Maturation State-Specific Alternative Splicing in FLT3-ITD and NPM1 Mutated AML

High Mitochondrial Protein Expression as a Potential Predictor of Relapse Risk in Acute Myeloid Leukemia Patients with the Monocytic FAB Subtypes M4 and M5

AML is a highly aggressive and heterogeneous form of hematological cancer. Proteomics-based stratification of patients into more refined subgroups may contribute to a more precise characterization of the patient-derived AML cells. Here, we reanalyzed liquid chromatography-tandem mass spectrometry (LC-MS/MS) generated proteomic and phosphoproteomic data from 26 FAB-M4/M5 patients. The patients achieved complete hematological remission after induction therapy. Twelve of them later developed chemoresistant relapse (RELAPSE), and 14 patients were relapse-free (REL_FREE) long-term survivors. We considered not only the RELAPSE and REL_FREE characteristics but also integrated the French-American-British (FAB) classification, along with considering the presence of nucleophosmin 1 (NPM1) mutation and cytogenetically normal AML. We found a significant number of differentially enriched proteins (911) and phosphoproteins (257) between the various FAB subtypes in RELAPSE patients. Patients with the myeloblastic M1/M2 subtype showed higher levels of RNA processing-related routes and lower levels of signaling related to terms like translation and degranulation when compared with the M4/M5 subtype. Moreover, we found that a high abundance of proteins associated with mitochondrial translation and oxidative phosphorylation, particularly observed in the RELAPSE M4/M5 NPM1 mutated subgroup, distinguishes relapsing from non-relapsing AML patient cells with the FAB subtype M4/M5. Thus, the discovery of subtype-specific biomarkers through proteomic profiling may complement the existing classification system for AML and potentially aid in selecting personalized treatment strategies for individual patients.

A review on multimodal machine learning in medical diagnostics

Nowadays, the increasing number of medical diagnostic data and clinical data provide more complementary references for doctors to make diagnosis to patients. For example, with medical data, such as electrocardiography (ECG), machine learning algorithms can be used to identify and diagnose heart disease to reduce the workload of doctors. However, ECG data is always exposed to various kinds of noise and interference in reality, and medical diagnostics only based on one-dimensional ECG data is not trustable enough. By extracting new features from other types of medical data, we can implement enhanced recognition methods, called multimodal learning. Multimodal learning helps models to process data from a range of different sources, eliminate the requirement for training each single learning modality, and improve the robustness of models with the diversity of data. Growing number of articles in recent years have been devoted to investigating how to extract data from different sources and build accurate multimodal machine learning models, or deep learning models for medical diagnostics. This paper reviews and summarizes several recent papers that dealing with multimodal machine learning in disease detection, and identify topics for future research.

A Bayesian model for unsupervised detection of RNA splicing based subtypes in cancers

Identification of cancer sub-types is a pivotal step for developing personalized treatment. Specifically, sub-typing based on changes in RNA splicing has been motivated by several recent studies. We thus develop CHESSBOARD, an unsupervised algorithm tailored for RNA splicing data that captures "tiles" in the data, defined by a subset of unique splicing changes in a subset of patients. CHESSBOARD allows for a flexible number of tiles, accounts for uncertainty of splicing quantification, and is able to model missing values as additional signals. We first apply CHESSBOARD to synthetic data to assess its domain specific modeling advantages, followed by analysis of several leukemia datasets. We show detected tiles are reproducible in independent studies, investigate their possible regulatory drivers and probe their relation to known AML mutations. Finally, we demonstrate the potential clinical utility of CHESSBOARD by supplementing mutation based diagnostic assays with discovered splicing profiles to improve drug response correlation.

Chasing leukemia differentiation through induction therapy, relapse and transplantation

Despite rapid advances in our understanding of acute myeloid leukemia (AML), the disease remains challenging to treat with 5-year survival for adult patients 20 years or older estimated to be 26% (Cancer 2021). The use of new targeted therapies including BCL2, IDH1/IDH2, and FLT3 inhibitors has revolutionized treatment approaches but also changed the disease trajectory with unique modes of resistance. Recent studies have shown that stem cell maturation state drives expression level and/or dependence on various pathways, critical to determining drug response. Instead of anticipating these changes, we remain behind the curve chasing the next expanded clone. This review will focus on current approaches to treatment in AML, including defining the significance of blast differentiation state on chemotherapeutic response, signaling pathway dependence, metabolism, immune response, and phenotypic changes. We conclude that multimodal treatment approaches are necessary to target both the immature and mature clones, thereby, sustaining drug response.