Digital cytopathology streamlines workflows by fostering synergy between physicians and technology. With AI managing time-consuming, quantifiable tasks, human experts apply their clinical judgment to verify and interpret results. To support interpretation, Cellbytes is the first platform to also provide reference values for healthy subjects. Remote collaboration becomes effortless, enabling seamless sample evaluation and colleague consultations. Cellbytes brings digital cytopathology to life.
Capture cytomorphological slides using your microscope.
Images are securely uploaded to the cloud-based file storage.
Cellbytes analyzes key features from whole-slide and single-cell images.
Pre-classification results are reviewed and verified by a physician.
Share a standardized digital report directly to your laboratory information system.
Prioritize urgent samples with a multi-parameter risk model.
Access whole-slide images, numerical analysis, and visual masks.
View single-cell images with classifications and potential dysplasia indicators.
Compare results with population-scale reference ranges.
Assess and share clinically essential findings in a structured report.
Easily share slides and cell images with colleagues for consultations and meetings.
Our AI aims to perform nearly all the same diagnostic tasks as doctors, enhancing every step of the analysis process rather than just optimizing isolated aspects. By combining cell classification with morphology, dysplasia assessment, and sample quality analysis, we provide a fresh angle for research.
Triage score: Assessment of the urgency of the bone marrow sample. Samples can be prioritized based on their triage score, ensuring that the most critical cases are analyzed first.
Scarcity estimate: Assessment of whether the sample is representative enough for cytomorphological assessment. Scarcity is determined based on the quantity of cellular material in the sample slide.
Cellularity estimate: Estimate of the density of hematopoietic cells in the bone marrow.
Out-of-focus: Identification of the out-of-focus areas within the whole-slide image. Users are warned if the proportion of out-of-focus area exceeds a critical threshold indicating that clinically relevant parameter values might be unreliable.
Sample region: Identification of the region within the whole-slide image of the sample that contains the specimen. The identified region is visualized as an overlay mask on the whole-slide image.
Center region: Identification of the possible center region of the sample within the whole-slide image. The identified region is visualized as an overlay mask on the whole-slide image.
Megakaryocytes in the whole slide: Quantification of megakaryocytes within the sample slide. Megakaryocytes are counted per 10x field of view, and counts are categorized as "low," "intermediate," or "high" for easier interpretation. The identified megakaryocytes are visualized as an overlay mask on the whole-slide image.
Quantity and distribution of red blood cells, eukaryotic cells and stromal tissue within the sample. The application provides numerical data detailing their quantities across various regions of the sample and visualizes them as an overlay mask on the whole-slide image.
Lipid droplets: Quantity and distribution of lipid droplets within the sample. The application provides numerical data detailing their quantities across various regions of the sample and visualizes them as an overlay mask on the whole-slide image.
Cells: Detection and classification of individual bone marrow cells in the selected regions of the sample, categorized into 17 distinct cell types. The application provides differential counts for each cell type, along with their locations, visualized on the images.
Dysplasia: Assessment of dysplastic alterations in individual bone marrow cells. Different forms of dysplasia are assessed in various cell types, including promyelocytic dysplasia, megakaryocyte dysplasia, and erythroblast dysplasia. Users are warned if the quantity of dysplastic changes exceeds a critical threshold.
Cellbytes can be used by uploading high-resolution images of bone marrow aspirates to the platform. The application analyzes these images using advanced machine learning algorithms, delivering quantitative results and visual annotations. It is accessible via a secure web interface—no installation required.
Yes, Cellbytes offers a trial version for institutions and professionals interested in evaluating its capabilities. You can request trial access by contacting the support team through the website.
Cellbytes's algorithms have been trained on a comprehensive real-world dataset comprising different pathologies and over 60,000 samples. The system achieves high accuracy in detecting and classifying various blood cell types and their abnormalities, with performance metrics benchmarked against expert diagnoses.
No, Cellbytes is designed to assist hematopathologists. It serves as a diagnostic aid that enhances efficiency and consistency. Final clinical decisions should always be made by a qualified medical professional.
Cellbytes accelerates and enhances hematological analysis by providing AI-assisted interpretation of bone marrow cell images. Cellbytes will facilitate remote working and consultation but also reduce diagnostic turnaround time and improve standardization, particularly in high-volume settings.
Cellbytes supports high-resolution digital images of May-Grünwald-Giemsa–stained bone marrow aspirate slides. Whole-slide images must be captured at 10x magnification and saved in an open-source format such as MIRAX. Region-of-interest images should be acquired at 100x magnification and saved in JPEG or MIRAX format.
Being cloud-based ensures seamless updates, scalable computing resources, and consistent algorithm performance across users. It also allows for secure data storage and easy collaboration between institutes, laboratories and specialists across geographic locations.
Pricing depends on your usage volume. Contact the support team to test a trial version and discuss your needs.
Absolutely. Cellbytes is increasingly used in research settings to study hematologic disorders, develop new biomarkers, and support clinical trials. It offers data export and integration with other medical data domains.