Fig. 5: Proposed clinical workflow for patient prescreening using the image based FGFR prediction device. | Nature Communications

Fig. 5: Proposed clinical workflow for patient prescreening using the image based FGFR prediction device.

From: Development and deployment of a histopathology-based deep learning algorithm for patient prescreening in a clinical trial

Fig. 5

Clinical workflow shows three parties involved: clinical study sites, central laboratory, and cloud platform with image-based AI device. Central laboratory routinely tests tissue for clinical study qualifying FGFR alterations (Standard workflow). Tissue is also stained, scanned, then an automated tool sends the image to a cloud platform to run the FGFR Device (Machine Learning (ML) workflow). The physicians can then evaluate the FGFR Device prediction on an online-portal and notify the central laboratory to cancel the molecular test within 48 h.

Back to article page