The annual meeting of the American Society of Hematology (ASH) is considered the most important congress on hematologic diseases worldwide. International experts again exchanged information on current research results and the most important innovations in diagnostics and therapy in December 2022. A wide range of hematology topics were covered – from classification to artificial intelligence to
Toward prognostic makers.
According to the new WHO classification of hematologic neoplasms, genetics is playing an increasingly important role. Therefore, genetics-based definitions are assigned a supporting function. For example, TP53 alterations resulting from deletion, mutation, or copy-neutral loss of heterozygosity should be considered. A single-hit (sh) or double-hit (dh; two or more than two changes) can be distinguished. In a cohort study of 1520 patients with MDS and AML, TP53dh was identified as the most important prognostic factor [1]. MDS cases with <5% blasts differ from MDS cases with ≥5% blasts and from AML cases by the predominance of TP53sh and by being the only subgroup in which a complex karyotype (CK) showed an independent negative impact on overall survival OS. Thus, the presence of a CK generally plays a minor prognostic role. The remaining subgroups (MDS ≥5%, <10% blasts; ≥10%, <20% blasts; AML) shared a number of similarities, with TP53dh occurring in high frequency and only TP53alt, but not a CK, independently affecting OS. In the overall cohort, TP53dh was the strongest prognostic factor; however, blast count also influences OS independent of TP53 allele status.
The time factor in AML relapses
Although most patients with newly diagnosed acute myeloid leukemia (AML) achieve complete remission (CR) after induction chemotherapy, most subsequently relapse. Previous studies have demonstrated the importance of several prognostic factors for relapse-free survival, including age, cytogenetics, prior hematologic disease (AHD), and prior cancer therapy. However, additional factors such as measurable residual disease (MRD) status, time to count recovery, and type of remission (CR vs. CR with incomplete count recovery [CRi]) are now considered when assessing treatment response. In patients in complete remission, the impact of additional initial response variables on relapse or death was examined at various time points up to three years, a time point after which AML relapse is very rare [2].
Of 972 patients with a confirmed diagnosis of AML or another high-grade myeloid neoplasm (≥10% blasts), 656 achieved morphologic remission (CR or CRi) and were included in the analysis. The primary end points were survival without relapse at 1, 2, or 3 years. The median age at diagnosis was 60 years. De novo AML was identified in 373 patients (57%), and the remainder were a mixture of antecedent AHD and treatment-related AML. ELN2017 risk classification was favorable in 181 patients (28%), intermediate in 206 (41%), and unfavorable in 203 (31%). Most patients received high- or medium-intensity induction therapy (88%). After induction, 540 patients (82%) achieved CR and 116 patients (18%) achieved CRi. MRD after induction therapy was observed in 173 patients (26%). In multivariable analysis, age, type of remission (CRi), and MRD status were the strongest predictors of survival without relapse at years 1, 2, and 3. In the model for year 3, CR MRD+ was associated with the largest decrease in the probability of surviving without relapse after induction. The 3-year relapse-free survival (RFS) was 18.5% for MRD+ patients versus 36.4% for MRD- patients. Ultimately, remission category (CR vs. CRi), MRD status, and age were found to be the strongest predictors of survival without relapse at years 1, 2, or 3 after induction in patients with AML and other high-grade myeloid neoplasms.
Landscape of single cell pathology
Diffuse large B-cell lymphoma is a heterogeneous disease with established patterns of recurrent molecular and genetic features. However, to date, little is known about the cellular composition of DLBCL tumors and the underlying biology of the tumor microenvironment (TME). Therefore, imaging mass cytometry was used to simultaneously quantify two groups of 34 protein markers in serial sections of tumor tissue [3]. A total of 57 unique markers identifying major cell lineages (CD3, CD20, PDPN, CD68, etc.), immune functions (IDO, PD1, granzyme B, etc.), and tumor phenotypes (IRF4, BCL6, p53, etc.) were quantified at single cell resolution in 545 tumor cores from 328 primary DLBCL tumors.
Single cell clustering and cell community analysis identified 27 different cell types and 19 cell communities with 5 overarching patterns of tumor composition and structure. “Cold” microenvironments, i.e., tumors poor in infiltrating immune and stromal cells, were the most common, accounting for 35.1% of the imaged tumors. “Cytotoxic” TMEs (20.7%) were significantly enriched in CD8 T cells expressing granzyme B and LAG3, and CD163+ M2-like macrophages expressing IDO and S100A9. “Stromal” tumors (21.3%) were enriched in PDPN-positive stroma and M1-like macrophages. Separately, it was also found that 98% of cases with replicated tumor cores had no statistically significant differences in tumor composition between cores.
