Abstract
At the point of cancer diagnosis, molecular biomarkers aim to stratify patients into precise disease subtypes predictive of outcome independent of standard clinical parameters such as tumour ...stage. Although prognostic gene expression signatures have been derived for many cancer types, seldom have they been shown to improve therapeutic decision making, limiting their clinical use. While intra-tumour transcriptomic heterogeneity (RNA-ITH) has been shown to bias existing biomarkers, efforts to control for this biological parameter have not been considered in biomarker development. Here, we analyse multi-region RNA-seq and whole-exome data for 156 tumour regions from 48 TRACERx patients to explore RNA-ITH in NSCLC. We show that chromosomal instability is a major driver of RNA-ITH, through the generation of heterogeneous copy number events within tumours, and that existing prognostic gene expression signatures are vulnerable to sampling bias. To address this issue, we develop the Outcome Risk Associated Clonal Lung Expression (ORACLE) assay, comprised of genes expressed homogeneously within individual tumours but heterogeneously between patients. These genes are enriched in modules associated with cell proliferation, such as mitosis and nucleosome assembly, that are often selected for through copy number gain events occurring early in tumour evolution. Our approach to identify “clonal” transcriptomic biomarkers in NSCLC overcomes tumour sampling bias, improves survival risk forecasting over current clinicopathological risk factors, and may be generalised to other cancer types, whilst revealing the early evolutionary selection of high risk DNA copy number events driving poor clinical outcome.
Citation Format: Dhruva Biswas, Nicolai J. Birkbak, Rachel Rosenthal, Crispin T. Hiley, Emilia L. Lim, Krisztian Papp, Marcin Krzystanek, Dijana Djureinovic, Yin Wu, David A. Moore, Marcin Skrypski, Christopher Abbosh, Maise Al Bakir, Thomas B. Watkins, Selvaraju Veeriah, Gareth A. Wilson, Mariam Jamal-Hanjani, Arul M. Chinnaiyan, Patrick Micke, Jiri Bartek, Istvan Csabai, Zoltan Szallasi, Javier Herrero, Nicholas McGranahan, Charles Swanton. A clonal expression biomarker improves prognostic accuracy: TRACERx lung 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 2678.
Abstract
Proliferation is a key phenotypic feature of cancer, with higher rates associated with poorer clinical outcomes. Thus far, proliferation rates have been measured using pathological or ...experimental techniques on bulk tumor samples. However, tumors are heterogeneous compositions of distinct clones. Measuring the proliferation of individual clones has been unfeasible to date since proliferation and clonal diversity cannot easily be measured for the same set of cells, but it is potentially important as it may allow the identification of clones that develop more aggressive phenotypes (e.g., metastatic potential or treatment resistance). We have developed SPRINTER, a novel algorithm that uses single-cell whole-genome DNA sequencing (scDNA-seq) data to enable the accurate identification of actively replicating cells in both the S and G2 phases of the cell cycle and their assignment to distinct tumor clones, thus providing a proxy to estimate clone-specific proliferation rates. To evaluate SPRINTER’s accuracy, we generated a ground truth dataset of 8,844 diploid and tetraploid cancer cells by coupling scDNA-seq with 5-Ethynyl-2-deoxyuridine (EdU) labeling, and demonstrated that SPRINTER can accurately distinguish clone proliferation rates in contrast to previous approaches. We further generated a longitudinal, primary-metastasis matched dataset of 23,001 cancer cells obtained from 5 samples from the primary tumor and 5 samples from distinct metastases from a patient with non-small cell lung cancer, allowing us to integrate analyses of proliferation and cancer evolution through the metastatic disease course. We revealed widespread heterogeneity in clone proliferation rates both between and within samples, supported by multiple orthogonal analyses including Ki-67 pathology, nuclei microscopy imaging, and patient clinical imaging, with high proliferation seen in fast-growing metastatic lesions. We demonstrated an association between clones with high proliferation and increased metastatic potential, as well as increased shedding of circulating tumor DNA. We further illustrated SPRINTER’s broad applicability on previous datasets of 42,009 breast cancer cells and 19,905 ovarian cancer cells, revealing an association between high proliferation and increased rates of different genetic variants. In conclusion, SPRINTER infers the proliferation rates of distinct tumor clones from scDNA-seq data, allowing the identification of clones with potentially aggressive phenotypes, such as metastatic potential.
