The Lymphoma Channel on VJHemOnc is an independent medical education platform, supported with funding from AstraZeneca (Diamond), BMS (Gold), Johnson & Johnson (Gold), Takeda (Silver) and Galapagos (Bronze). Supporters have no influence on the production of content. The levels of sponsorship listed are reflective of the amount of funding given.
The Multiple Myeloma Channel on VJHemOnc is an independent medical education platform, supported with funding from BMS (Gold) and Legend Biotech (Bronze). Supporters have no influence on the production of content. The levels of sponsorship listed are reflective of the amount of funding given.
EHA 2025 | SmartCytoFlow 2.0: integrating discriminative and generative AI for flow cytometry analysis
Adrian Mosquera-Orgueira, MD, PhD, University Hospital of Santiago de Compostela, Santiago, Spain, comments on the latest version of SmartCytoFlow, an artificial intelligence (AI) platform for automated flow cytometry analysis. Dr Mosquera-Orgueira notes that this version integrates both discriminative and generative AI, providing not only predictions but also text descriptions that explain the reasoning behind the predictions. This interview took place at the 30th Congress of the European Hematology Association (EHA) in Milan, Italy.
These works are owned by Magdalen Medical Publishing (MMP) and are protected by copyright laws and treaties around the world. All rights are reserved.
Transcript
Smart CytoFlow is an AI platform designed for the automatization of flow cytometry analysis, particularly automated and standardized analysis of Euroflow standard tubes. So what Smart CytoFlow originally does is to apply discriminative AI models. These models do actually identify different subsets of cells in the samples and obviously it also links that with diagnosis so we have done some developments for lymphoma detection for myeloma detection and both for diagnosis and MRD and also for lymphoblastic leukemia...
Smart CytoFlow is an AI platform designed for the automatization of flow cytometry analysis, particularly automated and standardized analysis of Euroflow standard tubes. So what Smart CytoFlow originally does is to apply discriminative AI models. These models do actually identify different subsets of cells in the samples and obviously it also links that with diagnosis so we have done some developments for lymphoma detection for myeloma detection and both for diagnosis and MRD and also for lymphoblastic leukemia. But in this second version of SmartCytoFlow, we have added a new functionality. So rather than just relying only on the predictions made by the discriminative AI, that is, whether your sample is tumor or normal and the uncertainty of that prediction, we also provide some text descriptions based on generative AI, which link those predictions to the phenotypes that are being detected as abnormal in the samples and makes a relevant explanation on why the model is actually making that prediction. So this second version of the SmartCytoFlow is an integration of discriminative AI and generative AI for immunophenotypic automatization, and it’s really promising.
This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.
Read more...
