Country / Region
EMEA
Tags
Artificial intelligence, Clinical Practice, Data analytics, Patient safety
Fluoropyrimidines, a class of antimetabolite drugs used in chemotherapy, can cause severe cardiac complications including coronary artery spasms with potential fatal outcomes. Current risk assessment approaches are hindered by insufficient data on predisposing factors. Here, we present a retrospective audit examining cardiac risk in patients receiving fluoropyrimidine chemotherapy (5-fluorouracil and Capecitabine). Our study leverages MedCAT natural language processing technologies and SNOMED CT to analyse clinical records, identify risk factors, and evaluate outcomes in these patients. The findings aim to improve early identification of high-risk individuals, facilitate timely referrals to cardio-oncology specialists, and optimize clinical decision-making through better utilization of existing healthcare data.
Description
Hospital-wide risk factor and outcome data for a historical cohort of 3446 patients that have undergone chemotherapy with Fluoropyrimidines is surfaced using an AI model trained to identify SNOMED CT concepts.
Scope
SNOMED CT was selected for its level of granularity and its concept orientation. Its structure, non-ambiguity and non-redundancy facilitate good performance in named entity recognition and linking methods used by natural language processing models.
How SNOMED CT will be used
SNOMED CT concepts are used by a natural language processing model (developed using MedCAT) enabling it to identify comorbidity and medical history risk factors and cardiac outcomes of patients undergoing Fluoropyrimidine chemotherapy.
Why SNOMED CT will be used
Contact
