6570 Background: Genomic analysis is critical for accurate diagnosis and risk stratification for patients with myeloid malignancies. Whole genome sequencing (WGS) has the ability to capture a breadth of alterations and has the potential to complement or replace current genomic profiling techniques. We sought to validate the molecular findings from Tempus xH, a WGS based assay, with standard of care [SOC] results from a major academic institution (MDACC). Methods: We retrospectively sequenced 43 patient samples from MDACC using the Tempus xH WGS assay. SOC data from cytogenetics (karyotype +/- FISH), 81-gene panel sequencing, and/or optical genome mapping OGM (n=10), was regarded as the source of truth. We performed a two-part study: an initial unblinded analysis (n=10) comparing WGS to SOC, followed by a blinded unbiased comparative review (n=33). Clinically relevant alterations were compared between the SOC and xH reportable range workflows. We excluded mutations with <= 10% VAF and filtered copy number alterations (CNAs) < 5Mb, and combined OGM and cytogenetics (10>= metaphases). Structural variants (SVs) were included if they overlapped orthogonal truth or were recurrent myeloid fusions. Results: In the unblinded analysis (n=10), WGS detected 53 clinically reported alterations, yielding 94% sensitivity. Single nucleotide variants (SNVs) had 100% sensitivity, and 18/20 CNA events and 6/7 SVs were detected. The single missed SV event was a derivative chromosome (chr), between chr1 and chr20, identified by cytogenetics with breakpoints over the masked centromere region of the WGS. However, WGS detected CNAs in chr arms of 1 and 20, signalling the event was captured yet not resolved as a derivative chr. Two copy losses on chr 19 (complete loss and p-arm) and an ASXL1 frameshift were detected by WGS but not by SOC. In the blinded study (n=33), WGS detected 143 events, including 124 true positives and no alterations from SOC went undetected by WGS demonstrating 100% sensitivity. WGS identified 24/24 of the CNAs detected by cytogenetics. Among all specimens, 21 novel, high confidence variants were identified in 11 specimens (39%) by WGS that were not detected by SOC; including 14 CNAs, two SNVs, and five SVs, including diagnostically and/or prognostically critical alterations: 2 MECOM rearrangements, 1 KMT2A -PTD, an 11MB deletion in chr 17p ( TP53 locus), and a focal RUNX1 deletion highlighting the potential clinical benefit of WGS. Conclusions: WGS accurately recapitulates results from targeted NGS panels, OGM and cytogenetics and advances our understanding of MDS biology. xH WGS uncovered a significant number of novel, clinically relevant and potentially targetable findings undetected by SOC. WGS provides a complementary view of the genome, identifying actionable and prognostic markers helping clinicians make more informed treatment decisions.

e18558 Background: Categorizing gene fusions as drivers or passengers is a data-intensive challenge requiring manual interventions in all but the most well-established biomarkers. The large numbers of structural variants identified from genomic assays require a combination of tools to prioritize clinically relevant mutations in a timely manner. Recent advances in artificial intelligence (AI) and AI-agents have enabled the summarization of large-scale biological data and are ideally suited for flagging important knowledge quickly and with limited manual intervention. Methods: This work introduces Peryton, a flexible AI-agent capable of analyzing the oncogenic potential of both RNA and DNA fusion events. By integrating genomic breakpoints, gene annotations, literature mining and RNA expression signatures, Peryton generates concise, fully referenced summary of a fusion’s biological function, oncogenic potential, and therapy implications. Peryton incorporates peer-reviewed articles from PubMed Central, publicly available fusion knowledgebases, and internal data (a curated gene fusion database and internal RNA sequencing expression values) to evaluate and prioritize the oncogenic potential of fusions identified within a sample. The chain-of-thought strategy does extensive pre-computation before prompting the large language model (LLM) with data on which protein features are retained or lost in individual gene fusions. A key output is an oncogenicity score, an AI-generated classification of a fusion's cancer-causing potential based on literature review. Results: We benchmarked Peryton against a dataset of 300 known positive events from COSMIC and 300 presumed passenger events. The agent demonstrated high curation accuracy, with an area under the receiver operating characteristic greater than 0.98 across the 600 randomly selected records. In a clinical validation study of 41 non-canonical gene fusions flagged by internal pathologists, which were evaluated for clinical reporting, Peryton correctly prioritized the reportability of 90% of fusions providing evidence for its use supporting non-canonical fusion curation. Furthermore, when applied to all structural variant calls from 2 WGS AML cases (202 and 188 respectively), the agent successfully prioritized with the top score in each of the reported clinical drivers (KMT2A::AFDN and ROCK1::PDGFRA). The next highest fusion events showed prognostic literature evidence further validating its utility in real-world scenarios. Conclusions: Peryton provides a precise assessment of gene fusion events, effectively prioritizing candidate fusions for further clinical and research evaluation. This retrospective analysis demonstrates the agent's potential to streamline fusion curation and ultimately improve patient care.

