Genomic Subtypes Research Articles

Abstract 3930: The therapeutic implications of consensus genomic subtypes of gastric adenocarcinoma

Abstract Background: Gastric adenocarcinoma (GAC) is heterogeneous lethal disease in genomic and clinical level. Although the clinical relevance of genomic and molecular subtypes of GAC has been demonstrated, differences in classification methods have made it difficult to use these findings in clinical applications. We aim to examine a consensus of genomic subtypes and correlate them with clinical outcomes. Method: We collected genomic data from 2527 GAC tumors and divided the data into discovery (n = 1427) and validation sets (n = 1100). A cluster of clusters approach (COCA) was used to find consensus subtypes of gastric tumors. We constructed the GAC predictor of the integrated consensus subtype with 120 genes (GPICS120) and applied it to the validation data set in order to validate the clinical significance of the new subtypes. By analyzing genomic and proteomic data we evaluated each subtype's potential response rate to standard and experimental treatments such radiation therapy, targeted therapy, and immunotherapy and further validated their functional significance in cell line models. Results: We identified 6 clinically and molecularly distinct genomic consensus subtypes (CGSs) by integrating 8 previously established genomic subtypes. CGS1 is characterized by worst prognosis, remarkably high stem cell characteristics, high IGF1 expression, and low genomic alterations. CGS2 showed canonical epithelial gene expression patterns. CGS3 and CGS4 are characterized by high copy number alterations and low immune activity. CGS5 has highest mutation burden and moderately elevated immune activity that is characteristics of MSI-high tumors. The majority of CGS6 tumors have extremely high methylation, high immune activity, and are EBV-positive. The most intriguing finding is that CGS1 is the most responsive to immunotherapy whereas CGS3 is significantly associated with benefit of chemoradiation therapy due to high basal level of ferroptosis. Furthermore, we showed that ferroptosis inducer statin sensitized radiation-resistant GAC cells to ionizing radiation, suggesting that statin can be used to enhance the efficacy of chemoradiation in GAC and it would warrant further investigation in preclinical models. Conclusion: Consensus subtype is robust classification system and can serve as the basis for future clinical trials as well as pre-clinical research into targeted subtype-based therapies. Citation Format: Yun Seong Jeong, Young-Gyu Eun, Sung Hwan Lee, Sang-Hee Kang, Sun Young Yim, Eui Hyun Kim, Joo Kyung Noh, Bo Hwa Sohn, Ju-Seog Lee. The therapeutic implications of consensus genomic subtypes of gastric adenocarcinoma [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 3930.

Abstract 3361: Surfaceome and cancer testis antigen profiling of lung adenocarcinoma by large-scale transcriptomic analysis

Abstract Background: The emergence of improved modalities to target tumor specific and tumor associated antigens in solid tumors (e.g. antibody-drug conjugates, bi-specific antibodies, and adoptive cell therapy) underscores the need to efficiently identify and validate new therapeutic targets. Methods: We examined differential expression of surfaceome genes (Fonseca et al., 2016) and cancer testis (CT) antigen genes (Almeida et al., 2009) among major lung adenocarcinoma genomic subtypes; conducting comprehensive surfaceome/CT antigen profiling of 9002 NSCLC adenocarcinoma samples using data generated by RNA sequencing (whole transcriptome) at Caris Life Sciences (Phoenix, AZ). NSCLC tumors were stratified into 5 subgroups based on the presence of specific driver mutations: ALK fusion (N=325), EGFR mutant (N=1739), KRAS mutant (no STK11 or KEAP1 mutation; N=2561), KRAS + STK or KEAP1 mutation (N=1264), and Pan-WT (no mutation in ALK, EGFR, KRAS, STK11, or KEAP1 detected; N=3113). Fold-change in expression was calculated by dividing the median value for a specific subgroup by the median of all other groups combined. Significance was tested by Mann-Whitney U test, with p-values adjust for multiple hypothesis testing (Benjamini-Hochberg). Results: Surfaceome genes with the highest median expression overall included CD74, TMBIM6, and CD44, while the highest median expression among CT antigens was observed for SPAG9, KIAA0100, and RQCD1, yet the expression of these genes across the 5 driver mutation-based subgroups was similar (fold-changes ranged 0.93-1.06). To identify driver-associated candidate genes with relatively high expression, genes with subgroup-specific median expression >75th-percentile and with significant differences when compared to all other groups were ranked by fold-change. While several candidate surfaceome genes were associated with a specific subgroup, CLDN10 was identified among the top 3 candidates for both ALK fusion and KRAS + STK11/KEAP1 mutant subgroups: ALK fusion (GRIN2A, 3.03-fold; CLDN10, 2.53-fold; SLITRK6, 1.86-fold), EGFR mutant (SFTPC, 3.82-fold; MUC21, 1.99-fold; ATP13A4, 1.60-fold), KRAS mutant (VSIG1, 1.86-fold; AREG, 1.47-fold; C15orf48, 1.25-fold), and KRAS + STK11/KEAP1 mutant (ABCC2, 4.03-fold; MUC13, 3.21-fold; CLDN10, 3.00-fold). The top driver-associated CT antigen candidates included LEMD1 (1.42-fold) in ALK fusion, TMEM108 (1.27-fold) in EGFR mutant, XAGE1D (1.7-fold) in KRAS mutant, and CABYR (1.52-fold) in KRAS + STK11/KEAP1 mutant subgroups. Conclusion: Our analysis revealed several surfaceome and CT antigen candidate genes with relatively high expression across lung adenocarcinoma, as well as differential expression of candidate genes within specific genomic subtypes. These findings can help prioritize the validation and development of novel therapeutic targets in lung adenocarcinoma. Citation Format: Swartz Andrew, Andrew Elliott, George W. Sledge, Stephen V. Liu, Balazs Halmos, Vincent K. Lam, Taofeek K. Owonikoko. Surfaceome and cancer testis antigen profiling of lung adenocarcinoma by large-scale transcriptomic analysis [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 3361.

