RESUMO
INTRODUCTION: Chronic Myelomonocytic Leukemia (CMML) and Myelodysplastic Syndromes (MDS) are increasingly represented in the general population. We propose a screening strategy based on algorithms calculated from quantitative and analytical data from the XN analyser. MATERIALS AND METHODS: We tested the performance of previously published MDS and CMML scores on an evaluation cohort of 749 individual eligible patients over 50 years of age. These patients were classified into 3 groups as follows: 713 patients without MDS or CMML, 18 patients with MDS, and finally 18 patients with CMML. In a second step, a routine cohort of 37 828 samples was studied to evaluate the impact of this approach. RESULTS: The concordance rate between cytology and the two scores is 92.1%. The sensitivity and specificity of the CMML score are 100% and 96.2%, respectively. For the MDS score, they are 83.3% and 89.6% respectively. The ratio of platelets measured by fluorescence on board (PLT-F) as reflex tests generated is 1.5% after 6 months. The additional smear ratio for suspected MDS is calculated at 0.6%. CONCLUSION: We propose a flowchart using embedded artificial intelligence to help the cytologist in an optimized smear review and thus improve guidance to the clinician and the patients in the diagnosis process. This strategy permits a more comprehensive approach to MDS and CMML detection fitting with the new definition of CMML according to the recommendations of the World Health Organization (WHO) published in 2022.
