Molecular Stratification Research Articles

Refining molecular subtypes and risk stratification of ovarian cancer through multi-omics consensus portfolio and machine learning.

Ovarian cancer (OC), known for its pronounced heterogeneity, has long evaded a unified classification system despite extensive research efforts. This study integrated five distinct multi-omics datasets from eight multicentric cohorts, applying a combination of ten clustering algorithms and ninety-nine machine learning models. This methodology has enabled us to refine the molecular subtyping of OC, leading to the development of a novel Consensus Machine Learning-driven Signature (CMLS). Our analysis delineated two prognostically significant cancer subtypes (CS), each marked by unique genetic and immunological signatures. Notably, CS1 is associated with an adverse prognosis. Leveraging a subtype classifier, we identified five key genes (CTHRC1, SPEF1, SCGB3A1, FOXJ1, and C1orf194) instrumental in constructing the CMLS. Patients classified within the high CMLS group exhibited a poorer prognosis and were characterized by a "cold tumor" phenotype, indicative of an immunosuppressive microenvironment rich in MDSCs, CAFs, and Tregs. Intriguingly, this group also presented higher levels of tumor mutation burden (TMB) and tumor neoantigen burden (TNB), factors that correlated with a more favorable response to immunotherapy compared to their low CMLS counterparts. In contrast, the low CMLS group, despite also displaying a "cold tumor" phenotype, showed a favorable prognosis and a heightened responsiveness to chemotherapy. This study's findings underscore the potential of targeting immune-suppressive cells, particularly in patients with high CMLS, as a strategic approach to enhance OC prognosis. Furthermore, the redefined molecular subtypes and risk stratification, achieved through sophisticated multi-omics analysis, provide a framework for the selection of therapeutic agents.

Identification of Genomic Signatures for Colorectal Cancer Survival Using Exploratory Data Mining.

Clinicopathological presentations are critical for establishing a postoperative treatment regimen in Colorectal Cancer (CRC), although the prognostic value is low in Stage 2 CRC. We implemented a novel exploratory algorithm based on artificial intelligence (explainable artificial intelligence, XAI) that integrates mutational and clinical features to identify genomic signatures by repurposing the FoundationOne Companion Diagnostic (F1CDx) assay. The training data set (n = 378) consisted of subjects with recurrent and non-recurrent Stage 2 or 3 CRC retrieved from TCGA. Genomic signatures were built for identifying subgroups in Stage 2 and 3 CRC patients according to recurrence using genomic parameters and further associations with the clinical presentation. The summarization of the top-performing genomic signatures resulted in a 32-gene genomic signature that could predict tumor recurrence in CRC Stage 2 patients with high precision. The genomic signature was further validated using an independent dataset (n = 149), resulting in high-precision prognosis (AUC: 0.952; PPV = 0.974; NPV = 0.923). We anticipate that our genomic signatures and NCCN guidelines will improve recurrence predictions in CRC molecular stratification.

Molecular Classifications in Gastric Cancer: A Call for Interdisciplinary Collaboration.

Gastric cancer (GC) is a heterogeneous disease, often diagnosed at advanced stages, with a 5-year survival rate of approximately 20%. Despite notable technological advancements in cancer research over the past decades, their impact on GC management and outcomes has been limited. Numerous molecular alterations have been identified in GC, leading to various molecular classifications, such as those developed by The Cancer Genome Atlas (TCGA) and the Asian Cancer Research Group (ACRG). Other authors have proposed alternative perspectives, including immune, proteomic, or epigenetic-based classifications. However, molecular stratification has not yet transitioned into clinical practice for GC, and little attention has been paid to alternative molecular classifications. In this review, we explore diverse molecular classifications in GC from a practical point of view, emphasizing their relationships with clinicopathological factors, prognosis, and therapeutic approaches. We have focused on classifications beyond those of TCGA and the ACRG, which have been less extensively reviewed previously. Additionally, we discuss the challenges that must be overcome to ensure their impact on patient treatment and prognosis. This review aims to serve as a practical framework to understand the molecular landscape of GC, facilitate the development of consensus molecular categories, and guide the design of innovative molecular studies in the field.

Pathway level subtyping identifies a slow-cycling biological phenotype associated with poor clinical outcomes in colorectal cancer

Molecular stratification using gene-level transcriptional data has identified subtypes with distinctive genotypic and phenotypic traits, as exemplified by the consensus molecular subtypes (CMS) in colorectal cancer (CRC). Here, rather than gene-level data, we make use of gene ontology and biological activation state information for initial molecular class discovery. In doing so, we defined three pathway-derived subtypes (PDS) in CRC: PDS1 tumors, which are canonical/LGR5+ stem-rich, highly proliferative and display good prognosis; PDS2 tumors, which are regenerative/ANXA1+ stem-rich, with elevated stromal and immune tumor microenvironmental lineages; and PDS3 tumors, which represent a previously overlooked slow-cycling subset of tumors within CMS2 with reduced stem populations and increased differentiated lineages, particularly enterocytes and enteroendocrine cells, yet display the worst prognosis in locally advanced disease. These PDS3 phenotypic traits are evident across numerous bulk and single-cell datasets, and demark a series of subtle biological states that are currently under-represented in pre-clinical models and are not identified using existing subtyping classifiers.

