Gemtuzumab ozogamicin plus midostaurin in combination with standard '7 + 3' induction therapy in newly diagnosed AML: Results from the SAL-MODULE phase I study.

We conducted a phase I trial in newly diagnosed acute myeloid leukaemia (AML) to investigate the combination of two novel targeted agents, gemtuzumab ozogamicin (GO) and midostaurin, with intensive chemotherapy in FLT3-mutated AML and CBF leukaemia. Three dose levels of midostaurin and one to three sequential doses of 3 mg/m2 GO in combination with '7 + 3' induction were evaluated. Based on safety findings in 12 patients, our results show that 3 mg/m2 GO on Days 1 + 4 and 100 mg midostaurin on Days 8-21 can be safely combined with IC in newly diagnosed AML.

Remission induction versus immediate allogeneic haematopoietic stem cell transplantation for patients with relapsed or poor responsive acute myeloid leukaemia (ASAP): a randomised, open-label, phase 3, non-inferiority trial.

BACKGROUND: Whether high-dose cytarabine-based salvage chemotherapy, administered to induce complete remission in patients with poor responsive or relapsed acute myeloid leukaemia scheduled for allogeneic haematopoietic stem-cell transplantation (HSCT) after intensive conditioning confers a survival advantage, is unclear. METHODS: To test salvage chemotherapy before allogeneic HSCT, patients aged between 18 and 75 years with non-favourable-risk acute myeloid leukaemia not in complete remission after first induction or untreated first relapse were randomly assigned 1:1 to remission induction with high-dose cytarabine (3 g/m2 intravenously, 1 g/m2 intravenously for patients >60 years or with a substantial comorbidity) twice daily on days 1-3 plus mitoxantrone (10 mg/m2 intravenously) on days 3-5 or immediate allogeneic HSCT for the disease control group. Block randomisation with variable block lengths was used and patients were stratified by age, acute myeloid leukaemia risk, and disease status. The study was open label. The primary endpoint was treatment success, defined as complete remission on day 56 after allogeneic HSCT, with the aim to show non-inferiority for disease control compared with remission induction with a non-inferiority-margin of 5% and one-sided type 1 error of 2·5%. The primary endpoint was analysed in both the intention-to-treat (ITT) population and in the per-protocol population. The trial is completed and was registered at ClinicalTrials.gov, NCT02461537. FINDINGS: 281 patients were enrolled between Sept 17, 2015, and Jan 12, 2022. Of 140 patients randomly assigned to disease control, 135 (96%) proceeded to allogeneic HSCT, 97 (69%) after watchful waiting only. Of 141 patients randomly assigned to remission induction, 134 (95%) received salvage chemotherapy and 128 (91%) patients subsequently proceeded to allogeneic HSCT. In the ITT population, treatment success was observed in 116 (83%) of 140 patients in the disease control group versus 112 (79%) of 141 patients with remission induction (test for non-inferiority, p=0·036). Among per-protocol treated patients, treatment success was observed in 116 (84%) of 138 patients with disease control versus 109 (81%) of 134 patients in the remission induction group (test for non-inferiority, p=0·047). The difference in treatment success between disease control and remission induction was estimated as 3·4% (95% CI -5·8 to 12·6) for the ITT population and 2·7% (-6·3 to 11·8) for the per-protocol population. Fewer patients with disease control compared with remission induction had non-haematological adverse events grade 3 or worse (30 [21%] of 140 patients vs 86 [61%] of 141 patients, χ2 test p<0·0001). Between randomisation and the start of conditioning, with disease control two patients died from progressive acute myeloid leukaemia and zero from treatment-related complications, and with remission induction two patients died from progressive acute myeloid leukaemia and two from treatment-related complications. Between randomisation and allogeneic HSCT, patients with disease control spent a median of 27 days less in hospital than those with remission induction, ie, the median time in hospital was 15 days (range 7-64) versus 42 days (27-121, U test p<0·0001), respectively. INTERPRETATION: Non-inferiority of disease control could not be shown at the 2·5% significance level. The rate of treatment success was also not statistically better for patients with remission induction. Watchful waiting and immediate transplantation could be an alternative for fit patients with poor response or relapsed acute myeloid leukaemia who have a stem cell donor available. More randomised controlled intention-to-transplant trials are needed to define the optimal treatment before transplantation for patients with active acute myeloid leukaemia. FUNDING: DKMS and the Gert and Susanna Mayer Stiftung Foundation.

