Skip to main content
. 2022 Sep 3;35(12):1882–1887. doi: 10.1038/s41379-022-01141-4

Fig. 2. Integration of AI-based solutions in the clinical pipeline.

Fig. 2

Schematic overview of the integration of the AI-solution into the routine pathological biomarker diagnostic practice. After the biopsy is cut and stained with H&E, it is digitally scanned and transferred for pathological examination. Samples are further processed by the traditional route including IHC screening as well as FISH and NGS, to assess ALK and ROS1 status. A final biomarker report is produced after an average of four weeks. The AI software (Imagene-AI. LTD) automatically analyses the WSI directly after the scanning of the H&E slide, taking an average of 2 min for ALK/ROS1 status inference. Flagging of samples carrying ALK or ROS1 alterations can be viewed immediately after this process by the pathologist and assist in their further analysis prioritization. After the results of the routine test panel are ready, they can be compared to the AI classifier inference. Discrepant cases can then be subjected to additional testing and clarification. AI artificial intelligence, WSI whole image slide, IHC Immunohistochemistry, FISH Fluorescence in situ hybridization, NGS Next-generation sequencing.