Schematic of the data extraction and normalization pipeline. (A) Unstructured medical text was parsed, and text segments pertaining to biomarker information (region of interest [ROI]) were extracted.