Statistical tests for intra-tumour clonal co-occurrence and exclusivity.

Tumour progression is an evolutionary process in which different clones evolve over time, leading to intra-tumour heterogeneity. Interactions between clones can affect tumour evolution and hence disease progression and treatment outcome. Intra-tumoural pairs of mutations that are overrepresented in a co-occurring or clonally exclusive fashion over a cohort of patient samples may be suggestive of a synergistic effect between the different clones carrying these mutations. We therefore developed a novel statistical testing framework, called GeneAccord, to identify such gene pairs that are altered in distinct subclones of the same tumour. We analysed our framework for calibration and power. By comparing its performance to baseline methods, we demonstrate that to control type I errors, it is essential to account for the evolutionary dependencies among clones. In applying GeneAccord to the single-cell sequencing of a cohort of 123 acute myeloid leukaemia patients, we find 1 clonally co-occurring and 8 clonally exclusive gene pairs. The clonally exclusive pairs mostly involve genes of the key signalling pathways.

CombiFlow: Flow cytometry-based identification and characterization of genetically and functionally distinct AML subclones.

Many cancers, including leukemias, are dynamic oligoclonal diseases. Tools to identify and prospectively isolate genetically distinct clones for functional studies are needed. We describe our CombiFlow protocol, which is a combinatorial flow cytometry-based approach to identify and isolate such distinct clones. CombiFlow enables the visualization of clonal evolution during disease progression and the identification of potential relapse-inducing cells at minimal residual disease (MRD) time points. The protocol can be adapted to various research questions and allows functional studies on live sorted cell populations. For complete details on the use and execution of this protocol, please refer to de Boer et al. (2018).

Advances in acute myeloid leukemia.

Acute myeloid leukemia (AML) is an uncommon but potentially catastrophic diagnosis with historically high mortality rates. The standard of care treatment remained unchanged for decades; however, recent discoveries of molecular drivers of leukemogenesis and disease progression have led to novel therapies for AML. Ongoing research and clinical trials are actively seeking to personalize therapy by identifying molecular targets, discovering patient specific and disease specific risk factors, and identifying effective combinations of modalities and drugs. This review focuses on important updates in diagnostic and disease classifications that reflect new understanding of the biology of AML, its mutational heterogeneity, some important genetic and environmental risk factors, and new treatment options including cytotoxic chemotherapy, novel targeted agents, and cellular therapies.

3+7 Combined Chemotherapy for Acute Myeloid Leukemia: Is It Time to Say Goodbye?

PURPOSE OF REVIEW: With the recent approval of multiple new drugs for the treatment of acute myeloid leukemia (AML), the relevance of conventional treatment approaches, such as daunorubicin and cytarabine ("3+7") induction chemotherapy, has been challenged. We review the AML risk stratification, the efficacy of the newly approved drugs, and the role of "3+7". RECENT FINDINGS: Treatment of AML is becoming more niched with specific subtypes more appropriately treated with gemtuzumab, midostaurin, and CPX-351. Although lower intensity therapies can yield high response rates, they are less efficient at preventing relapses. The only curative potential for poor-risk AML is still an allogeneic stem cell transplant. The number of AML subtypes where 3+7 alone is an appropriate therapeutic option is shrinking. However, it remains the backbone for combination therapy with newer agents in patients suitable for intensive chemotherapy.

Association Between Measurable Residual Disease in Patients With Intermediate-Risk Acute Myeloid Leukemia and First Remission, Treatment, and Outcomes.

