Integrating CT-Derived Body Composition and Real-Time Injection Pressure Monitoring for Predicting Contrast Media Extravasation: A Retrospective Cohort Study with Implications for Nurse-Led Risk Stratification

Purpose: Contrast media extravasation can complicate CT angiography (CTA), particularly when high-flow injection protocols are used. Current risk assessment relies mainly on clinical factors and injection pressure monitoring, whereas opportunistic CT-derived body composition has rarely been evaluated for this purpose. Methods: This single-center retrospective cohort study analyzed 12,460 consecutive adult patients who underwent contrast-enhanced abdominal CTA. The L3 skeletal muscle index (SMI) was extracted from routine CT images, and pressure metrics were obtained from power injector logs. Logistic-regression models incorporating clinical variables, pressure parameters, and body composition variables were evaluated and internally validated. Results: Extravasation occurred in 380 patients (3.05%). Patients with extravasation had lower L3 SMI, a higher prevalence of sarcopenia, higher peak pressures, and elevated injection pressure-to-rate ratios (IPIRs). In multivariable analysis, lower L3 SMI, sarcopenia, peak pressure, IPIR, age, catheter gauge, and injection site were associated with extravasation. The combined model achieved an AUC of 0.941 (95% CI 0.927–0.955), showing higher discrimination than the pressure-only model (AUC 0.926, 95% CI 0.910–0.942), body composition-only model (AUC 0.840, 95% CI 0.819–0.861), and clinical-only model (AUC 0.796, 95% CI 0.772–0.820). The bootstrap optimism-corrected AUC was 0.928. Conclusions: Integrating CT-derived body composition with real-time pressure monitoring may improve individualized extravasation risk stratification during abdominal CTA. Prospective multi-center validation is needed before clinical implementation.