Gibson, Sarah E.

Abstract Objectives Diagnosing adult T-cell leukemia/lymphoma (ATLL) remains challenging due to its clinical and histopathologic heterogeneity. While HTLV-1 serology is a key diagnostic criterion, it is often unavailable at the time of diagnosis and does not confirm viral integration in tumor cells. The HTLV-1–encoded HBZ RNA is a reliable diagnostic biomarker for ATLL. We evaluated the diagnostic performance of RNAscope in situ hybridization (ISH) for HBZ RNA in formalin-fixed, paraffin-embedded (FFPE) tissues to distinguish ATLL from other T-cell lymphomas. Methods HBZ ISH was evaluated in 72 cases, including 52 ATLLs and 20 non-ATLL T-cell lymphomas, to determine diagnostic sensitivity and specificity, interobserver agreement, interlaboratory reproducibility, and comparison with digital gene expression. Results This assay demonstrated 86.5% sensitivity and 100% specificity, with sensitivity improving to 97% when excluding bone marrow clot sections and FFPE blocks older than 5 years. Interobserver agreement among 3 pathologists was excellent (κ = 0.92, percent agreement: 94%). Interlaboratory reproducibility of HBZ ISH was high, with 93% to 95% concordance across these Clinical Laboratory Improvement Amendments–certified laboratories. HBZ RNA ISH–positive cases demonstrated significantly higher HBZ gene expression levels by digital gene expression profiling. Conclusions We established HBZ ISH as a reliable tissue-based surrogate for HTLV-1 infection in ATLL. This assay will enhance diagnostic accuracy, expedite definitive diagnosis, and improve patient selection for targeted therapies and clinical trials, particularly in regions with a high prevalence of HTLV-1.

Abstract Background: Recent advances in computer vision have given rise to image foundation models (FMs), large-scale neural networks pre-trained on millions of unlabeled histopathology patches, which can learn generalizable representations of tissue morphology. In many cancers, these FMs have predicted genetic mutations and prognoses directly from histology, suggesting that morphology encodes “hidden” molecular information. In lymphoma, this field is nascent but growing.General-purpose histopathology FMs have limited representation of lymphoid morphology. We developed and evaluated LymphoVision, a lymphoma-specialized image foundation model pretrained with self-supervised learning, to extract clinically meaningful features from routine hematoxylin and eosin (H&E) histology. Methods: LymphoVision was pre-trained using knowledge distillation and consistency (DINOv2) on over 37 million multi-resolution patches (5x, 10x, 20x, 40x) sampled using a density-aware strategy from 31,211 H&E whole slide images (WSIs) from 9,155 archival lymphoma and reactive lymphoproliferative cases. To ensure diverse, representative samples, our patch sampling method uses an efficient graph-based clustering algorithm and prioritizes morphologically informative regions by clustering patch embeddings and performing density sampling within each cluster. The pretrained encoder was evaluated using a clustering-constrained attention multiple instance learning (MIL) framework on three diagnostic tasks: (1) cell-of-origin (COO) classification in diffuse large B-cell lymphoma (DLBCL), distinguishing germinal center B-cell (GCB) from non-GCB subtypes based on the Hans immunohistochemistry algorithm; (2) follicular lymphoma (FL) grading, differentiating low-grade (grades 1–2) from high-grade (grade 3); and (3) multiclass lymphoma subtype classification across six categories: DLBCL, FL, marginal zone lymphoma (MZL), mantle cell lymphoma (MCL), classical Hodgkin lymphoma (CHL), and benign/reactive lymphoid conditions. WSI-level labels were derived from pathology reports, and none of the evaluation cases were included in pretraining. Models used a ViT-Giant LymphoVision backbone to generate 20x patch embeddings, followed by MIL using 50% training data and 2-fold cross-validation (25% validation, 25% test). Performance was