Gurpreet Kaur Sagoo: The Power of AI Based Morphological Fingerprinting in Predicting Myelofibrosis Progression
Gurpreet Kaur Sagoo, Professor at Armed Forces Medical College, shared a post on LinkedIn about a recent article she and her colleagues co-authored, published in Indian Journal of Hematology and Blood Transfusion, adding:
“Morphometric analysis of megakaryocytes plays a central role in diagnosing and subclassifying Philadelphia-chromosome negative Myeloproliferative Neoplasms (MPNs), as specific cellular and spatial abnormalities distinguish Essential Thrombocythemia (ET), Polycythemia Vera (PV), and Primary Myelofibrosis (PMF).
While manual histopathological evaluation is a standard World Health Organization (WHO) requirement, digital morphometry and machine learning workflows on bone marrow trephine biopsies provide precise, reproducible, and automated quantification of these parameters to avoid subjective variation.
Primary Morphometric Parameters Analyzed Digital image analysis systems typically measure several core quantitative indicators from H&E or immunohistochemically stained sections:
- Cellular Size and Area: Calculated via the cytoplasmic major diameter and total cytoplasmic area.
- Nuclear Area and Perimeter: Measures the total footprint of the nucleus and its boundary length.
- Nuclear-to-Cytoplasmic (N:C) Ratio: Derived by dividing the nuclear area by the overall cytoplasmic area.
- Nuclear Shape Indices: Quantified via the Nuclear Roundness Factor and Nuclear Contour Ratio to capture the irregularity or lobation profile.
- Topographical Clustering and Density: Quantitative spatial alignment mapping whether cells form loose or tight clusters (3 or more juxtaposed cells) and their spatial distance from bone trabeculae or blood sinuses.
- Advanced and Predictive Morphometric markers:
Megakaryocytic Activation (M-ACT Index):
A specialized morphometric index scored on the coexistence of megakaryocytic emperipolesis (intact neutrophils passing through megakaryocyte cytoplasm), tight cluster count, and dense collagen fiber perimeters.
High M-ACT indices mathematically correlate with higher JAK2 V617F allele burden and predict rapid progression to secondary myelofibrosis.
AI-Based Morphological Fingerprinting: Advanced neural networks extract up to 9 distinct megakaryocytic phenotypes from bone marrow images based on shape variations, generating cohort-wide radar plots.
This high-throughput tool isolates true MPNs from reactive mimics (like immune thrombocytopenia or reactive thrombocytosis) with an area under the curve (AUC) accuracy of up to 0.95.”
Title: Quantitative Evaluation of Megakaryocytes in Bone Marrow Biopsy of Myeloproliferative Neoplasms by Morphometry
Authors: Gurpreet Kaur Sagoo, Dibyajyoti Boruah, Ankur Ahuja, Barun Kumar Chakrabarty, S. Venkatesan
Stay updated on all scientific advances with Hemostasis Today.
-
May 21, 2026, 16:22Konstantin Yenkoyan: Neuroscience Week 2026 Continues in Yerevan
-
May 21, 2026, 16:21Ifeanyichukwu Ifechidere: Rebalanced Haemostasis in Liver Disease
-
May 21, 2026, 16:20Stéphanie Forté: Gene Therapy in Sickle Cell Disease from Molecular Advances to Clinical Care
-
May 21, 2026, 15:22Divyaswathi Citla Sridhar: Addressing the Gaps in Inherited Platelet Function Disorders
-
May 21, 2026, 14:35Wolfgang Miesbach: From genes to mechanisms – An Outstanding Update on Inherited Platelet Disorders
-
May 21, 2026, 14:04Bastu Odoka։ The World of Transfusion Medicine Keeps Evolving
-
May 21, 2026, 13:53A New Story of Hope for Children With ITP – ITP Support Association
-
May 21, 2026, 13:40Sonal Sonu: Less Than 50 People in the World Have Golden Blood
-
May 21, 2026, 13:33Danny Hsu: Honoured to Speak at the International Society of Hematology Congress in Bangkok