Background: Flow cytometry (FCM) is a co‐criterion in myelodysplastic syndromes (MDS) diagnostics according to the WHO classification. The presented study compared diagnostic power and prognostic impact of different FCM‐based scores.Methods: A total of 807 bone marrow (BM) samples of patients with cytopenia (543 MDS, 153 non‐clonal cytopenias, 111 non‐MDS myeloid malignancies) and 78 healthy controls have been investigated using a standardized 8‐color‐FCM procedure. FCSS, Ogata‐score, iFS, RED‐score, and ELN‐NEC were analyzed for sensitivity and specificity in comparison to standard diagnostic tools. Median follow up for patients was 26 month (range: 0.2–89).Results: The iFS showed the highest accuracy (80%) with the best balance between sensitivity (79%) and specificity (86%). This was also valid in MDS with very low IPSS‐R and even in MDS without ring sideroblasts, with normal blast count and karyotype, where iFS could confirm diagnosis in 62% and 65% of patients. Besides the high diagnostic power, the established iFS category “consistent with MDS” was associated with inferior overall survival (OS) independent from WHO classification (median: 51 month vs. not reached,p < 0.0001). Remarkably, this iFS category redefined a subgroup of patients with worse OS within IPSS‐R low‐risk category (73 month vs. not reached,p = 0.0433). Finally, multivariable analysis showed that iFS added independent prognostic information regarding OS besides IPSS‐R.Conclusions: The iFS separates non‐clonal cytopenias and MDS with the highest accuracy, provided information in addition to standard diagnostic procedures, and refined established prognostic tools for outcome prediction.

By simultaneous DNA staining and immunophenotyping, the aneuploid blasts of acute lymphoblastic leukemia (ALL) can be quantitated by flow cytometry. The present study evaluates the application of this method for the detection of minimal residual disease (MRD). For this purpose leukemia cells with known aneuploidy were serially diluted with diploid bone marrow cells obtained after chemotherapy. These mixed samples were immunophenotyped using antibodies CD34, CD10, or CD19 and processed for DNA staining with propidium iodide. Fifty thousand cells were measured for each experiment and the results were compared with the analysis of control sample without added aneuploid cells. Aneuploid cells diluted to 1% were not detectable in 6/15 cases when DNA content analysis was used alone, and in dilutions of < 1%, these cells were not detectable. When DNA content and immunophenotype were simultaneously measured, aneuploidy was detectable in 15/15 cases at 1.00%, 6/6 at 0.50%, 16/16 at 0.10%, and 16/17 at 0.05%. Therefore this method may be useful for the detection of minimal residual disease in cases of aneuploid lymphoblastic leukemia. Cytometry 30:47–53, 1997.

Typical state-of-the-art flow cytometry data samples typically consist of measures of 10 to 30 features of more than 100,000 cell “events”. Artificial intelligence (AI) systems are able to diagnose such data with almost the same accuracy as human experts. However, such systems face one central challenge: their decisions have far-reaching consequences for the health and lives of people. Therefore, the decisions of AI systems need to be understandable and justifiable by humans. In this work, we present a novel explainable AI (XAI) method called algorithmic population descriptions (ALPODS), which is able to classify (diagnose) cases based on subpopulations in high-dimensional data. ALPODS is able to explain its decisions in a form that is understandable to human experts. For the identified subpopulations, fuzzy reasoning rules expressed in the typical language of domain experts are generated. A visualization method based on these rules allows human experts to understand the reasoning used by the AI system. A comparison with a selection of state-of-the-art XAI systems shows that ALPODS operates efficiently on known benchmark data and on everyday routine case data.