The multimodality cell segmentation challenge: toward universal solutions.

Cell segmentation is a critical step for quantitative single-cell analysis in microscopy images. Existing cell segmentation methods are often tailored to specific modalities or require manual interventions to specify hyper-parameters in different experimental settings. Here, we present a multimodality cell segmentation benchmark, comprising more than 1,500 labeled images derived from more than 50 diverse biological experiments. The top participants developed a Transformer-based deep-learning algorithm that not only exceeds existing methods but can also be applied to diverse microscopy images across imaging platforms and tissue types without manual parameter adjustments. This benchmark and the improved algorithm offer promising avenues for more accurate and versatile cell analysis in microscopy imaging.

Mimicking clinical trials with synthetic acute myeloid leukemia patients using generative artificial intelligence.

Clinical research relies on high-quality patient data, however, obtaining big data sets is costly and access to existing data is often hindered by privacy and regulatory concerns. Synthetic data generation holds the promise of effectively bypassing these boundaries allowing for simplified data accessibility and the prospect of synthetic control cohorts. We employed two different methodologies of generative artificial intelligence - CTAB-GAN+ and normalizing flows (NFlow) - to synthesize patient data derived from 1606 patients with acute myeloid leukemia, a heterogeneous hematological malignancy, that were treated within four multicenter clinical trials. Both generative models accurately captured distributions of demographic, laboratory, molecular and cytogenetic variables, as well as patient outcomes yielding high performance scores regarding fidelity and usability of both synthetic cohorts (n = 1606 each). Survival analysis demonstrated close resemblance of survival curves between original and synthetic cohorts. Inter-variable relationships were preserved in univariable outcome analysis enabling explorative analysis in our synthetic data. Additionally, training sample privacy is safeguarded mitigating possible patient re-identification, which we quantified using Hamming distances. We provide not only a proof-of-concept for synthetic data generation in multimodal clinical data for rare diseases, but also full public access to synthetic data sets to foster further research.

Prognostic impact of CEBPA mutational subgroups in adult AML.

Despite recent refinements in the diagnostic and prognostic assessment of CEBPA mutations in AML, several questions remain open, i.e. implications of different types of basic region leucin zipper (bZIP) mutations, the role of co-mutations and the allelic state. Using pooled primary data analysis on 1010 CEBPA-mutant adult AML patients, a comparison was performed taking into account the type of mutation (bZIP: either typical in-frame insertion/deletion (InDel) mutations (bZIPInDel), frameshift InDel or nonsense mutations inducing translational stop (bZIPSTOP) or single base-pair missense alterations (bZIPms), and transcription activation domain (TAD) mutations) and the allelic state (single (smCEBPA) vs. double mutant (dmCEBPA)). Only bZIPInDel patients had significantly higher rates of complete remission and longer relapse free and overall survival (OS) compared with all other CEBPA-mutant subgroups. Moreover, co-mutations in bZIPInDel patients (e.g. GATA2, FLT3, WT1 as well as ELN2022 adverse risk aberrations) had no independent impact on OS, whereas in non-bZIPInDel patients, grouping according to ELN2022 recommendations added significant prognostic information. In conclusion, these results demonstrate bZIPInDel mutations to be the major independent determinant of outcome in CEBPA-mutant AML, thereby refining current classifications according to WHO (including all dmCEBPA and smCEBPA bZIP) as well as ELN2022 and ICC recommendations (including CEBPA bZIPms).

