AI100 with Shonit

The AI100 with Shonit™ is an automated cell-locating device intended to improve efficiency and standardization in hematological reviews by digitally imaging peripheral blood smears and pre-classifying white blood cells, red blood cells, and platelets. Its methodology relies on microscopic imaging, motorized slide scanning, and software-based image analysis to present cells for final verification by a trained operator, with studies demonstrating strong performance in cell identification and morphological assessment compared to manual microscopy techniques.

The AI100 with Shonit™ is an in-vitro diagnostic device designed for clinical laboratories to aid in hematological analysis. Its primary functions include performing differential counts of White Blood Cells (WBCs), characterizing Red Blood Cell (RBC) morphology, and assessing Platelet (PLT) morphology from peripheral blood smears. The device automates the process of locating blood cells on a prepared slide and presents digital images of these cells for review by a skilled operator, who then identifies and classifies each cell. The system utilizes a high-resolution microscope with LED illumination, a motorized XYZ stage for slide manipulation, and a camera to capture images. Software controls hardware operations, manages image data, analyzes images to pre-classify cells, and facilitates report generation. The operational principle of the AI100 with Shonit™ mirrors manual microscopy, beginning with the device scanning a peripheral blood smear (prepared with Romanowsky stain from a K2EDTA anticoagulated sample) to identify an optimal area for analysis. It then captures multiple Field of View (FOV) images in a spiral pattern, applying image processing algorithms to extract and pre-classify WBCs (into types like neutrophils, lymphocytes, blasts), RBCs (by size and shape like normocytes, macrocytes, target cells), and Platelets (normal, large, giant, clumps). Analytical performance studies demonstrated acceptable repeatability and reproducibility for WBC, RBC, and platelet parameters, meeting predefined criteria. Method comparison studies involving 882 samples showed strong agreement with manual microscopy for WBC differentials, RBC morphological characterization (anisocytosis, macrocytosis, poikilocytosis), and platelet morphology assessment, with high sensitivity and specificity for various abnormalities.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the AI100's 40X objective lens compare to systems using 100X oil immersion for resolving fine morphological details critical for identifying subtle dysplastic changes or rare cell types?; What is the typical processing time per slide, from loading to the availability of pre-classified images for operator review, and how does this compare to fully automated hematology analyzers or traditional manual review?; Given the device uses an outward spiral scanning pattern, how does it ensure comprehensive coverage of the optimal viewing area and avoid missing clinically significant cells that might be sparsely distributed?; The system currently does not interact with Laboratory Information Systems (LIS); what are the plans and timelines for LIS integration, and what data standards will be supported for future interoperability?; How does the image processing algorithm differentiate between various RBC poikilocytes (e.g., echinocytes vs. acanthocytes, or different types of fragmented cells) and what is its performance on samples with multiple co-existing RBC abnormalities?