assessed using area under the receiver operating characteristic curve (AUC), accuracy, recall (sensitivity) and F1 score. Results: For DLBCL COO classification (total n=358 cases), LymphoVision achieved a mean test AUC of 0.93 (range 0.92–0.94), accuracy of 84.5% and F1 of 0.85 across two independent test folds (n=91 per fold, 50% GCB). These results outperformed general-purpose FMs on the same task and cases, including Virchow-v2 (AUC 0.85) and UNI-v2 (AUC 0.88), despite those models being pretrained on 10–100 times more images. For FL grading (total n=660 cases), the test set in each fold included 166 cases (18% grade 3), yielding a mean test AUC of 0.88 (range 0.87–0.90), accuracy of 89.2% and F1 of 0.80. In multiclass lymphoma subtype classification (total n=1,176 cases), the test set included 296 cases per fold. LymphoVision achieved a mean test AUC of 0.98 (range 0.97-0.99), mean test accuracy of 87.1% and mean F1 of 0.84. Class-wise recall was 92.0% for DLBCL (n=112), 88.2% for FL (n=72), 77.2% for MZL (n=29), 70.8% for MCL (n=35), 87.5% for CHL (n=24), and 95.8% for benign/reactive lymphoid conditions (n=24). Conclusion: LymphoVision is a lymphoma-specialized FM trained using self-supervision that achieves high diagnostic accuracy across multiple clinically meaningful tasks using weak supervision and standard H&E images, suggesting the encoder has learned fine-grained lymphoma morphology. Based on historical benchmarks from the 1997 International Lymphoma Study Group classification project, expert hematopathologists achieved diagnostic agreement rates of 55% to 84% using H&E morphology alone for the non-Hodgkin lymphoma (NHL) entities included in our multiclass task. LymphoVision approached or exceeded these benchmarks. These findings underscore the value of our approach using disease-specialized pretraining and intelligent patch sampling for data-efficient histopathology learning. LymphoVision has the potential to augment diagnostics, enable scalable AI-powered classification in real-world workflows, and serve as a foundation for prognostic and theragnostic applications. External validation and multimodal integration are ongoing.

Evaluation of T-cell receptor (TCR) β-chain constant region 1 (TRBC1) expression is an alternative T-cell clonality assessment method. However, evaluation of TRBC1 immunohistochemistry (IHC) for distinguishing mycosis fungoides (MF)/Sezary syndrome (SS) from reactive inflammatory infiltrates in skin is incomplete. We evaluated the utility of a novel dual TRBC1-CD3 IHC stain in skin biopsies with MF/SS (n=40), reactive (n=24), or atypical T-cell infiltrates indeterminant for MF/SS (n=9). Twenty of 24 reactive cases showed clear polytypic TRBC1 expression (median percent TRBC1 positivity among CD3-positive T cells [%TRBC1+] 50% in dermis, 42.5% in epidermis among all cases), while 34/40 MF/SS were clearly monotypic (%TRBC1+ either ≤20% or ≥85%) (sensitivity 85.0%, specificity 91.7%, positive predictive value 94.4%, negative predictive value 78.6%). Discordance between TRBC1 expression and diagnosis was associated with few neoplastic/monotypic T cells and/or lack of physical separation between neoplastic and non-neoplastic populations. Among patch/plaque MF/SS, TRBC1 showed similar diagnostic sensitivity (80.0% vs. 86.7%) and high categorical correlation (monotypic vs. not monotypic, 80%) with TCR gene rearrangement results. TRBC1 interpretation was reproducible, with ≥2/3 and 3/3 pathologists rendering identical interpretations in 100% and 86% of cases, respectively, after independent review and 100% agreement following collective review. Digital image analysis confirmed visual %TRBC1+ accuracy ( r =0.9749, P <0.0001). On the basis of these results, we propose %TRBC1+ cutoffs of <25% or >75% for establishing T-cell monotypia in skin. Considering such thresholds, TRBC1-CD3 evaluation facilitated diagnostic refinement in 7/9 (78%) atypical cases. Overall, TRBC1-CD3 IHC clearly and rapidly aids diagnosis of cutaneous T-cell proliferations.