6532 Background: The menin inhibitor revumenib was recently FDA approved for treating patients with relapsed or refractory KMT2A-rearranged (rKMT2A) acute leukemias. Cytogenetics, FISH, and targeted next-generation sequencing (NGS) frequently miss KMT2A 11q23 partial tandem duplications (KMT2A-PTD) ). Although KMT2A-PTDs have expression signatures similar to rKMT2A, they were excluded from revumenib’s registration trial. Preclinical models have shown that menin inhibitors may also be effective for KMT2A-PTD, highlighting the need for precise breakpoint and KMT2A fusion product detection. Here, we evaluated the effectiveness of high-resolution WGS to identify a diverse array of KMT2A-PTD. Methods: Using a WGS assay (Tempus xH) optimized for comprehensive profiling of myeloid neoplasms, we capture the entire KMT2A locus at base pair resolution. DNA was extracted from blood or bone marrow aspirates and was used to construct paired-end libraries via tagmentation. Sequencing was performed on the Illumina NovaSeq-X platform, achieving a mean coverage of 80X. Data were analyzed using the DRAGEN Platform with custom post-processing filters. Exon copy number calls from a targeted NGS assay and exon capture RNAseq NGS assay (Tempus xT and xR, respectively) were used for verification. Results: WGS from 230 hematopoietic neoplasms (68% AML, 18% MDS, 12% CML, and 2% others) identified 13 specimens (5.6%) containing a KTM2A-PTD, with variant allele frequencies (VAFs) between 9-66%. All PTDs contained breakpoints within known intron boundaries: one breakpoint in intron 1 (13/13) with terminal breakpoints located in intron 8 (6/13) or intron 10 (7/13). RNA data was available for 11 of 13 specimens and contained direct support for the presence of all the KMT2A-PTDs (100%). Using an NGS-targeted panel, exon-level copy calls were assessed for all 13 specimens with PTDs. Although unvalidated, we observed exon level gains in 9 of the 13 specimens (69%). As shown in prior studies, KMT2A-PTDs were mutually exclusive to other translocations, including rKMT2A. However, other high-frequency mutations for myeloid disease were present in select samples including mutations in IDH1, DNTM3A, WT1, and RUNX1. Conclusions: WGS is an effective tool for detecting KMT2A alterations that may be missed by traditional techniques such as NGS targeted capture, FISH or cytogenetics. 100% concordance was observed between WGS and RNAseq for KMT2A-PTDs, supporting the reliability of WGS. The FDA approval of menin inhibitors for KMT2A-rearranged AML/ALL suggests potential clinical opportunities for broad tests (WGS) to identify other rearrangements, including KMT2A-PTDs, highlighting the need for further research into targeted anti-leukemia therapies.