Multivariate Cox proportional hazard analysis showed that TME clusters were significantly associated with both overall survival and progression-free survival. Notably, cytotoxic TMEs were found to have the most unfavorable clinical outcomes, while cold tumors showed a slight trend toward worse survival.
A total of 23 phenotypic markers of B-cell malignant populations were examined for associations with TME. Malignant B cells in the cytotoxic TMEs showed significantly increased MHC-I, PD-L1, and vimentin expression and were enriched in ABC-like DLBCLs. Cold TMEs had significantly increased BCL2 and Ki67 proliferation index and decreased MHC-I expression. Stromal TMEs were enriched with GCB-like DLBCLs.
Taken together, these results suggest that cold and cytotoxic TMEs may represent a dichotomy of two distinct pathways to aggressive disease-either extreme immune dysfunction (cytotoxic) or evasion (cold). Identification of TME composition in patients will improve stratification and may provide helpful context for the milieu of immunologic therapies currently being investigated for the treatment of DLBCL.
Tumor architecture in multiple myeloma
The spatial aspects of immune infiltration have been extensively studied in the context of solid tumors and are related to response to immunotherapy. Immune-based approaches have begun to change the therapeutic landscape in hematologic malignancies such as multiple myeloma (MM) and its precursors, monoclonal gammopathy of undetermined significance (MGUS) and smoldering myeloma (SMM). The bone marrow (BM) immunological microenvironment was studied in detail in these cases with single cell analyses of BM aspirates. However, analysis of the spatial patterns of tumor/immune infiltration in trephine biopsies and the underlying mechanisms are poorly understood. Therefore, high-dimensional spatial analysis was combined with machine learning to analyze BM biopsies from patients with newly diagnosed MM, MGUS, and SMM. The mechanisms underlying the infiltration of T cells into the tumor were investigated using in vitro models. Growth patterns of tumor cells and mechanisms underlying infiltration of neoantigen-specific T cells have also been studied using the humanized MISTRG6 mouse model [4].
MM biopsies, but not MGUS biopsies, showed a multifocal pattern of tumor growth. Dense clusters of tumor cells were observed in MM that appeared to form regions of T-cell exclusion. Compared with MGUS, MM biopsies showed a lower density of TCF1+ stem-like T cells, an increased proportion of granzyme B+ CD8+ effector T cells, and increased expression of myeloid cells. Multifocal growth of MM, but not MGUS, was reproduced in humanized MISTRG6 mice. This suggests that this is an intrinsic property associated with malignancy.
Consistent with the IHC data, MM tumors were resistant to T cell infiltration in in vitro models. Prior activation of T cells resulted in enhanced T cell infiltration and was optimally achieved after anti-CD3/CD2/CD28 stimulation. T cell infiltration also required CD2/CD58 interactions and was abrogated by disruption of these interactions. Neoantigen-specific T cells readily recognize MM cells in culture. However, entry of neoantigen-specific T cells into clusters of antigen-expressing MM tumors was antigen-specific and required in situ stimulation with antigen-presenting dendritic cells (DCs). After transfer into humanized mice, neoantigen-specific T cells can be easily localized to the tumor. However, entry of these T cells into antigen-expressing tumor masses in vivo again depends on tumor-associated DCs.
Artificial intelligence in action
Artificial intelligence is increasingly finding its way into medicine to facilitate workflows or verify diagnostics. Hematopoietic neoplasms, for example, are diagnosed using a combination of different methods. They require complex equipment and highly skilled clinical laboratory scientists and technicians, which increase both turnaround times and costs. Now, a web-based classifier has been developed that can detect 33 different hematopoietic neoplasms and normal peripheral blood/bone marrow. To evaluate the performance of the model, 325 samples were prospectively sequenced using both WGS and WTS on NovaSeq instruments [5]. Single nucleotide variants (SNV), structural variants (SV), and copy number alterations (CNA) were extracted from WGS data with a tumor-without-normal pipeline, and gene fusions (GF) and gene expression (GE) from WTS. SNVs were filtered using common databases. In parallel, an independent final routine diagnosis was made based on gold standard procedures (GST) according to WHO guidelines.