Abstract 6248: Multilayered molecular interactions underlying glioblastoma pathogenesis through multi-omics analysis

Abstract Glioblastoma, IDH-wild type (GBM) is the most common malignant and treatment-refractory brain tumor with a dismal prognosis. Despite the identification of various driver genes in previous large-scale genomic analyses, there has been limited progress in the development of novel treatments. Acquiring an extensive understanding of the molecular mechanisms underlying GBM pathogenesis is crucial to improve prognostic outcomes. We analyzed 289 whole-genome sequencing data (WGS) including 159 unpublished deep WGS (≥ ×120 coverage) along with RNA-seq, DNA methylation array, and assay for transposase-accessible chromatin with sequencing (ATAC-seq) to uncover the molecular mechanisms regulating development and progression of GBM. Our deep WGS enables us to delineate a fine view of clonal architecture, where mutational signature varies between clonal and subclonal mutations, suggesting that different mutational processes contribute to GBM pathogenesis depending on its developmental stage. The differentiation status of tumor cells detected by transcriptional deconvolution analysis is associated with multi-layer profiles including genomic alterations, expression subtypes, DNA methylation, and open chromatin status. Tumors predominantly comprised of differentiated cells display genetic and epigenetic profiles that align with the classical subtype, whereas tumors predominantly composed of stem-like cells exhibit profiles consistent with the proneural subtype. ATAC-seq unveils genome-wide chromatin accessibility features associated with gene expression subtypes. Motif enrichment analysis of differentially accessible sites among the subtypes identified specific transcription factors. The proneural subtype is enriched for the SOX10 binding motif, which is associated with regulating cell states, and exhibits higher expression of SOX10 compared to other subtypes. Conversely, the mesenchymal and classical subtypes are enriched for the CREB1 binding motif, which promotes cell proliferation and angiogenesis through TGFβ2 upregulation, wherein TGFβ2 expression is significantly elevated. These findings support a model in which the difference in chromatin structure also regulates the progression of GBM.Our analysis reveals the multi-layer molecular interactions underlying the progression of GBM enhancing our understanding of the pathogenesis. Citation Format: Takuma Nakashima, Yusuke Funakoshi, Ryo Yamamoto, Yuriko Sugihara, Shohei Nambu, Yoshiki Arakawa, Shota Tanaka, Joji Ishida, Ryuta Saito, Ryosuke Hanaya, Koji Yoshimoto, Yoshitaka Narita, Hiromichi Suzuki. Multilayered molecular interactions underlying glioblastoma pathogenesis through multi-omics analysis [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 6248.

Abstract 6186: Histopathology-derived molecular portrait of serous ovarian cancer

Abstract Background: The progression of ovarian cancer (OV) is often evaluated through clinical staging and molecular profiling of genomic alterations and subtypes, with the molecular profiling of tumors playing a crucial role in cancer management by informing therapeutic decisions. However, collecting molecular data is expensive, time-consuming, and can be sensitive to sample handling protocols. These factors prevent routine implementation at many care centers, thereby limiting benefit to patients. This work examines the viability of an image-centric method for inferring molecular endophenotypes and predicting overall survival in OV, leveraging commonly gathered histopathology scans. Methods: Embedding representations of digitized Hematoxylin and Eosin slides were produced using a self-supervised histology model. First, we assessed the ability of the embeddings to predict overall survival using Cox proportional hazards models by calculating the concordance index. Next, these embeddings served as input features to machine learning models trained to predict genomic patterns within tumors, including molecular subtypes, driver gene alterations, and mutational processes, leading to a more comprehensive analysis of the prognostic potential of our approach. Results: Histopathological embeddings were found to be more predictive of patient survival than traditional factors such as age, histological grade, and clinical stage. Embedding risk combined with clinicopathological risk had a higher predictive value than clinicopathological risk alone (0.578 vs 0.569, p = 0.025). Moreover, embedding risk alone was more predictive of survival (0.577 vs 0.569, p = 0.038). Risk scores derived from embeddings and all clinicopathological features were uncorrelated (R = 0.03), indicating that embeddings provide unique prognostic information. Motivated by these findings, we investigated the reasons behind the enhanced accuracy in survival prediction. We found that models trained on embeddings predicted previously defined molecular subtypes (TCGA, 2011), achieving area under the receiver operating characteristic curve (AUROC) scores as follows: Differentiated (0.86), Immunoreactive (0.85), Mesenchymal (0.9), and Proliferative (0.89). Similarly, our models identified crucial oncogenic events, including the loss of TP53 and CDKN2A, and instances of whole genome duplication (AUROCs of 0.85, 0.99 and 0.8). Conclusions: Taken together, this work showcases the potential utility of standard of care digitized histopathology scans for imputing genomic patterns and predicting overall survival in OV. Our results suggest that imputation could provide a fast and scalable solution for deploying patient screening biomarkers, thereby playing a pivotal role in shaping the future of precision medicine. Citation Format: Yajas Shah, Zachary R. McCaw, Anna Shcherbina, Christopher Probert. Histopathology-derived molecular portrait of serous ovarian cancer [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 6186.

Abstract 5672: Transcriptomic profiling of lung adenocarcinoma from never-smokers reveals molecular subtypes with clinical implications

Abstract Lung cancer is the leading cause of cancer mortality worldwide and about 10-25% of lung cancers are from never-smokers (LCINS). Previous genomic studies of LCINS identified multiple genomic subtypes with different genetic drivers and evolutionary processes. To dissect the profile of LCINS cell states and tumor microenvironment, and their relationship with genomic lesions, we sequenced and assembled a large RNA-seq data set of 685 samples of lung adenocarcinoma from never-smokers and integrated them with whole genome sequencing data, pathological features and clinical outcomes from the same subjects. We identified three transcriptomic subtypes defined by distinct gene expression patterns. The three subtypes were associated with different pathway activities and showed remarkable heterogeneity in cell composition, lineage fidelity and pathological features. The genomic driver events and mutational signatures were significantly enriched in specific subtypes. Clinical outcomes also differed across the subtypes. For example, one subtype had prolonged overall survival and was associated with predicted response to immune checkpoint blockade. This study emphasizes the importance of transcriptome-based classification of LCINS, which has profound clinical implications beyond those provided by genomic and pathological assessment. Citation Format: Wei Zhao, Tongwu Zhang, Phuc H. Hoang, Jian Sang, Maria Teresa Landi. Transcriptomic profiling of lung adenocarcinoma from never-smokers reveals molecular subtypes with clinical implications [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 5672.

Machine learning-based extrachromosomal DNA identification in large-scale cohorts reveals its clinical implications in cancer.