Tracing back primed resistance in cancer via sister cells

Exploring non-genetic evolution of cell states during cancer treatments has become attainable by recent advances in lineage-tracing methods. However, transcriptional changes that drive cells into resistant fates may be subtle, necessitating high resolution analysis. Here, we present ReSisTrace that uses shared transcriptomic features of sister cells to predict the states priming treatment resistance. Applying ReSisTrace in ovarian cancer cells perturbed with olaparib, carboplatin or natural killer (NK) cells reveals pre-resistant phenotypes defined by proteostatic and mRNA surveillance features, reflecting traits enriched in the upcoming subclonal selection. Furthermore, we show that DNA repair deficiency renders cells susceptible to both DNA damaging agents and NK killing in a context-dependent manner. Finally, we leverage the obtained pre-resistance profiles to predict and validate small molecules driving cells to sensitive states prior to treatment. In summary, ReSisTrace resolves pre-existing transcriptional features of treatment vulnerability, facilitating both molecular patient stratification and discovery of synergistic pre-sensitizing therapies.

Detailed DNA methylation characterisation of phyllodes tumours identifies a signature of malignancy and distinguishes phyllodes from metaplastic breast carcinoma.

Phyllodes tumours (PTs) are rare fibroepithelial lesions of the breast that are classified as benign, borderline, or malignant. As little is known about the molecular underpinnings of PTs, current diagnosis relies on histological examination. However, accurate classification is often difficult, particularly for distinguishing borderline from malignant PTs. Furthermore, PTs can be misdiagnosed as other tumour types with shared histological features, such as fibroadenoma and metaplastic breast cancers. As DNA methylation is a recognised hallmark of many cancers, we hypothesised that DNA methylation could provide novel biomarkers for diagnosis and tumour stratification in PTs, whilst also allowing insight into the molecular aetiology of this otherwise understudied tumour. We generated whole-genome methylation data using the Illumina EPIC microarray in a novel PT cohort (n = 33) and curated methylation microarray data from published datasets including PTs and other potentially histopathologically similar tumours (total n = 817 samples). Analyses revealed that PTs have a unique methylome compared to normal breast tissue and to potentially histopathologically similar tumours (metaplastic breast cancer, fibroadenoma and sarcomas), with PT-specific methylation changes enriched in gene sets involved in KRAS signalling and epithelial-mesenchymal transition. Next, we identified 53 differentially methylated regions (DMRs) (false discovery rate < 0.05) that specifically delineated malignant from non-malignant PTs. The top DMR in both discovery and validation cohorts was hypermethylation at the HSD17B8 CpG island promoter. Matched PT single-cell expression data showed that HSD17B8 had minimal expression in fibroblast (putative tumour) cells. Finally, we created a methylation classifier to distinguish PTs from metaplastic breast cancer samples, where we revealed a likely misdiagnosis for two TCGA metaplastic breast cancer samples. In conclusion, DNA methylation alterations are associated with PT histopathology and hold the potential to improve our understanding of PT molecular aetiology, diagnostics, and risk stratification. © 2024 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

P062 Characterization of UC1 and UC2 patients: towards personalized medicine in Ulcerative Colitis

Abstract Background Ulcerative colitis (UC) is a highly heterogeneous chronic inflammatory bowel disease (IBD) that primarily affects the colon. Although there are some available classifications, mostly based on clinical parameters, the lack of a molecular stratification prevents full comprehension of the diagnostic and prognostic heterogeneity of patients, and thus from correctly predicting their response to available treatments. We have recently developed an unbiased stratification of UC patients (UC1/UC2) based on their transcriptomic profiles (Czarnewski P; et al. Nature Communications, 2019), which holds great potential for the design of personalized medicine in IBD. However, further characterization is necessary to validate it and consolidate its clinical application. Methods To achieve this goal, we used bulk and single-cell RNA sequencing (sc-RNAseq) datasets from a prospective cohort of Swedish IBD patients to understand whether UC1 and UC2 are different diseases or states of the same disease, and to deeply characterize the cellular composition of UC1 and UC2 patients. In addition, we took advantage of murine spatial transcriptomics datasets of colonic tissues at steady state or during mucosal healing, to get insights into the spatiotemporal dynamics of UC1/UC2 transcriptional profiles. Results A longitudinal analysis of UC1/UC2 transcriptomic profiles as well as gene module analysis revealed that UC1 and UC2 profiles are fluctuant during the course of disease, eventually representing different stages of the same disease. scRNAseq analysis allowed us to identify specific cell types characterizing UC2 patients.[FC1] Finally, while genes defining UC2 patients distributed homogenously within the murine mouse colon at steady state and during mucosal healing, UC1 genes primarily mapped within damage/regeneration regions. Conclusion This study details the differences in immune cell composition and molecular pathways in UC1 and UC2 patients, contributing to a more comprehensive understanding of the new molecular stratification system we have developed for ulcerative colitis patients.

Abstract A077: The EZH2 epigenetic factor orchestrates oncogenic properties in ATM-deficient pancreatic cancer

