Morphogo
- Bailey
- 0
Abstract
Introduction: The differential count of nucleated cells in bone marrow aspirate smears is necessary for the clinical diagnosis of haematological malignancy. The manual bone marrow differential count is time-consuming and lacks consistency. In this study, a new system based on artificial intelligence (AI) was developed to perform automatic cellular classification of bone marrow cells and determine their possible clinical applications.
Materials and methods:
Bone marrow aspirate smears were obtained from Xinqiao Hospital of the Army Medical University. First, an automated analysis system (Morphogo) scanned and generated complete digital images from bone marrow smears. The software then analyzed the nucleated bone marrow cells in the selected areas of the smears at a magnification of × 1,000 using an AI-based platform.
Cell sorting results were independently reviewed and confirmed by 2 experienced pathologists. The automatic cell sorting performance of the system was evaluated using 3 categories: precision, sensitivity, and specificity. Correlation coefficients and linear regression equations were calculated between the automatic cell classification by the AI-based system and the concurrent manual differential count.
See It Here: Roche Diagnostics Roche-09365397023 believes that broad access to such high-quality and reliable, instrument-less Point of Care tests can contribute significantly to better management of the current pandemic crisis.
Results:
In 230 cases, the classification precision was greater than 85.7% for cells of hematopoietic lineage. The averages of sensitivity and specificity of the system were found to be 69.4 and 97.2%, respectively. The cell percentage differential from the automated count based on 200-500 cell counts correlated with the cell percentage differential provided by pathologists for granulocytes, erythrocytes, and lymphocytes (r ≥ 0.762, p <0.001).
Discussion / conclusion:
This pilot study confirmed that the Morphogo system is a reliable tool for automatic analysis of bone marrow cell differential count and has potential for clinical applications. Current large-scale multicenter validation studies will provide further information to further confirm the clinical utility of the system.
Keywords:
Automatic cell classification; Bone marrow aspirate smear; Cell morphology; Digital image.