Citation Format: Olivia Lucas, Sophia Ward, Rija Zaidi, Abigail Bunkum, Alexander M. Frankell, David A. Moore, Mark S. Hill, Wing Kin Liu, Daniele Marinelli, Emilia L. Lim, Sonya Hessey, Cristina Naceur-Lombardelli, Andrew Rowan, Sukhveer Mann, Haoran Zhai, Michelle Dietzen, Boyue Ding, Gary Royle, Nicholas McGranahan, Mariam Jamal-Hanjani, Nnennaya Kanu, Charles Swanton, Simone Zaccaria. Linking proliferation rate to the evolution of single-cell primary and metastatic tumor clones abstract. In: Proceedings of the AACR Special Conference in Cancer Research: Translating Cancer Evolution and Data Science: The Next Frontier; 2023 Dec 3-6; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(3 Suppl_2):Abstract nr PR010.
The introduction of the International Association for the Study of Lung Cancer grading system has furthered interest in histopathological grading for risk stratification in lung adenocarcinoma. ...Complex morphology and high intratumoral heterogeneity present challenges to pathologists, prompting the development of artificial intelligence (AI) methods. Here we developed ANORAK (pyrAmid pooliNg crOss stReam Attention networK), encoding multiresolution inputs with an attention mechanism, to delineate growth patterns from hematoxylin and eosin-stained slides. In 1,372 lung adenocarcinomas across four independent cohorts, AI-based grading was prognostic of disease-free survival, and further assisted pathologists by consistently improving prognostication in stage I tumors. Tumors with discrepant patterns between AI and pathologists had notably higher intratumoral heterogeneity. Furthermore, ANORAK facilitates the morphological and spatial assessment of the acinar pattern, capturing acinus variations with pattern transition. Collectively, our AI method enabled the precision quantification and morphology investigation of growth patterns, reflecting intratumoral histological transitions in lung adenocarcinoma.
INTRODUCTION
Acute myeloid leukemia (AML) is a hematopoietic malignancy that comprises almost 25% of pediatricleukemiasand is characterized by genetic and epigenetic alterations that lead to ...impairment of differentiation and clonal expansion. Despite intensive chemotherapy, more than half of children with AML either fail to achieve a complete remission (CR) or relapse after initial response.As such, the availability of a predictor of treatment failure at diagnosis may allow early institution of alternative therapies to improve outcome. MicroRNAs (miRNAs), a class of small non-coding RNAs that regulate the translation of their target mRNAs, are frequentlydysregulatedin cancers, and thus may serve as robust biomarkers of patient outcome.
METHODS
To identifymiRNAbiomarkers associated with treatment failure and candidate therapeutic targets, we performed a comprehensive sequence-based characterization of the pediatric AMLmiRNAexpression landscape usingmiRNA-seqdata from 637 primary samples. AmiRNAexpression-based model for EFS separating patients into low, intermediate, andhigh riskgroups was produced using penalized Cox regression. The model was designed usingmiRNAexpression data obtained from a training cohort, which consisted of two-thirds of our study cohort (n=425), and then tested on the remaining one-third of our study cohort (n=212). The training and test cohorts were derived by random selection.
RESULTS
A 36-miRNA EFS predictive model was generated. This model was comprised of16miRNAsthat were over-expressed and 20 that were under-expressed in patients who experienced an event (death, relapse or IF). Among the 36miRNAtranscripts were miR-155, miR-335, miR-139 and miR-375, which have been previously individually associated with survival in pediatric AML.
To demonstrate the potential clinical utility of the model, we determined 2 miRNAmodel score thresholds in the training cohort to separate patients into low, intermediate and high miRNAmodel score risk groups. The miRNAmodel score groupings were significantly associated with EFS in both the training cohort and test cohort (Figure 1A; P<0.001). Specifically, within the training cohort, the model identified 108 (25%) patients as high risk (5-year EFS: 7.36%; HR: 2.83, P<0.001), and 106 (25%) patients as low risk (5-year EFS: 81.4%; HR: 0.32, P<0.001). The training and test cohorts were combined for further subclass evaluation. In the combined cohort, multi-variate Cox regression analysis, which includedmiRNA expression risk status and conventional cytogenetic and molecular (CM) risk groups, demonstrated that themiRNA risk classification remains an independent predictor of high risk (HR: 2.44, P<0.001) and low risk (HR: 0.34, P<0.001).