The clinical implications of extrachromosomal DNA (ecDNA) in cancer therapy remain largely elusive. Here, we present a comprehensive analysis of ecDNA amplification spectra and their association with clinical and molecular features in multiple cohorts comprising over 13,000 pan-cancer patients. Using our developed computational framework, GCAP, and validating it with multifaceted approaches, we reveal a consistent pan-cancer pattern of mutual exclusivity between ecDNA amplification and microsatellite instability (MSI). In addition, we establish the role of ecDNA amplification as a risk factor and refine genomic subtypes in a cohort from 1015 colorectal cancer patients. Importantly, our investigation incorporates data from four clinical trials focused on anti-PD-1 immunotherapy, demonstrating the pivotal role of ecDNA amplification as a biomarker for guiding checkpoint blockade immunotherapy in gastrointestinal cancer. This finding represents clinical evidence linking ecDNA amplification to the effectiveness of immunotherapeutic interventions. Overall, our study provides a proof-of-concept of identifying ecDNA amplification from cancer whole-exome sequencing (WES) data, highlighting the potential of ecDNA amplification as a valuable biomarker for facilitating personalized cancer treatment.

Characterizing Immune Infiltration in Esthesioneuroblastoma Subtypes Through Gene Expression Deconvolution

BackgroundEsthesioneuroblastoma (ENB) is a rare cancer deriving from the olfactory mucosa. Among the basal or neural genomic subtypes, the basal subtype is associated with poorer survival, and poor differentiation, and higher levels of tumor-infiltrating immune cells (TIICs). The immune microenvironment of these ENB subtypes remains unclear. We used an established machine-learning algorithm on ENB transcriptomic profiles. MethodsThe authors characterized 22 immune cell populations using the CIBERSORTx deconvolutional machine-learning pipeline on RNAseq data from 18 ENB cases. The characterization aimed to elucidate differences in relative proportions and populations of TIICs between basal and neural ENB. ResultsNo differences in age, Hyams, Dulguerov, IDH2 mutation, or PD-L1 expression were seen between basal and neural subtypes of ENB (p>0.05). No difference in median overall survival was also appreciated (52.0±13.1 vs. 50.0±43.2 months, p=0.5). As a cohort, M2 macrophages were the most abundant subpopulation (14%) followed by naïve B-cells (13%) and CD4 memory resting T cells (12%). No gross differences on CD20, CD4, or CD8 cells/mm2 were apparent on gross histology (p>0.05). However, further analysis showed that activated CD4 memory T cells were significantly increased in the basal ENBs, whereas resting dendritic cells were increased in the neural ENB subtype. The TIIC profiles alone could not differentiate between basal and neural ENB but did suggest immunoprofile differences. ConclusionBasal and neural subtypes display distinct TIIC involvement, which may impact their difference in outcome. These findings provide the framework for further investigation in novel immunomodulation strategies for ENB.

Real-World comprehensive genomic profiling data for diagnostic clarity in pulmonary Large-Cell neuroendocrine carcinoma

BackgroundPulmonary large-cell neuroendocrine carcinoma (LCNEC) is an uncommon subtype of lung cancer believed to represent a spectrum of tumors sharing characteristics of both small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). Other groups have proposed genomic LCNEC subtypes, including small cell-like, non-small cell-like, and carcinoid-like subtypes. The primary goal of this study was to better define the NSCLC-like subtype with comprehensive genomic profiling (CGP). MethodsAn institutional database was queried to identify tissue specimens (TBx, N = 1,426) and liquid biopsies (LBx, N = 39) submitted for CGP during routine clinical care (8/2014 – 7/2023) with a disease ontology of LCNEC. TBx were profiled with FoundationOne® (F1) or F1CDx, using hybrid-capture technology to detect genomic alterations (GAs). Results1,426 LCNEC samples were genomically profiled. The presence of RB1 and TP53 genomic alterations (GAs) were used to define a SCLC-like subtype (n = 557). A carcinoid-like group was defined by the presence of MEN1 mutation in the absence of TP53 GAs (n = 25). The remaining 844 samples were compared to the SCLC-like group and GAs enriched relative to the SCLC-like samples with a false discovery rate (FDR) < 0.0001 were used to define a NSCLC-like group. These NSCLC-like subtype-defining GAs included SMARCA4, KRAS, FGF3/4/19, STK11, CDKN2A/B, MTAP, and CCND1. Under this schema, 530 samples were classified as NSCLC-like and 314 remained unclassified. ConclusionsLarge-scale CGP can better characterize biologically distinct molecular subtypes in LCNEC. Further studies to define how these molecular subtypes may help inform treatment decisions in this complex and challenging malignancy are warranted.

Cervical Cytology Preserves Histologically Detected Surface Epithelial Slackening, Unique to the POLE Mutation-subtype in Endometrial Cancer.

Among the four genomic subtypes of endometrial cancer, distinguishing between the DNA polymerase epsilon mutation (POLEmut) and no specific molecular profile (NSMP) subtypes requires genomic profiling owing to the lack of surrogate immunohistochemical markers. We have previously found that, histologically, the POLEmut-subtype exhibits surface epithelial slackening (SES). Therefore, to improve subtype identification, we aimed to extract cytological features corresponding to SES in POLEmut-subtype cervical cytology specimens. We analyzed 104 endometrial cancer cervical cytology specimens, with integrative diagnosis confirmation via histology, immunohistochemistry, and genomic profiling. Cytological features were evaluated for the presence of atypical glandular cells, atypical cell appearance in single cells and clusters, and cytological SES and the presence of tumor-infiltrating inflammatory cells in clusters. Based on cervical cytology, the POLEmut- and p53mut-subtypes exhibited more frequent atypical cells in smaller clusters, giant tumor cells, and cytological SES patterns than the NSMP-subtype. Tumor-infiltrating lymphocytes were frequent in the POLEmut- and mismatch repair-deficient subtypes. Histologically-detected SES as well as other endometrial cancer features may be preserved in the atypical cell clusters observed in cervical cytology specimens. Cytological detection of SES and of smaller clusters of atypical cells and inflammatory cells with moderate atypia are suggestive of POLEmut-subtype. Integrative diagnosis including genomic profiling remains critical for diagnostic confirmation.

Promising and Minimally Invasive Biomarkers: Targeting Melanoma.

The therapeutic landscape of malignant melanoma has been radically reformed in recent years, with novel treatments emerging in both the field of cancer immunotherapy and signalling pathway inhibition. Large-scale tumour genomic characterization has accurately classified malignant melanoma into four different genomic subtypes so far. Despite this, only somatic mutations in BRAF oncogene, as assessed in tumour biopsies, has so far become a validated predictive biomarker of treatment with small molecule inhibitors. The biology of tumour evolution and heterogeneity has uncovered the current limitations associated with decoding genomic drivers based only on a single-site tumour biopsy. There is an urgent need to develop minimally invasive biomarkers that accurately reflect the real-time evolution of melanoma and that allow for streamlined collection, analysis, and interpretation. These will enable us to face challenges with tumour tissue attainment and process and will fulfil the vision of utilizing "liquid biopsy" to guide clinical decisions, in a manner akin to how it is used in the management of haematological malignancies. In this review, we will summarize the most recent published evidence on the role of minimally invasive biomarkers in melanoma, commenting on their future potential to lead to practice-changing discoveries.

