PURPOSE:Germline likely pathogenic and pathogenic variants (LPV/PVs) in POT1 have been associated with an increased risk of various cancers including angiosarcoma, melanoma, glioma, thyroid cancer, and chronic lymphocytic leukemia (CLL). However, to date, most of the published data regarding POT1 PVs involve cohorts of patients selected for specific cancer types, or stem from commercial laboratory cohorts from patients selected for germline clinical testing, which may cause ascertainment biases. The objective was to identify germline POT1 variants in a pan-cancer cohort and describe the associated phenotypes. METHODS:Germline exome data available for 19,315 patients with cancer from the Oncology Research Information Exchange Network (ORIEN) were assessed for POT1 LPV/PV. Data regarding cancer diagnoses were obtained for those with and without POT1 variants. Associations were assessed. RESULTS:POT1 LPV/PVs were identified in 23 patients. The cancer types seen in more than one patient include CLL (n = 7), papillary thyroid cancer (PTC, n = 5), colorectal cancer (n = 3), lung cancer (n = 3), glioblastoma (n = 2), and neuroendocrine tumors (n = 2). Compared with POT1-negative patients, those with POT1 LPV/PVs were 5.5-fold more likely to be diagnosed with PTC (95% CI, 1.9 to 15.1; P = .004) and 16.6-fold more likely to be diagnosed with CLL (95% CI, 6.4 to 41.9; P < .001). Patients with POT1 LPV/PVs had a younger median age of first cancer diagnosis compared with POT1-negative patients (P = .008). CONCLUSION:To our knowledge, this study is the largest investigation of POT1 germline variants in a pan-cancer cohort. We identify and confirm specific associations with CLL and PTC in this cohort. Differences in results between analysis of largely unselected cohorts compared with clinical testing cohorts highlight the need to study gene-cancer associations in more unselected populations.

T-lineage acute lymphoblastic leukaemia (T-ALL) is a high-risk tumour1 that has eluded comprehensive genomic characterization, which is partly due to the high frequency of noncoding genomic alterations that result in oncogene deregulation2,3. Here we report an integrated analysis of genome and transcriptome sequencing of tumour and remission samples from more than 1,300 uniformly treated children with T-ALL, coupled with epigenomic and single-cell analyses of malignant and normal T cell precursors. This approach identified 15 subtypes with distinct genomic drivers, gene expression patterns, developmental states and outcomes. Analyses of chromatin topology revealed multiple mechanisms of enhancer deregulation that involve enhancers and genes in a subtype-specific manner, thereby demonstrating widespread involvement of the noncoding genome. We show that the immunophenotypically described, high-risk entity of early T cell precursor ALL is superseded by a broader category of 'early T cell precursor-like' leukaemia. This category has a variable immunophenotype and diverse genomic alterations of a core set of genes that encode regulators of hematopoietic stem cell development. Using multivariable outcome models, we show that genetic subtypes, driver and concomitant genetic alterations independently predict treatment failure and survival. These findings provide a roadmap for the classification, risk stratification and mechanistic understanding of this disease. Comprehensive genomic and transcriptomics analyses of more than 1,300 cases of childhood T-lineage acute lymphoblastic leukaemia identify 15 distinct subtypes that are associated with specific outcomes.

10015 Background: While cure rates for childhood acute lymphoblastic leukemia (ALL) exceed 90%, half of relapses arise in those originally classified with standard risk (SR) disease. Methods: We performed genome/transcriptome sequencing of diagnostic and germline samples of children with SR (n=1381) B-ALL or high-risk (HR) B-ALL with favorable cytogenetics ( ETV6: RUNX1 or double trisomy (DT) of chromosomes (chr) 4+10; n=115) to identify predictors of relapse. We used a case-control study to analyze 439 patients who relapsed and 1057 who remained in complete remission for > 5 years. Results: Genomic subtype was associated with relapse. Unbalanced ETV6:RUNX1 translocations were more common than balanced in relapse patients (OR=2.01, CI=1.25-3.20, P=0.002). Conversely, balanced TCF3:PBX1 translocations were more often associated with relapse than unbalanced in TCF3:PBX1 ALL (OR=0.11, CI=0.01-0.50, P=0.003). A striking finding was the high relapse rate in PAX5 altered ALL (57 of 116 cases (49%); OR=3.29, CI=2.16-5.01, P=3.49x10 -8 ). The nature of the heterogeneous PAX5 driver alterations of this subtype influenced relapse risk, with internal PAX5 amplifications and biallelic PAX5 alterations associated with the highest risk. Specific chr gains influenced outcome in hyperdiploid ALL, with gain of chr 10 and disomy of chr 7 associated with favorable outcome (OR=0.27, CI=0.17-0.42, P=8.02x10 -10 , St Jude Children’s Research Hospital (SJCRH) validation cohort: OR=0.22, CI=0.05-0.80, P=0.009), while disomy of chr 10 and 17 and gain of chr 6 were enriched in patients that relapsed (OR=7.16, CI=2.63-21.51, P=2.19x10 -5 ; SJCRH cohort: OR=21.32, CI=3.62-119.30, P=0.0004). Genomic alterations were also associated with relapse in a subtype-dependent manner, including alterations of INO80 in ETV6:RUNX1, IKZF1 and CREBBP in hyperdiploid, and FHIT in Ph-like ALL. Conclusions: Genetic subtype, aneuploidy patterns, and secondary genomic alterations influence risk of relapse in children otherwise classified with SR ALL, or HR ALL with favorable genetics. Comprehensive genomic analysis is required for optimal risk stratification and treatment allocation, and particularly to study reduction of therapy in the lowest risk patients. [Table: see text]

