Clinical Summary
Thought leadership
Performance Evaluation of a Novel Artificial Intelligence-Assisted Digital Microscopy System for the Routine Analysis of Bone Marrow Aspirates
BACKGROUND
BMA smear analysis is essential for diagnosing and monitoring benign and neoplastic hematological conditions. Traditionally performed manually through microscopy, this method is labor-intensive, subjective, and time-consuming, requiring the evaluation of at least 500 cells. We propose a digital solution to assist the hematopathologist in reviewing BMA specimens by scanning samples at 100x magnification and utilizing the AI-DSS to streamline analysis. 1 d p2 Bagg
METHODS
We analyzed 795 BMA specimens (615 Romanowsky-stained and 180 Prussian blue-stained) from three renowned clinical sites: the University of Pennsylvania, Oregon Health & Science University, and Tel Aviv Sourasky Medical Center. The study included patients with neoplastic and other clinical conditions, comparing the Scopio Labs X100 Full Field BMA system (test method) to manual microscopy (reference method) in terms of efficiency, sensitivity, specificity, as well as repeatability and reproducibility.
RESULTS
Efficiency, Sensitivity and Specificity
BMA primary parameters (normal or abnormal) and secondary parameters (increased/decreased/ normal)* were evaluated, showing high correlation with manual methods.
- Efficiency: 90.85% for Romanowsky-stained samples and 90.00% for Prussian blue-stained samples
- Sensitivity: 81.61% and 81.94%, respectively
- Specificity: 92.88% and 93.38% respectively
Standardization and Reliability
Overall agreement was 91.1%, with repeatability and reproducibility within predefined criteria.
Specimen Analysis
The system analyzed an average of 1,385 (± 536) cells per specimen.. 1 e p1,b p2 Bagg Additionally, 87% (± 21%) of samples were accepted by operators. 1 e p1 Bagg This resulted in 17.62 (± 7.24) ROIs per specimen.


CONCLUSION
The strong correlation between the digital system and manual microscopy demonstrates its credibility for accurate, high-quality BMA analysis, accelerating expert review and diagnosis. It also supports ** remote evaluation and new research into normal and neoplastic hematopoiesis.
“This technology allows hematopathologists to evaluate BMA remotely, without needing a physical microscope or being on-site, enabling timely assessments in remote or under-resourced healthcare settings.” 1 a p9 Bagg
AI-Artificial Intelligence; BMA: Bone Marrow Aspirate; DSS- Decision Support System ICSH: International Council for Standardization in Haematology ROIs: Regions of Interest
*Area selection for digital BMA analysis was performed in 2 phases. First, the whole slide image was scanned for clinically relevant areas, and the number of particles, stripped cells, and hemodilution were assessed for adequate sample quality. Within the clinically relevant areas, a second analysis was performed on ROIs, including a differential of all cells within the prespecified area. These cells were then preclassified into the 27 primary and 13 secondary characteristics for morphological assessment.
Reference:
Bagg A, Raess PW, Rund D, Bhattacharyya S, Wiszniewska J, Horowitz A, et al. Performance Evaluation of a Novel Artificial Intelligence-Assisted Digital Microscopy System for the Routine Analysis of Bone Marrow Aspirates. Mod Pathol. 2024 Sep;37(9):100542. doi: 10.1016/j.modpat.2024.100542
* Scopio’s Full-Field Bone Marrow Aspirate Application is CE-marked and cleared for sale in CE countries as well as in additional regions. Not commercially available in the US for in vitro diagnostic procedures.
** Scopio’s Full-Field remote capabilities are available through the secure hospital network.
*** BMA is compatible both with Scopios’ X100 and X100HT.