Furthermore, to demonstrate the strength of our predictive model, we evaluated the clinical significance of the model in each of the low, standard and high CM risk cohorts. In this analysis, our model was capable of further stratifying patients in each of the 3 CM risk cohorts into 3 distinct miRNAmodel score-based risk categories (Figure 1B, P<0.001). Of particular interest is the ability of the model to identify patients with poorer outcomes in the CM low risk cohort and those with more favorable outcomes in the CMhigh risk cohort.
CONCLUSIONS
We present amiRNAexpression-based predictor of outcome in pediatric AML, comprised of 36miRNAtranscripts. Our predictive model was developed and tested on a large cohort of primary patient samples (n=637), and demonstrated that diagnosticmiRNAexpression profiles can identify risk groups in patients independent of other CM risk factors. Moreover, this model is applicable to RNA from samples that are routinely obtained for diagnosis, and thus has the potential to impact clinical practice.
Display omitted
No relevant conflicts of interest to declare.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
AML is a malignancy of hematopoietic progenitor cells characterized by differentiation arrest leading to hematopoietic insufficiency secondary to the accumulation of hematopoietic progenitors. While ...much is known about the karyotypic and immunophenotypic makeup of AML, comprehensive genomic data is still emerging. The Therapeutically Applicable Research to Generate Effective Treatments (TARGET) AML project, a cooperative effort between the National Cancer Institute (NCI) and the Children's Oncology Group (COG), has sought to deepen the knowledge-base of genomic alterations that are present in childhood cancers.
Sequence data from 918 children with AML was available, including 227 whole genome, 145 transcriptome, and 691 targeted exome capture. We identified 164 patients diagnosed with AML before the age of 2 years, which contributed to our study. From these data, the presence of fusions, somatic SNVs, and copy number alterations (for those with whole genome data) were assessed and evaluated in a pathway-centric analysis (TargetMine, http://targetmine.mizuguchilab.org). The OncoPrinter function (www.cbioportal.org) was used to visualize the mutations.
Of the 164 patients <2 years of age, all but 3 (98%) had at least one detectable genomic alteration. In contrast to older patients, structural alterations were the predominant variants in this age group with 142 patients (87%) harboring translocations. The most prevalent fusion involved KMT2A (MLL) with 14 different partners in 45%, NUP98 fusions with 6 different fusion partners in 8%, and CBFA2T3-GLIS2 fusions in 7%. Other notable fusions include CBFB-MYH11 (6%), RUNX1 with 3 fusion partners (3%), RBM15-MKL1 (3%), and translocations involving ETV6 and KAT6A (2% each). In addition to fusions, copy number variations (CNVs) including deletions, duplications and copy-neutral LOH were observed in 33 patients (20%). Regions of involvement of these CNVs ranged from segmental intragenic CNVs to larger chromosomal gains, trisomy 19 or 21 among the most frequent (7% and 6%, respectively). Intragenic exon 8 and 9 deletion of the CBLgene was the most common intragenic abnormality observed in 5 patients (3%).
Sequence variants including Single Nucleotide variants (SNVs) as well as insertions/deletions were detected in 102 patients (62%). Activating mutations of the RAS/MAPK signaling pathway were the predominant variants with prevalence of 46% (NRAS (18%), KRAS (14%), FLT3 (7%), KIT (3%), CBL and PTPN11 (5% each). Overall, 75 patients had a concomitant structural alteration and a RAS/MAPK variant, suggesting cooperation between these two class of variants. Additional sequence variants included mutations in WT1 (3%), CREBBP (2%), TET2 (2%), and GATA1 (2%).