Importance: Measurable residual disease (MRD) is widely used as a therapy-stratification factor for acute myeloid leukemia (AML), but the association of dynamic MRD with postremission treatment (PRT) in patients with intermediate-risk AML (IR-AML) has not been well investigated. Objective: To investigate PRT choices based on dynamic MRD in patients with IR-AML. Design, Setting, and Participants: This cohort study examined 549 younger patients with de novo IR-AML in the South China Hematology Alliance database during the period from January 1, 2012, to June 30, 2016, including 154 who received chemotherapy, 116 who received an autologous stem cell transplant (auto-SCT), and 279 who received an allogeneic SCT (allo-SCT). Subgroup analyses were performed according to dynamic MRD after the first, second, and third courses of chemotherapy. The end point of the last follow-up was August 31, 2020. Statistical analysis was performed from December 1, 2019, to September 30, 2020. Exposures: Receipt of chemotherapy, auto-SCT, or allo-SCT. Main Outcomes and Measures: The primary end points were 5-year cumulative incidence of relapse and leukemia-free survival. Results: Subgroup analyses were performed for 549 participants (314 male participants [57.2%]; median age, 37 years [range, 14-60 years]) according to the dynamics of MRD after 1, 2, or 3 courses of chemotherapy. Comparable cumulative incidences of relapse, leukemia-free survival, and overall survival were observed among participants who had no MRD after 1, 2, or 3 courses of chemotherapy. Participants who underwent chemotherapy and those who underwent auto-SCT had better graft-vs-host disease-free, relapse-free survival (GRFS) than those who underwent allo-SCT (chemotherapy: hazard ratio [HR], 0.35 [95% CI, 0.14-0.90]; P = .03; auto-SCT: HR, 0.07 [95% CI, 0.01-0.58]; P = .01). Among participants with MRD after 1 course of chemotherapy but no MRD after 2 or 3 courses, those who underwent auto-SCT and allo-SCT showed lower cumulative incidence of relapse (auto-SCT: HR, 0.25 [95% CI, 0.08-0.78]; P = .01; allo-SCT: HR, 0.08 [95% CI, 0.02-0.24]; P < .001), better leukemia-free survival (auto-SCT: HR, 0.26 [95% CI, 0.10-0.64]; P = .004; allo-SCT: HR, 0.21 [95% CI, 0.09-0.46]; P < .001), and overall survival (auto-SCT: HR, 0.22 [95% CI, 0.08-0.64]; P = .005; allo-SCT: HR, 0.25 [95% CI, 0.11-0.59]; P = .001) vs chemotherapy. In addition, auto-SCT showed better GRFS than allo-SCT (HR, 0.45 [95% CI, 0.21-0.98]; P = .04) in this group. Among participants with MRD after 1 or 2 courses of chemotherapy but no MRD after 3 courses, allo-SCT had superior cumulative incidence of relapse (HR, 0.10 [95% CI, 0.06-0.94]; P = .04) and leukemia-free survival (HR, 0.18 [95% CI, 0.05-0.68]; P = .01) compared with chemotherapy, but no advantageous cumulative incidence of relapse (HR, 0.15 [95% CI, 0.02-1.42]; P = .10) and leukemia-free survival (HR, 0.23 [95% CI, 0.05-1.08]; P = .06) compared with auto-SCT. Among participants with MRD after 3 courses of chemotherapy, allo-SCT had superior cumulative incidences of relapse, leukemia-free survival, and overall survival compared with chemotherapy (relapse: HR, 0.16 [95% CI, 0.08-0.33]; P < .001; leukemia-free survival: HR, 0.19 [95% CI, 0.10-0.35]; P < .001; overall survival: HR, 0.29 [95% CI, 0.15-0.55]; P < .001) and auto-SCT (relapse: HR, 0.25 [95% CI, 0.12-0.53]; P < .001; leukemia-free survival: HR, 0.35 [95% CI, 0.18-0.73]; P = .004; overall survival: HR, 0.54 [95% CI, 0.26-0.94]; P = .04). Among participants with recurrent MRD, allo-SCT was also associated with advantageous cumulative incidence of relapse, leukemia-free survival, and overall survival compared with chemotherapy (relapse: HR, 0.12 [95% CI, 0.04-0.33]; P < .001; leukemia-free survival: HR, 0.24 [95% CI, 0.10-0.56]; P = .001; overall survival: HR, 0.31 [95% CI, 0.13-0.75]; P = .01) and auto-SCT (relapse: HR, 0.28 [95% CI, 0.09-0.81]; P = .02; leukemia-free survival: HR, 0.30 [95% CI, 0.12-0.76]; P = .01; overall survival: HR, 0.26 [95% CI, 0.10-0.70]; P = .007). Conclusions and Relevance: This study suggests that clinical decisions based on dynamic MRD might be associated with improved therapy stratification and optimized PRT for patients with IR-AML. Prospective multicenter trials are needed to further validate these findings.