Mutated IKZF1 is an independent marker of adverse risk in acute myeloid leukemia.

Genetic lesions of IKZF1 are frequent events and well-established markers of adverse risk in acute lymphoblastic leukemia. However, their function in the pathophysiology and impact on patient outcome in acute myeloid leukemia (AML) remains elusive. In a multicenter cohort of 1606 newly diagnosed and intensively treated adult AML patients, we found IKZF1 alterations in 45 cases with a mutational hotspot at N159S. AML with mutated IKZF1 was associated with alterations in RUNX1, GATA2, KRAS, KIT, SF3B1, and ETV6, while alterations of NPM1, TET2, FLT3-ITD, and normal karyotypes were less frequent. The clinical phenotype of IKZF1-mutated AML was dominated by anemia and thrombocytopenia. In both univariable and multivariable analyses adjusting for age, de novo and secondary AML, and ELN2022 risk categories, we found mutated IKZF1 to be an independent marker of adverse risk regarding complete remission rate, event-free, relapse-free, and overall survival. The deleterious effects of mutated IKZF1 also prevailed in patients who underwent allogeneic hematopoietic stem cell transplantation (n = 519) in both univariable and multivariable models. These dismal outcomes are only partially explained by the hotspot mutation N159S. Our findings suggest a role for IKZF1 mutation status in AML risk modeling.

UBTF tandem duplications are rare but recurrent alterations in adult AML and associated with younger age, myelodysplasia, and inferior outcome.

Tandem-duplication mutations of the UBTF gene (UBTF-TDs) coding for the upstream binding transcription factor have recently been described in pediatric patients with acute myeloid leukemia (AML) and were found to be associated with particular genetics (trisomy 8 (+8), FLT3-internal tandem duplications (FLT3-ITD), WT1-mutations) and inferior outcome. Due to limited knowledge on UBTF-TDs in adult AML, we screened 4247 newly diagnosed adult AML and higher-risk myelodysplastic syndrome (MDS) patients using high-resolution fragment analysis. UBTF-TDs were overall rare (n = 52/4247; 1.2%), but significantly enriched in younger patients (median age 41 years) and associated with MDS-related morphology as well as significantly lower hemoglobin and platelet levels. Patients with UBTF-TDs had significantly higher rates of +8 (34% vs. 9%), WT1 (52% vs. 7%) and FLT3-ITD (50% vs. 20.8%) co-mutations, whereas UBTF-TDs were mutually exclusive with several class-defining lesions such as mutant NPM1, in-frame CEBPAbZIP mutations as well as t(8;21). Based on the high-variant allele frequency found and the fact that all relapsed patients analyzed (n = 5) retained the UBTF-TD mutation, UBTF-TDs represent early clonal events and are stable over the disease course. In univariate analysis, UBTF-TDs did not represent a significant factor for overall or relapse-free survival in the entire cohort. However, in patients under 50 years of age, who represent the majority of UBTF-mutant patients, UBTF-TDs were an independent prognostic factor for inferior event-free (EFS), relapse-free (RFS) and overall survival (OS), which was confirmed by multivariable analyses including established risk factors such as age and ELN2022 genetic risk groups (EFS [HR: 2.20; 95% CI 1.52-3.17, p < 0.001], RFS [HR: 1.59; 95% CI 1.02-2.46, p = 0.039] and OS [HR: 1.64; 95% CI 1.08-2.49, p = 0.020]). In summary, UBTF-TDs appear to represent a novel class-defining lesion not only in pediatric AML but also younger adults and are associated with myelodysplasia and inferior outcome in these patients.

Unsupervised meta-clustering identifies risk clusters in acute myeloid leukemia based on clinical and genetic profiles.