Abstract Despite decades of extensive basic and translational research improving our understanding of pancreatic ductal adenocarcinoma (PDAC), its complex biology remains a significant challenge and limits overcoming its dismal prognosis. Due to its resistance to conventional therapies and heterogeneous mutational landscape, molecular stratification and the development of targeted therapeutic options are essential for improving survival of PDAC patients. Previous work from our group investigated loss-of-function mouse models for the serine/threonine kinase ATM causing homologous recombination deficiency (HRD) in PDAC. Specifically, ATM loss caused certain oncogenic features in the arising tumors, including the acquisition of an invasive phenotype along with the emergence of a prominent desmoplastic reaction, but also sensitizes toward platinum-based therapies and PARP1 inhibition (Russell, Nat Commun; Perkhofer, Cancer Res; Gout, Gut). However, the mechanisms orchestrating these oncogenic features remain unclear. Here, we identified the histone-methyltransferase EZH2 as a direct ATM substrate to operate as an epithelial-mesenchymal gatekeeper. In detail, ATM-deficient tumors accumulated EZH2, suggesting that ATM status contributes to rewiring the epigenetic landscape of pancreatic tumors, subsequently supporting oncogenic features. To investigate whether and how EZH2 impacts ATM-deficient PDAC pathogenesis, we generated conditional double ATM and EZH2-knockout in the KrasG12D/+; Ptf1aCre/+ mice. Double-null mice showed extended mouse survival, reduced metastasis burden, and decreased undifferentiated malignant cell features. Interestingly, both in vivo and in vitro investigations revealed that ATM status significantly impacts EZH2 activity via a post-translational regulation axis. Specifically, the ATM kinase activity was critical to promote Culin1 E3-ubiquitin ligase recruitment and subsequent EZH2 ubiquitination and degradation. Finally, we demonstrated that a combinatorial therapy involving a potent EZH2 inhibitor promotes genomic instability and exacerbates the cytotoxic effects of PARP interference in ATM-deficient HRD PDAC. Overall, our study demonstrates that ATM deficiency significantly impacts EZH2 epigenetic functions and pinpoints the molecular stratification according to ATM status prior to pharmacological EZH2 interference in PDAC therapies as a valuable approach to target ATM-deficient HRD tumors. Citation Format: Elodie Roger, Anna Härle, Johann Gout, Lukas Perkhofer, Elisabeth Hessmann, Alexander Kleger. The EZH2 epigenetic factor orchestrates oncogenic properties in ATM-deficient pancreatic cancer [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Pancreatic Cancer; 2023 Sep 27-30; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(2 Suppl):Abstract nr A077.

Abstract A117: Molecular stratification of therapeutic targets in pancreatic cancer

Abstract Precision medicine approaches using molecular tumor profiling have resulted in striking clinical benefits in certain patient populations. However, targeted therapy in pancreatic cancer has been unable to displace multi-agent chemotherapy, in part due to lack of actionable biomarkers to anchor stratification. To address this need, we have conducted a biomarker focused analysis of cancer dependencies and cell surface targets to understand specific vulnerabilities present in molecular subgroups. We started by interrogating genome scale CRISPR screening and expression data from The Cancer Dependency Map to identify lineage-enriched, selective dependencies in pancreatic cancer. Our results reveal synthetic lethal interactions such as paralog pairs (e.g., VPS4A/B), dependencies predictable by overexpression (e.g., KLF5), and dependencies predictable by genetic alterations (e.g., RAB10 with SMAD4 deletion). Our results reaffirm the importance of KRAS signaling but also highlight a diversity of downstream essential mechanisms including DOCK5 and RAF1. We also identified vulnerabilities specific to transcriptional subtypes, finding ITGA3 to be a selective dependency of not only the basal-like state in pancreas cancer but also across the spectrum of epithelial malignancies. Given the central role of KRAS mutations in tumorigenesis and the emergence of KRAS inhibitors, we studied predictors of KRAS dependency. GSEA analysis of genes overexpressed in cell lines with lower dependency on KRAS identified gene sets related to focal adhesion, EMT, and neuron development, while lines with high KRAS dependency were enriched in genes related to epithelial differentiation/development and the canonical Hallmark KRAS up signature. These gene signatures also predict resistance and sensitivity to small molecule inhibition of KRAS with a high degree of accuracy (Pearson r = 0.68, 0.94 respectively). In addition to vulnerabilities from intracellular target inhibition, cancer cells may be targeted selectively based on surface proteins using cellular and antibody based therapeutic strategies. We used bulk gene expression data from TCGA and GTEx as well as single cell expression data from The Human Cell Atlas, to calculate enriched genes in pancreas cancer versus critical normal tissues. Our approach captures targets including MSLN, TROP2, MUC1, and CLDN18, which have established tumor specificity, but also identifies PSCA, GPRC5A, and CEACAM5 as additional emerging pancreas tumor specific antigens. We charted interpatient and intratumoral heterogeneity in spatial transcriptomics data using a cohort of 21 treated and untreated pancreatic cancer tissue specimens, identifying MUC1 and GPRC5A as both expressed homogenously across patients and throughout individual tumors while scarcely expressed in normal tissues. Ultimately, these findings may prompt prioritization of new molecularly defined dependencies, and construction of rational combinatorial regimens. Citation Format: Dennis Gong, Jimmy A. Guo, Carina Shiau, Ananya D. Jambhale, Steven Wang, Scott Ginebaugh, Patrick Z. Yu, Kevin S. Kapner, Seema Chugh, Laleh Abbassi, Daniel Zhao, Westley W. Wu, Peter Chen, Harshabad Singh, William L. Hwang, Andrew J. Aguirre. Molecular stratification of therapeutic targets in pancreatic cancer [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Pancreatic Cancer; 2023 Sep 27-30; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(2 Suppl):Abstract nr A117.

Early Detection of Molecular Residual Disease and Risk Stratification for Children with Acute Myeloid Leukemia via Circulating Tumor DNA.