Abstract Purpose: R,R-1,2 cyclohexanediamine-pyrophosphato-platinum(II) (PT-112) is a novel immunogenic cell death–inducing small molecule under phase II development in several cancer types. It inhibits ribosome biogenesis and causes organelle stresses, leading to selective immunogenic cell death in cancer cells. The possibility of PT-112’s pyrophosphate moiety driving high drug concentrations to bone sites of disease has led to an interest in PT-112’s use in multiple myeloma. In this study, we present findings from phase I and in vivo studies for PT-112 in relapsed or refractory multiple myeloma. Experimental Design: PT-112 biodistribution was analyzed in mice via laser ablation inductively coupled plasma mass spectrometry. The activity of PT-112 was assessed in de novo and transplantable Tg(Igkv3-5*-MYC)#Plbe (Vk*MYC) multiple myeloma mouse models as monotherapy or combination therapies. M-spike levels and survival were measured. A phase I dose escalation study of PT-112 monotherapy was conducted using a 3 + 3 design in patients with heavily pretreated relapsed or refractory multiple myeloma with exhausted available therapies. Results: In vivo biodistribution imaging revealed high concentrations in the bone, kidney, lung, skin, and liver. PT-112 was active in Vk*MYC multiple myeloma mouse models, both alone and in combination. Phase I data showed that PT-112 monotherapy was safe and well-tolerated, establishing a recommended phase II dose of 360 mg/m2 on days 1, 8, and 15 of a 28-day cycle. Confirmed responses and other signals of activity were observed. Conclusions: These results suggest a lack of cross-resistance with the standard of care and support the translational value of the Vk*MYC model system. Further clinical investigation of PT-112 is warranted in multiple myeloma.

Abstract Objectives High-grade B-cell lymphoma (HGBL), introduced in the 2016 World Health Organization (WHO) revised fourth edition classification, included cases defined by MYC and BCL2 and/or BCL6 rearrangements or by high-grade morphology. Diagnostic criteria and nomenclature for these lymphomas were refined in the 2022 WHO fifth edition (WHO-5) classification and International Consensus Classification (ICC). This review describes our approach to the diagnosis of HGBL. Methods Two cases are presented illustrating how we diagnose HGBL, including 1 case harboring MYC and BCL6 rearrangements and a second showing TdT expression in an HGBL with MYC and BCL2 rearrangements. The ways in which these cases are distinguished from other lymphomas with high-grade features and the appropriate nomenclature using WHO-5 and ICC classifications are emphasized. Results An HGBL diagnosis requires integration of morphology, immunophenotype, and genetics and exclusion of other lymphomas with high-grade morphology, including Burkitt lymphoma, B-lymphoblastic leukemia/lymphoma (B-LBL/ALL), and blastoid mantle cell lymphoma. A diagnosis of HGBL/large B-cell lymphoma with 11q aberration should also be considered in certain patient populations. Conclusions High-grade B-cell lymphomas are subclassified based on morphologic and genetic features. There are differences in the nomenclature and definition of these lymphomas in the WHO-5 and ICC classifications. Distinguishing HGBLs from other mature B-cell lymphomas and B-LBL/ALL is critical so that patients receive appropriate treatment.

Abstract Objectives VEXAS syndrome is an adult-onset autoinflammatory disease caused by a somatic pathogenic mutation in the UBA1 (ubiquitin-like modifier activating enzyme 1) gene. Patients present with rheumatologic manifestations and cytopenias and may have an increased predisposition to myelodysplastic syndrome (MDS) and plasma cell neoplasms. Prior studies have reported on the peripheral blood and bone marrow findings in patients with VEXAS syndrome. Due to the protean clinical presentation and lack of specificity of morphologic features (eg, vacuoles in early erythroid and granulocytic precursors), an optimal screening methodology to identify these patients in a timely fashion is desirable. Methods To further evaluate and describe the salient diagnostic morphologic features in VEXAS syndrome, we carried out a comprehensive study of the largest single-institution cohort to date. Diagnostic and follow-up bone marrow biopsy specimens from 52 male patients with molecularly identified VEXAS syndrome underwent central review. Results Cytopenias were common in all cases, primarily macrocytic anemia, monocytopenia, and thrombocytopenia. Bone marrow aspirate and biopsy were often hypercellular, with an increased myeloid/erythroid ratio, granulocytic hyperplasia with left shift, erythroid left shift, and megakaryocyte hyperplasia, which exhibited a range of striking morphologic findings. Distinctly vacuolated myeloid and erythroid precursors were seen in more than 95% of cases. Conclusions Our data reveal potential novel diagnostic features, such as a high incidence of monocytopenia and distinct patterns of atypical megakaryopoiesis, that appear different from dysmegakaryopoiesis typically associated with MDS. In our experience, those findings are suggestive of VEXAS, in the appropriate clinical context.