Background: The BCR::ABL1 fusion gene is a hallmark of chronic myeloid leukemia (CML) and is also present in a subset of acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), and mixed phenotype acute leukemia (MPAL) cases. Traditional methods for identifying this important biomarker, such as FISH and RT-PCR, are limited in the ability to detect novel or complex variants and also do not assess additional genetic alterations that could impact therapeutic decision making. Whole Genome Sequencing (WGS) is a comprehensive alternative, with the ability to detect a wide range of genetic alterations, including single nucleotide variants, insertions, deletions, and structural rearrangements such as BCR::ABL1 fusions. Methods: We retrospectively analyzed 215 hematological clinical samples sent for testing at Tempus that received both WGS via Tempus xH and RNA-seq via Tempus xR. Patients had historical diagnoses of AML (n=146), MDS (n=41), CML (n=25), and other blood cancers (n=3). Results: We detected BCR::ABL1 fusions in 29 samples (20 CML, 7 AML, 0 MDS, 2 others) using a combination of WGS and RNA-seq. Of these, 90% (n=26) were detected by both methods. For the three discordant samples, one was detected by WGS and not RNA-seq and 2 were detected by RNA-seq with a low total read support not detected by the WGS assay. Of the 26 samples that had support from both WGS and RNA-seq, BCR breakpoints largely aligned with expectation based on the known p210 (major) and p190 (minor) variants, which differ between AML and CML. RNA-seq confirmed all (100%) of the exon 1-3 breakpoints in p190 and exons 13-14 in p210. Most notably, our analysis uncovered 2 abnormal breakpoints with clinical implications, which were confirmed via RNA-seq. One specimen had a breakpoint within exon 2 of ABL1 resulting in an in frame transcript with a shortened exon 2. The other abnormal breakpoint was in intron 2 of ABL1 that results in exon 2 loss and exon 3 retention. It is known that the next generation allosteric tyrosine kinase inhibitor asciminib loses efficacy without exon 3, suggesting the use of an alternative therapeutic inhibitor is more appropriate highlighting the clinical utility of breakpoint resolution of WGS. Lastly, we used the WGS data to assess the presence of any pathogenic or likely pathogenic mutations in ASXL1 and RUNX1, both of which may impact survival and treatment outcomes. In this small cohort (n=26), we found ASXL1 alterations in 27% and RUNX1 in 12%, providing additional insights WGS can play in identifying key molecular markers. Conclusions: Our findings demonstrate that WGS can not only identify known BCR::ABL1 fusion events but also uncover novel breakpoints and variants of clinical significance, particularly when used in conjunction with RNA-seq. These data suggest that WGS is a powerful tool for the comprehensive genomic profiling of leukemia, with potential implications for personalized medicine and targeted therapy.

Germline pathogenic variants in TP53 cause Li-Fraumeni syndrome, with significantly elevated cancer risk from infancy. Accurate classification of TP53 variants is essential to guide clinical management and surveillance, yet many variants remain classified as variants of uncertain significance (VUS). To improve classification accuracy and reduce the proportion of VUS, the ClinGen TP53 Variant Curation Expert Panel (VCEP) has updated its specifications. The updated specifications incorporate the latest ClinGen recommendations and methodological advances, providing greater granularity for multiple evidence types, and also introduce the novel use of variant allele fraction as evidence of pathogenicity, particularly in the context of clonal hematopoiesis. Whenever feasible, the VCEP followed a data-driven approach using likelihood ratio-based quantitative analyses to guide code application and determine strength modifications, while also factoring in expert judgment. Proposed modifications were first discussed in working group meetings and then subjected to comprehensive review during monthly general VCEP meetings to reach consensus. The performance of new specifications was compared to that of the old specifications for 43 pilot variants, and led to both decreased VUS and increased certainty, with clinically meaningful classifications for 93 https://cspec.genome.network/cspec/ui/svi/svi/GN009 .