Patient-tailored minimal residual disease (MRD) monitoring based on circulating tumor DNA (ctDNA) sequencing of leukemia-specific mutations enables early detection of relapse for pre-emptive treatment, but its utilization in pediatric acute myelogenous leukemia (AML) is scarce. Thus, we aim to examine the role of ctDNA as a prognostic biomarker in monitoring response to the treatment of pediatric AML. A prospective longitudinal study with 50 children with AML was launched, and sequential bone marrow (BM) and matched plasma samples were collected. The concordance of mutations by next-generation sequencing-based BM-DNA and ctDNA was evaluated. In addition, progression-free survival (PFS) and overall survival (OS) were estimated. In 195 sample pairs from 50 patients, the concordance of leukemia-specific mutations between ctDNA and BM-DNA was 92.8%. Patients with undetectable ctDNA were linked to improved OS and PFS versus detectable ctDNA in the last sampling (both P < 0.001). Patients who cleared their ctDNA post three cycles of treatment had similar PFS compared with persistently negative ctDNA (P = 0.728). In addition, patients with >3 log reduction but without clearance in ctDNA were associated with an improved PFS as were patients with ctDNA clearance (P = 0.564). Thus, ctDNA-based MRD monitoring appears to be a promising option to complement the overall assessment of pediatric patients with AML, wherein patients with continuous ctDNA negativity have the option for treatment de-escalation in subsequent therapy. Importantly, patients with >3 log reduction but without clearance in ctDNA may not require an aggressive treatment plan due to improved survival, but this needs further study to delineate.

Momelotinib expands the therapeutic armamentarium for myelofibrosis: Impact on hierarchy of treatment choices.

The primary objective of treatment in myelofibrosis (MF) is prolongation of life, which is currently accomplished only by allogeneic hematopoietic stem cell transplantation (AHSCT). Determination of optimal timing for AHSCT is facilitated by molecular risk stratification. Non-transplant treatment options in MF are palliative in scope and include Janus kinase 2 (JAK2) inhibitors (JAKi): momelotinib (FDA approved on September 15, 2023), ruxolitinib (November 16, 2011), fedratinib (August 16, 2019), and pacritinib (February 28, 2022); all four JAKi are effective in reducing spleen size and alleviating symptoms, considered a drug class effect and attributed to their canonical JAK-STAT inhibitory mechanism of action. In addition, momelotinib exhibits erythropoietic effect, attributed to alleviation of ineffective erythropoiesis through inhibition of activin A receptor type-I (ACVR1). In transplant-ineligible or deferred patients, the order of treatment preference is based on specific symptoms and individual assessment of risk tolerance. Because of drug-induced immunosuppression and other toxicities attributed to JAKi, we prefer non-JAKi drugs as initial treatment for MF-associated anemia that is not accompanied by treatment-requiring splenomegaly or constitutional symptoms. Otherwise, it is reasonable to consider momelotinib as the first-line JAKi treatment of choice, in order to target the triad of quality-of-life offenders in MF: anemia, splenomegaly, and constitutional symptoms/cachexia. For second-line therapy, we favor ruxolitinib, over fedratinib, based on toxicity profile. Pacritinib and fedratinib provide alternative options in the presence of severe thrombocytopenia or ruxolitinib-resistance/intolerance, respectively. Splenectomy remains a viable option for drug-resistant symptomatic splenomegaly and cytopenia.

An innovative electrohydrodynamics-driven SERS platform for molecular stratification and treatment monitoring of lung cancer.

The advancement of molecular diagnostics for lung cancer stratification and monitoring is essential for the strategic planning and prompt modification of treatments, aiming to enhance clinical results. To address this need, we suggest a nanocavity structure designed to sensitively analyze the protein signature on small extracellular vesicles (sEVs). This approach facilitates precise, noninvasive staging and treatment monitoring of lung cancer. The nanocavity is created through molecular recognition, involving the interaction of sEVs with nanobox-based core-shell surface-enhanced Raman scattering (SERS) barcodes and asymmetric, mirrorlike gold microelectrodes. By applying an alternating current to the gold microelectrodes, a nanofluidic shear force was generated, promoting the binding of sEVs and the effective assembly of the nanoboxes. This interaction induced a nanocavity between the nanobox and the gold microelectrode, which significantly amplified the electromagnetic field. This amplification enhanced Raman signals from four SERS barcodes simultaneously, allowing the generation of patient-specific molecular sEV signatures. When tested on a cohort of clinical samples (n = 76) using the nanocavity architecture, these patient-specific sEV molecular signatures accurately identified, stratified, and monitored lung cancer patients' treatment, demonstrating its potential for clinical application.

RNA-seq–based miRNA signature as an independent predictor of relapse in pediatric B-cell acute lymphoblastic leukemia

AbstractAberrant micro-RNA (miRNA) expression profiles have been associated with disease progression and clinical outcome in pediatric cancers. However, few studies have analyzed genome-wide dysregulation of miRNAs and messenger RNAs (mRNAs) in pediatric B-cell precursor acute lymphoblastic leukemia (BCP-ALL). To identify novel prognostic factors, we comprehensively investigated miRNA and mRNA sequencing (miRNA-seq and mRNA-seq) data in pediatric BCP-ALL samples with poor outcome. We analyzed 180 patients, including 43 matched pairs at diagnosis and relapse. Consensus clustering of miRNA expression data revealed a distinct profile characterized by mainly downregulation of miRNAs (referred to as an miR-low cluster [MLC]). The MLC profile was not associated with any known genetic subgroups. Intriguingly, patients classified as MLC had significantly shorter event-free survival (median 21 vs 33 months; log-rank P = 3 ×10−5). Furthermore, this poor prognosis was retained even in hyperdiploid ALL. This poor prognostic MLC profiling was confirmed in the validation cohort. Notably, non-MLC profiling at diagnosis (n = 9 of 23; Fisher exact test, P = .039) often changed into MLC profiling at relapse for the same patient. Integrated analysis of miRNA-seq and mRNA-seq data revealed that the transcriptional profile of MLC was characterized by enrichment of MYC target and oxidative phosphorylation genes, reduced intron retention, and low expression of DICER1. Thus, our miRNA-mRNA integration approach yielded a truly unbiased molecular stratification of pediatric BCP-ALL cases based on a novel prognostic miRNA signature, which may lead to better clinical outcomes.