As part of the advancement in therapeutic decision-making for brain tumor patients at St. Jude Children’s Research Hospital (SJCRH), we develop and compare the performance of three classification models: a deep learning neural network (NN), an exact bootstrap k-nearest neighbor (kNN), and a random forest classifier (RF) model to predict the 82 molecularly distinct central nervous system (CNS) tumor classes based on DNA-methylation profiles of 2,801 patients. We validate their classification accuracy, precision, and recall with 2,054 samples from two independent cohorts. Although all models perform robustly to missing data, the NN model achieves the highest classification accuracy and maintains better balance between precision and recall than kNN and RF. Average precision and recall of NN reduce to that of RF and kNN only when tumor purity was less than 50%. In conclusion, DNA-methylation based deep learning approach provides the most potential advancement toward precision medicine for brain tumors.

Abstract Background: Neoantigen-based personalized cancer vaccines carry significant promise in treating solid malignancies. However, there are uncertainties regarding the choice between the primary or the metastatic tumor for neoantigen prediction in individual patients. Here, we conducted a thorough examination of somatic variations in 676 patients who had paired primary and metastatic solid tumors. Methods: Patients were enrolled in the Total Cancer Care protocol (NCT03977402) to which patients provided an IRB-approved written informed consent within the Oncology Research Information Exchange Network (ORIEN). Whole-exome sequencing of 756 primary and metastatic tumor pairs was performed (N = 676 patients). These included Genitourinary (n=83), Gynecological (n=97), Gastrointestinal (n=213), Thoracic (n=33), Cutaneous (n=24), Breast (n=108), Endometrial (n=49), Sarcoma (n=35), Head-and-Neck (n=106) and others (n=8). The data was analyzed through the ORIEN AVATAR Molecular Analysis Pipeline for somatic mutation variant detection and variant annotation. In this analysis, we focused on somatic events that result in an in-frame alteration (such as missense, in-frame deletion and in-frame insertion) and out-of-frame protein-altering mutations (such as frameshifts, de novo start, out-of-frame, and nonstop gain). Clonal population structure was determined based on pyclone-vi. Results: For in-frame events, bladder cancer, melanoma, and gynecological cancers shared close to 75% of the mutations between paired primary and metastatic cases. In contrast, sarcoma and thyroid cancer had a low overlap (~ 25%) of variants. For out-of-frame events, these events tend to have a lower proportion of shared somatic variants between primary and metastasis than in-frame variants. Oncogenic drivers (e.g., BRAF V600E, KRAS G12A, and TP53 loss-of-function) were highly likely to be present in both paired primary and metastatic tumors. Next, we performed additional analysis on evolutionary selection of protein-coding variants via dN/dS calculation. We found no significant global shift in dN/dS ratio between paired primary and metastatic tumors across malignancies. However, we found increased selection of protein-coding variants in brain and liver metastatic sites, which correlated with increased homologous recombination deficiency within these sites. Conclusions: Our analysis demonstrates genetic variations that exist when comparing paired primary and metastatic tumors that appear to vary by histology. Variants are potentially undergoing negative selection supported by the preferential loss of out-of-frame events in metastatic tumors and positive selection in specific metastatic sites. Overall, understanding the clonal structure will be key to neoantigen prediction for effective neoantigen-based vaccines. Citation Format: Alyssa Obermayer, Timothy Shaw, Darwin Chang, Joshua Davis, Jamie K. Teer, Xiaoqing Yu, Xuefeng Wang, Dale Hedges, Aik Choon Tan, Robert Rounbehler, Abdul Rafeh Naqash, Margaret Gatti-Mays, Aakrosh Ratan, Martin McCarter, Howard Colman, Igor Puzanov, Susanne Arnold, Michelle Churchman, Patrick Hwu, William Dalton, George Weiner, Jose Conejo-Garcia, Paulo C. Rodriguez, Bodour Salhia, Ahmad A. Tarhini. Analysis of clonal heterogeneity within paired primary and metastatic tumor samples of patients with solid tumors and implications for neoantigen-based personalized cancer vaccines [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3886.

