Figure 2. Sample patient note with its extracted phenotypes using the cTAKES pipeline.

While cTAKES could correctly categorize most medical terms, it falsely detected rows marked with red signs with their mistaken tags marked in bold. In our customized query, we excluded cTAKES modifier tags that represented phenotypes identified based on family members or patients’ history, were negated, generic, conditional, or detected with uncertainty. These records are marked in orange.