Genomic and epigenomic integrative subtypes of renal cell carcinoma in a Japanese cohort

Renal cell carcinoma (RCC) comprises several histological types characterised by different genomic and epigenomic aberrations; however, the molecular pathogenesis of each type still requires further exploration. We perform whole-genome sequencing of 128 Japanese RCC cases of different histology to elucidate the significant somatic alterations and mutagenesis processes. We also perform transcriptomic and epigenomic sequencing to identify distinguishing features, including assay for transposase-accessible chromatin sequencing (ATAC-seq) and methyl sequencing. Genomic analysis reveals that the mutational signature differs among the histological types, suggesting that different carcinogenic factors drive each histology. From the ATAC-seq results, master transcription factors are identified for each histology. Furthermore, clear cell RCC is classified into three epi-subtypes, one of which expresses highly immune checkpoint molecules with frequent loss of chromosome 14q. These genomic and epigenomic features may lead to the development of effective therapeutic strategies for RCC.

Distinct Hodgkin lymphoma subtypes defined by noninvasive genomic profiling.

The scarcity of malignant Hodgkin and Reed-Sternberg cells hampers tissue-based comprehensive genomic profiling of classic Hodgkin lymphoma (cHL). By contrast, liquid biopsies show promise for molecular profiling of cHL due to relatively high circulating tumour DNA (ctDNA) levels1-4. Here we show that the plasma representation of mutations exceeds the bulk tumour representation in most cases, making cHL particularly amenable to noninvasive profiling. Leveraging single-cell transcriptional profiles of cHL tumours, we demonstrate Hodgkin and Reed-Sternberg ctDNA shedding to be shaped by DNASE1L3, whose increased tumour microenvironment-derived expression drives high ctDNA concentrations. Using this insight, we comprehensively profile 366 patients, revealing two distinct cHL genomic subtypes with characteristic clinical and prognostic correlates, as well as distinct transcriptional and immunological profiles. Furthermore, we identify a novel class of truncating IL4R mutations that are dependent on IL-13 signalling and therapeutically targetable with IL-4Rα-blocking antibodies. Finally, using PhasED-seq5, we demonstrate the clinical value of pretreatment and on-treatment ctDNA levels for longitudinally refining cHL risk prediction and for detection of radiographically occult minimal residual disease. Collectively, these results support the utility of noninvasive strategies for genotyping and dynamic monitoring of cHL, as well as capturing molecularly distinct subtypes with diagnostic, prognostic and therapeutic potential.

Abstract C012: Integration of social variables resolves subclasses of tumor microenvironment

Abstract Intro: Recent work has identified subclasses of the tumor microenvironment, with most work emphasizing a robust adaptive immune response associated with ER-negative and/or basal-like tumors. Since adaptive immune responses may influence response to chemotherapy and may be targetable by immunotherapy, understanding population-level heterogeneity in immune response is important. Methods: We leveraged the Carolina Breast Cancer Study (CBCS), a large population-based study that oversampled young (≤50 years) and Black women with invasive breast cancer (BC). We performed NanoString gene expression profiling on tissue from 1,329 BC patients, including 704 (53%) Black and 625 (47%) non-Black women, and assayed genes representative of at least 10 individual immune cell types from both adaptive and innate arms of the immune system, homologous recombination deficiency signature and TP53 mutation status. Using a latent class approach that included variables for self-reported race, triple negative status, individual measures of socioeconomic status and community-level measures of advantage and disadvantage, we identified latent classes that integrated tumor microenvironment with tumor biology, individual and community level exposures under a Cells-to-Society conceptual model of breast cancer outcomes. Results: Latent Class Analysis identified 6 integrated classes of the tumor microenvironment, further subdividing a previously identified and validated adaptive immune class into two subgroups. The two adaptive classes were distinguished by different expression levels of B-cells, CD8+ T cells, cytotoxic cells and PD-L1, and differing frequency of basal-like subtype (Adaptive Class 1 RFD [95% CI]: 19.0 [11.9, 26.2]; Adaptive Class 2 RFD [95% CI]: 32.8 [23.3,42.3]) and presence of genomic instability gene expression subtypes (Adaptive Class-1 RFD [95% CI]: 22.1 [14.0, 30.1]; Adaptive Class-2 RFD [95% CI]: 37.5 [28.1,46.2]) relative to the immune-quiet tumors. Discussion: Social variables have complex effects on tumor biology and microenvironment, but latent class models and other model-based approaches may help to integrate biological data with other social factors to identify susceptible subgroups and to optimize precision medicine in diverse populations. Citation Format: Alina M. Hamilton, Kara McCormack, Marc A. Emerson, Melissa A. Troester, Terry Hyslop. Integration of social variables resolves subclasses of tumor microenvironment [abstract]. In: Proceedings of the 16th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2023 Sep 29-Oct 2;Orlando, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2023;32(12 Suppl):Abstract nr C012.

Improved MRD Negativity Rates in Adverse Genomic Risk B-ALL Patients with Chemotherapy / Blinatumomab Induction: Experience from the Australasian Leukaemia Lymphoma Group (ALLG) ALL06/09 Studies