BACKGROUND: Increasingly large and complex biomedical data sets challenge conventional hypothesis-driven analytical approaches, however, data-driven unsupervised learning can detect inherent patterns in such data sets. METHODS: While unsupervised analysis in the medical literature commonly only utilizes a single clustering algorithm for a given data set, we developed a large-scale model with 605 different combinations of target dimensionalities as well as transformation and clustering algorithms and subsequent meta-clustering of individual results. With this model, we investigated a large cohort of 1383 patients from 59 centers in Germany with newly diagnosed acute myeloid leukemia for whom 212 clinical, laboratory, cytogenetic and molecular genetic parameters were available. RESULTS: Unsupervised learning identifies four distinct patient clusters, and statistical analysis shows significant differences in rate of complete remissions, event-free, relapse-free and overall survival between the four clusters. In comparison to the standard-of-care hypothesis-driven European Leukemia Net (ELN2017) risk stratification model, we find all three ELN2017 risk categories being represented in all four clusters in varying proportions indicating unappreciated complexity of AML biology in current established risk stratification models. Further, by using assigned clusters as labels we subsequently train a supervised model to validate cluster assignments on a large external multicenter cohort of 664 intensively treated AML patients. CONCLUSIONS: Dynamic data-driven models are likely more suitable for risk stratification in the context of increasingly complex medical data than rigid hypothesis-driven models to allow for a more personalized treatment allocation and gain novel insights into disease biology.

There are various ways in which clinicians can predict the risk of disease progression in patients with leukemia, helping them to treat the patients accordingly. However, these approaches are usually designed by human experts and might not fully capture the complexity of a patient's disease. Here, with a large cohort of patients with acute myeloid leukemia, we design an unsupervised machine learning model ­ a type of computer model that learns from patterns in data without human input­to separate these patients into subgroups according to risk. We identify four distinct groups which differ with regards to patient genetics, laboratory values, and clinical characteristics. These groups have differences in response to treatment and patient survival, and we validate our findings in another dataset. Our approach might help clinicians to better predict outcomes in patients with leukemia and make decisions on treatment.

Deciphering the genetic basis of immune thrombocytopenia: current evidence for genetic predisposition in adult ITP.

Immune thrombocytopenia (ITP) is the consequence of a complex, still incompletely understood immunological dysregulation. Proposed mechanisms include autoantibody-induced platelet destruction, impaired platelet production as well as abnormalities in T-cell immunity, such as T helper cells (Th1) polarization, a high proportion of Th17 cells, and a reduced number of regulatory T cells. Although the etiology of ITP is incompletely understood and considered multifactorial in most cases, genetic variants are thought to play a key role in susceptibility to ITP, especially in persistent or chronic ITP. Efforts are currently underway to uncover possible predisposing genetic factors for the development of ITP. Single-nucleotide polymorphisms and copy number variations have been identified in several immune-related genes, such as cytokine genes, Fcγ receptor genes or T-cell costimulation genes, and have been associated with patients' susceptibility to ITP. However, because of the clinical heterogeneity and low incidence of ITP it remains challenging to perform genetic analyses with sufficiently large sample size within informative patient populations, highlighting the need for collection of well-annotated biomaterials in clinical trials or registry projects. Another significant challenge is to go beyond performing association studies alone and to establish genotype-phenotype associations, thus proving causality between a genetic alteration and ITP pathogenesis. This review summarizes our current knowledge on genetic alterations identified as potential predisposing factors for the development of ITP in adults, thereby addressing signaling pathways considered critical for ITP pathogenesis.

[Novel Strategies to Treat Acute Myeloid Leukemia]. / Neue Therapieoptionen bei der Akuten Myeloischen Leukämie.

