Paired Primary Tumors Research Articles

Epidermal growth factor receptor mutations and brain metastases in non-small cell lung cancer.

Studies have revealed that non-small cell lung cancer (NSCLC) with epidermal growth factor receptor (EGFR) mutations has a high incidence of brain metastases (BMs). However, the association between EGFR mutations and BMs remains unknown. This review summarizes detailed information about the incidence of BMs, clinical and imaging characteristics of BMs, brain surveillance strategies, influence of treatments on BMs, prognosis after BMs, and differences in EGFR mutations between paired primary tumors and BMs in EGFR-mutated NSCLC. The prognostic results demonstrate that patients with mutated EGFR have a higher incidence of BMs, EGFR tyrosine kinase inhibitors (EGFR-TKIs) (afatinib and osimertinib) delay the development of BMs, and patients with mutated EGFR with synchronous or early BMs have better overall survival after BMs than those with wild-type EGFR. The EGFR mutation status of BM sites is not always in accordance with the primary tumors, which indicates that there is heterogeneity in EGFR gene status between paired primary tumors and BMs. However, the EGFR gene status of the primary site can largely represent that of BM sites. Among patients developing synchronous BMs, patients with mutated EGFR are less likely to have central nervous system (CNS) symptoms than patients with wild-type EGFR. However, the possibility of neuro-symptoms is high in patients with metachronous BMs. Patients with mutated EGFR tend to have multiple BMs as compared to patients with wild-type EGFR. Regarding very early-stage NSCLC patients without neuro-symptoms, regular neuroimaging follow-up is not recommended. Among advanced NSCLC patients with EGFR mutation, liberal brain imaging follow-up in the first several years showed more advantages in terms of cost.

Programmed Death-Ligand 1 Expression in Lung Cancer and Paired Brain Metastases—a Single-Center Study in 190 Patients

IntroductionExpression of programmed death-ligand 1 (PD-L1) is the only routinely used tissue biomarker for predicting response to programmed cell death protein 1/PD-L1 inhibitors. It is to date unclear whether PD-L1 expression is preserved in brain metastases (BMs). MethodsIn this single-center, retrospective study, we evaluated PD-L1 expression using the SP263 assay in consecutively resected BMs of lung carcinomas and paired primary tumors, diagnosed from 2000 to 2015, with correlation to clinicopathological and molecular tumor and patient characteristics. ResultsPD-L1 tumor proportional score (TPS) could be evaluated on whole tissue slides in 191 BMs and 84 paired primary lung carcinomas. PD-L1 TPS was less than 1% in 113 of 191 (59.2%), 1% to 49% in 34 of 191 (17.8%), and greater than or equal to 50% in 44 of 191 (23.0%) BMs. TPS was concordant between BMs and paired primary lung carcinomas in most cases, with discordance regarding the clinically relevant cutoffs at 1% and 50% in 18 of 84 patients (21.4%). Four of 18 discordant cases had no shared mutations between the primary lung carcinoma and BM. Intratumoral heterogeneity, as assessed using tissue microarray cores, was only significant at the primary site (pWilcoxon signed rank = 0.002) with higher PD-L1 TPS at the infiltration front (mean = 40.4%, interquartile range: 0%–90%). Neither TPS greater than or equal to 1% nor TPS greater than or equal to 50% nor discordance between the primary lung carcinoma and BMs had prognostic significance regarding overall survival or BM-specific overall survival. ConclusionsPD-L1 expression was mostly concordant between primary lung carcinoma and its BM and between resections of BM and stereotactic biopsies, mirrored by tissue microarray cores. Differences in PD-L1 TPS existed primarily in cases with TPS greater than 10%, for which also human assessment tends to be most error prone.

Transient cell-in-cell formation underlies tumor relapse and resistance to immunotherapy.

Despite the remarkable successes of cancer immunotherapies, the majority of patients will experience only partial response followed by relapse of resistant tumors. While treatment resistance has frequently been attributed to clonal selection and immunoediting, comparisons of paired primary and relapsed tumors in melanoma and breast cancers indicate that they share the majority of clones. Here, we demonstrate in both mouse models and clinical human samples that tumor cells evade immunotherapy by generating unique transient cell-in-cell structures, which are resistant to killing by T cells and chemotherapies. While the outer cells in this cell-in-cell formation are often killed by reactive T cells, the inner cells remain intact and disseminate into single tumor cells once T cells are no longer present. This formation is mediated predominantly by IFNγ-activated T cells, which subsequently induce phosphorylation of the transcription factors signal transducer and activator of transcription 3 (STAT3) and early growth response-1 (EGR-1) in tumor cells. Indeed, inhibiting these factors prior to immunotherapy significantly improves its therapeutic efficacy. Overall, this work highlights a currently insurmountable limitation of immunotherapy and reveals a previously unknown resistance mechanism which enables tumor cells to survive immune-mediated killing without altering their immunogenicity.