As part of the advancement in therapeutic decision-making for brain tumor patients at St. Jude Children’s Research Hospital (SJCRH), we developed three robust classifiers, a deep learning neural network (NN), k-nearest neighbor (kNN), and random forest (RF), trained on a reference series DNA-methylation profiles to classify central nervous system (CNS) tumor types. The models’ performance was rigorously validated against 2054 samples from two independent cohorts. In addition to classic metrics of model performance, we compared the robustness of the three models to reduced tumor purity, a critical consideration in the clinical utility of such classifiers. Our findings revealed that the NN model exhibited the highest accuracy and maintained a balance between precision and recall. The NN model was the most resistant to drops in performance associated with a reduction in tumor purity, showing good performance until the purity fell below 50%. Through rigorous validation, our study emphasizes the potential of DNA-methylation-based deep learning methods to improve precision medicine for brain tumor classification in the clinical setting.

Purpose: The promise of immune checkpoint inhibitor (ICI) therapy underlines the importance of comprehensively investigating the rationale for combinations with diverse immune modulators across different cancer types. Given the progress made with PD1 blockade to date, we examined mRNA co-expression levels of PD-1 with 13 immune checkpoints, including co-inhibitory receptors (LAG3, CTLA4, PD-L1, TIGIT, TIM3, VISTA, BTLA) and co-stimulatory molecules (CD28, OX40, GITR, CD137, CD27, HVEM), using RNA-Seq by Expectation-Maximization (RSEM). Methods: We analyzed real-world clinical and transcriptomic data from the Total Cancer Care Protocol (NCT03977402) and Avatar® project of patients with cancer treated within the Oncology Research Information Exchange Network (ORIEN) network. Using anti-PD1 as a backbone, we intended to investigate the rationale for combinations in different cancers. Pearson's R coefficients and associated P-values were calculated using SciPy 1.7.0. Results: The co-expression of PD1 with 13 immune checkpoints and PD-L1 varies across selected malignancies included. In cutaneous melanoma, PD1 expression correlated significantly with four co-inhibitory receptors (LAG3, TIM3, TIGIT, VISTA) and one co-stimulatory molecule (CD137). In urothelial carcinoma, PD1 expression significantly correlated with four co-inhibitory (TIGIT, CTLA4, LAG3, VISTA) and four co-stimulatory (OX40, CD27, CD137, HVEM) molecules. In pancreatic adenocarcinoma, only CD28 showed a significant correlation with PD1 expression. No significant correlations with PD1 expression were found in the ovarian cancer cohort. Notably, melanoma and urothelial carcinoma exhibited a dominant co-expression of co-inhibitory molecules with PD1, indicative of exhausted T cells, in contrast to the co-stimulatory molecule dominance in ovarian and pancreatic cancers, suggesting less differentiated T cells. Conclusions: Our findings highlight the potential for diverse combination strategies in immunotherapy, particularly with PD1 blockade, across various cancers.