BACKGROUND: While the "7+3" regimen of cytarabine + anthracycline has been the backbone of acute myeloid leukemia (AML) treatment for four decades, several novel drugs have been approved in the past five years. Despite these promising novel therapeutic options, treatment of AML remains challenging, given the biologically heterogenous character of the disease. AIM: This review provides an update on novel treatment strategies for AML. MATERIAL AND METHODS: This article is based on the current European LeukemiaNet (ELN) recommendations and the DGHO «Onkopedia¼ guideline on AML treatment. RESULTS AND CONCLUSION: The treatment algorithm is based on patient-related and disease-specific factors, such as patient age and fitness as well as AML molecular profile. Younger patients considered fit for intensive chemotherapy receive 1-2 courses of induction therapy ("7+3" regimen, eg. cytarabine/daunorubicin, or CPX-351 for patients with myelodysplasia-related AML or therapy-related AML). For CD33+ patients or those with evidence of an FLT3 mutation "7+3" in combination with Gemtuzumab-Ozogamicin (GO) or Midostaurin is recommended, respectively. For consolidation, patients receive either high-dose chemotherapy (± GO/± Midostaurin) or undergo allogeneic hematopoietic cell transplantation (HCT), based on ELN risk stratification. In some cases, maintenance therapy with oral azacytidine or FLT3 inhibitor is indicated. Patients experiencing relapse should receive chemotherapy-based re-induction therapy or, in case of an FLT3 mutation, Gilteritinib and subsequently undergo allogeneic HCT. For older patients or those considered unfit for intensive therapy, azacytidine in combination with Venetoclax is a promising novel treatment strategy. Although not yet approved by the European Medical Agency (EMA), for patients with IDH1IDH1 or IDH2 mutations treatment with the IDH1 and IDH2 inhibitors Ivosidenib and Enasidenib should be considered.

An overview and a roadmap for artificial intelligence in hematology and oncology.

BACKGROUND: Artificial intelligence (AI) is influencing our society on many levels and has broad implications for the future practice of hematology and oncology. However, for many medical professionals and researchers, it often remains unclear what AI can and cannot do, and what are promising areas for a sensible application of AI in hematology and oncology. Finally, the limits and perils of using AI in oncology are not obvious to many healthcare professionals. METHODS: In this article, we provide an expert-based consensus statement by the joint Working Group on "Artificial Intelligence in Hematology and Oncology" by the German Society of Hematology and Oncology (DGHO), the German Association for Medical Informatics, Biometry and Epidemiology (GMDS), and the Special Interest Group Digital Health of the German Informatics Society (GI). We provide a conceptual framework for AI in hematology and oncology. RESULTS: First, we propose a technological definition, which we deliberately set in a narrow frame to mainly include the technical developments of the last ten years. Second, we present a taxonomy of clinically relevant AI systems, structured according to the type of clinical data they are used to analyze. Third, we show an overview of potential applications, including clinical, research, and educational environments with a focus on hematology and oncology. CONCLUSION: Thus, this article provides a point of reference for hematologists and oncologists, and at the same time sets forth a framework for the further development and clinical deployment of AI in hematology and oncology in the future.

Allogeneic Hematopoietic Cell Transplantation vs Standard Consolidation Chemotherapy in Patients With Intermediate-Risk Acute Myeloid Leukemia: A Randomized Clinical Trial.