Multi-scale characterization of tumor-draining lymph nodes in resectable lung cancer treated with neoadjuvant immune checkpoint inhibitors

SummaryBackgroundRegional lymph node (LN) acts as a pivotal organ for antitumor immunity. Paradoxically, tumor-draining LNs (TDLNs) are usually the first site of tumor metastasis in lung cancer. It is largely unknown about the association between the status of TDLNs and the response of primary tumor beds to immune checkpoint inhibitors (ICIs) in lung cancer patients. Also, studies characterizing the TDLNs in response to ICIs are scarce.MethodsWe characterized and compared the radiological, metabolic (18F-FDG) and pathologic responses between primary tumor beds and paired TDLNs (invaded/non-invaded) from 68 lung cancer patients who underwent neoadjuvant ICIs plus surgery. Additionally, we performed the spatial profiling of immune and non-immune cells within TDLNs using multiplexed immunofluorescence. Therapy responses (e.g., pathologic complete (pCR) or major response (MPR)) of primary lung tumor beds and paired TDLNs were investigated separately.FindingsWe observed that responses of TDLNs to ICIs markedly differ from their paired primary lung tumors regarding the radiological, metabolic (18F-FDG uptake), and pathologic alterations. Neoadjuvant ICIs therapy specifically decreased 18F-FDG-reflected metabolic activity in the primary tumor beds with pCR/MPR but not their TDLNs counterparts. Furthermore, the presence of invaded TDLNs was associated with poor pathologic responses in the matched primary tumor beds and predictive of rapid post-treatment tumor relapse. Spatial profiling demonstrated exclusion of T cell infiltrates within the metastatic lesions of invaded TDLNs, and diminished multiple immune and non-immune compositions in non-involved regions surrounding the metastatic lesions.InterpretationThese results provide the first clinically-relevant evidence demonstrating unique response patterns of TDLNs under ICIs treatment and revealing the underappreciated association of TDLNs status with the response of their paired primary tumors to ICIs in lung cancer.FundingThis work was supported by the National Natural Science Foundation of China (82072570 to F. Yao; 82002941 to B. Sun), the excellent talent program of Shanghai Chest Hospital (to F.Y), the Basic Foundation Program for Youth of Shanghai Chest Hospital (2021YNJCQ2 to H.Yang), and the Innovative Research Team of High-level Local Universities in Shanghai (SHSMU-ZLCX20212302 to F. Yao).

Overexpression in metastatic breast cancer supports Syndecan-1 as a marker of invasiveness and poor prognosis

BackgroundMetastasis is the main cause of breast cancer (BC) mortality. Increasing evidence points to a role of syndecan-1 (CD138) expression as a prognostic marker involved in BC tissue and leptomeningeal metastasis. Aim of this study was to investigate and compare syndecan-1 tissue expression and localization in primary and secondary BC, focusing on brain metastases.MethodsSyndecan-1 expression was determined by immunohistochemistry. Focal vs diffuse (< or > 50% of cancer cells, respectively) pattern of expression, cellular localization (cytoplasm vs membrane) and intensity of immunostaining on neoplastic cells were evaluated. Moreover, the extent and pattern of expression of syndecan-1 were compared between primary tumors and paired metastases and correlated with the tumor intrinsic subtype.ResultsA total of 23 cases, 10 with paired primary and metastatic tumor and 13 brain metastases, were evaluated. Syndecan-1 was expressed in both primary and metastatic BC. A diffuse cytoplasmic expression was observed in most primary BCs; by contrast, all metastatic lesions showed a membrane pattern of expression, suggesting a shift in cellular localization of syndecan-1 during the metastatic process. Concerning the extent of expression, we observed in metastatic lesions, a trend of association between intrinsic subtypes and extent of positivity. In particular, both BC characterized by overexpression of HER2 and triple-negative tumors were correlated with a diffuse pattern of expression with a moderate to strong intensity.ConclusionA diffuse cytoplasmic expression was observed in most primary BCs; by contrast, all metastatic lesions showed a membrane pattern of expression, suggesting a shift in cellular localization of syndecan-1 during the metastatic process.

N-cadherin protects oral cancer cells from NK cell killing in the circulation by inducing NK cell functional exhaustion via the KLRG1 receptor

BackgroundCirculating tumor cells (CTCs) can survive in the circulation and return to primary tumors through a self-seeding process. However, the mechanisms underlying CTCs escape from natural killer (NK) cell-mediated immune...

