Application of Single-Cell Approaches to Study Myeloproliferative Neoplasm Biology

Abstract

Philadelphia-negative myeloproliferative neoplasms (MPNs) are an excellent tractable disease model of a number of aspects of human cancer biology, including genetic evolution, tissue-associated fibrosis, and cancer stem cells. In this review, we discuss recent insights into MPN biology gained from the application of a number of new single-cell technologies to study human disease, with a specific focus on single-cell genomics, single-cell transcriptomics, and digital pathology.

Continuous Indexing of Fibrosis (CIF): Improving the Assessment and Classification of MPN Patients

“…Statistical descriptions of bone marrow morphological features using enhanced image analysis techniques have only recently been described, and application to fibrosis complements our recent work describing megakaryocyte morphology and topology in MPNs 40,45,46 . Of note, while the specific ML strategies employed for detecting and quantitating fibrosis in the form of a continuous score (CIF score) are distinct from those previously employed in our megakaryocyte analysis, they draw upon shared technical and infrastructural processes and deliver outputs that are readily integrated into shared analytical workflows.…”

From cell to cell: Identification of actionable targets in bone marrow fibrosis using single-cell technologies

“…Improved techniques, such as TARGET-Seq and genotyping of transcriptomes, now also allow reliable genotyping of the captured single cells to definitely distinguish malignant MPN progeny from (atypical) bystander hematopoietic cells and to unravel heterogeneity within the population of MPN cells itself [10,11]. A comprehensive overview of the accomplishments and potential future applications of single-cell approaches to the hematopoietic compartment in MPNs has recently been published by Royston et al [12] and will thus not be discussed in more detail here.…”

Perspective: Pivotal translational hematology and therapeutic insights in chronic myeloid hematopoietic stem cell malignancies

“…Diagnosis depends upon careful integration of clinical, genetic, and histological features and is enshrined in the revised 2016 World Health Organization classification scheme of myeloid malignancies 48 . Despite significant advances over recent years in genomic technologies relevant to diagnosis and disease monitoring in MPN, the key elements of morphological assessment remain largely unchanged 49,50 . Indeed, key morphologic features relating to marrow cellularity, megakaryocyte pleomorphism/atypia, and fibrosis are firmly embedded in current MPN classification schemes but remain subjective and largely qualitative.…”

“…In response, several investigators demonstrate the utility of an automated machine‐learning image analysis pipeline that uses image analysis/machine learning techniques to extract and interrogate important cytomorphological and topographic features of megakaryocytes from digitized images of BM biopsies (Figure 2). 50,51 This allowed them to differentiate reactive samples from common MPN subtypes and assisted in disease classification. Using the machine‐learned features from extracted megakaryocytes we identified discrete cellular subtypes beyond the sensitivity of detection by specialist pathologists.…”

“…Inconsistencies in the interpretation of key morphologic features may lead to inaccurate diagnosis and disease classification, with multiple studies suggesting significant intra-and inter-observer variability among pathologists. Although this appears to be partly attributable to experience and training, the subjective and qualitative nature of routine BM biopsy reporting remains a fundamental limiting factor in any classification scheme incorporating morphology-based assessment of marrow tissue.In response, several investigators demonstrate the utility of an automated machine-learning image analysis pipeline that uses image analysis/machine learning techniques to extract and interrogate important cytomorphological and topographic features of megakaryocytes from digitized images of BM biopsies (Figure2) 50,51. This allowed them to differentiate reactive samples from common MPN subtypes and assisted in disease classification.…”