Introduction: Despite the benefits conferred by pediatric treatment protocols, outcomes in adolescent and young adult (AYA) acute lymphoblastic leukemia (ALL) remain inferior to children. Poor outcomes in AYA ALL have been associated with reduced tolerability of intensive cytotoxics and higher incidence of adverse risk disease. Our earlier ALLG ALL06 study demonstrated intensive pediatric induction/consolidation could be delivered in a similar time frame as in children suggesting tolerability was not a major contributor to poor outcome in AYA ALL. ALL06 established post consolidation (TP2=d79) minimal residual disease (MRD) positivity (MRD+) and high BMI was predictive of inferior disease free- and overall survival (OS). In order to improve TP2 MRD negativity (MRD-) in B-cell disease, the ALLG ALL09 study substituted standard consolidation for blinatumomab (blin) in phase I and II of a BFM-style protocol . We present the impact of blin on outcomes in genomic cohorts from ALL09 and compare these to similar cohorts from ALL06 treated with standard chemotherapy. Methods: Between Sep 2020 and Apr 2022, 55 patientswere enrolled to ALL09. Of these, 45 had appropriate samples for genomic analyses and 39/45 had an MRD marker identified. These 39 patients formed the genomic cohort presented below. Genomic subtyping was performed using whole transcriptome sequencing, MLPA, karyotype and FISH analyses, and disease was classified into genomic adverse (AR) vs standard (SR) risk based on contemporary literature classifiers. Bone marrow MRD was assessed using ASO-PCR for IGH and TCR rearrangements, at d33 post induction (TP1), and TP2 post blin consolidation. Differences in TP2 MRD were assessed, with estimated 3 yr relapse free survival (RFS) and OS calculated using the Kaplan-Meier method. Results: In the ALL09 cohort, median follow up was 629 days with 3 yr RFS 84.5% and OS 87.1%. In the genomic cohort, median follow up was 602 days with 3 yr RFS 88.8% and OS 93.2%, respectively. These compared favorably to 3 yr RFS and OS (both 67.5%) in the B-cell genomic cohort from ALL06. Between ALL09 and ALL06 there were similar proportions of AR patients (73% and 62.5%, respectively). In ALL09, the most frequent subtypes were PAX5alt (18%) and PAX5 p.P80R (11%) versus KMT2Ar and DUX4r (both 17.5%) in ALL06. In ALL09, Ph-like disease represented 13% vs 12.5% in ALL06. In the ALL09 genomic cohort, the MRD+ rate reduced between TP1 and TP2 (p&amp;lt; 0.001), with the reduction in the rate of positivity being more pronounced in the AR cohort (AR MRD+ TP1=85% vs AR TP2=39%; p&amp;lt; 0.001). In contrast, in ALL06 MRD+ in the AR cohort reduced by only 28% between TP1 and TP2 (p &amp;gt;0.05). This suggests a positive impact of blin consolidation in AR patients in ALL09. Despite improved AR MRD+ rates overall, not all genomic subtypes in ALL09 followed this trend (Table 1). To date, there are 5 relapses in the ALL09 cohort, 4 from the genomic cohort. Of these, 1 was hypodiploid and strikingly the remaining 3 patients harbored a TCF3 rearrangement ( TCF3r). All 4 relapses were significantly associated with MRD+ at TP2 (RFS: MRD-, 100% (n=14) vs MRD+, 71% (n=25); p= 0.013), and all were observed in the AR genomic group. Of interest, the fifth case, outside of the genomic cohort, was classified as PAX5alt. In ALL09, high-risk (HR) patients proceeded to intensive HR block chemotherapy with the aim of eliminating MRD prior to allogeneic transplant (SCT). In the genomics cohort 15 patients proceeded to HR therapy, n=12 were considered AR. Sustained MRD+ through HR blocks 1 and 2 was noted in patients harboring PAX5 p.P80R (3/3), KMT2Ar (2/2) and TCF3r (2/3). Of the 9 patients who proceeded to SCT, 3 had PAX5 p.P80R 1 had PAX5alt, 2 were KMT2Ar and there was 1 each of DUX4r, hypodiploid and TCF3r. The 3 patients who remained MRD+ in the SR cohort had PAX5 p.P80R and all 3 proceeded to SCT. Conclusion: This is the first report of the impact of blin consolidation on genomic subsets of de novo AYA ALL. Our study suggests blin sensitivity as assessed by TP2 MRD response may be genomic subtype dependent, with TCF3r associated with sustained MRD+, leading to eventual relapse (3/3 patients). In addition, the PAX5 p.P80R subset was also associated with sustained high levels of TP2 MRD+ (3/4 with an MRD marker) and SCT (3/5 patients). These results may assist with identification of AYA ALL subsets that could benefit from alternative approaches as blin becomes increasingly incorporated into front-line protocols.

STAG2/LMO2 Gamma-Delta (γδ) T-ALL: Identification and Characterization of an Extremely High Risk Group of T-ALL in the Very Young

