Analytical and Morphological Performance of a Digital Image-Based Morphology Analyzer Versus a Conventional Flow-Cytometry-Based Automated Hematology Analyzer

Abstract

A method-comparison study (n=417) benchmarked miLab™ BCM against the Sysmex XN-series for CBC and WBC differentials. BCM showed high concordance for core CBC indices (r >0.97), superior lymphocyte and monocyte agreement with expert manual review, and strong blast detection performance (r=0.921; sensitivity 88%, NPV 96%). Its AI-driven dual-imaging workflow positions BCM as an efficient verification layer in routine hematology workup.

Figure 1. WBC differential correlation (r) and blast detection performance of the digital analyzer. Error bars: 95% CI. The digital analyzer showed improved lymphocyte and monocyte concordance with expert manual review compared to the automated analyzer alone.

Key Highlights

• miLab™ BCM demonstrated high concordance with the Sysmex XN-series across core CBC parameters (RBC r=0.975, WBC r=0.985, PLT r=0.972, Hgb r=0.977).
• WBC differential agreement with expert manual review surpassed Sysmex for lymphocytes (r=0.937 vs. 0.897) and monocytes (r=0.891 vs. 0.860), reflecting closer alignment with expert visual morphology classification.
• Blast detection showed strong correlation with expert microscopy (r=0.921), with a high NPV of 96% - supporting reliable rule-out of blast-negative samples.
• The entire workflow - CBC, smear preparation, staining, imaging, and AI-driven morphological classification - is completed in a single automated run, reducing manual smear review burden.

Keywords