Known fusions involving KMT2A were seen in 76 patients (46%). Where genomic data were available, the site of breakpoint varied significantly within KMT2A (16 different breakpoints) ranging from 5’ ATH motif to the 3’ SET domain with 6/16 fusions clustering in the ring finger domain of the gene. KMT2A-SEPT6 fusion was seen in 6 patients (8%) and was uniquely detected in those <2 years of age (P<0.0001). In those with KMT2A fusions, additional CNVs involving KMT2Awere seen in 3 patients. Concomitant mutations in RAS activating signaling pathway were seen in 43 cases (57%).
To determine whether these alterations were enriched in any functional pathways, genes involved in fusions/SNVs were uploaded in the TargetMine data warehouse and analyzed for pathway enrichment. Strikingly, 67% of all known fusions in this cohort contained a gene from the Chromatin Modifier classification including CREBBP, EP300, KAT6A, KDM5A, KMT2A, and NCOA2 and an additional 15% of fusion partners could be categorized as genes involved in transcriptional regulation (CBFB, ETV6, RUNX1, and PDGFRB). Likewise, 75% of all SNVs were enriched in pathways involving transcriptional regulation by TP53, (CREBBP, FLT3, GATA1, KIT, KMT2A, KRAS, NRAS, and WT1).
Although sequence variants are the most common variants in older AML, structural alterations dominate the genomic landscape in children with AML <2 years of age with a prevalence of >90% (translocation or CNV). Cooperating mutations of the RAS/MAPK pathway appear to be the dominant variants that in cooperation with the chromatin modifying alterations might contribute to the leukemic initiation and progression.
Display omitted
No relevant conflicts of interest to declare.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
The age distribution of acute myeloid leukemia is unusual among malignancies, with onset spanning from early infancy until past the 9th decade. Despite similar histology, cytogenetic abnormalities ...and recent identification of somatic mutations (e.g., DNMT3A mutations) have highlighted differences in the events driving adult compared to childhood de novo AML. However, the full extent of these differences remains unknown and is likely to have relevance to treatment approaches. The TARGET AML initiative is an effort of the Children's Oncology Group (COG) and the National Cancer Institute to comprehensively characterize the molecular abnormalities of pediatric AML. The dataset comprises 1) whole genome sequencing (WGS) of AML and matched remission bone marrow in 197 cases, 2) mRNA transcriptome sequencing of 284 cases, 3) miRNA sequencing of 692 cases, 4) methylation array data on 289 cases, and 5) targeted capture sequencing of 174 candidate genes identified from WGS in 800 diagnostic samples, including 182 with WGS (Figure 1). The majority of patients (93%) studied were uniformly treated on COG study AAML0531 or its pilot safety precursor study, AAML03P1. Relapsed specimen data (not shown) are available for a subset of these cases. All patient samples were obtained by written consent upon enrollment in the clinical trial.
Consistent with adult studies, we identified a relatively low mutational burden, with 2206 somatic tier 1 mutations resulting in a coding change in 1682 genes (median 6 per patient) from the WGS discovery data. We successfully verified 70-90% of variant calls by secondary methods. Also as with adult data, there were relatively few recurrently mutated genes, with fewer than 40 genes altered in >2% of samples. However, there were marked differences in somatic mutation frequencies in comparison to adult TCGA data, both by raw frequency and after adjustment for cytogenetic subtypes present among the two cohorts (Figure 2). Mutations in TP53, NPM1, IDH1, IDH2, TET2 and DNMT3A are more frequent in adult compared to pediatric disease; in contrast, mutations in NRAS, KRAS, WT1, FLT3, PTPN11, GATA2, ASXL2, MYC, SETD2, EZH2 and IKZF1 appear more common in pediatric AML. Mutations of several genes, including CEBPA, ASXL2 and KRAS are not only more common in pediatric AML, they show peak prevalence within specific pediatric age groups. In addition, several genes, including FLT3, WT1, and KIT show significant differences in mutational hotspots compared to adults.
