Acute myeloid leukemia (AML) is a heterogenous blood cancer with a dismal prognosis. It emanates from leukemic stem cells (LSCs) arising from the genetic transformation of hematopoietic stem cells (HSCs). LSCs hold prognostic value, but their molecular and immunophenotypic heterogeneity poses challenges: there is no single marker for identifying all LSCs across AML samples. We hypothesized that imaging flow cytometry (IFC) paired with artificial intelligence-driven image analysis could visually distinguish LSCs from HSCs based solely on morphology. Initially, a seven-color IFC panel was employed to immunophenotypically identify LSCs and HSCs in bone marrow samples from five AML patients and ten healthy donors, respectively. Next, we developed convolutional neural network (CNN) models for HSC-LSC discrimination using brightfield (BF), side scatter (SSC), and DNA images. Classification using only BF images achieved 86.96% accuracy, indicating significant morphological differences. Accuracy increased to 93.42% when combining BF with DNA images, highlighting differences in nuclear morphology, although DNA images alone were inadequate for accurate HSC-LSC discrimination. Model development using SSC images revealed minor granularity differences. Performance metrics varied substantially between AML patients, indicating considerable morphologic variations among LSCs. Overall, we demonstrate proof-of-concept results for accurate CNN-based HSC-LSC differentiation, instigating the development of a novel technique within AML monitoring.

Novel therapeutic tools are warranted to improve outcomes for children with acute myeloid leukemia (AML). Differences in the proteome of leukemic blasts and stem cells (AML-SCs) in AML compared with normal hematopoietic stem cells (HSCs) may facilitate the identification of potential targets for future treatment strategies. In this explorative study, we used mass spectrometry to compare the proteome of AML-SCs and CLEC12A+ blasts from five pediatric AML patients with HSCs and hematopoietic progenitor cells from hematologically healthy, age-matched controls. A total of 456 shared proteins were identified in both leukemic and control samples. Varying protein expression profiles were observed in AML-SCs and leukemic blasts, none having any overall resemblance to healthy counterpart cell populations. Thirty-four proteins were differentially expressed between AML-SCs and HSCs, including the upregulation of HSPE1, SRSF1, and NUP210, and the enrichment of proteins suggestive of protein synthesis perturbations through the downregulation of EIF2 signaling was found. Among others, NUP210 and calreticulin were upregulated in CLEC12A+ blasts compared with HSCs. In conclusion, the observed differences in protein expression between pediatric patients with AML and pediatric controls, in particular when comparing stem cell subsets, encourages the extended exploration of leukemia and AML-SC-specific biomarkers of potential relevance in the development of future therapeutic options in pediatric AML.

Megakaryocytes (MKs) are rare, large, polyploid bone marrow (BM) cells responsible for the production of platelets. The identification and characterization of MKs is widely recognized as challenging. Manual microscopy is especially difficult due to the rarity and complex morphology of MKs, while flow cytometry faces additional challenges from MKs’ large size, fragility, and platelet adhesion, causing false positives. We present a novel approach to accurately enrich MKs from human BM aspirates with a specific focus on the detection and quantification of microMKs. By integrating CD41⁺ cell enrichment, immunophenotyping, and morphometric analysis, we identified cells of the megakaryocytic lineage. To increase accuracy, a convolutional neural network was trained to identify CD41⁻ cells falsely displaying an MK-like immunophenotype due to adhesive CD41⁺ platelets. This allowed for exclusion of 94.9% of false positive events, considerably enhancing specificity. CD41 positive enrichment prior to imaging flow cytometry acquisition increased the MK frequency nearly 200-fold, yielding a population of both mature and immature MKs, thus supporting analysis of MK progenitors. Overall, this advanced approach enables enrichment of MKs from human BM, considerably increasing the accuracy and statistical power of the MK analysis. This may provide an important addition in the context of MK-related diagnostics and research.