Sjögren's syndrome: one year in review 2023.

Primary Sjögren's syndrome (pSS) is a systemic autoimmune disorder characterised by the T-cell-mediated hyperactivation of B-cells and cytokine production. The condition may evolve from an asymptomatic, indolent course, with glandular involvement, to extra-glandular systemic manifestations up to lymphoma development. On tissue level, the typical feature is the lymphocytic infiltration of the salivary gland by B-, T- and antigen presenting cells, as mirrored by the diagnostic cornerstone role of minor salivary gland (MSG) biopsy. Recently, increasing research focused on the investigation of mechanisms underlying the complex pathogenesis of the disease and highlighted the multi-factorial nature of SS consisting of concomitant involvement of environmental, genetic, neuroendocrine and immune factors. In particular, many aspects have been investigated regarding genetic and epigenetics, the role of specific B- and T-cell phenotypes and the investigation of disease-specific biomarkers as predictors of disease development, activity, and lymphomagenesis. Surely, a deeper understanding of these multiple mechanisms may facilitate earlier diagnosis, enable subphenotyping of patients and open novel therapeutic possibilities to address the unmet needs of the disease in the upcoming years.In this review, following the others of this series, we will summarise the most recent literature on pSS pathogenesis and clinical features focusing in particular on new insights into pSS molecular stratification and therapeutic advances in the era of precision medicine.

Effectiveness and safety of a 26-week taper regimen of glucocorticoid in GCA patients: Results from a prospective cohort study

To assess the effectiveness and safety of the 26-week tapering regimen of glucocorticoids (GC) used in the GiACTA trial in a prospective cohort of treatment-naive, biopsy-proven GCA patients. Patients with a new diagnosis of biopsy-proven GCA enrolled in the GC arm of the START project (molecular stratification of patients with GCA to tailor GC and tocilizumab therapy) were included. All patients were treated with the 26-week taper regimen of GC used in the GiACTA trial. The primary endpoint was the rate of relapse-free remission at week 52. The secondary endpoints were the proportion of patients with incident aortic damage, cumulative GC doses and GC-related adverse events (AE). 22 patients were included between December 2018 and February 2022. At week 52, 10 patients (45%, 95% CI 24-68) were in relapse-free remission. After a median (IQR) follow-up of 35 (22-40) months, 7 patients (32%, 95% CI 14-55) were in relapse-free remission. 18 patients with baseline large-vessel imaging underwent CT angiography at the end of the follow-up. No patients had evidence of new aortic dilation, significant progression of aortic damage or large vessel stenosis. 15/22 patients (68%) had at least one relapse during follow-up. No patients developed visual or cerebrovascular manifestations during relapses. 15/22 (68%) patients had at least one GC-related AE. A 26 week taper regimen of GC was effective and safe in inducing and maintaining remission in a sizeable proportion of newly diagnosed GCA patients. However, the frequency of GC-related adverse events was high.

Molecular Classification of Relapsed DLBCL Reveals Novel Biologic Subgroups

Background Diffuse large B cell lymphoma (DLBCL) is the most common aggressive lymphoid malignancy in adults. Though standard immunochemotherapy regimens can result in clinical remission and cure in a majority of patients, approximately 30% of patients are primary treatment resistant or eventually relapse (rrDLBCL), and their prognosis is very poor. Recent molecular classification by large scale genomic and transcriptomic profiling of newly diagnosed DLBCL tumors has resulted in a deeper understanding of disease drivers, however, the molecular heterogeneity of rrDLBCL is yet to be fully explored. Therefore, in this study, we used an unsupervised approach to better define transcriptional and genetic programs in rrDLBCL samples hypothesizing that molecular stratification within this high-risk group may lead to a better understanding of this disease, which remains an unmet need, and a more personalized therapeutic approach. Methods Data used in this study included RNA (n=143) and whole exome (n=126) sequencing data from available FFPE tumor samples at the time of a relapse (any line of treatment, r1-r10 relapse timepoints included in analysis, one per patient), consented to the Molecular Epidemiology Resource (n=61), banked in the Mayo Lymphoma Biobank (n=50), or consented to the CC-122-ST-001 clinical trial (n=32, NCT01421524). Unsupervised clustering was performed on protein coding gene expression values using the non-negative matrix factorization (NMF) approach. NMF was carried out by performing 200 runs on cluster sizes from 2 to 6. Cell of origin (COO) was determined using the method described by Reddy et al. The tumor microenvironment was analyzed using CIBERSORTx and Lymphoma Microenvironment Classification (LME). Genetic classification was done using LymphGen. Results In order to identify unique transcriptional programs in rrDLBCL, we implemented NMF and consensus clustering to define an ideal quantity of stable patient clusters. We identified 4 rrDLBCL patient clusters (rrC1-rrC4) as the most stable solution, based on consensus clustering selection criteria of iterative NMF runs. While relapse timepoint was not associated with individual clusters, we did observe significant differences in tumor characteristics between the clusters. COO classification found that rrC1 and rrC4 were enriched for GCB-DLBCL while rrC3 was enriched for ABC-DLBCL classification (P &amp;lt; .001). To explore the biologic programs in each cluster, we used pathFindR, which leverages protein-protein interaction networks to identify disease related pathways, rrC1 and rrC4 were defined by pathways involved in cell cycle and mismatch repair; oxidative phosphorylation in rrC2; and MAPK and NF-kB in rrC3. Next, LME classification was done to define patterns in the tumor microenvironment and we found a significant difference between groups (P &amp;lt; .001), with rrC2 and rrC4 tumors enriched for a depleted TME, while rrC1 and rrC3 had a more immune rich or inflammatory TME. This was supported by CIBERSORTx analysis identifying significant differences (P &amp;gt; .001) in abundance of CD8 T cells, memory CD4 T cells, and M1 macrophages between clusters, with rrC4 having the lowest abundance. We then explored genetic patterns in each cluster. Gene level mutation enrichment analysis showed BCL7A mutation exclusively in C4 (6/6, P = .03). The clusters were not associated with LymphGen classification. Lastly, our recently published high risk signature classification that is associated with early clinical failure and is characterized by metabolic dysregulation, a depleted TME, T53 alterations, and poor outcome, was significantly different between the groups (P = .03) with enrichment for high-risk cases in rrC4. In summary, we show for the first time that rrDLBCL patients can be classified into four gene expression clusters that are associated with distinct pathway, TME, and genetic programs. These clusters should now be tested to learn if they can help select patients for newer therapies for rrDLBCL such as CAR-T and bispecific antibodies.