Single-Cell and Bulk RNA Sequencing Reveal Malignant Epithelial Cell Heterogeneity and Prognosis Signatures in Gastric Carcinoma.

Gastric carcinoma (GC) heterogeneity represents a major barrier to accurate diagnosis and treatment. Here, we established a comprehensive single-cell transcriptional atlas to identify the cellular heterogeneity in malignant epithelial cells of GC using single-cell RNA sequencing (scRNA-seq). A total of 49,994 cells from nine patients with paired primary tumor and normal tissues were analyzed by multiple strategies. This study focused on the malignant epithelial cells, which were divided into three subtypes, including pit mucous cells, chief cells, and gastric and intestinal cells. The trajectory analysis results suggest that the differentiation of the three subtypes could be from the pit mucous cells to the chief cells and then to the gastric and intestinal cells. Lauren’s histopathology of GC might originate from various subtypes of malignant epithelial cells. The functional enrichment analysis results show that the three subtypes focused on different biological processes (BP) and pathways related to tumor development. In addition, we generated and validated the prognostic signatures for predicting the OS in GC patients by combining the scRNA-seq and bulk RNA sequencing (bulk RNA-seq) datasets. Overall, our study provides a resource for understanding the heterogeneity of GC that will contribute to accurate diagnosis and prognosis.

Lipopolysaccharide-binding protein expression is associated to the metastatic status of osteosarcoma patients

Osteosarcoma (OS) is a rare malignant primary bone tumours characterized by a high genetic and cell composition heterogeneity. Unfortunately, despite the use of drug combinations and the recent development of immunotherapies, the overall survival has not improved in the last four decades. Due to the key role of the tumour microenvironment in the pathogenesis of OS, a better understanding of its microenvironment is mandatory to develop new therapeutic approaches. From retrospective biological cohorts of OS, we analysed by immunohistochemistry the presence of lipopolysaccharide (LPS)-binding protein (LBP) in diagnostic biopsies with local disease and compared their level of infiltration to patients suffering from metastatic status. LBP is considered as a marker of LPS exposure and can indirectly reflect the presence of Gram-negative microbiota. LBP were detected in the cytoplasm of OS cells as well as in tumour-associated macrophage. Tumour samples of patients with local disease were significantly enriched in LBP compared to tumour tissues of patients with metastatic status. Lung metastatic tissues showed similar level of LBP compared to paired primary tumours. Overall, this study strongly suggests the presence of Gram-negative bacteria in OS tissues and demonstrated their significant differential level according the metastatic status. This tumour-associated microbiome may help in the conceptualisation of new therapeutic approach to trigger efficient therapeutic responses against cancer.

Morphological features of 52 cases of breast phyllodes tumours with local recurrence

Typical phyllodes tumours (PTs) of the breast are fibroepithelial neoplasms characterised histologically by stromal hypercellularity and leaf-like structures. However, morphological changes may be atypical in some cases, increasing the difficulty of diagnosis and the likelihood of misdiagnosis. To provide more morphological clues for pathological diagnosis of PTs, we retrospectively analysed 52 PT cases with typical morphological features after recurrence, and summarized the clinicopathological characteristics of the paired primary tumours. We found five special histological features in the primary tumours distinct from classic PTs, namely (1) PTs with epithelioid feature (three cases); (2) PTs with gland-rich feature (eight cases); (3) PTs with fibroadenoma-like feature (20 cases); (4) PTs with myxoid fibroadenoma-like feature (five cases); and (5) PTs with pseudohemangiomatoid stromal hyperplasia–like feature (four cases). All the features can exist independently, and a few cases displayed more than two distinctive features at the same time. In this cohort of recurrent PTs, all the primary tumours were absent of recognisable stromal hypercellularity and leaf-like structures that are the critical diagnostic criteria of PTs; however, they showed some other non-classic characteristics which may provide significant clues for the diagnosis of PTs. Particularly, tumours with epithelioid feature displayed high grade at earlier stages, tumours with fibroadenoma-like feature were most likely to be confused with classical fibroadenomas, and tumours with myxoid feature were prone to be neglected because of their hypocellularity.