Abstract Sarcomas encompass a group of malignant diseases arising from mesenchymal origins. Given their rarity and diversity, a fundamental understanding of the genomic underpinnings for many sarcoma subtypes is still lacking. Studies are often limited to one or several of the more common subtypes or a narrow evaluation of a broader sampling. We therefore report on one of the largest comprehensive omics evaluation in sarcomas to date, including whole exome sequencing (WES; n = 1170) and RNA-sequencing (n = 983) of tissues from 29 different sarcoma histologic subtypes collected at 13 institutions in the US as part of the Oncology Research Information Exchange Network (ORIEN). We identified recurrent somatic mutations previously identified in sarcomas (e.g. TP53, KIT) as well as other cancer types (e.g. BRCA1). The burden of putatively pathogenic driver point mutations was higher in metastatic samples (median = 3) as compared to primary tumor samples (median = 2; p < 0.001). We observed frequent copy number alterations including whole genome doubling more commonly in metastatic compared to primary tumors (23.4% vs. 16.9%; p = 0.0.25). Inspection of gene expression dimensionality reduction (UMAP) showed separation of gastrointestinal stromal tumors (GISTs), leiomyosarcomas, myxoid liposarcomas, and well/de-differentiated liposarcomas from the other histologies. Differential expression analysis for these four histologies with gene set enrichment analysis highlights the diversity of disease-specific pathways and need for sarcoma subtype-specific translational focus. Estimation of immune cell abundances based on RNA-seq followed by hierarchical clustering identified five immune subtypes. The subtypes ranged from low (clusters A, B) to high (clusters D, E) immune infiltration with higher abundance of T, B, Natural Killer (NK), and myeloid cells (FDR < 0.01). Intermediate immune group C was predominantly composed by GISTs and marked by an enrichment for NK cells (FDR < 0.01) compared to all groups except the immune “hot” group E; however, this immune group exhibited modest infiltration by other immune cell types. Notably, we observed significant differences in the overall survival of patients with sarcomas in immune enriched (C, D, E) compared to immune depleted clusters (A, B; p = 0.002). In summary, we report the genomic and expressional landscape of over 1000 sarcomas, representing one of the largest comprehensive profiling efforts in this disease. We identify the mutational and copy number variation landscape and observe differences between primary and metastatic samples. We highlight expression pathways that are enriched in histologic subtypes that cluster most distinctly from others, providing a subtype-specific roadmap for further translational efforts. Finally, we define immune enriched or depleted sarcoma subgroupings that carry a prognostic impact. Citation Format: Alex C. Soupir, Oscar E. Ospina, Dale Hedges, Jamie K. Teer, Michael D. Radmacher, David M. McKean, Nathan Seligson, Martin McCarter, Breelyn Wilkey, Greg Riedlinger, John Groundland, Benjamin J. Miller, Bryan Schneider, Reema Patel, Abdul Rafeh-Naqash, Stephen Edge, Bodour Salhia, Chris Moskaluk, Maggy Johns, Michelle L. Churchman, Oliver Hampton, David Liebner, Brooke L. Fridley, Andrew S. Brohl. Genomic landscape and estimation of immune infiltration of soft tissue sarcoma histology subtypes from the ORIEN network [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3928.

Trisomy 21, the genetic cause of Down syndrome (DS), is the most common congenital chromosomal anomaly. It is associated with a 20-fold increased risk of acute lymphoblastic leukemia (ALL) during childhood and results in distinctive leukemia biology. To comprehensively define the genomic landscape of DS-ALL, we performed whole-genome sequencing and whole-transcriptome sequencing (RNA-Seq) on 295 cases. Our integrated genomic analyses identified 15 molecular subtypes of DS-ALL, with marked enrichment of CRLF2-r, IGH::IGF2BP1, and C/EBP altered (C/EBPalt) subtypes compared with 2257 non-DS-ALL cases. We observed abnormal activation of the CEBPD, CEBPA, and CEBPE genes in 10.5% of DS-ALL cases via a variety of genomic mechanisms, including chromosomal rearrangements and noncoding mutations leading to enhancer hijacking. A total of 42.3% of C/EBP-activated DS-ALL also have concomitant FLT3 point mutations or insertions/deletions, compared with 4.1% in other subtypes. CEBPD overexpression enhanced the differentiation of mouse hematopoietic progenitor cells into pro-B cells in vitro, particularly in a DS genetic background. Notably, recombinationactivating gene-mediated somatic genomic abnormalities were common in DS-ALL, accounting for a median of 27.5% of structural alterations, compared with 7.7% in non-DS-ALL. Unsupervised hierarchical clustering analyses of CRLF2-rearranged DS-ALL identified substantial heterogeneity within this group, with the BCR::ABL1-like subset linked to an inferior event-free survival, even after adjusting for known clinical risk factors. These results provide important insights into the biology of DS-ALL and point to opportunities for targeted therapy and treatment individualization.

3126 Background: Neoantigen-based personalized cancer vaccines carry significant promise in treating solid malignancies. For the purpose of neoantigen prediction, questions remain on the suitability of the primary versus the metastatic tumor of an individual patient (pt). Here we performed an in-depth analysis of somatic variants of 45 pts with paired primary and metastatic solid tumors. Methods: Pts were enrolled in the Total Cancer Care protocol across the Oncology Research Information Exchange Network (ORIEN). Whole-exome sequencing (WES) of primary and metastatic tumor pairs was performed for 45 pts. These included head and neck (n = 8), renal cell carcinoma (n = 7), non-small cell lung cancer (n = 6), melanoma (n = 5), bladder (n = 3), sarcoma (n = 3), ovary (n = 3), esophageal (n = 2), colorectal (n = 2) and other singletons (n = 5). The data was analyzed through the ORIEN AVATAR Molecular Analysis Pipeline, which consisted of adapter trimming, single sample variant detection, somatic mutation detection and filtering, variant annotation, and filtering of variants from a panel of normal samples. In this analysis, we focused on somatic events that result in an in-frame alteration (such as missense, in-frame insertion/deletion) and out-of-frame protein-altering mutations (such as frameshifts, de novo start and nonstop gain). To assign the clones based on the variant allele frequency, we performed complete linkage clustering. Each cluster was then assigned based on the identified oncogenic driver variant. Results: For in-frame events, we noticed that bladder cancer, melanoma, and gynecological cancers shared close to 75% of the mutations between paired primary and metastatic cases. In contrast, esophageal, brain, and endometrial cancers had a low overlap (< 25%) of variants. For out-of-frame events, we found that these events tend to have a lower proportion of shared somatic variants between primary and metastasis than in-frame variants. Through a closer evaluation of the variant allele frequency in variants shared between primary and metastatic disease, we were able to track several cancer drivers, such as oncogenic drivers of BRAF V600E, RB1 R358*, NRAS G13R, KRAS G12A, and TP53 loss-of-function that were consistently present in paired primary and metastatic tumors. Here, we were able to assign the putative clone based on the driver event in 39 out of 45 pts. Conclusions: Our analysis demonstrates genetic variations that exist when comparing paired primary and metastatic tumors that appear to vary by histology. Variants are potentially undergoing negative selection supported by the preferential loss of out-of-frame events in metastatic tumors. Understanding the clonal structure will be key to neoantigen prediction for effective neoantigen-based vaccines. Additional steps in variant prioritization are ongoing and will be reported at the meeting.