The Human Leukemia Immunopeptidome Is Driven By Pathogenic Molecular Variant Biotypes

Background Historically, acute myelogenous leukemia (AML) has been characterized as having a low somatic mutational burden and is predicted to have low neoantigen levels compared to solid tumors. However, recent data has shown that AML is an immunogenic malignancy and oncogenic driver mutations promote immune escape through immunoediting. Downregulation of human leukocyte antigen (HLA) class II HLA-DRB1, HLA-DQB1, HLA-DPA1 and HLA-DPB1 gene expression on leukemia cells has previously been associated with immune escape in AML following allogeneic hematopoietic stem cell transplantation (allo-HCT). While allo-HCT and cellular therapies are well-known potentially curative treatment options in AML, the human leukemia immunopeptidome has still not been well characterized. Here, we sought to uncover the human leukemia immunopeptidome by detecting differential major histocompatibility complex (MHC) class I neopeptide expression using molecular biotype subgroups. Methods We conducted retrospective analysis on 31 patients with hematopoietic malignancies including AML, chronic myelomonocytic leukemia (CMML), myelodysplastic syndrome/myeloproliferative neoplasm (MDS/MPN), MDS, chronic myelogenous leukemia (CML) and other hematopoietic malignancies (i.e. NK cell leukemia) at MedStar Georgetown University Hospital. Mutations were identified using next generation sequencing and separated into the following annotated molecular biotypes based on the mechanism of pathogenesis: Group 0 (G0) DNA repair, G1 splicing, G2 transcription factor, G3 kinase activation, G4 epigenetic modification, G5 chromatin remodeling, G6 cohesin, G7 tumor suppressor (i.e. p53), G8 telomere maintenance, G9 histone modification, G10 nucleotransport (i.e. NPM1) and G11 cell cycle regulation. FASTA nucleotide sequences were obtained by subjecting mutations to transcript expression and cDNA alignment in the National Center for Biotechnology Information (NCBI) gene database. MHC class I binding neopeptides were identified by mapping patient FASTA nucleotide sequences and HLA class I alleles in the Net MHC pan 4.1 database. Neopeptides were grouped by % Rank and categorized as weak binding (WB) if &amp;gt; 0.5 and strong binding (SB) if ≤0.5. Descriptive and inferential statistics were performed using SAS. Results Thirty-one patients were evaluated including 17 AML (54.8%), 5 MDS (16.1%), 3 CMML (9.7%), 2 MDS/MPN (6.5%), 1 CML (3.2%) and 3 hematopoietic malignancies (9.7%). The median age was 65 years (range 28 -89); 43.8% were female and 43.3% were African American (AA). 144mutations were identified; thus far, 82 mutations have been subjected to FASTA NP_ protein extraction. The G3 biotype had the highest frequency of neopeptide bindings (n=59) followed by G2 (n=50) and G4 biotypes (n=42) ( Fig.1). Among patients with MDS, smokers were more likely to exhibit neopeptide binding compared to non-smokers (p=0.032). Patients with neopeptide binding had a longer median overall survival (mOS) compared to patients without neopeptide binding (mOS = 566 vs 114 days, p=0.0174) (Fig 2). Data exploration allowed isolation of a 54-year-old Asian male with Li-Fraumeni syndrome and p53 mutated AML who demonstrated high affinity neopeptide binding to HLA-B*39:01 with %Rank 0.08. Additionally, a 34 -year-old AA female with FLT3 mutated AML showed very high affinity neopeptide binding to HLA-A*29:02 with %Rank 0.007 as well as evidence of somatic HLA-A*03:01 deletion. Conclusions Pathogenic molecular variant biotypes constitute a novel immune signature, the human leukemia immunopeptidome. Smoking increased the likelihood of MHC class I neopeptide binding among patients with MDS. The presence of neopeptide binding predicted improved mOS in patients with myeloid malignancies. Further investigation of the human leukemia immunopeptidome may provide additional prognostic significance above that of standard molecular and cytogenetic risk stratification.