Background The prognosis of pediatric T-cell acute lymphoblastic leukemia (T-ALL) has improved with minimal residual disease (MRD)-stratified therapy, however, gamma delta T cell receptor positive (γδ) T-ALL remains a high-risk (HR) group. Limited studies have explored the clinical and genomic characteristics of γδ T-ALL, prompting us to conduct a comprehensive analysis of this entity and to identify determinants of outcome. Methods Through a consortium of 13 groups, we assembled a cohort of 200 patients up to 25 years of age with γδ T-ALL enrolled in clinical trials between 2000 and 2018. Clinical data of patients with non-γδ T-ALL enrolled on the same clinical trials were collected (n = 1,067). Complete remission (CR) was defined when bone marrow (BM) showed M1 cytomorphology and/or MRD &amp;lt;1% without evidence of extramedullary disease at end of induction/consolidation (EOI/EOC) and failure to achieve CR was considered treatment failure. A total of 76 γδ T-ALL samples were analyzed by whole genome (WGS) and/or transcriptome (RNAseq) sequencing. Results The frequency of γδ T-ALL was 8.0% of T-ALL cases. Patients with γδ T-ALL exhibited a higher rate of poor prednisone response ( P&amp;lt;0.01), MRD &amp;gt;1% at day 15 ( P&amp;lt;0.01), at EOI ( P&amp;lt;0.01) and EOC ( P&amp;lt;0.01), compared to non-γδ T-ALL cases. Furthermore, patients with γδ T-ALL showed significantly worse 5-year event free survival (EFS, 65% v. 78%, P&amp;lt;0.01) and overall survival (OS, 77% vs 83%, P=0.048). Almost all relapses of γδ T-ALL were isolated BM, while the central nervous system was the main site of relapse in non-γδ T-ALL, suggesting slow treatment response and chemo-resistance to the current treatment in γδ T-ALL. However, γδ T-ALL showed a higher rate of toxic death during treatment (7.6% vs 4.0%, P&amp;lt;0.01), suggesting the need for different therapeutic strategies and risk-classification, rather than treatment intensification. Strikingly, patients less than 3 years of age with γδ T-ALL exhibited significantly poor EFS (33% v. 70% [3-10 years] and 73% [&amp;gt;10], P&amp;lt;0.01) and OS (49% v. 78% [3-10] and 82% [&amp;gt;10], P&amp;lt; 0.01) ( Fig. A), a difference not observed in non-γδ T-ALL. MRD &amp;gt;1% at EOI showed poor EFS (51% v. 96% [MRD&amp;lt;0.01%] and 91% [1%&amp;gt;MRD&amp;gt;0.01%], P&amp;lt;0.01) and OS (66%). Integrated analysis of WGS and RNAseq identified enrichment of several genomic subtypes in γδ T-ALL, including STAG2/LMO2, hyperdiploidy with recurrent gains of chromosomes 8, 10, 11, 13q and 19, a recently identified “LMO2 γδ-like” subtype with distinct gene expression and LMO2/MYC/MYCN alterations, TLX3-rearranged (-R), and PICALM::MLLT10. No TAL1 nor TLX1-R were detected. STAG2/LMO2 was associated with age at diagnosis before 3 years, and extremely poor outcome, with 4 out of 5 cases dying within three years of diagnosis ( Fig. B). Of 24 STAG2/LMO2 T-ALL (additional 5 non-γδ, 13 TCR unknown cases), 22 of which were diagnosed by age three. All STAG2/LMO2 cases had alterations resulting in LMO2 activation and STAG2 inactivation, most commonly a single rearrangement between these two genes, and upregulation of HBE1, the LIN28-let7 pathway and stem cell proliferation pathways, suggesting a fetal hematopoietic origin. STAG2 has a critical role in the maintenance of enhancer-promoter looping mediated by the cohesin complex. To examine the consequences of STAG2 alterations, we performed integrated genomic/epigenomic analysis of the STAG2/LMO2 (MOLT-14 and PER-117) and STAG2 knockout (KO)/addback T-ALL lines. Chromatin loop sizes defined by H3K27ac HiChIP was highest in STAG2/LMO2 lines compared to other T-ALL. Following restoration of STAG2 expression in MOLT-14, CD34 and ID1/2 were down-regulated and H3K27ac was enriched in pathways related to T-cell differentiation. STAG2 KO in the non- STAG2/LMO2, LMO2-activated line PF382 identified genes also upregulated in STAG2/LMO2 primary samples, including CDK4 and STAG1. STAG2 KO lines exhibited partial compensation of STAG2 binding sites by STAG1 and upregulation of γδ-related genes, RORC and ID1/3. High throughput screening of 2,050 small molecules identified efficacy of HDAC, CDK and PARP inhibitors in STAG2/LMO2 lines. Conclusion Very young onset γδ T-ALL, but not non-γδ T-ALL, is enriched for the STAG2/LMO2 subtype and is a very high risk form of T-ALL. STAG2 loss perturbs chromatin organization and hematopoietic differentiation. Moreover, we demonstrate efficacy of novel therapeutic approaches that are needed to cure this form of leukemia.

Distinct Circulating Genomic Features of Classical Hodgkin Lymphoma of Older Adults

Introduction: Despite significant progress in curing younger patients with classical Hodgkin lymphoma (cHL), older patients continue to have substantially inferior outcomes. The bimodal distribution of age at cHL diagnosis has suggested distinct underlying pathogenesis in older adults. For example, among older patients where EBV+ disease is more prevalent, EBV+ status is associated with worse outcomes, unlike younger patients where it confers more favorable prognosis (Keegan, JCO 2005). The reasons for this disparity remain enigmatic. To address these challenges, we used circulating tumor DNA (ctDNA) profiling, enabling noninvasive genotyping, risk stratification, and classification into distinct molecular subtypes (Spina, Blood 2018; Alig, ASH 2022). We leveraged several non-invasive techniques to systematically characterize the genomic peculiarities of elderly cHL patients distinguishing them from younger counterparts, with special attention to EBV status. Methods: We comprehensively profiled newly diagnosed cHL patients from 3 prospective clinical trials: LYSA-PVAB (NCT02414568) for subjects ≥60 years (Ghesquières, ASH 2019), AHL2011 (NCT01358747) for patients aged 16-59 years with advanced cHL, and BIO-LYMPH (NCT04417803) for patients ≥18 years. We used age ≥60 as a threshold for stratification of older versus younger patients, a common cutoff in clinical trials. We identified somatic single nucleotide variants using CAPP-Seq (Newman et al, Nature Biotech 2016); and analyzed copy number alterations with CANARy (Chabon, Nature 2020). Patients were categorized into high vs low pretreatment ctDNA levels using a predefined threshold (2.5 log hGE/mL; Kurtz, JCO 2018). We determined EBV status using EBER and/or LMP1 staining of FFPE tumor tissue or plasma EBV levels (&amp;gt;32 copies/mL) by VirCAPP-Seq (Garofalo, ASH 2022). Results: A total of 215 newly diagnosed cHL patients were evaluable, with a median age of 51 years, and 35% being older (≥60 years). In this cohort, 82% had advanced stage disease, 68% had high IPS (≥3), and 57% had extra nodal involvement. At 4-year follow-up, OS and PFS were 84% and 74%, respectively. At presentation, older patients (≥60y) had more advanced stage (93% vs 73%, p&amp;lt;0.001), higher IPS (78% vs 61%, p&amp;lt;0.001), and greater EBV+ prevalence (63% vs 36%, p&amp;lt;0.001). Older patients also experienced significantly more relapses (39% vs 17%, p&amp;lt;0.001). and mortality (29% vs 6%, p&amp;lt;0.001). The median pre-treatment ctDNA level was 982 hGE/mL, but not significantly different in older vs younger patients (p=0.28). Higher ctDNA levels were positively associated with risk-associated features including B symptoms (p=0.01), IPS (p&amp;lt;0.001), bulky disease (p=0.04), extra nodal disease (p=0.02) and correlated with TMTV (R=0.47, p&amp;lt;.001). Older patients had significantly fewer somatic mutations than younger counterparts (Median 107 vs 192 SNVs within targeted genomic regions, p&amp;lt;0.001; Fig. 1A top); among older patients, EBV+ cases had even fewer mutations than EBV- counterparts (Median 83 vs 142 SNVs, p=0.05). Older cHL cases harbored significantly more BCL2 mutations and recurrent copy number anomalies, but had significantly fewer mutations in STAT6, ITPKB, B2M, GNA13, PAX5 and XPO1 among others(all p&amp;lt;0.05 Fig. 1A bottom). The H2 genomic subtype prevalence, characterized by genomic instability as well as mutations in TP53, KMT2D, and BCL2 (Alig et al, ICML 2023), was higher among older patients (56%) than younger patients (21%) (p=0.01). Older H2 cHL showed no significant difference in EBV+ status compared to H1 counterparts (p=0.9). In older patients, multivariate Cox regression (including EBV status, ctDNA, and IPS) identified EBV+ status (HR=2.2, p=0.02) as an independent predictor of worse PFS and DFS outcomes (HR=2.5, p=0.02 and HR=2.7, p=0.04). When older cHL patients were stratified by baseline ctDNA level and EBV status, high-risk cases (EBV+ and ctDNA High) had particularly poor outcomes in terms of both PFS and DFS ( Fig. 1B). Conclusion: Older cHL patients exhibit unique genomic characteristics compared to younger counterparts, including more prevalent H2 tumor genetic subtype and EBV positivity. Identifying of a high-risk subgroup of older cHL underscores the importance of ctDNA levels and EBV status. We envision future trials of personalized biomarker driven treatment strategies could therapeutically target this risk, including by PD1 blockade.