EGFR, HER2, and HER3 protein expression in paired primary tumor and lymph node metastasis of colorectal cancer

Due to the difficulty in sampling of metastatic tumors, patient selection is commonly based on results of primary tumor samples when metastatic samples are not available. However, due to tumor heterogeneity, metastatic tumors may be different from primary tumors in their phenotypes. The aim of this study was to investigate the expression of EGFR, HER2, and HER3 between primary and lymph node metastatic lesions of colorectal cancer. Paired primary tumors and lymph node metastases from 79 patients with colorectal cancer were retrospectively collected and analyzed for EGFR, HER2, and HER3 expression. High EGFR, HER2, and HER3 expression (2+ and 3+) was found in 64.2%, 66.0%, and 85.0% of primary tumors, and 56.8%, 46.0%, and 76.0% of lymph node metastases, respectively. Correlation rates between primary and metastatic lesions were 67.1%, 63.3%, and 74.7% for EGFR, HER2, and HER3, respectively. Stage IV tumors (with distant metastasis) had higher correlation rates of HER2 expression compared to stage III tumors (without distant metastasis) (P = 0.050). Moderate correlation rates in EGFR, HER2, and HER3 expression were observed between primary and metastatic lesions of colorectal cancer. Tumor stage or existence of distant metastasis could serve as potential predictive markers for the correlation of HER2 expression between primary tumors and lymph node metastases of colorectal cancer.

Programmed Death Ligand-1 Combined Positive Score Concordance and Interrater Reliability in Primary Tumors and Synchronous Lymph Node Metastases in Resected Cases of p16+ Oropharyngeal Squamous Cell Carcinoma.

Pembrolizumab is used in patients with metastatic head and neck squamous cell carcinoma contingent upon the programmed death ligand-1 (PD-L1) combined positive score (CPS). To compare PD-L1 CPS scores derived from paired resected primary tumors (PTs) and lymph node metastases (LMs) in patients with p16+ oropharyngeal squamous cell carcinoma (OPSCC). We identified 38 resected p16+ OPSCCs for which paired PTs and LMs were available. PD-L1 immunohistochemistry using the SP263 antibody clone was done on both the PT and the LM. CPS scoring was performed by 4 observers, and data were analyzed at the CPS cut points of greater than or equal to 1 and 20 in regard to interobserver and interspecimen agreement. Overall agreement between consensus CPS scoring of PT and LM was seen in 76% of paired specimens (κ = 0.53). No specimen received a negative consensus score. Interobserver agreement for both PT and LM was fair to substantial (κ = 0.54 and 0.51, respectively) and was inferior to that seen in a prospective series of unselected head and neck squamous carcinoma cases evaluated at our institution (κ = 0.84). Given the high rates of interobserver and interspecimen variability, evaluation of additional material or by additional observers may be of value in performing CPS scoring in cases of p16+ OPSCC. This is particularly the case when a negative or low-positive result is being evaluated in a patient who is otherwise a good candidate for immunotherapy.

DNA methylation alterations across time and space in paediatric brain tumours

DNA methylation is increasingly used for tumour classification and has expanded upon the > 100 currently known brain tumour entities. A correct diagnosis is the basis for suitable treatment for patients with brain tumours, which is the leading cause of cancer-related death in children. DNA methylation profiling is required for diagnosis of certain tumours, and used clinically for paediatric brain tumours in several countries. We therefore evaluated if the methylation-based classification is robust in different locations of the same tumour, and determined how the methylation pattern changed over time to relapse. We sampled 3–7 spatially separated biopsies per patient, and collected samples from paired primary and relapse brain tumours from children. Altogether, 121 samples from 46 paediatric patients with brain tumours were profiled with EPIC methylation arrays. The methylation-based classification was mainly homogeneous for all included tumour types that were successfully classified, which is promising for clinical diagnostics. There were indications of multiple subclasses within tumours and switches in the relapse setting, but not confirmed as the classification scores were below the threshold. Site-specific methylation alterations did occur within the tumours and varied significantly between tumour types for the temporal samples, and as a trend in spatial samples. More alterations were present in high-grade tumours compared to low-grade, and significantly more alterations with longer relapse times. The alterations in the spatial and temporal samples were significantly depleted in CpG islands, exons and transcription start sites, while enriched in OpenSea and regions not affiliated with a gene, suggesting a random location of the alterations in less conserved regions. In conclusion, more DNA methylation changes accumulated over time and more alterations occurred in high-grade tumours. The alterations mainly occurred in regions without gene affiliation, and did not affect the methylation-based classification, which largely remained homogeneous in paediatric brain tumours.