Importance: The ideal postremission strategy in intermediate-risk acute myeloid leukemia (AML) in first complete remission (CR) has been a matter of debate. Objective: To explore the optimal therapy for patients with intermediate-risk AML after first complete remission. Design, Settings, and Participants: This investigator-initiated, open-label, 2-armed, phase 3 randomized clinical trial assessed patients at 16 hospitals in Germany from February 2, 2011, until July 1, 2018. Key eligibility criteria included cytogenetically defined intermediate-risk AML according to Medical Research Council classification, first CR or CR with incomplete blood cell count recovery after conventional induction therapy, age of 18 to 60 years, and availability of a human leukocyte antigen (HLA)-matched sibling or unrelated donor. A detailed statistical analysis plan was written and finalized on July 7, 2020. Data were exported for analysis on April 13, 2021. Interventions: Patients were randomized 1:1 to receive allogeneic hematopoietic cell transplantation (HCT) or high-dose cytarabine for consolidation and salvage HCT only in case of relapse. Strata for randomization included age (18-40 vs 41-60 years), NPM1 and CEBPA variation status, and donor type (unrelated vs related). Main Outcomes and Measures: End points included overall-survival as the primary outcome and disease-free survival, cumulative incidence of relapse, treatment-related mortality, and quality of life measured according to the Medical Outcomes Study 36-Item Short-Form Health Survey as secondary outcomes. Results: A total of 143 patients (mean [SD] age, 48.2 [9.8] years; 81 [57%] male) with AML who fulfilled the eligibility criteria were randomized. In the intention-to-treat analysis, the probability of survival at 2 years was 74% (95% CI, 62%-83%) after primary allogeneic HCT and 84% (95% CI, 73%-92%) after consolidation chemotherapy (P = .22). Disease-free survival after HCT at 2 years was 69% (95% CI, 57%-80%) compared with 40% (95% CI, 28%-53%) after consolidation chemotherapy (P = .001). Allogeneic HCT during the first CR was associated with a cumulative incidence of relapse at 2 years of 20% (95% CI, 13%-31%) compared with 58% (95% CI, 47%-71%; P < .001). Nonrelapse mortality at 2 years after primary allogeneic HCT was 9% (95% CI, 5%-19%) and 2% (95% CI, 0%-11%) after consolidation chemotherapy (P = .005). Similar outcomes were observed when analyses were confined to the 96 patients at intermediate risk according to the European Leukemia Network classification. Most importantly, all 41 patients relapsing after consolidation chemotherapy (36 hematologic, 4 molecular, and 1 extramedullary) proceeded to allogeneic HCT. No significant differences in health-related quality of life measures were observed between groups. Conclusions and Relevance: Primary allogeneic HCT during first CR was not associated with superior overall survival compared with consolidation chemotherapy in patients 60 years or younger with intermediate-risk AML during the first CR and an available donor. Trial Registration: ClinicalTrials.gov Identifier: NCT01246752.

Alterations of cohesin complex genes in acute myeloid leukemia: differential co-mutations, clinical presentation and impact on outcome.

Functional perturbations of the cohesin complex with subsequent changes in chromatin structure and replication are reported in a multitude of cancers including acute myeloid leukemia (AML). Mutations of its STAG2 subunit may predict unfavorable risk as recognized by the 2022 European Leukemia Net recommendations, but the underlying evidence is limited by small sample sizes and conflicting observations regarding clinical outcomes, as well as scarce information on other cohesion complex subunits. We retrospectively analyzed data from a multi-center cohort of 1615 intensively treated AML patients and identified distinct co-mutational patters for mutations of STAG2, which were associated with normal karyotypes (NK) and concomitant mutations in IDH2, RUNX1, BCOR, ASXL1, and SRSF2. Mutated RAD21 was associated with NK, mutated EZH2, KRAS, CBL, and NPM1. Patients harboring mutated STAG2 were older and presented with decreased white blood cell, bone marrow and peripheral blood blast counts. Overall, neither mutated STAG2, RAD21, SMC1A nor SMC3 displayed any significant, independent effect on clinical outcomes defined as complete remission, event-free, relapse-free or overall survival. However, we found almost complete mutual exclusivity of genetic alterations of individual cohesin subunits. This mutual exclusivity may be the basis for therapeutic strategies via synthetic lethality in cohesin mutated AML.

Large case-control study indicates no association of KIR genotype and risk of developing acute myeloid leukemia.