Identification of NCCN and WHO guideline-indicated genetic variation is crucial for diagnosis, risk assessment, and therapeutic decision-making in patients with myeloid malignancies. Current practice often utilizes small mutation panels, cytogenetics (CG), and/or fluorescence in situ hybridization—independent tests that require a separate specimen and have timeline and performance limitations. Whole-genome sequencing (WGS) is a promising alternative with rapid turnaround times and low sample requirements, but evidence showing concordance with other methods is still limited. Here, we report the performance of a WGS assay on patients with acute myeloid leukemia (AML), chronic myelogenous leukemia (CML), myelodysplastic syndromes (MDS) and other hematological malignancies (HM). We show high concordance with matched findings using Tempus xT-Heme (a 648 gene targeted DNA-seq panel), Tempus xR (a whole transcriptome RNA-seq assay), and available cytogenetic results. All 230 patients in our cohort had targeted DNA panel results, 215 had RNA-seq results, and 10 had cytogenetic data. For the WGS assay, DNA obtained from blood or bone marrow aspirates was used to construct paired-end libraries via tagmentation and WGS to 80X mean coverage with the Illumina NovaSeq-X platform. Data were analyzed using the DRAGEN Platform with custom post-processing filters. We filtered SNV/Indel alterations to 40 genes with VAF ≥10%, 608 recurrent rearrangements, and CNAs greater than 5MB. In our cohort (68% AML, 18% MDS, 12% CML and 2% other HM) the WGS assay identified 504 reportable SNV/indels, 119 SVs, and 14 large CNAs, which were highly concordant with other assays. Specifically, 99.4% of SNVs/indels and 95.1% of SVs were concordant between xT-Heme and WGS. A subset of guideline-recommended SVs (e.g. RUNX1-RUNX1T1, ELN adverse risk fusion) detected by WGS were not identified by xT-Heme (10.7% [13/121]) but were confirmed via RNA-seq. Large CNAs detected by WGS were 100% concordant (14/14) with available clinically reported CG. Notably, our cohort included 26 FLT3 internal tandem duplications (ITDs) ranging from 12 to 97 nucleotides in length, and 24 of these were identified by WGS —the missed ITDs had xT-Heme VAFs of less than 4%. WGS was less sensitive at low VAFs (targeted sequencing depths of ∼600x) that were likely subclonal. We demonstrate high concordance (97%) between our WGS assay and conventional methods in identifying guideline recommended genomic alterations, including large CNAs, in myeloid malignancies. The ability to obtain large CNA results, historically reserved for CG testing, has the potential to save costs and fill an unmet clinical need globally where cytogenetic resources are limited. Additionally, WGS can identify unique SVs that may be missed by conventional methods and may enhance personalized treatment strategies. Robert Huether, Derick Hoskinson, Pavana Anur, Raul Torres, Karl R. Beutner, Kristiyana Kaneva, Yan Yang, Kelly A. Potts, Andrew Frazier, Iris Braunstein, Brett M. Mahon, Michael A. Thompson, Kate Sasser, Halla Nimeiri, Lewis J. Kraft, Francisco M. De La Vega, Guillermo Garcia-Manero. Comprehensive whole genome sequencing (WGS) assay provides diagnostic insight into clinically relevant genomic alterations across myeloid malignancies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 7172.