Introduction: We previously showed that HIV associated EBV negative (-) diffuse large B-cell lymphomas (DLBCL) of germinal center origins (GCB) have enhanced genomic stability and increased expression of DNA repair genes compared to their HIV negative counterparts (Maguire et al 2019). Given that immune deficiency can be induced virally or therapeutically, including the administration of immunosuppressive drugs post solid organ or hematopoietic stem cell transplant, we questioned whether these observations are specific to HIV or if they represent a more global immune impaired signature. Here we assessed the genomic stability of monomorphic DLBCL post-transplant lymphoproliferative disorders (PTLD) and the potential role of DNA repair. Methods: A total of 49 monomorphic DLBCL PTLDs (26 EBV(-), 23 EBV positive (+)) and 170 immune competent (IC) DLBCL (EBV status unknown) were studied. All were pathologically reviewed and samples with &lt;60% tumor content were macro-dissected before RNA/DNA extraction. Cell of origin (COO) profiles were assessed using the DLBCL90 molecular assay which can distinguish primary mediastinal B-cell lymphoma (PMBL) from GCB, Activated B-cell (ABC) and unclassified (UNC) DLBCL (Ennishi et al 2019). NanoString gene expression profiling (GEP) was performed using the PanCancer Pathways panel and analyzed using NanoString nSolver Advanced Analysis. Copy number (CN) analysis was performed on a subset of samples (45 PTLD, 41 DLBCL) using the OncoScan CN assay and analyzed using Chromosome Analysis Suite (ChAS) v4.1. All p-values listed were Benjamini-Hochberg adjusted and Mann-Whitney tested for the GEP and CN data respectively. Results: In contrast to IC-DLBCL which are GCB enriched (50%), COO profiling revealed PTLD to be ABC enriched (43%). CN analyses revealed that the median CN/sample for PTLD (0.57, p&lt;0.001), EBV(+)PTLD (0.25, p&lt;0.001) and EBV(-)PTLD (0.80, p = 0.050) tumors were significantly less aberrant than IC-DLBCL (1.01). The data also showed that EBV(+)PTLD are less aberrant than EBV(-)PTLD tumors (0.25 vs 0.80, p&lt;0.001). GEP pathway analysis identified DNA repair as downregulated in PTLD and EBV(+)PTLD but upregulated in EBV(-)PTLD compared IC-DLBCL. Nine DNA repair genes ( RAD51, CCNA2a, CHEK1b, POLR2Da, PCNAa, WHSC1a, GTF2H3a, ATM c, EP300 c) were differentially expressed (DE) in EBV(-)PTLD compared to IC-DLBCL (7 increased, 2 decreased, p ≤ 0.0425): where the most overexpressed gene RAD51, a critical mediator of Homologous Recombination (HR) and the Fanconi Anemia (FA) DNA repair pathways, was not DE in PTLD or EBV(+)PTLD compared to IC-DLBCL. Five a genes were increased in PTLD (p ≤ 0.0204) but not EBV(+)PTLD, 1 b was decreased in EBV(+)PTLD (p = 0.01) but not PTLD and 2 c were reduced in PTLD regardless of EBV status compared to IC-DLBCL (p ≤ 0.0024). Twenty-8 genes were DE in EBV(+)PTLD compared to IC-DLBCL. Of these, 23 were downregulated (p ≤ 0.0418) and included HR, FA and mismatch repair genes ( CHEK1, CHEK2, BRAC1/FANCS, FANCA, FANCB, FANCE, MSH2, MSH6). All but 2/23 were similarly expressed in PTLD compared to IC-DLBCL of which 11 were significant (p ≤ 0.0356). The five upregulated genes ( ZBTB32, DDB2, NBN, POLB, RBX1, p ≤ 0.0092) were also upregulated in PTLD (p ≤ 0.0042) but not EBV(-)PTLD compared to IC-DLBCL. Two additional genes ( SKP1, FBXW7) were also identified as increased in EBV(+)PTLD and PTLD (p ≤ 0.0204) compared to