DNA Methylation Signatures Predict Cytogenetic Subtype and Outcome in Pediatric Acute Myeloid Leukemia (AML).

Pediatric acute myeloid leukemia (AML) is a heterogeneous disease composed of clinically relevant subtypes defined by recurrent cytogenetic aberrations. The majority of the aberrations used in risk grouping for treatment decisions are extensively studied, but still a large proportion of pediatric AML patients remain cytogenetically undefined and would therefore benefit from additional molecular investigation. As aberrant epigenetic regulation has been widely observed during leukemogenesis, we hypothesized that DNA methylation signatures could be used to predict molecular subtypes and identify signatures with prognostic impact in AML. To study genome-wide DNA methylation, we analyzed 123 diagnostic and 19 relapse AML samples on Illumina 450k DNA methylation arrays. We designed and validated DNA methylation-based classifiers for AML cytogenetic subtype, resulting in an overall test accuracy of 91%. Furthermore, we identified methylation signatures associated with outcome in t(8;21)/RUNX1-RUNX1T1, normal karyotype, and MLL/KMT2A-rearranged subgroups (p < 0.01). Overall, these results further underscore the clinical value of DNA methylation analysis in AML.

Machine learning integrates genomic signatures for subclassification beyond primary and secondary acute myeloid leukemia.

Although genomic alterations drive the pathogenesis of acute myeloid leukemia (AML), traditional classifications are largely based on morphology, and prototypic genetic founder lesions define only a small proportion of AML patients. The historical subdivision of primary/de novo AML and secondary AML has shown to variably correlate with genetic patterns. The combinatorial complexity and heterogeneity of AML genomic architecture may have thus far precluded genomic-based subclassification to identify distinct molecularly defined subtypes more reflective of shared pathogenesis. We integrated cytogenetic and gene sequencing data from a multicenter cohort of 6788 AML patients that were analyzed using standard and machine learning methods to generate a novel AML molecular subclassification with biologic correlates corresponding to underlying pathogenesis. Standard supervised analyses resulted in modest cross-validation accuracy when attempting to use molecular patterns to predict traditional pathomorphologic AML classifications. We performed unsupervised analysis by applying the Bayesian latent class method that identified 4 unique genomic clusters of distinct prognoses. Invariant genomic features driving each cluster were extracted and resulted in 97% cross-validation accuracy when used for genomic subclassification. Subclasses of AML defined by molecular signatures overlapped current pathomorphologic and clinically defined AML subtypes. We internally and externally validated our results and share an open-access molecular classification scheme for AML patients. Although the heterogeneity inherent in the genomic changes across nearly 7000 AML patients was too vast for traditional prediction methods, machine learning methods allowed for the definition of novel genomic AML subclasses, indicating that traditional pathomorphologic definitions may be less reflective of overlapping pathogenesis.

Isoform-specific and signaling-dependent propagation of acute myeloid leukemia by Wilms tumor 1.

Acute myeloid leukemia (AML) is caused by recurrent mutations in members of the gene regulatory and signaling machinery that control hematopoietic progenitor cell growth and differentiation. Here, we show that the transcription factor WT1 forms a major node in the rewired mutation-specific gene regulatory networks of multiple AML subtypes. WT1 is frequently either mutated or upregulated in AML, and its expression is predictive for relapse. The WT1 protein exists as multiple isoforms. For two main AML subtypes, we demonstrate that these isoforms exhibit differential patterns of binding and support contrasting biological activities, including enhanced proliferation. We also show that WT1 responds to oncogenic signaling and is part of a signaling-responsive transcription factor hub that controls AML growth. WT1 therefore plays a central and widespread role in AML biology.

Risk adapted therapeutic strategy in newly diagnosed acute myeloid leukemia: Refining the outcomes of ELN 2017 intermediate-risk patients.