Childhood B-cell acute lymphoblastic leukaemia (B-ALL) is characterised by recurrent genetic abnormalities that drive risk-directed treatment strategies. Using current techniques, accurate detection of such aberrations can be challenging, due to the rapidly expanding list of key genetic abnormalities. Whole genome sequencing (WGS) has the potential to improve genetic testing, but requires comprehensive validation. We performed WGS on 210 childhood B-ALL samples annotated with clinical and genetic data. We devised a molecular classification system to subtype these patients based on identification of key genetic changes in tumour-normal and tumour-only analyses. This approach detected 294 subtype-defining genetic abnormalities in 96% (202/210) patients. Novel genetic variants, including fusions involving genes in the MAP kinase pathway, were identified. WGS results were concordant with standard-of-care methods and whole transcriptome sequencing (WTS). We expanded the catalogue of genetic profiles that reliably classify PAX5 alt and ETV6::RUNX1 -like subtypes. Our novel bioinformatic pipeline improved detection of DUX4 rearrangements ( DUX4 -r): a good-risk B-ALL subtype with high survival rates. Overall, we have validated that WGS provides a standalone, reliable genetic test to detect all subtype-defining genetic abnormalities in B-ALL, accurately classifying patients for the risk-directed treatment stratification, while simultaneously performing as a research tool to identify novel disease biomarkers.

Identification of specific leukemia subtypes is a key to successful risk-directed therapy in childhood acute lymphoblastic leukemia (ALL). Although RNA sequencing (RNA-seq) is the best approach to identify virtually all specific leukemia subtypes, the routine use of this method is too costly for patients in resource-limited countries. This study enrolled 295 patients with pediatric ALL from 2010 to 2020. Routine screening could identify major cytogenetic alterations in approximately 69% of B-cell ALL (B-ALL) cases by RT-PCR, DNA index, and multiplex ligation-dependent probe amplification. STIL-TAL1 was present in 33% of T-cell ALL (T-ALL) cases. The remaining samples were submitted for RNA-seq. More than 96% of B-ALL cases and 74% of T-ALL cases could be identified based on the current molecular classification using this sequential approach. Patients with Philadelphia chromosome-like ALL constituted only 2.4% of the entire cohort, a rate even lower than those with ZNF384-rearranged (4.8%), DUX4-rearranged (6%), and Philadelphia chromosome-positive (4.4%) ALL. Patients with ETV6-RUNX1, high hyperdiploidy, PAX5 alteration, and DUX4 rearrangement had favorable prognosis, whereas those with hypodiploid and KMT2A and MEF2D rearrangement ALL had unfavorable outcomes. With the use of multiplex ligation-dependent probe amplification, DNA index, and RT-PCR in B-ALL and RT-PCR in T-ALL followed by RNA-seq, childhood ALL can be better classified to improve clinical assessments.