Analysis of CD58 As a Predictive Marker of Morphologic Response to Inotuzumab Ozogamicin in Patients with B-Cell Acute Lymphoblastic Leukemia

BACKGROUND Clinical outcomes for B-cell Acute Lymphoblastic Leukemia (B-ALL) have significantly improved in the last decade, owing to the introduction of newer agents such as inotuzumab ozogamicin (INO) and blinatumomab. Both of these antibody-based therapies are effective in adults with relapsed or refractory (R/R) B-ALL and have shown promising results in the frontline setting as well. Given the utility of these therapies, there have been efforts to identify predictive markers for these treatments in B-ALL patients. Recently, CD58 was identified as a modulator of response to blinatumomab due to its role as a CD2 ligand in T-cell activation (Li et al. Sci Adv 2022). This costimulatory relationship between CD58 and CD2 has also been linked to Natural Killer cell mediated lysis, which has been described as a mechanism of action of INO (Fu et al. Signal Transduct Target Ther 2022; Huntington et al. Nat Rev Cancer 2020). In this study, we evaluated the correlation of CD58 expression with achievement of morphologic Complete Remission (CR) or Complete Remission with incomplete blood count recovery (CRi) after treatment with INO. METHODS Retrospective chart review was performed to identify patients in our institution's B-ALL database who received INO between 09/15/2011 and 03/10/2023 with blood or bone marrow samples available for flow cytometry analysis prior to INO infusion. Because all patients had measurable CD58 expression by flow cytometry, mean fluorescence intensity (MFI) was used instead of percent positivity to perform a simple logistic regression analysis to evaluate the relationship between CD58 MFI and achievement of CR/CRi after treatment with INO. We then calculated the odds of achieving CR/CRi by age, molecular risk at diagnosis, and R/R disease status. Two patients lacked diagnostic cytogenetic data. Molecular risk stratification and assessment of morphologic response were performed according to National Comprehensive Cancer Network (NCCN) guidelines. To account CD58 MFI outliers, we then confirmed our results by comparing response rates between patients with a CD58 MFI above our median (CD58 MFI-high) to patients with a CD58 MFI below our median (CD58 MFI-low) using Fisher's exact test. RESULTS We found 34 patients who met our criteria for analysis, 2 of which had previously received INO (Table1). Median Overall Survival (OS) was 21.5 months (range, 1-58) with a median leukemia-free survival (LFS) of 13.5 months (range, 1-52). Response rates after INO were high, including 28 patients (82.4%) achieving CR, 30 patients (88.2%) achieving CR/CRi, and 27 patients (79.4%) achieving Measurable Residual Disease (MRD) negativity. Both patients with a history of prior treatment with INO attained CR with MRD negativity. The chi-square test statistic for our simple logistic regression model was not statistically significant (Figure1, p=0.915). The logistic regression coefficient showed no statistically significant association between CD58 MFI and CR/CRi rates after treatment with INO ( p=0.913). Our model predicted an 88.6% probability of achieving CR/CRi for a CD58 MFI of 8,727.97 (Q1), 88.4% for 11,719.88 (Q2), 88.0% for 19,622.85 (Q3), and 85.7% for 55,217.74 (Q4). There was no statistically significant difference in the odds of achieving CR/CRi in patients with poor molecular risk at diagnosis (OR=0.06, 95% CI [0.01, 1.22]), age &amp;gt;35 years (OR=2.75, 95% CI [0.33, 22.9]), or R/R disease (OR=0.98, 95% CI [0.01, 1.97]). When comparing response rates between CD58 MFI-high and CD58 MFI-low patients, each group contained 15 patients that achieved CR/CRi and 2 patients with refractory disease. This confirmed an equal response rate to INO between CD58 MFI-high and CD58 MFI-low patients ( p=1.000). CONCLUSION Overall, we did not find a statistically significant correlation between CD58 MFI and achievement of CR/CRi. Our study is limited by our small sample size, single-center location, and retrospective design. However, it is important to note the high rates of CR/CRi that were seen, regardless of the intensity of CD58 expression. Further work is needed to evaluate the role of CD58 as a predictive marker for other therapies in B-ALL. These data suggest that INO is likely to remain an effective option for these patients, regardless of the intensity of CD58 expression.

Optimizing Management of AML: Insights from a Community-Based Quality Improvement Initiative

Background There has been groundbreaking progress in the molecular diagnosis, risk stratification, and treatment landscape for acute myeloid leukemia (AML), allowing providers to tailor a personalized therapy approach for individual patients. Despite these advances, many barriers to optimal AML treatment exist in integrating targeted therapies, selecting guideline-aligned treatments and managing adverse events. To address these challenges, we conducted a quality improvement (QI) study in community oncology clinics to assess current practice patterns and develop strategies to overcome barriers to optimal care for patients with AML. Methods The QI initiative comprised baseline provider surveys and pre- and post-surveys of providers participating in small-group, team-based audit-feedback (AF) sessions. Survey questions were designed to assess knowledge, confidence, and experiences with targeted therapies for AML. Care teams from each clinic, along with an expert AML physician, participated in AF sessions to (a) assess system-specific practice gaps identified via the provider surveys, (b) prioritize areas for improvement, and (c) develop action plans for addressing root causes. Between 1/2022-6/2022, 104 hematology/oncology providers completed baseline surveys and 54 care team members from the enrolled clinics participated in the AF sessions (Table 1). Results Provider Surveys: Providers' top reported challenges encountered in managing patients with AML were individualizing treatment plans (33%) and supportive care counseling (21%). Moreover, the most challenging issue encountered in selecting therapies for patients with AML was selecting optimal therapies for individual patients (27%). Factors considered most important in treatment decision-making, in addition to clinical guidelines, included effects on quality of life (51%) and treatment effectiveness (51%). Additionally, top parameters guiding treatment decision-making include age (63%) and karyotype/molecular abnormalities (60%). Clinicians reported the most difficulty in managing myelosuppression (34%) associated with novel AML therapies. When asked their top barriers to engaging patients with AML in shared decision-making, providers reported patient's low health literacy (29%) and not enough time to engage in SDM (25%). AF Sessions: Providers participating in the small group AF sessions reported testing for and interpreting results of molecular and genetic features (30%) and individualizing treatment plans (28%) as top challenges. Following the AF session, clinicians reported improvements in high/very high confidence in their ability to select guideline-aligned treatment for patients with AML (70% to 96%) and ability to manage adverse effects of novel AML treatments (76% to 96%). Care teams also set goals to improve individualized treatment plans and adverse event management, stay up to date with clinical guidelines and knowledge of novel therapies, and increase multidisciplinary collaborative care for patients with AML. Action plans to achieve these goals include enhancing knowledge of clinical evidence for new and emerging therapies for AML and developing personalized treatment plans incorporating patient-specific factors and preferences (Figure 1). Conclusions Through this QI initiative, clinical teams identified barriers to optimal AML care and created and implemented specific action plans focused on increasing knowledge of AML therapies, optimizing personalized treatment plans and improving adverse events management. Meaningful confidence gains were observed in aligning treatment plans to guideline recommendations and managing adverse effects of treatment. These data underscore clinical practice gaps to address in future initiatives to support safe, effective, and evidence-based integration of targeted therapies into AML care to optimize patient outcomes. Study Sponsor Statement The study reported in this abstract was funded by educational grants from AbbVie, Inc. and Astellas. The grantors had no role in the study design, execution, analysis, or reporting.