CRISPR-mediated MECOM depletion retards tumor growth by reducing cancer stem cell properties in lung squamous cell carcinoma

Targeted therapy for lung squamous cell carcinoma (LUSC) remains a challenge due to the lack of robust targets. Here, we identified MECOM as a candidate of therapeutic target for LUSC by screening 38 genes that were commonly amplified in three pairs of primary tumors and patient-derived xenografts (PDXs) using a clustered regularly interspaced short palindromic repeats (CRISPR)-mediated approach. High MECOM expression levels were associated with poor prognosis. Forced expression of MECOM in LUSC cell lines promoted cancer stem cell (CSC) properties, and its knockout inhibited CSC phenotypes. Furthermore, systemic delivery of CRISPR-mediated MECOM depletion cassette using adenovirus with an adaptor, which is composed of a single-chain fragment variable (scFv) against epithelial cell adhesion molecules (EpCAM) fused to the ectodomain of coxsackievirus and adenovirus receptor, and a protector, which consists of the scFv connected to the hexon symmetry of the adenovirus, could specifically target subcutaneous and orthotopic LUSC and retard tumor growth. This study could provide a novel therapeutic strategy for LUSC with high efficacy and specificity.

Dysregulated Immune and Metabolic Microenvironment Is Associated with the Post-Operative Relapse in Stage I Non-Small Cell Lung Cancer.

Simple SummaryThe underlying mechanism of post-operative relapse of non-small cell lung cancer (NSCLC) remained poorly understood. This study highlights that both tumors and adjacent tissues from stage I NSCLC with relapse show a dysregulated immune and metabolic environment. Immune response shifts from an active state in primary tumors to a suppressive state in recurrent tumors. A model based on the enriched biological features in the primary tumors with relapse could effectively predict recurrence for stage I NSCLC. These results provide insights into the underpinning of the post-operative relapse and suggest that identifying NSCLC patients with a high risk of relapse could help the clinical decision of applying appropriate therapeutic interventions.The underlying mechanism of post-operative relapse of non-small cell lung cancer (NSCLC) remains poorly understood. We enrolled 57 stage I NSCLC patients with or without relapse and performed whole-exome sequencing (WES) and RNA sequencing (RNA-seq) on available primary and recurrent tumors, as well as on matched tumor-adjacent tissues (TATs). The WES analysis revealed that primary tumors from patients with relapse were enriched with USH2A mutation and 2q31.1 amplification. RNA-seq data showed that the relapse risk was associated with aberrant immune response and metabolism in the microenvironment of primary lesions. TATs from the patients with relapse showed an immunosuppression state. Moreover, recurrent lesions exhibited downregulated immune response compared with their paired primary tumors. Genomic and transcriptomic features were further subjected to build a prediction model classifying patients into groups with different relapse risks. We show that the recurrence risk of stage I NSCLC could be ascribed to the altered immune and metabolic microenvironment. TATs might be affected by cancer cells and facilitate the invasion of tumors. The immune microenvironment in the recurrent lesions is suppressed. Patients with a high risk of relapse need active post-operative intervention.

Abstract 760: Concordance of DNA copy number profiles between primary and metastatic colorectal carcinoma

Abstract The most common form of genomic instability in sporadic colorectal cancer (CRC) is chromosomal instability, a phenotype characterized by a high frequency of DNA copy number gain (CNG) and loss (CNL). Given the fundamental role of chromosomal instability in the pathogenesis of CRC, this study aimed to investigate the concordance of DNA copy number variants (CNV) between primary and metastatic CRC to better understand how the genomic structure differs during tumor evolution. In this study, the SurePrint G3 Human Genome CGH+SNP 2x400K Microarray (Agilent, SC, California, USA) was used for genome-wide screening of CNV in 9 pairs of primary tumors and metachronous liver metastases obtained during surgical resection from patients diagnosed with advanced CRC. Data from the Microarray-based Comparative Genomic Hybridization was analyzed by COSMIC Cancer Gene Census (CGC) and an independent non-CGC panel to map hotspot regions harboring genes causally implicated in cancer in the course of the metastatic cascade. Also, unique CNV in metastases were further investigated to identify loci advantageous for tumor progression and those presumably associated with therapy resistance. Overall, CNG observed in primary tumors were highly concordant with those in liver metastases. The CNG hotspot regions for both tissues represented chr7 (p22.3-p13), chr8 (q21.13-q24.3), chr13 (q12.11-q13.3), chr20 (q11.1-q13.33) and chrX (q11.1-q28) loci, while the most frequent CNL was at chr8 (p21.3-p12) locus. All identified loci underwent CNV events with a higher frequency in metastases (67-100%) than in primary tumors (55%). The identified genomic loci with CNG encode DNA repair genes (such as PMS2, NBN, BRCA2, RAD54B, RAD21, RPA3,4, and TREX2), homeobox genes (HOXA9,11,13), genes encoding centrosome-associated proteins (CETN2, MPLKIP), proto-oncogenes (MYC), and also tumor suppressor (ATRX, ASXL1). Additionally, an analysis of unique CNV within distinct tissues revealed CNG at Chr6 (p25.1-p22.3) in 89% of metastatic samples and absent in all the primary tumor lesions. This particular genomic region contains proto-oncogene DEK. Although the analysis is still underway, we have identified several shared and unique regions of altered copy number in both, primary tumors and metastases. We conclude that despite the high inter- and intra-tumoral genomic heterogeneity, metachronous metastases “inherit” a similar profile of DNA CNV from their respective primary tumor tissues. We intend to further extend our studied group of patients and perform integrative analyses of CNV with gene expression. This study was supported by the Charles University (UNCE/MED/006) and the Grant Agency of the Czech Republic (21-04607X, 21-27902S) Citation Format: Michal Kroupa, Kristyna Tomasova, Katerina Saskova, Jana Drabova, Josef Horak, Ludmila Vodickova, Jachym Rosendorf, Vaclav Liska, Pavel Vodicka. Concordance of DNA copy number profiles between primary and metastatic colorectal carcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 760.

