BACKGROUND: Although optical microscopy (OM) remains the reference technique for analysis of ascitic (AF) and pleural (PF) fluids, novel hematological analyzers are equipped with modules for body fluid (BF) analysis. This study was aimed to analyze the performance of XN-BF module in Sysmex XN-9000, and to develop validation rules for automated cell counts in BFs. METHODS: The evaluation of XN-BF module included assessment of carryover, Limit of Blank (LoB), Limit of Detection (LoD), Limit of Quantitation (LoQ), linearity, data comparison with OM, and development of rules for assisting the validation of automated analysis of BFs and activating reflex testing. RESULTS: The carryover was negligible. The LoB, LoD, LoQ and linearity were always excellent. The comparison with OM was characterized by Pearson's correlations ranging from r=0.50 to r=0.99 (p<0.001), modest bias and high diagnostic concordance (Area Under the Curve between 0.85 and 0.99). The use of instrument-specific cut-offs further increased diagnostic concordance. The implementation of reflex testing rules based on XN-BF data increased sensitivity and specificity of BFs classification to 0.98 and 0.95. CONCLUSIONS: Our results suggest that the XN-BF module on Sysmex-9000 may be a suitable alternative to OM for screening BF samples, especially when specific validation rules are used.
BACKGROUND: Although optical microscopy (OM) remains the reference technique for analysis of ascitic (AF) and pleural (PF) fluids, novel hematological analyzers are equipped with modules for body fluid (BF) analysis. This study was aimed to analyze the performance of XN-BF module in Sysmex XN-9000, and to develop validation rules for automated cell counts in BFs. METHODS: The evaluation of XN-BF module included assessment of carryover, Limit of Blank (LoB), Limit of Detection (LoD), Limit of Quantitation (LoQ), linearity, data comparison with OM, and development of rules for assisting the validation of automated analysis of BFs and activating reflex testing. RESULTS: The carryover was negligible. The LoB, LoD, LoQ and linearity were always excellent. The comparison with OM was characterized by Pearson's correlations ranging from r=0.50 to r=0.99 (p<0.001), modest bias and high diagnostic concordance (Area Under the Curve between 0.85 and 0.99). The use of instrument-specific cut-offs further increased diagnostic concordance. The implementation of reflex testing rules based on XN-BF data increased sensitivity and specificity of BFs classification to 0.98 and 0.95. CONCLUSIONS: Our results suggest that the XN-BF module on Sysmex-9000 may be a suitable alternative to OM for screening BF samples, especially when specific validation rules are used.