IC-DLBCL. In contrast, SKP1 was increased (p = 2.7E-4) and FBXW7 was decreased (p = 0.3610) in EBV(-)PTLD compared to IC-DLBCL. Together the protein products of RBX1, SKP1 and FBXW7 comprise the functional cores of a complete UPS complex, the SCF FBXW7 E3 ubiquitin ligase which have been reported to regulate the DNA damage response and promote genomic stability. Conclusions: Like HIV associated DLBCL, these results show that PTLD is more genomically stable than IC-DLBCL, and that genomic stability is further enhanced in EBV(+)PTLD which have increased expression of UPS related genes. In contrast, EBV(-)PTLD have increased expression of DNA damage repair genes, including RAD51. In summary, monomorphic DLBCL PTLD tumors have enhanced genomic stability that appears to be mediated by two distinct mechanisms stratified by EBV status. Further work is required to assess if this stratification of mechanisms is unique to PTLD or reflects a more global immune impaired EBV related signature.

Introduction: Previously we demonstrated that HIV associated diffuse large B-cell lymphomas (DLBCL) of Germinal Center B-cell subtype (GCB) are more proliferative, with enhanced genomic stability and increased expression of DNA repair genes, particularly Fanconi anemia (FA) and homologous recombination (HR) genes compared to their HIV negative counterparts (Maguire et al, 2019). It is unknown if this signature is unique to the GCB subtype or if it is also observed in Activated B-cell (ABC) and Unclassified (UNC) DLBCL. Herein we validate our HIV associated GCB DLBCL findings in a larger independent cohort and report for the first time the aberrational profile of molecularly profiled HIV associated ABC DLBCL compared to their HIV negative counterparts respectively. Methods: An independent cohort of 353 DLBCL FFPE cases was assembled, including 125 HIV-DLBCL samples. All samples underwent hematopathological review to confirm diagnosis and determine tumor content. Samples with &lt;60% tumor content were macro-dissected before they were RNA/DNA extracted. Cell of origin (COO) profiles were assessed using the Lymph3Cx molecular assay which can distinguish primary mediastinal B-cell lymphoma (PMBL) like from GCB, ABC and UNC DLBCL (Mottok et al 2018). Copy number (CN) analyses were performed on samples with sufficient DNA (41 HIV neg, 32 HIV pos) using the OncoScan FFPE assay kit and the resulting data analyzed using Chromosome Analysis Suite (ChAS) v4.1 and the multisample viewer module of Reproductive Health Research Analysis Software (RHAS) v1.1. Additional digital gene expression profiling was performed using the Pan-Cancer panel (NanoString, Seattle, WA) customized to include 9 BCL2 family genes for a total of 739 genes and analyzed using the NanoString nSolver Advanced Analysis module. Results: GCB, ABC, UNC and PMBL-like proportions were 50%, 30%, 14% and 6% in HIV unassociated (HIV neg) and 50%, 23%, 25% and 2% in HIV pos tumors respectively, confirming both cohorts were GCB-enriched by molecular profiling. PMBL-like cases were excluded from further analyses. OncoScan CN analyses showed that HIV pos DLBCL were significantly less aberrant than HIV neg DLBCL (median CN per sample 0.56 vs 1.01, p&lt;0.001). HIV pos GCB tumors were significantly less aberrant than both HIV neg GCB tumors (median CN per sample 0.50 vs 1.11, p&lt;0.001), and HIV pos ABC tumors (median CN per sample 0.50 vs 0.81, p=0.009). HIV pos