Abstract 51: Multiomic profiling of patient-derived xenografts established from patients with malignant pleural mesothelioma proposes pathways associated with poor prognosis

Abstract Background: Despite recent treatment advances, malignant pleural mesothelioma (MPM) is an aggressive, recalcitrant malignancy. Currently, histologic subtype (epithelioid/non-epithelioid/biphasic) is the primary prognostic factor; other potential biomarkers to guide therapeutic strategies remain elusive. Even with multimodality therapies, recurrence is high in early-stage disease. In the unresectable/metastatic setting, there are only two FDA approved regimens, both in the first line setting: cisplatin/pemetrexed and ipilimumab/nivolumab. Unfortunately, most who respond to first line treatment experience disease progression within a year. A few established MPM cell lines, with inherent limitations, provide minimal preclinical insight. The relative lack of model systems that accurately reflect MPM tumorigenesis is a barrier to therapeutic and diagnostic advances in MPM. Methods: We developed a diverse library of 22 extensively annotated patient-derived xenograft (PDX) models from 22 patients with MPM. Multi-omic analyses including, targeted tumor next-generation sequencing by MSK-IMPACT, RNA-sequencing, and immunohistochemistry was performed. We deconvoluted the mutational landscapes, global expression profiles, and molecular subtypes of these MPM models and further compared the PDXs to MPM clinical specimens, including matched PDX and primary tumor pairs. Results: The mutational landscapes of PDX models strongly correlated with paired tumor samples. There were some differences in CDKN2A/B mutations and relative enrichment of NF2 with fewer BAP1 alterations, the significance of which is being investigated. When compared by histological subtype, we observed an upregulation of genes involved in NOTCH and EMT signaling in the epithelioid models. Models derived from patients with shorter overall survival or poor response to platinum doublet had higher expression of WNT/β-catenin signaling, hedgehog pathway, and epithelial-mesenchymal transition signaling as well as downregulation of immune-activation pathways, including type I and II interferon signaling and inflammatory response pathways. Conclusions: This library of MPM PDXs, the largest to date, effectively mimics human disease and provides unprecedented insight into the genomic, transcriptomic, and protein landscape of MPM. These PDX models will inform future clinical investigations and provide an important new preclinical resource. Citation Format: Michael Offin, Jennifer L. Sauter, Jacklyn V. Egger, Elisa deStanchina, John T. Poirier, Marjorie G. Zauderer, Charles Rudin, Triparna Sen. Multiomic profiling of patient-derived xenografts established from patients with malignant pleural mesothelioma proposes pathways associated with poor prognosis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 51.

Abstract 1596: Genomic heterogeneity as a barrier to precision oncology in urothelial cancer