Background: Identifying genetic alterations, as recommended by NCCN and WHO guidelines, is crucial for managing patients with myeloid malignancies including diagnosis, risk assessment, and therapeutic decisions. Traditional methods such as small mutation panels, cytogenetics, and fluorescence in situ hybridization require separate tests and have limitations. Whole-genome sequencing (WGS) is revolutionizing the assessment of genomic alterations in hematological malignancies (HM) by identifying the full suite of genomic changes with a single test. Studies by Duncavage et al. (2021) and re-analysis by Deshpande et al. (2023) showed that WGS offers a complementary, potentially superior, approach to traditional risk stratification in myeloid neoplasms. However, their study was partially constrained by focusing on risk-defining translocations, reducing the diversity of structural variants (SV) they report. To test the sensitivity of WGS in capturing a diverse array of mutation types, including SVs, we developed a comprehensive WGS assay optimized for clinically relevant alterations in Acute Myeloid Leukemia (AML), Chronic myelogenous leukemia (CML), Myeloproliferative neoplasm (MPN) and Myelodysplastic Syndromes (MDS). This tool is tailored to characterize recurrent SVs, extensive copy number alterations (CNAs), and single nucleotide variants/insertions and deletions (SNVs/indels). In this pilot study, we assess the performance of the WGS profiling assay in 135 patients with AML, MDS, CML, and small subset of other HM to evaluate the sensitivity of WGS identification of selected genetic alterations. We compared these results to matched data from targeted DNA panel, whole RNA transcriptome sequencing, and selected samples with available cytogenetic results. Method: To compare genomic alteration results and assess concordance on the same sample with the WGS assay, 135 patients were sequenced via Tempus xT-Heme (a 648 gene targeted DNA sequencing panel), 125 patients had additional RNA-seq results via Tempus xR (a whole transcriptome RNA sequencing assay). Ten samples also had cytogenetics data. For the WGS assay, DNA obtained from blood and bone marrow aspirates was used to construct paired-end libraries via tagmentation and whole-genome sequenced to 80X mean coverage with the Illumina NovaSeq-X platform (2x150bp reads). Data were analyzed in a tumor-only modality using the DRAGEN Bio-IT Platform with custom post-processing filters. We filtered SNV/Indel alterations to 40 genes (targeted by Duncavage et al.) with VAF>=10%, 608 recurrent rearrangements, and CNAs >5MB. Result: In our cohort (65% AML, 19% CML, 13% MDS and 2% other HM) the WGS assay identified 218 reportable SNVs/indels, 104 SVs, and 14 large CNAs, showing high concordance with xT-Heme, RNA or cytogenetic assays. Specifically, 99.5% of SNVs/indels and 98.9% of SVs were concordant between xT-Heme and WGS. A subset of guideline recommended SVs (e.g. RUNX1-RUNX1T1, ELN adverse risk fusion) were uniquely detected by the WGS assay alone and were confirmed with RNA (12.5% [13/104]). All large CNAs detected by WGS were 100% concordant with available clinically reported cytogenetics. Notably, our cohort included nine FLT3 internal tandem duplications (ITD) ranging from 12 to 97 nucleotides, (8/9) were identified by WGS-the missed ITD had an xT-Heme VAF of 1.9%. WGS missed some variants identified by xT-Heme with low VAFs <5% (targeted sequencing depths of ~600x) that were likely subclonal. WGS showed exceptionally high concordance to traditional techniques in identification of clinical relevant findings in myeloid neoplasms. Conclusions: We demonstrate high concordance (>98.9%) in identifying guideline recommended genomic alterations, including large CNAs, in myeloid malignancies by a single WGS test compared to parallel conventional methods. The ability to obtain large CNA results, historically reserved for cytogenetic testing, has the potential to save costs and fill an unmet clinical need globally where cytogenetic resources are limited. Additionally, WGS can identify unique SVs that may be missed by conventional methods and enables clinical benefits such as HLA typing for potential transplant (alloHCT) or diagnostic refinement by retroviral insertion (e.g. HTLV-1). These findings demonstrate the potential for integration of WGS into clinical practice to enhance personalized treatment strategies.

Colorectal cancer (CRC) is a leading cause of cancer-related death across the world. Irinotecan (IRI) is commonly used to treat metastatic CRC. The gene UGT1A1 encodes the enzyme responsible for the glucuronidation of SN-38, the active metabolite of IRI. Wild-type UGT1A1 contains six TA repeats [A(TA)6TAA] in its promoter region (also known as the *1 allele). Polymorphic UGT1A1 alleles with a higher number of TA repeats, such as UGT1A1 *28/(TA)7 and *37/(TA)8 alleles, cause decreased enzyme activity and are associated with severe toxicity in patients receiving IRI-based chemotherapy, for which dose reductions are recommended.

Some patients with therapy-related myeloid neoplasms (t-MN) may have unsuspected inherited cancer predisposition syndrome (CPS). We propose a set of clinical criteria to identify t-MN patients with high risk of CPS (HR-CPS). Among 225 t-MN patients with an antecedent non-myeloid malignancy, our clinical criteria identified 52 (23%) HR-CPS patients. Germline whole-exome sequencing identified pathogenic or likely pathogenic variants in 10 of 27 HR-CPS patients compared to 0 of 9 low-risk CPS patients (37% vs. 0%, p = 0.04). These simple clinical criteria identify t-MN patients most likely to benefit from genetic testing for inherited CPS.

