC with Wilcoxon paired test. Significance was set at 0.05 unless otherwise specified. Results: Analyses were performed on 60 BM out of which 23 cases had the matched PT and five single PT. The following GE signatures were statistically significantly lower in BM than in PT: PD-L1, PD-1, PD-L2, CD8+ T-Cells, Treg, B7-H3 or CD276 (type I membrane protein with a sequence similar to the extracellular domain of PD-L1), IFN Gamma, MHC2, apM (antigen processing machinery), TGF-Beta, TIS (tumour inflammation signature), TIGIT (T cell immunoreceptor with Ig and ITIM domains), Stroma, IDO1, Mast Cells, Endothelial Cells (all p&amp;lt; 0,001) and Inflammatory Chemokines (p=0,02). The following GE signatures were upregulated in BM compared with PT; SOX2 (p &amp;lt; 0,001), Genomic Risk (p &amp;lt;0,001), BRCA-ness (p &amp;lt; 0,001), Differentiation (p= 0,004), Hypoxia (p &amp;lt;0,001), BC Proliferation (p=0,002) and HRD (p=0,047). The IHC staining showed statistically significant lower expression of CD4+ T helper cells (p=0.007), cytotoxic CD8+ T cells (p=0.04), TILs (p=0.004) and a trend not reaching statistical significance of Treg (p=0.07). No difference was found on the levels of CD68+ (p=0.35). BM with TP53 mutations had higher expression of BC proliferation (p=0.001), HRD (p=0.007) and mast cells (p=0.001) compared with BM with wt TP53. BM with PIK3CA mutations had higher expression of MHC class 2 (p = 0.027) and BC proliferation (p = 0.033) compared with BM with wt PIK3CA. Conclusions: We found a downregulation of GE signatures related to the immune system and tumour ME in BM compared with PT. This was validated by the lower levels of TILs, CD4+ and CD8+ T cells in BM. The lower immunogenicity of BM may partly explain the reduced effect by immunotherapy intracranially. New treatment combinations that enhance the immunogenicity of BM deserve further investigations. Citation Format: Anna Thulin, Anika Kovacs, Anne-Vibeke Laenkholm, Henrik Fagma. Gene expression profiling of brain metastases and matched primary breast tumours with focus on the immune system and tumour microenvironment [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr PS14-06.

Purpose Brain metastases can induce life-threatening events and their immune microenvironment is still unrevealed. To elucidate the probability of immune-oncology therapy for brain metastases, we utilized immune profiling of resected brain metastases and paired non-small cell lung cancer (NSCLC). Methods To compare pure tumor-infiltrating lymphocytes (TILs) or tumor tissues between NSCLC and brain metastases, macro-dissection was performed before extracting DNA and RNA from 39 patients who received surgical resection of local NSCLC tumor lesions and brain metastases from matched patients. We applied gene expression analysis of mRNA by using a 770 gene panel related with cancer immune cells. Targeted sequencing of 50 cancer hotspot genes on extracted DNA was performed for genotyping. TCRß sequencing of extracted DNA was utilized for profiling of TCR repertoire. Results We identified 161 differentially expressed genes, which reflected inhibition of dendritic cell maturation pathway, Th1 pathway, and leukocyte extravasation signaling pathway in brain metastases (p &amp;lt; 0.01). VCAM1 was one of highly inter-connected network genes with low expression level in brain metastases. Immune cell profiling analysis illustrated low T-cell infiltration and higher ratio of macrophages against TILs in the brain metastases (p &amp;lt; 0.001). The detected mutations in genotyping analysis were completely shared in 28 patients (71%). TCR sequencing analysis revealed that T-cell clones were expanded in 64% of patients with brain metastases. The majority of TCR repertoire in brain metastases were principally shared with the paired NSCLC, while the density of T cells was sparse in the brain metastases. Conclusion The microenvironment in brain metastases from NSCLC was revealed to be immunosuppressive, with the majority of T-cell clones in brain metastases being shared with the primary lesion. Citation Format: Yujin Kudo, Cara Haymaker, Jiexin Zhang, Alexandre Reuben, Dzifa Yawa Duose, Junya Fujimoto, Sinchita Roy-Chowdhuri, Luisa Maren Solis, Hitoshi Dejima, Edwin Roger Cuentas, Barbara Mino, Ronald Abraham, Norihiko Ikeda, Rajyalakshmi Luthra, Jack J. Lee, Humam Kadara, Jason T. Huse, Ignacio I. Wistuba. Characterization of the immune microenvironment in brain metastasis from non-small cell lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2790.