Immunogenetic association studies may give rise to new hypotheses on the immune surveillance of cancer. We hypothesized that certain combinations of killer immunoglobulin-like receptor (KIR) and HLA genotypes may enhance natural killer (NK) cell immunity against nascent acute myeloid leukemia (AML) and, thereby, lead to a skewed genotype distribution among patients. For this purpose, we analyzed KIR and HLA genotypes of 1767 German patients with AML and compared the results with that of the data of 51 890 German volunteers who had registered with German bone marrow donor file (DKMS). Patient samples were retrieved from the Collaborative Biobank and the biorepository of the Study Alliance Leukemia. All samples were genotyped with high-resolution amplicon-based next-generation sequencing. Because of the large number of controls, this study was very sensitive to detect the impact of KIR genotype. Knowledge on KIRs and their cognate HLA ligands allowed for testing of several hypotheses of NK cell-mediated endogenous leukemia surveillance. We did not find significant differences between the 2 cohorts in regard to the presence or absence of single KIR genes. When grouped based on telomeric or centromeric gene content, the major haplotypes A/A, A/B, and B/B were equally distributed among patients and control subjects. Using information on KIRs and their HLA ligands, we further tested receptor-ligand models and summation models without revealing markedly significant differences between patients and controls, albeit we observed a trend pointing at a minor protective effect of a low number of inhibitory KIR/KIR-ligand pairs. The results suggest that the KIR/KIR-ligand genotype has no effect on the susceptibility for the development of de novo AML.

Prediction of complete remission and survival in acute myeloid leukemia using supervised machine learning.

Achievement of complete remission signifies a crucial milestone in the therapy of acute myeloid leukemia (AML) while refractory disease is associated with dismal outcomes. Hence, accurately identifying patients at risk is essential to tailor treatment concepts individually to disease biology. We used nine machine learning (ML) models to predict complete remission and 2-year overall survival in a large multicenter cohort of 1,383 AML patients who received intensive induction therapy. Clinical, laboratory, cytogenetic and molecular genetic data were incorporated and our results were validated on an external multicenter cohort. Our ML models autonomously selected predictive features including established markers of favorable or adverse risk as well as identifying markers of so-far controversial relevance. De novo AML, extramedullary AML, double-mutated CEBPA, mutations of CEBPA-bZIP, NPM1, FLT3-ITD, ASXL1, RUNX1, SF3B1, IKZF1, TP53, and U2AF1, t(8;21), inv(16)/t(16;16), del(5)/del(5q), del(17)/del(17p), normal or complex karyotypes, age and hemoglobin concentration at initial diagnosis were statistically significant markers predictive of complete remission, while t(8;21), del(5)/del(5q), inv(16)/t(16;16), del(17)/del(17p), double-mutated CEBPA, CEBPA-bZIP, NPM1, FLT3-ITD, DNMT3A, SF3B1, U2AF1, and TP53 mutations, age, white blood cell count, peripheral blast count, serum lactate dehydrogenase level and hemoglobin concentration at initial diagnosis as well as extramedullary manifestations were predictive for 2-year overall survival. For prediction of complete remission and 2-year overall survival areas under the receiver operating characteristic curves ranged between 0.77-0.86 and between 0.63-0.74, respectively in our test set, and between 0.71-0.80 and 0.65-0.75 in the external validation cohort. We demonstrated the feasibility of ML for risk stratification in AML as a model disease for hematologic neoplasms, using a scalable and reusable ML framework. Our study illustrates the clinical applicability of ML as a decision support system in hematology.

Perturbations of mesenchymal stromal cells after allogeneic hematopoietic cell transplantation predispose for bone marrow graft-versus-host-disease.

