DeepPath™ – LYDIA is an AI-powered decision support system for detecting metastases on digitised microscope slides of H&E stained Lymph Nodes. The system was CE-Marked in May 2022 as a Class I in-vitro diagnostic product as per IVDD.
DeepPath™ – LYDIA analyses Whole Slide Images (WSI) and generates probabilistic heatmaps of tumour presence.
Segments are generated through a thresholding operation applied on those heatmaps. The segments outlining the tumour area are presented to the histopathologist for inspection.
High Performance across Multiple Cancer Types
A single sensitive and accurate AI-pipeline is trained to detect tumour hotspots from four different cancer types, namely breast, colon, lung and melanoma (expansion to additional types TBA). DeepPath™ – LYDIA processes a microscope slide digitised at 40X magnification within 15 minutes with Sensitivity > 99% at the tumor-segment-level and a False Positive Rate of 9 False Positive tumor segments per Lymph Node, as measured on blind samples from different labs digitised on different scanners.
Decision Support that Accelerates Diagnosis
Recent pilot runs have indicated that system utilisation accelerates the diagnostic process by up to 5 times ensuring at the same time that the number of missed metastatic lesions is minimised.
Flexible Deployment as per Client Requirements
DeepPath™ – LYDIA can be deployed either on the cloud or locally and currently supports Phillips Tiff, Aperio, Hamamatsu and 3D-Histech WSI formats. Multiple digitised microscope slides are processed in parallel by specialised GPUs. As such, the system is scalable limited only by the number of available GPUs. Finally the system is offered either as standalone software (includes basic case management), or as an integrated component to third party Picture Archiving and Communication Systems (PACS).
Try LYDIA on your samples for free and if happy with the results we will provide options for:
- A yearly licence with unlimited users based on system usage as estimated from past data
- Our Pay as you Use scheme with a fixed cost per microscope slide analysed