Comparison of Sysmex XN-V body fluid mode and deep-learning-based quantification with manual techniques for total nucleated cell count and differential count for equine bronchoalveolar lavage samples.
Authors: Lapsina Sandra, Riond Barbara, Hofmann-Lehmann Regina, Stirn Martina
Journal: BMC veterinary research
Summary
# Editorial Summary Bronchoalveolar lavage remains essential for diagnosing equine lower airway disease, yet the routine manual counting of cells in these samples is labour-intensive and inconsistent between operators. Lapsina and colleagues evaluated two automated approaches against traditional manual methods across 69 clinical samples: the Sysmex XN-V analyser (a haematology instrument operating in body fluid mode) for total nucleated cell counts and basic two-part differentials, and the Olympus VS200 digital slide scanner paired with deep-learning software for more detailed four-part cell classification including alveolar macrophages, lymphocytes, neutrophils, and mast cells. The Sysmex XN-V showed reasonable concordance for total counts but had limitations in accurately segregating mononuclear from polymorphonuclear cells; the deep-learning approach demonstrated promising capability for identifying specific cell types, though the paper implicitly suggests variable reliability depending on cell category. For practitioners, these findings indicate that whilst automated methods could substantially reduce turnaround time in respiratory diagnostics, neither technology has entirely replaced manual review, making them most valuable as screening tools or for high-volume practices rather than as standalone replacements for careful microscopy in diagnostically challenging cases.
Read the full abstract on PubMed
Practical Takeaways
- •Consider implementing Sysmex XN-V automated analysis for routine BAL TNCC and basic differential counts to reduce laboratory turnaround time without sacrificing accuracy
- •Deep-learning slide scanning may offer a practical alternative for detailed four-part BAL differentials, allowing faster reporting of respiratory disease diagnostics
- •Automated BAL analysis can free up technician time for other laboratory tasks while providing clinically reliable results for respiratory disease assessment
Key Findings
- •Sysmex XN-V body fluid mode showed acceptable agreement with manual TNCC and two-part differential (mononuclear vs polymorphonuclear) in equine BAL samples
- •Deep-learning-based algorithm for four-part differential (alveolar macrophages, lymphocytes, neutrophils, mast cells) demonstrated comparable performance to manual counting methods
- •Automated methods provide significantly faster processing times compared to manual techniques while maintaining diagnostic accuracy