Functional impairment of the bone marrow (BM) niche has been suggested as a major reason for prolonged cytopenia and secondary graft failure after allogeneic hematopoietic cell transplantation (alloHCT). Because mesenchymal stromal cells (MSCs) serve as multipotent progenitors for several niche components in the BM, they might play a key role in this process. We used collagenase digested trephine biopsies to directly quantify MSCs in 73 patients before (n = 18) and/or after alloHCT (n = 65). For the first time, we demonstrate that acute graft-versus-host disease (aGvHD, n = 39) is associated with a significant decrease in MSC numbers. MSC reduction can be observed even before the clinical onset of aGvHD (n = 10). Assessing MSCs instantly after biopsy collection revealed phenotypic and functional differences depending on the occurrence of aGvHD. These differences vanished during ex vivo expansion. The MSC endotypes observed revealed an enhanced population of donor-derived classical dendritic cells type 1 and alloreactive T cells as the causing agent for compartmental inflammation and MSC damage before clinical onset of aGvHD was ascertained. In conclusion, MSCs endotypes may constitute a predisposing conductor of alloreactivity after alloHCT preceding the clinical diagnosis of aGvHD.

Impact of IDH1 and IDH2 mutational subgroups in AML patients after allogeneic stem cell transplantation.

BACKGROUND: The role of allogeneic hematopoietic cell transplantation (alloHCT) in acute myeloid leukemia (AML) with mutated IDH1/2 has not been defined. Therefore, we analyzed a large cohort of 3234 AML patients in first complete remission (CR1) undergoing alloHCT or conventional chemo-consolidation and investigated outcome in respect to IDH1/2 mutational subgroups (IDH1 R132C, R132H and IDH2 R140Q, R172K). METHODS: Genomic DNA was extracted from bone marrow or peripheral blood samples at diagnosis and analyzed for IDH mutations with denaturing high-performance liquid chromatography, Sanger sequencing and targeted myeloid panel next-generation sequencing, respectively. Statistical as-treated analyses were performed using R and standard statistical methods (Kruskal-Wallis test for continuous variables, Chi-square test for categorical variables, Cox regression for univariate and multivariable models), incorporating alloHCT as a time-dependent covariate. RESULTS: Among 3234 patients achieving CR1, 7.8% harbored IDH1 mutations (36% R132C and 47% R132H) and 10.9% carried IDH2 mutations (77% R140Q and 19% R172K). 852 patients underwent alloHCT in CR1. Within the alloHCT group, 6.2% had an IDH1 mutation (43.4% R132C and 41.4% R132H) and 10% were characterized by an IDH2 mutation (71.8% R140Q and 24.7% R172K). Variants IDH1 R132C and IDH2 R172K showed a significant benefit from alloHCT for OS (p = .017 and p = .049) and RFS (HR = 0.42, p = .048 and p = .009) compared with chemotherapy only. AlloHCT in IDH2 R140Q mutated AML resulted in longer RFS (HR = 0.4, p = .002). CONCLUSION: In this large as-treated analysis, we showed that alloHCT is able to overcome the negative prognostic impact of certain IDH mutational subclasses in first-line consolidation treatment and could pending prognostic validation, provide prognostic value for AML risk stratification and therapeutic decision making.

Semi-supervised learning in cancer diagnostics.

In cancer diagnostics, a considerable amount of data is acquired during routine work-up. Recently, machine learning has been used to build classifiers that are tasked with cancer detection and aid in clinical decision-making. Most of these classifiers are based on supervised learning (SL) that needs time- and cost-intensive manual labeling of samples by medical experts for model training. Semi-supervised learning (SSL), however, works with only a fraction of labeled data by including unlabeled samples for information abstraction and thus can utilize the vast discrepancy between available labeled data and overall available data in cancer diagnostics. In this review, we provide a comprehensive overview of essential functionalities and assumptions of SSL and survey key studies with regard to cancer care differentiating between image-based and non-image-based applications. We highlight current state-of-the-art models in histopathology, radiology and radiotherapy, as well as genomics. Further, we discuss potential pitfalls in SSL study design such as discrepancies in data distributions and comparison to baseline SL models, and point out future directions for SSL in oncology. We believe well-designed SSL models to strongly contribute to computer-guided diagnostics in malignant disease by overcoming current hinderances in the form of sparse labeled and abundant unlabeled data.