Pediatric-adult differences in AML were not limited to somatic gene mutations, but extended to focal and chromosomal copy number alterations (CNA), translocations, miRNA expression, and methylation-induced gene silencing. We identified recurrent focal CNAs in multiple regions not reported in adult AML including 15 heterozygous focal deletions impacting ELF1, an ETS-family transcriptional regulator of hematopoiesis and leukemia driver as well as deletions of the splicing regulator MBNL1 in 10 cases, 8 of which co-occurred with focal deletions of the hematopoietic transcriptional regulator, ZEB2. De novo assembly of mRNA sequencing data identified fusion transcripts in 63% of cases compared to 45% of TCGA LAML. In addition to cytogenetically evident fusions with well-described enrichment for MLL translocations in pediatrics, we identified 29 diagnostic samples (10%) with nucleoporin family fusions (NUP98 with NSD1, KDM5A, PHF23, HOXD13, HMGB3, BRWD3, and CLINT; NUP214 with DEK and SET), CBFA2T3-GLIS2 fusions in 5, and rare fusions of ETS transcription factor genes (FUS-FEV, ETV6-INO8D). Comparison of miRNA expression patterns between adult and pediatric specimens similarly showed marked differences in expression of key regulatory miRNAs including let-7 family members. Finally, analysis of mRNA expression and DNA methylation for the identification of epigenetically silenced genes suggested that, although specific events favor silencing in adults or children, an overall pattern of gene silencing was more prevalent in pediatric compared to adult cases.
This work extends our understanding of the heterogeneity of AML, demonstrates fundamental differences in the biology of pediatric- and adult-onset disease, and suggests important age-related differences within "pediatric" AML. This rich dataset should provide a foundation for the establishment of biologically-guided treatment in children with AML.
Display omitted Display omitted
No relevant conflicts of interest to declare.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Acute myeloid leukemia (AML) carries a poor prognosis across age groups. In children, AML has become the leading cause of leukemia mortality, with only 60% of cases securing long-term remission. In ...adults, outcomes are far worse, with 5 year survival approaching 24%. The mutational and transcriptional characterization of AML1has not yet translated into improved outcomes for most patients.
The TARGET AML project is an effort of Children's Oncology Group (COG) and the National Cancer Institute to characterize molecular abnormalities in pediatric AML. 197 cases were selected for whole genome sequencing (WGS) of diagnostic specimens, 284 cases for mRNA sequencing, 289 cases for DNA methylation arrays, and 721 cases for targeted sequencing (182 assayed by WGS). Most patients (93%) were uniformly treated on COG study AAML0531 or AAML03P1. The Cancer Genome Atlas (TCGA) AML project1characterized 177 comparable adult AMLs with identical assays.
DNA methylation changes radically during differentiation of blood cells2, and recurrent pre-leukemic mutations in adult AML3affect DNA methylation and chromatin modifiers. We thus investigated whether differences in cell-of-origin, immune signalling, and regulatory aberrations were captured by focal or regional differences in DNA methylation, within or between adult and pediatric AML patients.
In cytogenetically similar TARGET and TCGA AML cases, striking differences in DNA methylation emerge (fig. 1). Pediatric FLT3-mutant AMLs dominate a cluster with normal-progenitor-like DNA methylation. Mutant DNMT3A, RUNX1, and TP53, which selectively favor preleukemic hematopoietic stem cells3,4,5 (HSCs), are common in adult AML, rare in pediatric AML, and tend towards HSC-like hypermethylation. Transcriptional & epigenetic signatures of the cell of origin persist even after leukemic transformation6. Thus we sought to identify the most likely cell of origin for each case. Previous studies of mRNA7 and DNA methylation8 differences in HSCs and progenitor cells (HSPCs), leukemic stem cells (LSCs), and AML blasts allowed us to model these differences in TCGA and TARGET AMLs. RNAseq results revealed many LSC-like cases with aberrant β-catenin signaling and TP53 regulation, distinct from blasts and normal HSPCs (fig. 2a). DNA methylation segregated cases resembling granulocyte/monocyte progenitors (GMPs) from those resembling other HSPC subsets (fig. 2b). DNMT3A mutants strongly associated with HSC/LSC-like mRNA expression, as did most MLL-rearranged AMLs. Nearly all TP53 and RUNX1 mutants presented LSC-like mRNA expression and retained HSC-like methylomes. These results suggest that decades of selective HSC attrition enable cooperating adult-specific mutations to initiate leukemia, while the timescales in pediatric AML favor fusion genes capable of transforming progenitors as well as HSCs.