Germline pathogenic variants in TP53 are associated with Li-Fraumeni syndrome, a cancer predisposition disorder inherited in an autosomal dominant pattern associated with a high risk of malignancy, including early-onset breast cancers, sarcomas, adrenocortical carcinomas, and brain tumors. Intense cancer surveillance for individuals with TP53 germline pathogenic variants is associated with reduced cancer-related mortality. Accurate and consistent classification of germline variants across clinical and research laboratories is important to ensure appropriate cancer surveillance recommendations. Here, we describe the work performed by the Clinical Genome Resource TP53 Variant Curation Expert Panel (ClinGen TP53 VCEP) focused on specifying the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP) guidelines for germline variant classification to the TP53 gene. Specifications were developed for 20 ACMG/AMP criteria, while nine were deemed not applicable. The original strength level for the 10 criteria was also adjusted due to current evidence. Use of TP53-specific guidelines and sharing of clinical data among experts and clinical laboratories led to a decrease in variants of uncertain significance from 28% to 12% compared with the original guidelines. The ClinGen TP53 VCEP recommends the use of these TP53-specific ACMG/AMP guidelines as the standard strategy for TP53 germline variant classification.

Heterozygous de novo variants in the eukaryotic elongation factorEEF1A2have previously been described in association with intellectual disability and epilepsy but never functionally validated. Here we report 14 new individuals with heterozygousEEF1A2variants. We functionally validate multiple variants as protein-damaging using heterologous expression and complementation analysis. Our findings allow us to confirm multiple variants as pathogenic and broaden the phenotypic spectrum to include dystonia/choreoathetosis, and in some cases a degenerative course with cerebral and cerebellar atrophy. Pathogenic variants appear to act via a haploinsufficiency mechanism, disrupting both the protein synthesis and integrated stress response functions of EEF1A2. Our studies provide evidence thatEEF1A2is highly intolerant to variation and that de novo pathogenic variants lead to an epileptic-dyskinetic encephalopathy with both neurodevelopmental and neurodegenerative features. Developmental features may be driven by impaired synaptic protein synthesis during early brain development while progressive symptoms may be linked to an impaired ability to handle cytotoxic stressors.

Abstract Background Patient derived tumor organoids (TOs) are emerging as potential models to elucidate mechanisms of tumor biology and therapeutic response. Here, we establish pan-cancer metrics for validation of genetic and transcriptomic recapitulation, and concordance of an organoid to its native tumor. Methods/Results We sequenced 50 tumor/TO pairs from 12 cancer types using the Tempus xT DNAseq panel and transcriptome RNAseq platforms. Concordance metrics between tumors and TOs were derived for genomic variants called by the DNA xT platform and comparative ratios were calculated for all detected somatic variants. Across all sequenced pairs, somatic variant detection concordance between any mutation identified in primary tissues and tumor organoids resulted in a mean value of 88.1%. Somatic primary tumor variant recapitulation, the percent of somatic variants identified in the primary tumors that were also detected in the TO, averaged 96.3%. In addition to genetic concordance, DNAseq can identify and track clonal and subclonal diversity from source material to TO. In particular, we observed that >90% of source tumor/TO pairs harbor variants with allelic fractions <40% in both the sequenced TO and primary tumor tissue, suggesting intra-tumor heterogeneity in subclonal cell populations is maintained. TOs and primary tumor transcriptomic profiles were compared by dimensionality reduction approaches (i.e. Uniform Manifold Approximation and Projection (UMAP), and Principal Components Analysis (PCA)) as well as differential expression analysis between cancer types. Overall, TOs recapitulated expected transcriptional programs of their tumor type as evidenced by UMAP and PCA as well as upregulation of defining pathways, such as estrogen receptor pathways in breast cancer TOs (ssGSEA p-values ranging from 0.004 to 5 × 10−5 for 5 gene ontology estrogen response pathways when compared to non-breast cancer TOs). Conclusion Determining genomic and transcriptomic concordance of TOs to source tumors is essential to confirm the validity of a given patient derived model. Our approach establishes metrics through key genomic features identified from routine next-generation sequencing data and can be extended beyond model validation to tracking clonal evolution over time in the presence or absence of therapeutic selection pressures. Our metrics may also serve as a critical quality control step if TOs are utilized in the clinical setting for personalized medicine. Citation Format: Brandon L. Mapes, Joshua SK Bell, Lee F. Langer, Robert Huether, Catherine Igartua, Veronica Sanchez-Freire, Robert Tell, Jeffrey A. Borgia, Ashiq Masood, Ameen A. Salahudeen. Universal genetic and transcriptomic concordance metrics to validate patient-derived tumor organoid models [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3908.