Outcome of Risk- and Response-Adapted Post-Remission Therapy in Patients with NPM1 AML Treated Intensively

Introduction: Molecular assessment of measurable residual disease (MRD) is a powerful prognostic tool to identify high risk of relapse. Specifically, large studies have shown that MRD negativity of NPM1 after two cycles of intensive chemotherapy (IC) predicts long term remission. The translation of MRD results to treatment decisions such as treatment augmentation (MRD erasing) and referral to an allogeneic stem cell transplant (allo-SCT) is still lacking. The purpose of this study is to report the outcome of AML patients with NPM1 mutation who received IC and achieved remission, in whom NPM1 MRD post cycle 2 of IC, combined with FLT3-ITD status, were used systematically to decide whether patients were to complete intensive consolidations or refer to alloSCT. Methods: This is a retrospective analysis from 3 large hospitals in Israel. Patients were treated with 7+3 for induction and a repeat induction based on day 14 marrow. Midostaurin was added for FLT3mut leukemia since 2018. Consolidation consisted of 3-4 HiDAC cycles at doses of 1-3 g/msqrX6. Gemtuzumab ozogamicin and maintenance with FLT3 inhibitors or CC486 could be added. NPM1 transcript levels (normalized to ABL copies) were collected at diagnosis, after 2 cycles of IC, at the end of IC cycles and during follow-up. Patients were transplanted if a donor was available and any of the following: NPM1 transcript level was not reduced by 3 logs by the end of second cycle chemotherapy, absolute level was &amp;gt;1%, poor risk AML (ELN 2017 guidelines) or high AR FLT3-ITD mutation was present. Patients were followed for relapse (type and timing) and whether a second remission was achieved, as well as for EFS and overall survival. Results: A database search yielded 90 patients with NPM1 AML who were treated according the abovementioned protocol between 5/2015 and 11/2022 and achieved remission, with a median follow-up of 872 days (range 103-2751). Forty-seven patients were to receive consolidation cycles without transplantation in CR1, based on the aforementioned criteria. All but one patient (death from sepsis) completed treatment cycles. During follow-up, 21 (45%) patients relapsed at a median of 343 days. Two patients died while in CR1 during follow-up (extensive burns; sudden death at home). In a subgroup analysis of patients &amp;lt;60y (N=28), there were 11 events (9 relapses and 2 deaths), whereas in patients &amp;gt;60y (N=19) there were 13 events (12 relapses and 1 death during follow-up). Of note, 4 patients &amp;gt;60y were treated with CC486, of whom only 1 relapsed. Of the 18 patients who relapsed and received further treatment (3 had palliative care only), only 1 patient did not enter CR2 (died during salvage). The other 17 patients entered CR2 after salvage regimen (N=7, 4 were qPCR MRDneg, 5 were subsequently transplanted in CR2) or after upfront alloSCT without any salvage regimen (N=10, 4 were qPCR MRDneg). Forty-three patients were planned for alloSCT at CR1 (27 - inadequate molecular response, 16 - ELN risk group). Only one of these patients failed to undergo alloSCT in CR1 as she relapsed. The other 42 patients underwent alloSCT at CR1. In this group of patients, the median progression-free survival was 5 years, with no difference in patients above or below the age of 60. One half of the events (9/18) were relapses and the other half were deaths in remission, almost all were GVHD associated (figure1). Discussion: NPM1 mut AML without FLT3 is still considered by the ELN to have favorable outcome and most patients are not being transplanted at CR1. Yet, many relapse. This study evaluated the outcome of AML patients according to molecular risk stratification and alloSCT consideration by the end of cycle 2 of IC. In line with previous publications, younger patients with standard-risk AML and with adequate molecular response had a favorable outcome, with &amp;gt;60% of them having a disease-free long-term survival without transplantation. Moreover, even in a case of relapse, patients can achieve CR2 with MRDneg and attain long-term remission (90% OS in long-term follow-up). Older patients have a less favorable outcome even when they achieve adequate molecular response with two thirds of them experiencing relapse. Hence, the potential long-term outcome without transplantation needs to be balanced by the high risk of relapse. Nevertheless, even these patients can be successfully treated with either a salvage regimen or immediate transplantation, and &amp;gt;40% of them can be long-term survivors.