INTRODUCTION: Despite advances in the treatment of acute myeloid leukemia (AML), cytotoxic chemotherapy remains the standard induction regimen. PATIENTS AND METHODS: In this single center retrospective study, we assessed outcomes of 99 consecutive adult AML patients treated with a risk-adapted strategy with a median follow-up of 35.5 months. RESULTS: We identified 24 (24 %), 55 (56 %) and 20 (20 %) patients classified as favorable-, intermediate-, and adverse- risk group respectively, according to the European LeukemiaNet (ELN) 2017 classification. Patients either received idarubicin and cytarabine induction chemotherapy with or without FLT3 inhibitors or hypomethylating agents based on age and comorbidity. The complete response (CR) rate was 76 % (82 % and 61 % in patients aged < 60 and ≥ 60, respectively). For the whole cohort, the 3-year overall survival (OS) was 53 %, being 62 % and 30 % in patients aged < 60 and ≥ 60, respectively. The 3-year leukemia-free survival (LFS) was 54 %, with 56 % and 45 % in patients aged < 60 and ≥ 60, respectively. The 3-year LFS were 58 %, 62 % and 25 % for patients within ELN favorable-, intermediate-, and adverse-risk groups respectively. Twenty-seven (36 %) out of 75 patients with intermediate- and adverse-risk disease underwent allogeneic hematopoietic cell transplantation (allo-HCT) in first CR with 92 % of them receiving post-transplant maintenance consisting of azacitidine in 19 (76 %) patients or sorafenib in 6 (24 %) patients. Of these patients younger than 60 years, the 3-year OS and LFS were 85 % and 69 %, respectively. CONCLUSION: These results indicate an improved OS for AML patients especially in intermediate-risk category who were treated with a total therapy consisting of induction chemotherapy followed by allo-HCT and post-transplant maintenance.

The transcriptome-wide landscape of molecular subtype-specific mRNA expression profiles in acute myeloid leukemia.

Molecular classification of acute myeloid leukemia (AML) aids prognostic stratification and clinical management. Our aim in this study is to identify transcriptome-wide mRNAs that are specific to each of the molecular subtypes of AML. We analyzed RNA-sequencing data of 955 AML samples from three cohorts, including the BeatAML project, the Cancer Genome Atlas, and a cohort of Swedish patients to provide a comprehensive transcriptome-wide view of subtype-specific mRNA expression. We identified 729 subtype-specific mRNAs, discovered in the BeatAML project and validated in the other two cohorts. Using unique proteomics data, we also validated the presence of subtype-specific mRNAs at the protein level, yielding a rich collection of potential protein-based biomarkers for the AML community. To enable the exploration of subtype-specific mRNA expression by the broader scientific community, we provide an interactive resource to the public.

Clinical features and outcomes of hypocellular acute myeloid leukemia in adults: A Korean AML registry data.

ABSTRACT: The hypocellular variant of acute myeloid leukemia (AML) is defined as bone marrow cellularity of <20% in a biopsy specimen at presentation. We performed a retrospective analysis of the clinical features and survival outcomes of hypocellular AML in a Korean population. We reviewed the medical records of all patients diagnosed with AML at nine hospitals participating in the Korean AML registry from 2006 to 2012. Overall survival (OS) and event-free survival (EFS) rates were calculated from the time of diagnosis until death or an event, respectively. In total, 2110 patients were enrolled and 102 (4.8%) were identified as having hypocellular AML. Patients with hypocellular AML were older than those with non-hypocellular AML (median age: 59 vs 49 years; P < .001) and presented with leukopenia more frequently (mean white blood cell count: 5810/µL vs 40549/µL; P < .001). There was no difference between patients with and without hypocellular AML in terms of the presence of antecedent hematologic disorders (5.9% vs 5.3%; P  = .809). FLT3-ITD and NPM1 mutations were less common in hypocellular than non-hypocellular AML (FLT3-ITD mutations: 1.2% vs 14.3%, P < .001; NPM1 mutations: 0% vs 9.5%, P = .019). No differences were seen between the hypocellular and non-hypocellular AML groups in the complete remission rate (53.9% vs 61.3%, P = .139) or early death rate (defined as any death before 8 weeks; 14.7% vs 13.0%, P = .629). The OS and EFS did not differ between the hypocellular and non-hypocellular AML groups (median OS: 16 vs 23 months, P = .169; median EFS: 6 vs 9 months, P = .215). Hypocellular AML is more frequently observed in older-aged patients and have fewer FLT3-ITD and NPM1 mutation, but the clinical outcomes of hypocellular AML do not differ from those of non-hypocellular AML.