Abstract Background: Precision oncology relies on the accurate molecular characterization of individual cancer patients at the time of treatment initiation. However, tumor molecular profiles are not static, and cancers continually evolve because of ongoing mutagenesis and clonal selection, with actionable genomic alterations potentially gained or lost during disease progression. Methods: To define the concordance of potentially actionable genomic alterations between primary and metastatic disease sites in patients with urothelial cancer and to identify drivers of metastatic disease progression, we performed an integrated analysis of clinical and genomic data from 2,199 urothelial carcinoma patients (2,732 tumor and 331 plasma cfDNA samples) prospectively profiled at Memorial Sloan Kettering Cancer Center (MSKCC) from 2014 to 2021. Paired primary and metastatic tumor samples from individual patients were sequenced with MSK-IMPACT and/or whole-exome sequencing (WES). Plasma samples were analyzed using the MSK-ACCESS cfDNA platform. Results: Among potentially actionable mutations (defined as OncoKB level 1-4), ERBB2 and ARID1A were associated with higher tumor grade and stage (p-value &amp;lt;0.001, p-value = 0.03 respectively). WES analysis of primary and metastatic tumor sites was consistent with early branched evolution with on average 42% of mutations shared between disease sites. Among chromatin-modifying genes, ARID1A has a discordant rate of 15.8%. In the 6 discordant cases, ARID1A mutation was seen exclusively in the metastatic tumor samples. This suggests a role for ARID1A mutations in metastatic disease progression. While known and likely oncogenic mutations were more likely to be concordant than variants of unknown functional significance, we observed a high degree of mutational discordance among potentially actionable genomic alterations with 23% discordant between primary and metastatic disease sites. 24% of mutations were exclusively identified by plasma cell-free DNA sequencing. Actionable mutations in cell-free DNA were more likely to be concordant with metastatic tumor biopsies than tumor tissue collected from the primary tumor site. Conclusions: In sum, our analysis of patient-matched primary and metastatic urothelial carcinomas revealed a high degree of lesion-to-lesion genomic heterogeneity that may be a barrier to precision oncology approaches for this disease. Our data also provide a rationale for the use of cell-free DNA sequencing to guide targeted therapy selection in patients with metastatic urothelial cancer. Citation Format: Ziyu Chen, Timothy N. Clinton, Soonbum Park, Andrew T. Lenis, Mark Donoghue, Jonathan E. Rosenberg, Bajorin F. Dean, Min Y. Teo, Bernard H. Bochner, Irina Ostrovnaya, Eugene J. Pietzak, Gopakumar V. Iyer, Sizhi Gao, Wenhuo Hu, Michael M. Shen, Michael F. Berger, Hikmat A. Al-Ahmadie, Corinne T. Abate-Shen, David B. Solit. Genomic heterogeneity as a barrier to precision oncology in urothelial cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1596.

Abstract 758: Differential gene expression profiling of matched primary and metastatic triple negative breast cancer

Abstract Background: Triple negative breast cancer (TNBC) is a molecularly complex and heterogeneous subtype of breast cancer with distinct biological features and clinical behavior. TNBC is associated with increased risk of metastasis and recurrence. To identify the underlying transcriptional changes and differentially expressed genes in metastatic progression of TNBC, we compared the transcriptomic profile of primary and matched metastatic tumors using massively parallel RNA sequencing. Methods: We performed gene expression profiling using paraffin embedded formalin fixed tissues of 38 TNBC patients obtained from University Hospital Zurich, Switzerland and Stavanger University Hospital, Norway. Of these, 18 were primary TNBC tumors that did not metastasize; 20 patients metastasized to lymph nodes or to distant organs (16 paired primary and metastatic tumors). Absolute fold change of ≥ 2 and a p-value &amp;lt;0.05 was used as the cutoff to identify significant differentially expressed genes (DEGs). Functional analysis was performed on DEGs. We also compared gene-expression profiles for hypoxia related- and EMT-related genes between our comparison groups. CIBERSORT analysis was used to estimate relative fractions of 22 immune cell types in each RNA-seq sample. Results: We identified 2442 DEGs between primary TNBC tumors and corresponding metastatic lesions and 1561 DEGs between metastasis-free and metastatic primary tumors. Of the 2442 DEGs between primary and metastatic tumors, we found 579 genes to be upregulated and 1863 genes to be downregulated in primary tumors compared to matched metastatic tumors. A total of 530 genes were found to be upregulated and 1031 genes were downregulated in metastasis-free compared to metastatic primary TNBC tumors. Of the 27 gene hypoxia signature, only 1 gene (MPRS17) was significantly different between primary and corresponding metastatic tumors and 3 hypoxia-related genes (KCTD11, FOSL1 and SDC1) showed significant differences between primary tumors with and without metastasis. We found expression for 1 EMT related gene (TWIST1) to be different between primary and matched metastatic tumors whereas 4 EMT related genes (CDH2, VIM, SNAI1 and FOXC2) were found to be significantly different between primary tumors that metastasized compared to primary tumors that remained metastasis-free. The relative fractions of 3 immune cell types, T cells CD4 memory resting, M2 macrophages and M0 macrophage were similar between the primary tumor and their matched metastatic lesions. Conclusions: Our results show that distinct patterns of gene expression exist between primary TNBCs and matched metastatic tumors. Further studies are warranted to explore whether these discrete expression profiles underlie or derive from disease status. Citation Format: Jaspreet Kaur. Differential gene expression profiling of matched primary and metastatic triple negative breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 758.