Discovery and Development of CD70 As a Cellular Therapy Target in High Risk Multiple Myeloma

Background: Even in the BCMA CAR-T era, multiple myeloma (MM) patients with high-risk genotypes have the poorest outcomes. We thus leveraged MMRF CoMMpass data to identify potential novel surface immunotherapeutic targets in high-risk MM. As described below, we focused on CD70, an immunotherapeutic target known to be upregulated on many hematologic and solid tumors but minimally expressed on normal tissue. While CD70 was previously explored in MM (McEarchern et al, Clin Cancer Res (2008)), it was not further investigated due to heterogeneous expression in newly diagnosed samples. Here we revisit this target to suggest cellular therapies against CD70 could be highly beneficial in high-risk MM patients, currently the greatest unmet need in the field. Methods: Primary patient samples were obtained under an IRB-approved protocol. CAR-T lentiviral transduction was performed on paired CD4/CD8 human T cells from at least 3 healthy donors. Effector function was assessed by cytotoxicity against MM cell lines and CAR-T profiling for memory and exhaustion markers. In vivo efficiency was assessed using bioluminescence. ATAC-seq peaks were called using Promo and HOMER (data from Jin et al. Blood (2018)). Machine learning was performed using XGBoost approach applied to CoMMpass RNA-seq data with 80% of samples for training and 20% as test set. Results: Analyzing CoMMpass mRNA data (release IA19) we first evaluated surface protein encoding-genes upregulated in tumors at relapse compared to diagnosis. Among &amp;gt;100 significantly upregulated genes, we focused on CD70 given known promise as an immunotherapy target. Further analysis across genomic subtypes revealed a specific, marked upregulation of CD70 in t(4;14) and gain 1q ( Fig. 1), as well as R-ISS 3 at diagnosis. Flow cytometry profiling of 8 cell line models, including t(4;14) models KHM11, KMS26, and LP1, revealed consistent CD70 protein expression with antigen density similar to BCMA. Profiling of relapsed primary MM samples confirmed that 19 of 28 samples had CD138+ plasma cells positive for CD70 expression. Of 5 patients in whom both BCMA and CD70 were measured, antigen densities were also similar. To target CD70, we then designed 10 unique anti-CD70 CAR-T constructs. 9 of these were scFv-based binders derived from therapeutic antibody fragments currently in preclinical or clinical development. By a structure-guided design based on the CD70:CD27 co-crystal structure (PDB: 7KX0), we also derived a unique truncated fragment of CD27 as an alternative CD70 binding element. All CD70 CARs were efficacious in in vitro cytotoxicity assays targeting MM cell lines AMO1, MM1.S, LP1, and RPMI8226. We then selected the top 3 performing scFv based-CARs as well as our natural ligand (CD27) CAR-T for in vivo studies. All 4 CD70 CAR-Ts prolonged survival of MM.1S-bearing mice compared to empty CAR ( n=5 mice/arm, p&amp;lt;0.005 by log-rank), similar to BCMA CAR-T controls. Remarkably, the CD27-based CAR showed superior expansion (50-85x, p&amp;lt; 0.0001 by ANOVA) and persistence compared to scFv-based ( Fig. 2). These results underscore that targeting CD70 using a natural ligand-based CAR-T represents a promising preclinical candidate for treatment of relapsed and high risk MM. We further investigated why CD70 may be upregulated in high risk MM. Analysis of ChIP-seq in ENCODE and primary MM tumor ATAC-seq was used to nominate a set of transcription factors that bind the CD70 locus. Extending from CoMMpass data, we built a predictive machine learning model for CD70 expression as a function of transcription factor gene expression, nominating key transcription factors TFAP2A, LEF1, MYB, and PAX5 as CD70 regulators; validation is now underway. Additionally, we showed that knockout of NSD2 in t(4;14) models KMS34 and KMS11 led to significantly decreased CD70 surface expression, implicating this central epigenetic driver in CD70 overexpression in this high-risk MM subtype. Finally, we also demonstrated that treatment with the hypomethylating agent azacytidine can drive increased CD70 expression, similar to that shown in AML models. Conclusion: Taken together, targeting CD70 appears to be a highly promising cellular therapeutic strategy for MM, with particularly high utility in high-risk patients after BCMA CAR-T relapse. We are continuing preclinical and future clinical development of this CAR-T strategy, using our natural ligand CD27 design with extended in vivo persistence.