1081 Background: In patients with brain metastasis (BM) of melanoma or lung cancer, significant activity of immune checkpoint inhibitors has been reported. However, details of the immune microenvironment in BM has not been unveiled. In this study, we used immunohistochemistry (IHC) to compare primary breast tumors and BM tumor samples with respect to tumor infiltrating lymphocytes (TILs) and tumor characteristics related to the immune system. Methods: We retrospectively identified 107 patients with breast cancer, diagnosed with BM, who had undergone surgery between 2001 and 2012 at 8 institutions. We collected 191 samples which included both BM samples alone and pair-matched samples (primary and BM). Hematoxylin and eosin (H&amp;E) stained slides were evaluated for stromal TILs in 10% increments (0–1%, &gt; 1– &lt; 10%, 10%–100%). IHC was performed using the following primary antibodies: CD4, CD8, Foxp3, PD-L1, PD-L2 and HLA class I. The cells positive for each antibody signal were counted automatically using ImageJ (NIH). The expression of PD-L1, PD-L2, and HLA on the tumor cells was scored as 0 (negative), 1 (weak or focal), or 2 (strong). Results: The median category of TILs of BM tumors was &gt; 1– &lt; 10% (range: 1–30%). Forty-six pair-matched samples were analyzed and the percentage of TILs in the primary breast tumor was significantly higher than that in BM samples (paired t-test, P &lt; 0.01). The number of CD4/CD8/Foxp3 positive cells in primary breast tumor was also significantly higher than in BM samples (paired t-test, P &lt; 0.05 for all categories). The negative/positive conversion occurred with the expression of HLA/PD-L2 on tumor cells (paired t-test, P = 0.03/0.06, respectively). No significant difference was observed in the overall survival (OS) of patients, from initial BM, based on high or low TILs (log-rank test, P = 0.131). However, triple negative breast cancer patients with low TILs had significantly shorter OS compared with patients with high TILs (log-rank test, P = 0.04). Conclusions: We demonstrated that TILs in BM tumors was significantly lower as compared to primary breast tumors. The expression of immune related molecules on tumor cells was converted in BM tumors.

BackgroundMitochondrial dysfunction contributes to many types of human disorders and cancer progression. Inner membrane mitochondrial protein (IMMT) plays an important role in the maintenance of mitochondrial structure and function. The aims of this study were to examine IMMT expression in lung adenocarcinoma and evaluate its correlation with clinicopathological parameters and patient prognosis.MethodsIMMT expression was immunohistochemically studied in 176 consecutive lung adenocarcinoma resection tissues, and its correlations with clinicopathological parameters were evaluated. Kaplan‐Meier survival analysis and Cox‐proportional hazards models were used to estimate the effect of IMMT expression on survival.ResultsHigh‐IMMT expression was detected in 84 of 176 (47.7%) lung adenocarcinomas. Levels were significantly correlated with advanced disease stage (stage II and III; P = 0.024), larger tumor size (>3 cm; P = 0.002), intratumoral vascular invasion (P < 0.001), and poorer adenocarcinoma patient prognosis (P = 0.002). Based on 176 patients with adenocarcinoma, multivariate analysis revealed that IMMT expression was an independent predictor of poorer survival (HR, 1.99; 95% confidence interval [CI], 1.06–3.74; P = 0.031). Further, treating A549 cells derived from lung adenocarcinoma, with IMMT siRNA resulted in significantly decreased proliferation.ConclusionHere, we first demonstrated that high‐IMMT expression is related to some clinicopathological parameters, and that its expression is an independent prognostic predictor of poorer survival in patients with lung adenocarcinoma; further studies are required to clarify the biological function of IMMT in lung adenocarcinoma. However, results suggest that this protein could be a novel prognostic indicator and therapeutic target.