Background PD1 immune checkpoint inhibitors (ICIs) have resulted in significant improvements in the care of patients with advanced malignancies. PD1 based combinations including with CTLA4 and LAG3 represent the future of ICI immunotherapy and there is a need to better understand the rationale for investigating combinations involving other immune modulator therapeutic candidates across cancer types. Methods We utilized real-world clinical and transcriptomic data collected under the Total Cancer Care Protocol (NCT03977402) and Avatar® project within the Oncology Research Information Exchange Network (ORIEN) of 18 cancer centers to which all included subjects provided an IRB-approved written informed consent at their participating institutions. Using RSEM we analyzed mRNA co-expression levels of PD-1 with 12 immune checkpoints including the co-inhibitory receptors LAG3, CTLA4, TIGIT, TIM3 (HAVCR2), VISTA (VSIR), BTLA and the positive co-signaling molecules CD28, OX40 (TNFRSF4), GITR (TNFRSF18), CD137 (TNFRSF9), CD27, HVEM (TNFRSF14) as well as PD-L1 (CD274). Pearson's R coefficients and associated P values were calculated using SciPy 1.7.0. We defined Pearson's coefficient > 0.5 and p < 10E-10 as significant correlation. Results Co-expression of PD1 along with the 12 immune checkpoints and PD-L1 across select malignancies included in our analysis is shown in (table 1), including Pearson's correlation and the associated P values, sorted by the level of correlation. In cutaneous melanoma, in terms of co-inhibitory receptors that suppress T cell activation, the expression of PD1 was significantly correlated with four molecules: LAG3, TIM3, TIGIT and VISTA; and only significantly correlated with one costimulatory molecule CD137. For urothelial carcinoma, there were 4 co-inhibitory (TIGIT, CTLA4, LAG3 and VISTA) and 4 co-stimulatory (OX40, CD27, CD137 and HVEM) molecules statistically correlated with PD1 expression. For pancreatic adenocarcinoma, only CD28 was deemed to be correlated with PD1 expression. No immune checkpoints were deemed significantly correlated with PD1 expression in the ovarian cancer cohort. Overall, in melanoma and to a certain extent in urothelial carcinoma, the co-expression of co-inhibitory molecules with PD1 was more dominant reflecting late exhausted T cells, as compared to co-stimulatory molecules likely more dominant in ovarian and pancreatic carcinomas reflecting less differentiated T cells. Conclusions With PD1 blockade as a backbone for immune checkpoint targeting combinations, our interrogations of pan-cancer transcriptomic data provide support for multiple potential combination strategies in the tested malignancies that warrant further investigation. Melanoma and urothelial carcinoma as more immunogenic tumors reflected a PD1+immunoinhibitory dominant phenotype, while less immunogenic ovarian and pancreatic carcinomas reflected a trend toward a PD1+immunostimulatory phenotype. Acknowledgements We are grateful to the participating patients and their family members as well as all research staff supporting the conduct of the Total Cancer Care protocol. Trial Registration NCT03977402 Ethics Approval We utilized real-world clinical and transcriptomic data collected under the Total Cancer Care Protocol (NCT03977402) and Avatar® project within the Oncology Research Information Exchange Network (ORIEN) of 18 cancer centers to which all included subjects provided an IRB-approved written informed consent at their participating institutions.

Acute lymphoblastic leukemia (ALL) is the most common childhood cancer. Here, using whole-genome, exome and transcriptome sequencing of 2,754 childhood patients with ALL, we find that, despite a generally low mutation burden, ALL cases harbor a median of four putative somatic driver alterations per sample, with 376 putative driver genes identified varying in prevalence across ALL subtypes. Most samples harbor at least one rare gene alteration, including 70 putative cancer driver genes associated with ubiquitination, SUMOylation, noncoding transcripts and other functions. In hyperdiploid B-ALL, chromosomal gains are acquired early and synchronously before ultraviolet-induced mutation. By contrast, ultraviolet-induced mutations precede chromosomal gains in B-ALL cases with intrachromosomal amplification of chromosome 21. We also demonstrate the prognostic significance of genetic alterations within subtypes. Intriguingly, DUX4 - and KMT2A -rearranged subtypes separate into CEBPA/FLT3 - or NFATC4 -expressing subgroups with potential clinical implications. Together, these results deepen understanding of the ALL genomic landscape and associated outcomes.

Gene fusions play a prominent role in the oncogenesis of many cancers and have been extensively targeted as biomarkers for diagnostic, prognostic, and therapeutic purposes. Detection methods span a number of platforms, including cytogenetics (e.g., FISH), targeted qPCR, and sequencing-based assays. Before the advent of next-generation sequencing (NGS), fusion testing was primarily targeted to specific genome loci, with assays tailored for previously characterized fusion events. The availability of whole genome sequencing (WGS) and whole transcriptome sequencing (RNA-seq) allows for genome-wide screening for the simultaneous detection of both known and novel fusions. RNA-seq, in particular, offers the possibility of rapid turn-around testing with less dedicated sequencing than WGS. This makes it an attractive target for clinical oncology testing, particularly when transcriptome data can be multi-purposed for tumor classification and additional analyses. Despite considerable efforts and substantial progress, however, genome-wide screening for fusions solely based on RNA-seq data remains an ongoing challenge. A host of technical artifacts adversely impact the sensitivity and specificity of existing software tools. In this chapter, the general strategies employed by current fusion software are discussed, and a selection of available fusion detection tools are surveyed. Despite its current limitations, RNA-seq-based fusion detection offers a more comprehensive and efficient strategy as compared to multiple targeted fusion assays. When thoughtfully employed within a wider ecosystem of diagnostic assays and clinical information, RNA-seq fusion detection represents a powerful tool for precision oncology.