With matched mRNA expression & DNA methylation data from 256 TARGET cases and 156 TCGA cases, we found over 100 genes where DNA methylation accompanied loss of transcription (silencing) in AML but not in normal HSPCs (fig. 3a). Many such genes lie in regions affected by recurrent copy number aberrations, most notably chromosome arms 5q and 19q. Recurrently mutated or deleted genes such as DNMT3A, TET2, SPRY4, and CDKN2A/B are silenced, some mutually exclusively with mutations or CNV. Functional enrichment analyses of silenced genes with DAVID9revealed 4 clusters: NK-cell signaling, innate immune response regulation, transcriptional regulation, and (on chromosome 19q) zinc finger genes involved in Toll-like receptor signaling. Some silencing co-occurs with specific molecular features, but no event was perfectly predicted by any molecular or cytogenetic feature (fig. 3b).
Drug-gene interaction mining with DGIDb10 suggests silencing may inform treatment. Silencing of the mitotic checkpoint gene CHFR may confer sensitivity to microtubule inhibitors11, silencing of MGMT suggests greater benefit from alkylating agents12, and demethylating agents may benefit cases with silenced immune response13. Biomarker driven clinical trials will be needed to evaluate these and other markers in pediatric and adult AML, but evidence of independent genetic and epigenetic evolution in AML14supports their continued investigation.
This work is dedicated to the late Robert J. Arceci, without whom none of this would have been possible.
Display omitted Display omitted Display omitted
No relevant conflicts of interest to declare.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Abstract
Background: The adaptive immune system plays an important role in tumor evolution. A key source of cytotoxic T cell response in cancer is neoantigens, cancer cell specific mutations that ...result in mutant peptides that elicit a T cell mediated immune response. However, a mutation can only engender a neoantigen if the associated mutant peptide is presented to T cells by HLA molecules, and, as such, transcriptional repression or loss of the HLA genes can have important implications for immune evasion.
Methods: We elucidate allele specific genomic and transcriptomic disruption to class I and II HLA genes. In whole exome sequencing (WES) data, our new tool (LOHHLA2.0) assesses loss of heterozygosity (LOH) status and somatic mutations. In RNA sequencing (RNAseq) data, LOHHLA2.0 quantifies allele specific expression and transcriptional repression referencing matched tumor adjacent normal samples. We applied LOHHLA2.0 to the TRACERx421 dataset, including 1554 WES tumor regions from 421 patients and 941 RNAseq regions from 357 tumor patients, 96 of which also had RNAseq from tumor adjacent normal samples.
Results: We find that our method is more accurate than existing tools (RSEM) at calling gene level expression, e.g. in HLA-DPB1, RSEM under-calls HLA expression by a factor of two. 36% of TRACERx421 primary tumors harbored HLA LOH of at least one class I HLA gene, validating our previous findings in the TRACERx100 cohort. Strikingly, we found that 74% (71/96) of primary tumors with matched tumor adjacent normal tissue exhibited transcriptional repression of one or more class I HLA alleles, and 81% (78/96) exhibited class II allele transcriptional repression. Class I HLA transcriptional repression was more likely to occur subclonally than LOH. In a subset of tumors, we observed convergence upon disruption of the same allele through alternative mechanisms; with genomic loss in one tumor region and transcriptional repression in another region of the same tumor. Across the tumor regions, we found a concordance between HLA expression and immune infiltrate levels, with immune hot tumors exhibiting higher HLA class I expression.
Conclusions: In this study, we find significant disruption to class I and II HLA expression adding to the diversity of immune evasion processes evident in early stage treatment naive NSCLC.
Citation Format: Clare Puttick, Oriol Pich, Michelle Leung, Carlos Martinez-Ruiz, Robert Bentham, Rachel Rosenthal, Sonya Hessey, James R. Black, Emilia L. Lim, Katey Enfield, Emma Colliver, Krijn Dijkstra, Crispin T. Hiley, Takahiro Karasaki, Ariana Huebner, Maise Al Bakir, Thomas B. Watkins, Alexander M. Frankell, Simone Zaccaria, Mariam Jamal-Hanjani, Nicholas McGranahan, Charles Swanton. Pervasive allele specific transcriptional repression of the class I and II HLA genes in TRACERx non-small cell lung 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 1394.
PDF