DNA Liquid Biopsies for Cancer-Associated Venous Thromboembolism Prediction

Introduction Cancer-associated venous thromboembolism (VTE) is a major source of oncologic cost, morbidity and mortality. Identifying an optimal population for prophylactic anticoagulation is challenging and adds to clinician overhead. Circulating tumor (ct)DNA sequencing assays (“liquid biopsies”) are increasingly deployed in clinic, with multiple US FDA approvals for matching to molecularly targeted therapy. Preliminary data suggest that cell-free (cf)DNA, which is also measured as part of liquid biopsies and may consist of tumor or wild-type DNA, is thrombogenic due to its association with neutrophil extracellular traps. However, clinical validation of cfDNA for VTE prediction is lacking. ctDNA detection is associated with aggressive tumor biology and worse survival; however, whether it is associated with VTE is unknown. To address these gaps, we studied the clinical utility of ctDNA and cfDNA from liquid biopsies for predicting VTE. Methods We conducted an observational study in two cohorts: 1) Patients with any cancer type at Memorial Sloan Kettering Cancer Center (MSK, New York USA), an academic medical center, with plasma ctDNA sequencing using MSK-ACCESS, a NY State-approved, 129-gene assay (N=4,141). 2) A nonoverlapping cohort of patients at MSK or GenesisCare (Sydney, Australia), a community oncology setting, with stage IV or recurrent non-small cell lung cancer and ctDNA sequencing using a separate commercial assay (Agilent ctDx Lung, N=463). We studied the independent contribution of ctDNA, cfDNA and related variables for predicting VTE using Cox proportional hazards. We studied the potential utility of DNA liquid biopsies for predicting VTE within six months using machine learning models, comparing their dynamic area under the receiver operating curve, precision, and recall to those of the Khorana score, the most widely implemented risk metric. We tested the impact of ongoing anticoagulation for non-VTE indications on future VTE in patients with and without detectable ctDNA using nonrandomized, real-world evidence. Results ctDNA detection was associated with an increased risk of VTE in a dose-dependent manner (Figure 1). This association was observed across assays and treatment sites and regardless of genomic subtype. ctDNA was associated with VTE risk in non-small cell lung, breast, pancreatic, prostate, and less represented cancers as well as melanoma but not in bladder, hepatobiliary, or colorectal cancer. By contrast, cfDNA concentration was associated with VTE risk in all cancer types. In multivariate analysis, ctDNA detection, cfDNA concentration, Khorana score, chemotherapy treatment within 30 days, and number of organ sites of metastasis were all independently associated with VTE. In a cohort of patients with treatment naive lung adenocarcinoma with segmented positron emission tomography scans, ctDNA was an independent predictor of VTE after controlling for metabolic tumor volume. A random survival forest machine learning model to predict VTE within six months of plasma draw using liquid biopsy data (i.e. cfDNA and ctDNA levels as well as genomic content), cancer type, and chemotherapy receipt (“LB+”) outperformed the Khorana score in 5-fold cross-validation and was similarly superior in the external lung cohort (Table 1). The superiority of the model was observed in patients treated or not treated with chemotherapy. Additional clinical and radiologic data, including demographics, metastatic sites of disease, and Khorana score components (“All”), did not improve the ctDNA assay-based LB+ model. Patients with anticoagulation had lower VTE rates if ctDNA was detectable (adjusted HR 0.54, 95%CI 0.33-0.87) but not if ctDNA was undetectable. Conclusion In patients with cancer, ctDNA and cfDNA are independent biomarkers for VTE. Risk models incorporating liquid biopsy data are feasible in academic or community settings and, if prospectively validated, would provide a means of VTE risk stratification at scale.

Feasibility and Accuracy of Nanopore Transcriptome Sequencing for Pediatric Acute Leukemia Lineage Classification in a Resource-Constrained Setting

Introduction: Gene expression profiling (GEP) has the potential to offer alternative approaches to conventional testing methodologies (flow cytometry, immunohistochemistry, karyotyping, and FISH) for cancer diagnostics. Possible advantages of this approach include unified protocols, automated analyses obviating high-end technical expertise, and perhaps lower training, capital, and skilled human resource requirements in comparison with complex traditional testing. These factors are of particular value in resource-constrained settings, where even leukemia lineage classification is unavailable or inconsistent in many locations, with critical implications for risk stratification and treatment selection. The goals of this study were to (a) ascertain and subsequently improve the accuracy of a lineage classifier in a different context than the pre-testing stage, and (b) determine feasibility of mRNA sequencing for cancer classification in a low-resource setting. The approach utilized PCR-amplified full-length cDNA sequencing using Oxford Nanopore Technologies (ONT) PCR-cDNA barcoding kit (SQK-PCB109) and MinION Mk1b. Methods: One hundred and ten pediatric acute leukemia specimens were collected and sequenced at Indus Hospital &amp; Health Network (IHHN), Karachi, Pakistan and analyzed at the University of North Carolina at Chapel Hill (UNC), USA. Specimens (isolated peripheral blood and / or bone marrow mononuclear cells), frozen in liquid nitrogen or Zymo DNA/RNA shield at -80 oC, were acquired from IHHN's pediatric biorepository. Specimen-associated data included clinical history, CBC, flow cytometry, FISH and karyotyping results from routine clinical testing at the IHHN's ISO 15189-accredited clinical laboratory. The study site did not have prior experience with NGS and did not require external hands-on training for this project. 100 ng of extracted RNA from each of 48 B-ALL, 30 T-ALL, 31 AML, and 1 mixed phenotype acute leukemia specimens was used to prepare libraries. Sequencing data were transferred to UNC and analyzed with an acute leukemia lineage classification algorithm. Results: Following delivery of reagents and equipment, a 6-week protocol optimization and self-training phase utilized open-source protocols and training videos. The next runs to sequence 110 specimens were completed in 4 months by 2 technical staff at the study site. Using a probability threshold for confident calling (&amp;gt;0.8), established during the development phase at UNC, 91 of 110 specimens (82.7%) were called with high confidence. Of the confident calls, 85 were correctly classified (42 B-ALL, 21 T-ALL, and 22 AML), representing an accuracy of 93.4%. The read N50 of confidently classified, concordant specimens ranged from 270 to 745. The total number of reads ranged from 10,000 to &amp;gt;900,000, while the number of aligned reads ranged from 5,000 to 1.5 million. On the other hand, 6 (6.6%) confidently called specimens were discordant. Of these, the mixed phenotype specimen was classified as AML, but mixed phenotype acute leukemia is not included in our training dataset. One misclassified T-ALL specimen had a read count below 10,000 - a threshold demonstrated in the development phase to be optimal for correct classification. Two additional misclassified samples had very low diversity of transcripts express, which is a quality indicator in development. Read counts of these specimens ranged from 22,000 to 311,000, while aligned reads were between 10,000 and 293,000. Further investigation of these discordant calls is ongoing. Low confidence calls (17% of total), with a classification probability between less than 0.8, were either concordant (13) or discordant (6). Further investigation and retesting of these specimens to ascertain reproducibility of low confidence calls has not been performed yet. Conclusions: Taken together, the approach tested - with automated analysis, limited space, training and technical staffing requirements - was feasible at the study site. It provides proof of concept for a low-cost, ONT sequencing based GEP approach for acute leukemia diagnostics whereby testing can be performed in peripheral laboratories while analysis can be centralized, even across continents. Next steps include (a) incorporating these data into a new training model with updated quality control parameters for the next validation run, and (b) developing the genomic subtype classification results.

EP06.03-10 HER3 Expression in Archived Tissue Samples from Patients with NSCLC Across Various Genomic Subtypes and Characteristics

EP06.03-10 HER3 Expression in Archived Tissue Samples from Patients with NSCLC Across Various Genomic Subtypes and Characteristics