ABC tumors were also less aberrant than HIV neg ABC tumors, however, the difference was not significant overall (median CN per sample 0.81 vs 1.03, p=0.162) or for gain/amplifications (median CN per sample 0.38 vs 0.42, p=0.814), but was borderline significant for loss/deletions (median CN per sample 0.31 vs 0.47, p=0.06). GEP analyses identified Cell Cycle and DNA repair as the top two upregulated pathways in HIV pos GCB tumors compared to HIV neg GCB tumors. In contrast, Cell Cycle and DNA repair were downregulated in HIV pos ABC tumors compared to HIV neg ABC tumors. However, one gene ( RBX1) was identified as significantly upregulated in both pathways in HIV pos GCB (adjusted p-value 0.0215) and HIV pos ABC (adjusted p-value 0.0046) tumors compared to their HIV neg counterparts. A second gene ( SKP1) was also significantly upregulated in HIV pos GCB (adjusted p-value 3.94E-5) and HIV pos ABC (adjusted p-value 0.0031) tumors. Together RBX1 and SKP1 encode the two principal functional subunits of the SKP1, CUL1, F-box protein (SCF) complex which is part of the ubiquitin proteasome system (UPS) and has been reported as essential for maintaining genome and chromosome stability (Thompson et al 2021). Conclusions: These results demonstrate that HIV pos DLBCL tumors are less aberrant than their HIV neg counterparts irrespective of COO subtype. However, the degree of genomic stability differs by COO, with HIV pos GCB tumors being more genomically stable than HIV pos ABC tumors. Although both have enhanced expression of UPS related SCF genes ( RBX1, SKP1), only HIV pos GCB tumors have enhanced expression of key DNA repair genes as well including Fanconi Anemia ( BRCA2/FANCD1, FANCL) and mismatch repair ( MSH2, MSH6) genes (adjusted p-values of ≤ 0.0263). Given the reported role of SCF in genomic stability maintenance, the HIV pos COO CN differences could be due to the contributions of both DNA repair and SCF mechanisms in HIV pos GCB and mainly SCF mechanisms in HIV pos ABC tumors.

Abstract We investigated the clinicopathologic features of 5 follicular lymphomas (FLs) that transformed (tFL) morphologically to diffuse large B-cell lymphomas (DLBCLs) and had a primary mediastinal large B-cell lymphoma (PMBL)–like gene expression profile (tFL-PMBLsig-pos). None of the tFL-PMBLsig-pos cases arose in the mediastinum, all cases tested had a germinal center B-cell phenotype, 20% were CD30+, 60% CD23+, 80% MAL+, 20% CD200+, and 0% CD273/PDL2+. Whole-exome sequencing detected alterations in genes associated with both FL/DLBCL (CREBBP, KMT2C, KMT2D, ARID1A, HIST1 members, and TNFRSF14) and PMBL (JAK-STAT pathway genes, B2M, and CD58). Copy number (CN) analysis detected gains/amplification of REL and STAT6 in 60%, gains of SOCS1 in 40%, and gains of chromosome 16, including IL4R, in 40% of the cases. CN gains/amplification of BCL6 and MYC and loss of TNFRSF14 and TNFAIP3 were identified in 20% of the cases. Three of 5 cases lacked a BCL2 rearrangement. Despite having some features that are less common in DLBCL (MAL and CD23 expression and JAK-STAT activation), these tFL-PMBLsig-pos cases lack the most characteristic CN alteration seen in PMBL (9p24.1 gain/amplification). This cohort expands the biologic heterogeneity of tFL, illustrating a subset with gene expression and some genetic features reminiscent of PMBL, with potential treatment implications that include the use of novel targeted therapies.