Abstract Gene fusions can serve as key drivers in the development of various cancers and represent important therapeutic targets and diagnostic biomarkers. Due to high detection of candidate fusions from RNA-sequencing data, there is a recognized need to build tools that will make reasonable and automated predictions to identify clinically or biologically relevant fusion events in a tumor sample. We developed a computational pipeline which scores and prioritizes all detected fusion transcripts within a sample to determine which fusions are likely driver events in the tumor. Specifically, the pipeline implements a categorization scheme that bins all scored fusion events into Low, Medium and High Confidence levels based on threshold read support levels and a DriverScore metric, which is derived from a binary classification algorithm using specific features, such as reading frame, breakpoint region, kinase domain and transcript isoform. We systematically analyzed 3200 fusion candidates from a previously published cohort of 500 paired tumor-normal samples sequenced with the Tempus xT assay. We found that 1.7% and 20.1% of fusion candidates were categorized in the High and Medium Confidence levels, respectively, while 78.2% of fusion events were deprioritized as Low Confidence calls. Of the 35 clinically-relevant fusions, 27 (77.1%) were captured in our prioritized set (High/Medium Confidence), including National Comprehensive Cancer Network actionable gene rearrangements involving RET, STAT6 and FUS, while the remaining 8 were assigned as Low Confidence due to an out of frame fusion transcript and insufficient read support. The frequency of prioritized fusions varied by cancer type, with prostate and breast cancer having the highest frequency of prioritized fusions. In addition to well-established canonical fusions, we also sought to characterize novel fusions, identifying a subset of 21 novel prioritized fusions which were also observed in The Cancer Genome Atlas tumor samples. Within this subset, 3% of fusion candidates contained a druggable domain such as a tyrosine kinase or Ras-binding domain, signifying the potential of categorization to enable novel fusion drug target discovery. Overall, our analysis highlights the utility of using an automated prioritization tool to detect known canonical fusion drivers and explore novel fusion drug targets and biomarkers. Citation Format: Sumaiya A. Islam, Robert Huether, Emily Kudalkar. Identification of novel druggable fusions enabled through the use of an automated RNA fusion prioritization pipeline [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5474.

Key Points Question Are filtering approaches an appropriate alternative to germline mutation subtraction for calculating tumor mutational burden (TMB)? Findings In this cohort study of 50 tumor samples comparing TMB calculated using 3 filtering approaches with germline-subtracted TMB, no strong association was found between TMB calculated using any filtering method and germline-subtracted TMB. Meaning These findings suggest that tumor-only methods of calculation may falsely overestimate TMB, potentially affecting patient care and treatment outcomes adversely; germline subtraction may more accurately measure TMB.

Objective/BackgroundWe performed a retrospective analysis of longitudinal real-world data (RWD) from patients with breast cancer to replicate results from clinical studies and demonstrate the feasibility of generating real-world evidence. We also assessed the value of transcriptome profiling as a complementary tool for determining molecular subtypes.MethodsDe-identified, longitudinal data were analyzed after abstraction from records of patients with breast cancer in the United States (US) structured and stored in the Tempus database. Demographics, clinical characteristics, molecular subtype, treatment history, and survival outcomes were assessed according to strict qualitative criteria. RNA sequencing and clinical data were used to predict molecular subtypes and signaling pathway enrichment.ResultsThe clinical abstraction cohort (n = 4000) mirrored the demographics and clinical characteristics of patients with breast cancer in the US, indicating feasibility for RWE generation. Among patients who were human epidermal growth factor receptor 2-positive (HER2+), 74.2% received anti-HER2 therapy, with ∼70% starting within 3 months of a positive test result. Most non-treated patients were early stage. In this RWD set, 31.7% of patients with HER2+ immunohistochemistry (IHC) had discordant fluorescence in situ hybridization results recorded. Among patients with multiple HER2 IHC results at diagnosis, 18.6% exhibited intra-test discordance. Through development of a whole-transcriptome model to predict IHC receptor status in the molecular sequenced cohort (n = 400), molecular subtypes were resolved for all patients (n = 36) with equivocal HER2 statuses from abstracted test results. Receptor-related signaling pathways were differentially enriched between clinical molecular subtypes.ConclusionsRWD in the Tempus database mirrors the overall population of patients with breast cancer in the US. These results suggest that real-time, RWD analyses are feasible in a large, highly heterogeneous database. Furthermore, molecular data may aid deficiencies and discrepancies observed from breast cancer RWD.