Background: Children with relapsed T-cell acute lymphoblastic leukemia (T-ALL) have poor prognosis, and identification of patients at risk for recurrence is required to prevent relapse. Prior genomic studies of T-ALL have failed to identify genetic alterations that are prognostic independent of minimal residual disease (MRD), in part due to limited cohort size, exclusion of patients with refractory disease, and lack of comprehensive, integrated genome and transcriptome sequencing, which is important as many T-ALL drivers occur in non-coding regions of the genome. Aim: To identify all coding and non-coding alterations and identify genomic predictors of relapse or refractory disease. Methods: We performed whole genome, exome, and transcriptome sequencing of 1,313 cases enrolled on the Children's Oncology Group AALL0434 trial of childhood T-ALL. ATAC-seq and HiChIP were used to investigate the consequences of non-coding mutations and structural variants (SV). We used transcriptome profiling by Uniform Manifold Approximation and Projection (UMAP), Leiden algorithm clustering, hierarchical clustering of oncogene expression, and analysis of sequence and structural DNA variants to identify subtypes of T-ALL cases, their driver and associated genomic alterations. Results: Using integrative genomic analysis, we identified the clonal subtype-defining putative drivers in 94% of samples, 60% of which were non-coding regions and required WGS for identification in 28% of cases. We identified 16 T-ALL subtypes, several of which were previously unrecognized and involved partitioning of known subgroups into groups with distinct gene expression and driver alterations. We identified 4 subtypes with deregulation of the TAL1/LMO1/LMO2 core transcriptional circuitry: TAL1-RA, TAL1-RB, TAL1-RB/RPL10, LMO2-refractory (15.5%, 20.4%, 2.8%, 0.9% of all cases). These subtypes were separated by the lower expression of T cell maturation genes CD4/CD8 and RAG in TAL1-RA and more mature double positive TAL1-RB. TAL1-RA was characterized by multiple activation mechanisms for TAL1, LMO1 and/or LMO2; and TAL1-RB with alterations of TAL1, LMO1, LMO2,LYL1 and, TAL2 including rearrangements of these genes to TCR enhancers, chimeric fusion oncoproteins and a diverse range of enhancer hijacking, amplification and SNV events. Activation mechanisms differed between the groups; TAL1-RA commonly had STIL::TAL1 fusion with frequently co-occurring LMO2/LMO1 SNV/Indels or LMO2 enhancer gains, whereas TAL1-RB had frequent TAL1/LMO2/TAL2/LYL1 TCR rearrangements, TAL1 enhancer gains or various other TAL1 rearrangements. Two additional subgroups included 39 cases with TAL1/LMO2 activation and RPL10 mutations (TAL1-RB/RPL10) and the other characterized by refractory disease (LMO2-refractory, day 29 MRD >5%, 8 cases) with deregulation of LMO2 by hijacking of the BCL11B enhancer. We observed striking association between early T-cell precursor phenotype and driver lesions, with several groups highly enriched for ETP-ALL cases. One group (11.5% of all cases, 41.9% of ETP ALL cases, Fig. 1A) had distinct gene expression and deregulation of HOXA13 by SVs, ZFP36L2 or ETV6 by chimeric fusions and MED12 by SNVs. A second group (7.3% cases, 30.4% ETP) had HOXA9 deregulation by HOXA9 SVs or NUP98/NUP214/KMT2/MLLT10 chimeric fusions. Selective HOXA13 deregulation could be explained by genomic breakpoints occurring in different chromatin compartments than rearrangements deregulating other HOXA genes such as HOXA9. Novel HOXA13 SVs include rearrangements to MIR181A1HG and MED13 loci. Event-free survival (EFS) varied by subtype (Fig1B. log rank P<0.0001). Adverse risk subtypes were SPI1-R (P<0.0001, subtype vs. rest of samples), HOXA9 (P<0.0001), LMO2-refractory (P<0.0001) and favorable risk TAL1-RB/RPL10 (P=0.046), NKX2-1 (P=0.0099), TLX1 (P=0.013), TLX3 (P=0.028) and KMT2A-R (P=0.003) subtypes (Fig. 1B). Outcome was associated with key oncogenic pathways: for example, absence of Notch pathway alterations was associated with poor outcome if D29 MRD≥0.01% (5-year EFS <75%), whereas, patients with Notch alterations had good outcome regardless of Day 29 MRD. Conclusion: Comprehensive definition of the genomic landscape of T-ALL requires integrated WGS and RNAseq, and identifies multiple subgroups and driver lesions associated with immunophenotype and outcome. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Background: T-cell acute lymphoblastic leukemia (T-ALL) is an aggressive hematopoietic malignancy including leukemias of early T cell precursor acute lymphoblastic leukemia (ETP-ALL) and transformed thymocytes. Prior genomic studies of T-ALL had limited cohort size, excluded refractory disease and focused on alterations in coding parts of the genome. Aims: Investigate the genomic basis of T-ALL by identifying both coding and non-coding alterations and defining T-ALL subtypes. Methods: We performed whole-genome sequencing (WGS), whole-exome sequencing (WES), RNA-sequencing of 1,313 cases enrolled on the Children’s Oncology Group AALL0434 trial. Results: Uniform Manifold Approximation and Projection (UMAP) and gene expression clustering analyses of RNA-seq data identified 16 subtypes, of which 4 have not been reported previously. Furthermore, we could divide existing subtypes into smaller subgroups with common subtype-defining alterations, such as structural variation (SV) and copy number variation (CNV). TLX1 activation was linked with TLX1-TCR SVs and deletions in the TLX1 chromosomal domain, whereas TLX3 deregulation was associated with TLX3-BCL11B enhancer hijacking, but also through TLX3-TCR, TLX3-CDK6 rearrangements. We discovered two separate NKX2-1 deregulated groups, one characterized by TCR rearrangements and RPL10 mutations and the other by NKX2-1 CNVs, chromosome 14 chromothripsis, or MYB-TCR rearrangements. We also observed a distinct group of 9 cases aged 1-2 years, with recurrent STAG2-LMO2 rearrangements and a patient with inactivating STAG2 mutation and CELF1 enhancer hijacking by LMO2. Moreover, we found a group of 22 cases that were highly enriched for ETP-ALL with the following hallmark lesions: BCL11B enhancer amplification, BCL11B locus SVs involving enhancer hijacking of ARID1B, CCDC26, and novel CD34+ enhancer hijacking of lincRNA locus in chromosome 6. Gene expression-based clustering was unable to stratify patients based on TAL1, TAL2, LMO1, LMO2, LYL1 expression, and their respective activation mechanisms. However, WGS enabled further characterization of these patients, by identifying several types of activation mechanisms and co-occurring alterations for each oncogene. We identified canonical events, such as TAL1-STIL fusions, TCR rearrangements, and activation by TAL1/LMO1/LMO2 regulatory region indels, but also novel events, such as CD34 specific enhancer duplications downstream of TAL1 and BCL11B enhancer hijacking by LMO2. We also observed two smaller clusters with TAL1/LMO2 activation, where one with 39 cases was associated with TAL1/LMO2 activation and RPL10 mutations and the other with 8 cases had refractory disease (Day 29 MRD >5%). HOXA gene expression-associated subtypes were defined by fusions involving KMT2A or MLLT10 and PICALM/DDX3X or HOXA9-TCR SVs. Furthermore, our analysis revealed segregation of patients by HOXA13 or ZFP36L2 rearrangements and NUP98/NUP214 fusions. Interestingly, HOXA locus breakpoints involving HOXA13 and HOXA9 were in different chromatin compartments and were associated with mutually exclusive activation of either HOXA13 or other HOXA genes. Interestingly, HOXA13-deregulation was associated with ETP-ALL and frequent BCL11B enhancer hijacking, whereas HOXA9 breakpoints typically involved TCRg and were non-ETP. Image:Summary/Conclusion: Large-scale analysis of all children enrolled on the AALL0434 study has identified new subtype-defining lesions in T-ALL, including candidate novel enhancer hijacking events and enhancer duplications that are likely to result in oncogene deregulation in T-ALL.

Radiation-induced high-grade gliomas (RIGs) are an incurable late complication of cranial radiation therapy. We performed DNA methylation profiling, RNA-seq, and DNA sequencing on 32 RIG tumors and an in vitro drug screen in two RIG cell lines. We report that based on DNA methylation, RIGs cluster primarily with the pediatric receptor tyrosine kinase I high-grade glioma subtype. Common copy-number alterations include Chromosome (Ch.) 1p loss/1q gain, and Ch. 13q and Ch. 14q loss; focal alterations include PDGFRA and CDK4 gain and CDKN2A and BCOR loss. Transcriptomically, RIGs comprise a stem-like subgroup with lesser mutation burden and Ch. 1p loss and a pro-inflammatory subgroup with greater mutation burden and depleted DNA repair gene expression. Chromothripsis in several RIG samples is associated with extrachromosomal circular DNA-mediated amplification of PDGFRA and CDK4. Drug screening suggests microtubule inhibitors/stabilizers, DNA-damaging agents, MEK inhibition, and, in the inflammatory subgroup, proteasome inhibitors, as potentially effective therapies.