The main focus for direct comparison in this prospective study was AML and ALL cases. AML was correctly classified with the highest probability in 90% of cases. 6/10 misclassified AMLs were classified as MDS. In these cases, the second most likely predicted disease was AML in 3/6 cases. Only one BCP-ALL was misclassified, with 98% of samples correctly predicted with an MP of 94%. 11/14 (79%) T-ALLs were correctly predicted, also with an MP of 94%. In the 3/14 false cases, the mp was 39%, with a second-highest prediction in a similar probability range (38% vs. 31%, 39% vs. 34%, and 61% vs. 24%). In addition, our cohort included 26 cases of multiple myeloma (MM), 24 of which were correctly predicted and two of which were misclassified as MGUS but with an almost equal probability of MM.
The role of clonal plasma cells in MM
Autologous hematopoietic stem cell transplantation (autoHCT) is considered the standard treatment for patients with multiple myeloma. However, most patients relapse, possibly due to the presence of clonal plasma cells (CPC) in the autograft. The role of CPC in autograft is not yet clear. MM patients with high-risk chromosomal abnormalities (HRMM) detected by florescence in situ hybridization (FISH) have worse outcomes compared with patients with standard-risk disease.
In a retrospective analysis, autograft products from 416 patients were studied: 75 CPC+ (18%) and 341 CPC- (82%) [6]. Fifty-seven percent of patients were male, and the mean age at auto-HCT was 62 years. Fewer patients in the CPC+ group than in the CPC- group received the VRD regimen as induction before transplantation (25% vs. 44%) and fewer patients achieved ≥VGPR after induction than in the CPC- group (32% vs. 62%). The median follow-up time for the entire cohort was 35.7 months. The 100-day rate and best CR rate after auto-HCT were 8% and 19% and 33% and 54% in the CPC+ and CPC- groups, respectively. The CPC+ group was also less likely to achieve MRD-negative CR after transplantation (11% vs. 42%). The median PFS in the CPC+ versus CPC- group was 12.8 and 32.1 months, respectively, and the median OS was 36.4 and 81.2 months, respectively.
The study showed a large impact of CPC in autograft on outcomes after auto-HCT in HRMM. Both the presence and degree of CPC in the autograft were highly predictive of worse PFS and OS, even in patients who had achieved ≥VGPR and MRD-negative CR/VGPR before auto-HCT. Novel strategies for ex vivo purification of CPC may improve patient outcomes.
Congress: 64th Annual Meeting of the American Society of Hematology (ASH)
Literature:
- Stengel A, Meggendorfer M, Walter W, et al: Interplay of TP53 Allelic State, Blast Count and Karyotype on Survival of Patients with AML and MDS. Blood 2022; 140(Supplement 1): 2073-2074.
- Lim J, Othus M, Shaw CM, et al: Time-Dependent Prediction of Relapse in Patients with Acute Myeloid Leukemia. Abstract 57. 10/12/2022, ASH 2022.
- McNally D, Lytle A, Ravichandran H, et al: The Single-Cell Pathology Landscape of Diffuse Large B Cell Lymphoma. Abstract 67. 10 Dec 2022, ASH 2022.
- Robinson H, Villa N, Jaye DL, et al: Tumor-Immune Architecture and the Regulation of Antigen-Specific T-Cell Infiltration in Multiple Myeloma and Premalignant Plasma Cell Disorders. Abstract 98. 10 Dec 2022, ASH 2022.
- Nadarajah N, Maschek S, Hutter S, et al: Evaluation of a Transparent Artificial Intelligence (AI) Disease Classification System with Whole Genome Sequencing (WGS) and Whole Transcriptome Sequencing (WTS) Data in a Prospective Study with 325 Cases. Blood 2022; 140(Supplement 1): 1915-1916.
- Pasvolsky O, Milton DR, Rauf M, et al: Impact of Presence and Amount of Clonal Plasma Cells in Autografts Affect Outcomes in High-Risk Multiple Myeloma Patients Undergoing Autologous Hematopoietic Stem Cell Transplant. Abstract 115. 10/12/2022, ASH 2022.
InFo ONCOLOGY & HEMATOLOGY 2023; 11(1): 20-22 (published 2/28-23, ahead of print).