ATRX loss is an independent predictor of poor survival in pancreatic neuroendocrine tumors

Although elevated neutrophil-to-lymphocyte ratio (NLR) has been associated with survival in some liver cancers, its prognostic relevance has not been studied in the context of combined hepatocellular cholangiocarcinoma CHCC-CC, a rare primary liver cancer. We investigated whether elevated NLR and a predominance of cholangiocarcinoma might predict poor prognosis in patients with resectable CHCC-CC. We retrospectively reviewed the clinicopathologic data of forty-two patients with CHCC-CC receiving hepatectomies at our hospital. We used Kaplan-Meier and Cox regression to analyze survival. Two-year disease-free survival and five-year overall survival rates were 43.2% and 32.9%, respectively. Univariate analyses showed that patients with NLR ≥3 had significantly worse 2-year DFS and 5-year OS rates. Univariant Kaplan-Meier survival analysis also associated these rates with a predominance in intrahepatic cholangiocarcinoma, AJCC tumor stage, pathological T stage and lymph-vascular invasion. However, our multivariate analysis found NLR ≥3 to be the only independent predictor of disease recurrence and poorer survival. Neutrophil-to-lymphocyte ratio was the most important independent predictor of poorer survival in patients with resectable CHCC-CC. Predominance of intrahepatic cholangiocarcinoma, advanced AJCC tumor stage and pathological T stage, and lymph-vascular invasion also may affect poor prognosis in patients receiving complete tumor resections.

Patients with active cancer experience ischemic stroke via cryptogenic mechanisms, with cancer-associated hypercoagulability being considered a major contributor to such strokes. Despite the remarkably shortened survival of these patients, the clinical predictors of survival are poorly understood. We determined the clinical factors including D-dimer levels serving as the predictors of overall survival in these patients. Retrospective study was conducted on cancer patients who visited our hospital for acute ischemic stroke with cryptogenic mechanisms from April 2012 through November 2014. Demographics, clinical characteristics, imaging and laboratory results including coagulation markers were collected, and overall survival was calculated from the patient medical records and a governmental national database. A high D-dimer level was defined as a D-dimer level exceeding the median value from the study population (>5.50μg/ml). A total of 93 patients were identified, with a median survival of 62days (interquartile range 32-223days). A high D-dimer level (p=0.004; hazard ratio [HR] 2.01, 95% confidence interval [CI] 1.26-3.21), systemic metastases (p=0.02; HR 2.08, 95% CI 1.11-3.90), and diabetes mellitus (p=0.03; HR 1.78, 95% CI 1.03-3.10) were identified as independent predictors of poor overall survival using multivariate Cox proportional hazard analysis. Most of the patients (87%) were primarily treated with low-molecular-weight heparin (dalteparin, n=49; enoxaparin, n=32). The type of low-molecular-weight heparin had no association with survival. A high D-dimer level, systemic metastases, and diabetes are independent predictors of poor survival in cancer patients with cryptogenic stroke.

Brain metastases (BrM) are the most common cancers in the brain and linked to poor prognosis. Given the high incidence and often limited treatment options, understanding the complexity of the BrM tumor microenvironment is crucial for the development of novel therapeutic strategies. We performed transcriptome-wide gene expression profiling combined with spatial immune cell profiling to characterize the tumor immune microenvironment in 95 patients with BrM from different primary tumors. We found that BrM from lung carcinoma and malignant melanoma showed overall higher immune cell infiltration as compared to BrM from breast carcinoma. RNA sequencing-based immune cell deconvolution revealed gene expression signatures indicative of tertiary lymphoid structures (TLS) in subsets of BrM, mostly from lung cancer and melanoma. This finding was corroborated by multiplex immunofluorescence staining of immune cells in BrM tissue sections. Detection of TLS signatures was more common in treatment-naïve BrM and associated with prolonged survival after BrM diagnosis in lung cancer patients. Our findings highlight the cellular diversity of the tumor immune microenvironment in BrM of different cancer types and suggest a role of TLS formation for BrM patient outcome.