Genomic characteristics and prognostic significance of co-mutated ASXL1/SRSF2 acute myeloid leukemia.

The ASXL1 and SRSF2 mutations in AML are frequently found in patients with preexisting myeloid malignancies and are individually associated with poor outcomes. In this multi-institutional retrospective analysis, we assessed the genetic features and clinical outcomes of 43 patients with ASXL1mut SRSF2mut AML and compared outcomes to patients with either ASXL1 (n = 57) or SRSF2 (n = 70) mutations. Twenty-six (60%) had secondary-AML (s-AML). Variant allele fractions suggested that SRSF2 mutations preceded ASXL1 mutational events. Median overall survival (OS) was 7.0 months (95% CI:3.8,15.3) and was significantly longer in patients with de novo vs s-AML (15.3 vs 6.4 months, respectively; P = .04 on adjusted analysis). Compared to ASXL1mut SRSF2wt and ASXL1wt SRSF2mut , co-mutated patients had a 1.4 and 1.6 times increase in the probability of death, respectively (P = .049), with a trend towards inferior OS (median OS = 7.0 vs 11.5 vs 10.9 months, respectively; P = .10). Multivariable analysis suggests this difference in OS is attributable to the high proportion of s-AML patients in the co-mutated cohort (60% vs 32% and 23%, respectively). Although this study is limited by the retrospective data collection and the relatively small sample size, these data suggest that ASXL1mut SRSF2mut AML is a distinct subgroup of AML frequently associated with s-AML and differs from ASXL1mut SRSF2wt /ASXL1wt SRSF2mut with respect to etiology and leukemogenesis.

Myeloid sarcoma, chloroma, or extramedullary acute myeloid leukemia tumor: A tale of misnomers, controversy and the unresolved.

The World Health Organization classification and definition of "myeloid sarcoma" is imprecise and misleading. A more accurate term is "extramedullary acute myeloid leukemia tumor (eAML)." The pathogenesis of eAML has been associated with aberrancy of cellular adhesion molecules, chemokine receptors/ligands and RAS-MAPK/ERK signaling. eAML can present with or without synchronous or metachronous intramedullary acute myeloid leukemia (AML) so a bone marrow evaluation is always recommended. Accurate diagnosis of eAML requires tissue biopsy. eAML confined to one or a few sites is frequently treated with local therapy such as radiotherapy. About 75-90% of patients with isolated eAML will develop metachronous intramedullary AML with a median latency period ranging from 4 to 12 months; thus, patients with isolated eAML may also be treated with systemic anti-leukemia therapy. eAML does not appear to have an independent prognostic impact; selection of post-remission therapy including allogeneic hematopoietic cell transplant (alloHCT) is typically guided by intramedullary disease risk. Management of isolated eAML should be individualized based on patient characteristics as well as eAML location and cytogenetic/molecular features. The role of PET/CT in eAML is also currently being elucidated. Improving outcomes of patients with eAML requires further knowledge of its etiology and mechanism(s) as well as therapeutic approaches beyond conventional chemotherapy, ideally in the context of controlled trials.

Laboratory quality assessment of candidate gene panel testing for acute myeloid leukaemia: a joint ALLG / RCPAQAP initiative.

Accurate classification of acute myeloid leukaemia (AML) has become increasingly reliant on molecular characterisation of this blood cancer. Throughout Australia and New Zealand massively parallel sequencing (MPS) is being adopted by diagnostic laboratories for the routine evaluation of patients with AML. This technology enables the surveying of many genes simultaneously, with many technical advantages over single gene testing approaches. However, there are many variations in wet and dry lab MPS procedures, which raises the prospect of discordant results between laboratories. This study compared the results obtained from MPS testing of ten diagnostic AML bone marrow aspirate samples sent to eight participating laboratories across Australasia. A reassuringly high concordance of 94% was observed with regard to variant detection and characterisation of pathogenicity. The level of discordance observed, although low, demonstrates the need for ongoing assessment of concordance between diagnostic testing laboratories through quality assurance programs.