Abstract 1724: Pancancer evaluation of tumor infiltrating lymphocytes (TILs) and PD-L1 in SHIVA-01 trial patients with different biopsy sites and histological types

Abstract Background: Immunologic characterization of tumors and their microenvironments had been proposed based on the combination of PD-L1 expression and the presence of tumor-infiltrating lymphocytes (TILs). This study aims to investigate in metastatic samples from patients with cancer, the distribution of TILs and expression of PD-L1 using Combined Positive Score (CPS), Tumor Proportion Score (TPS), as well as TILs. These variables were further evaluated in paired primary and metastatic samples and correlated with clinical outcome. Patients and methods: 550 pan-cancer patients of SHIVA01 trial (NCT01771458) with available contributive FFPE sample from a metastatic biopsy and 111 paired primary and metastatic samples were evaluated for TILs. The 550 metastatic specimens were biopsies from different anatomical sites subdivided into 7 groups: liver biopsies (n=179; 33%), visceral organ biopsies (n=92; 17%), lung biopsies (n=89; 16%), lymph node biopsies (n= 88; 16%), cutaneous biopsies (n=53; 10%), soft tissue biopsies (n=48; 9%) and brain biopsy (n=1; 0.2%). PD-L1 expression was assessed by Immunohistochemistry using the 22C3 antibody clone (Merck &amp; Co) and quantified using two scores: CPS and TPS, in 494 metastatic tumors and in 77 paired primary and metastatic tumors patients with contributive immunohistochemistry. The correlations of TILs and PD-L1 expression with clinical outcomes are ongoing. Results: In metastatic samples, we found no difference in TILs distribution according to histological subtype, primary system or biopsy site with a median of 10% [range: 0%-70%]. A significant decrease in the median percentage of TILs was found in metastases in comparison to their paired primary lesions (20% [5%-60%] versus 10% [0%-40%], p&amp;lt;0.0001). PD-L1 expression was homogenous in all metastatic tumors independently of primary system or biopsy site (median TPS = 2% ; CPS = 0 n=218, CPS≥1 n=265 ; p=0.056). There was a strong association between TPS count and histological subtypes (p=0.015) that was not observed with CPS (p=0.23). In paired primary/metastatic samples, we did not observe any changes in the CPS and TPS scores (p&amp;gt;0.3 for both). Conclusion: We show that metastatic sites are less infiltrated with lymphocytes as compared to their paired primary lesions independently of the initial primary tumor site, histological type or biopsy site. PD-L1 expression was similar in paired primary and metastatic samples. Citation Format: Zakhia El Beaino, Célia Dupain, Grégoire Marret, Xavier Paoletti, Laëtitia Fuhrmann, Charlotte Martinat, Ivan Bièche, Christophe Le Tourneau, Maud Kamal, Anne Vincent-Salomon. Pancancer evaluation of tumor infiltrating lymphocytes (TILs) and PD-L1 in SHIVA-01 trial patients with different biopsy sites and histological types [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1724.

Metastasis Related Epithelial-Mesenchymal Transition Signature Predicts Prognosis and Response to Immunotherapy in Gastric Cancer.

BackgroundIncreasing evidence has revealed the effect of epithelial-mesenchymal transition (EMT) on tumor microenvironment and cancer treatment. However, an EMT-based signature to predict the prognosis and therapeutic effect in gastric cancer (GC) has rarely been established.MethodsDifferentially expressed genes (DEGs) between paired primary gastric and ovarian metastatic tumors were identified through comparative RNA-seq analysis, followed by the construction of metastasis-related EMT signature (MEMTS) based on DEGs and EMT gene set. Then, both The Cancer Genome Atlas (TCGA) cohort and the Asian Cancer Research Group (ACRG) cohort were analyzed to explore the potential association between MEMTS and prognosis in GC. Samsung Medical Center (SMC) cohort and two individual immunotherapy treatment cohorts, including Kim cohort and Hugo cohort, were utilized to evaluate the predictive value of MEMTS on the response to adjuvant therapy and immunotherapy, respectively. Finally, the potential association of MEMTS with tumor environment and immune escape mechanisms was investigated.ResultsHigh MEMTS predicted a poor prognosis in patients with GC. Patients with low MEMTS potentially gained more benefits from adjuvant chemoradiotherapy than those with high MEMTS. MEMTS reliably predicted the response to immunotherapy in GC (area under the curve = 0.896). MEMTS was significantly associated with cancer-associated fibroblasts and stromal score in the aspect of the tumor microenvironment.ConclusionMEMTS serves as a potential biomarker to predict the prognosis and response to adjuvant therapy and immunotherapy in GC. MEMTS-based evaluation of individual tumors enables personalized treatment for GC patients in the future.