Risk Factors and Deep Learning Model for CI-AKI

Investigation and Early Risk Prediction of Acute Kidney Injury Caused by Iodinated Contrast Agents: Development of CI-AKIDPi Model Based on Interpretable Deep Learning

Contrast-associated acute kidney injury (CI-AKI) is a sudden and significant decline in renal function resulting from the use of a contrast agent.

Contrast-associated acute kidney injury (CI-AKI) is a sudden and significant decline in renal function resulting from the use of a contrast agent. Despite the frequent use of iodinated contrast agents in medical investigations, predictive models for CI-AKI are scarce. This study aimed to construct and validate interpretable deep learning models for the early risk prediction of acute kidney injury (AKI) associated with iodinated contrast agents.

Inclusion Criteria:

• patients who received iodinated contrast agents during hospitalization
• age ≥ 18 years
• hospitalization length ≥ 48 hours
• at least two serum creatinine tests during hospitalization

Exclusion Criteria:

• patients with stage 4 or 5 chronic kidney disease prior to admission
• patients undergoing hemodialysis, hemofiltration or peritoneal before admission
• a history of nephrectomy or kidney transplantation during hospitalization
• incomplete clinical information
• Serum creatinine always ≤ 40 μmol/L during hospitalization