How I diagnose and treat NPM1-mutated AML.

Mutations of the nucleophosmin (NPM1) gene, encoding for a nucleolar multifunctional protein, occur in approximately one-third of adult acute myeloid leukemia (AML). NPM1-mutated AML exhibits unique molecular, pathological, and clinical features, which led to its recognition as distinct entity in the 2017 World Health Organization (WHO) classification of myeloid neoplasms. Although WHO criteria for the diagnosis of NPM1-mutated AML are well established, its distinction from other AML entities may be difficult. Moreover, the percentage of blasts required to diagnose NPM1-mutated AML remains controversial. According to the European LeukemiaNet (ELN), determining the mutational status of NPM1 (together with FLT3) is mandatory for accurate relapse-risk assessment. NPM1 mutations are ideal targets for measurable residual disease (MRD) monitoring, since they are AML specific, frequent, very stable at relapse, and do not drive clonal hematopoiesis of undetermined significance. MRD monitoring by quantitative polymerase chain reaction of NPM1-mutant transcripts, possibly combined with ELN genetic-based risk stratification, can guide therapeutic decisions after remission. Furthermore, immunohistochemistry can be very useful in selected situations, such as diagnosis of NPM1-mutated myeloid sarcoma. Herein, we present 4 illustrative cases of NPM1-mutated AML that address important issues surrounding the biology, diagnosis, and therapy of this common form of leukemia.

AML, ALL, and CML classification and diagnosis based on bone marrow cell morphology combined with convolutional neural network: A STARD compliant diagnosis research.

Leukemia diagnosis based on bone marrow cell morphology primarily relies on the manual microscopy of bone marrow smears. However, this method is greatly affected by subjective factors and tends to lead to misdiagnosis. This study proposes using bone marrow cell microscopy images and employs convolutional neural network (CNN) combined with transfer learning to establish an objective, rapid, and accurate method for classification and diagnosis of LKA (AML, ALL, and CML). We collected cell microscopy images of 104 bone marrow smears (including 18 healthy subjects, 53 AML patients, 23 ALL patients, and 18 CML patients). The perfect reflection algorithm and a self-adaptive filter algorithm were first used for preprocessing of bone marrow cell images collected from experiments. Subsequently, 3 CNN frameworks (Inception-V3, ResNet50, and DenseNet121) were used to construct classification models for the raw dataset and preprocessed dataset. Transfer learning was used to improve the prediction accuracy of the model. Results showed that the DenseNet121 model based on the preprocessed dataset provided the best classification results, with a prediction accuracy of 74.8%. The prediction accuracy of the DenseNet121 model that was obtained by transfer learning optimization was 95.3%, which was increased by 20.5%. In this model, the prediction accuracies of the normal groups, AML, ALL, and CML were 90%, 99%, 97%, and 95%, respectively. The results showed that the leukemic cell morphology classification and diagnosis based on CNN combined with transfer learning is feasible. Compared with conventional manual microscopy, this method is more rapid, accurate, and objective.

Renin angiotensin system genes are biomarkers for personalized treatment of acute myeloid leukemia with Doxorubicin as well as etoposide.

Despite the availability of various treatment protocols, response to therapy in patients with Acute Myeloid Leukemia (AML) remains largely unpredictable. Transcriptomic profiling studies have thus far revealed the presence of molecular subtypes of AML that are not accounted for by standard clinical parameters or by routinely used biomarkers. Such molecular subtypes of AML are predicted to vary in response to chemotherapy or targeted therapy. The Renin-Angiotensin System (RAS) is an important group of proteins that play a critical role in regulating blood pressure, vascular resistance and fluid/electrolyte balance. RAS pathway genes are also known to be present locally in tissues such as the bone marrow, where they play an important role in leukemic hematopoiesis. In this study, we asked if the RAS genes could be utilized to predict drug responses in patients with AML. We show that the combined in silico analysis of up to five RAS genes can reliably predict sensitivity to Doxorubicin as well as Etoposide in AML. The same genes could also predict sensitivity to Doxorubicin when tested in vitro. Additionally, gene set enrichment analysis revealed enrichment of TNF-alpha and type-I IFN response genes among sensitive, and TGF-beta and fibronectin related genes in resistant cancer cells. However, this does not seem to reflect an epithelial to mesenchymal transition per se. We also identified that RAS genes can stratify patients with AML into subtypes with distinct prognosis. Together, our results demonstrate that genes present in RAS are biomarkers for drug sensitivity and the prognostication of AML.