Genomic analysis of paired tumor-normal samples and clinical data can be used to match patients to cancer therapies or clinical trials. We analyzed 500 patient samples across diverse tumor types using the Tempus xT platform by DNA-seq, RNA-seq and immunological biomarkers. The use of a tumor and germline dataset led to substantial improvements in mutation identification and a reduction in false-positive rates. RNA-seq enhanced gene fusion detection and cancer type classifications. With DNA-seq alone, 29.6% of patients matched to precision therapies supported by high levels of evidence or by well-powered studies. This proportion increased to 43.4% with the addition of RNA-seq and immunotherapy biomarker results. Combining these data with clinical criteria, 76.8% of patients were matched to at least one relevant clinical trial on the basis of biomarkers measured by the xT assay. These results indicate that extensive molecular profiling combined with clinical data identifies personalized therapies and clinical trials for a large proportion of patients with cancer and that paired tumor-normal plus transcriptome sequencing outperforms tumor-only DNA panel testing.

We developed and clinically validated a hybrid capture next generation sequencing assay to detect somatic alterations and microsatellite instability in solid tumors and hematologic malignancies. This targeted oncology assay utilizes tumor-normal matched samples for highly accurate somatic alteration calling and whole transcriptome RNA sequencing for unbiased identification of gene fusion events. The assay was validated with a combination of clinical specimens and cell lines, and recorded a sensitivity of 99.1% for single nucleotide variants, 98.1% for indels, 99.9% for gene rearrangements, 98.4% for copy number variations, and 99.9% for microsatellite instability detection. This assay presents a wide array of data for clinical management and clinical trial enrollment while conserving limited tissue.

280 Background: Pancreatic cancer is being increasingly associated with germline implications. Some large single-center studies have reported results ranging from 3.9% to 19.8% of patients found to have germline variants [Shindo, JCO 2017; Lowery, JNCI 2018]. Due to this wide range, we aim to further delineate prevalence of deleterious germline mutations in pancreatic cancer using a multi-institutional data set. We also aim to analyze predictive factors such as mutant allele frequency (MAF, in %) in germline versus somatic calls. Methods: We sequenced 23 genes in DNA prepared from clinical tissue and blood specimens submitted to Tempus Labs. Germline variants and somatic variants were processed separately. Germline variants were determined to be deleterious through the sum effect of a combination of in silico predictors, population databases, and internal evaluations. Tumor-normal comparisons were used to define somatic versus germline, and MAFs were calculated for each. Results: A total of 234 patient samples from 17 institutions were analyzed. Of these, 12 (5.1%) had predicted deleterious germline variants involving 8 different genes: BRCA1 (n = 3), CHEK2 (n = 3), ATM (n = 1), MLH1 (n = 1), MUTYH (n = 1), PALB2 (n = 1), SMAD4 (n = 1), TP53 (n = 1). For most somatic alterations, the MAFs were found to be greater than the germline deleterious alterations, with the latter approaching ~50% in most cases (Table). Conclusions: This multi-institutional study identifies 5% of patients with pancreatic cancer to have deleterious germline alterations. Somatic variant testing, particularly when paired with germline, can be used as a screening method for genetic counseling referrals, especially with MAF analyses of paired tumor-normal samples. [Table: see text]