Expression of putative leukemia stem cell targets in genetically-defined acute myeloid leukemia subtypes.

Although most acute myeloid leukemia (AML) patients achieve complete remissions, the majority still eventually relapse and die of their disease. Rare primitive leukemia cells, so-called leukemia stem cells (LSCs), represent one potential type of resistant cell subpopulation responsible for this dissociation between response and cure. Several LSC targets have been described, but there is limited evidence about their relative utility or that targeting any can prevent relapse. LSCs not only appear to be biologically heterogeneous, but the classic immunocompromised mouse transplantation model also has serious shortcomings as an LSC assay. Out data suggest that the most immature cell phenotype that can be identified within a patient's leukemia may be clinically relevant and represent the de facto LSC. Moreover, although phenotypically heterogeneous, these putative LSCs show consistent phenotypes within individual genetically defined groups. Using this LSC definition, we studied several previously described putative LSC targets, CD25, CD26, CD47, CD96, CD123, and CLL-1, and all were expressed across heterogeneous LSC phenotypes. In addition, with the exception of CD47, there was at most low expression of these targets on normal hematopoietic stem cells (HSCs). CD123 and CLL-1 demonstrated the greatest expression differences between putative LSCs and normal HSCs. Importantly, CD123 monoclonal antibodies were cytotoxic in vitro to putative LSCs from all AML subtypes, while showing limited to no toxicity against normal HSCs and hematopoietic progenitors. Since minimal residual disease appears to be a more homogeneous population of cells responsible for relapse, targeting CD123 in this setting may be most effective.

A deconvolution method and its application in analyzing the cellular fractions in acute myeloid leukemia samples.

BACKGROUND: The identification of cell type-specific genes (markers) is an essential step for the deconvolution of the cellular fractions, primarily, from the gene expression data of a bulk sample. However, the genes with significant changes identified by pair-wise comparisons cannot indeed represent the specificity of gene expression across multiple conditions. In addition, the knowledge about the identification of gene expression markers across multiple conditions is still paucity. RESULTS: Herein, we developed a hybrid tool, LinDeconSeq, which consists of 1) identifying marker genes using specificity scoring and mutual linearity strategies across any number of cell types, and 2) predicting cellular fractions of bulk samples using weighted robust linear regression with the marker genes identified in the first stage. On multiple publicly available datasets, the marker genes identified by LinDeconSeq demonstrated better accuracy and reproducibility compared to MGFM and RNentropy. Among deconvolution methods, LinDeconSeq showed low average deviations (≤0.0958) and high average Pearson correlations (≥0.8792) between the predicted and actual fractions on the benchmark datasets. Importantly, the cellular fractions predicted by LinDeconSeq appear to be relevant in the diagnosis of acute myeloid leukemia (AML). The distinct cellular fractions in granulocyte-monocyte progenitor (GMP), lymphoid-primed multipotent progenitor (LMPP) and monocytes (MONO) were found to be closely associated with AML compared to the healthy samples. Moreover, the heterogeneity of cellular fractions in AML patients divided these patients into two subgroups, differing in both prognosis and mutation patterns. GMP fraction was the most pronounced between these two subgroups, particularly, in SubgroupA, which was strongly associated with the better AML prognosis and the younger population. Totally, the identification of marker genes by LinDeconSeq represents the improved feature for deconvolution. The data processing strategy with regard to the cellular fractions used in this study also showed potential for the diagnosis and prognosis of diseases. CONCLUSIONS: Taken together, we developed a freely-available and open-source tool LinDeconSeq ( https://github.com/lihuamei/LinDeconSeq ), which includes marker identification and deconvolution procedures. LinDeconSeq is comparable to other current methods in terms of accuracy when applied to benchmark datasets and has broad application in clinical outcome and disease-specific molecular mechanisms.