566 - Point of Care Ultrasound Meets Artificial Intelligence: Evaluating Auto-IVC for Pediatric Inferior Vena Cava Assessment

Publication Number: 1543.566

Hannah L. Green, University of South Carolina School of Medicine Columbia, Lexington, SC, United States; Abigail McKenzie, University of South Carolina School of Medicine Columbia, Columbia, SC, United States; Kevin Weinberger, Prisma Children's Hospital/ University of SOuth Carolina, Columbia, SC, United States; Duncan L. Norton, University of South Carolina School of Medicine Columbia, Columbia, South Carolina, SC, United States; Kristopher Dixon, Prisma Children's Hospital - Midlands, Columbia, SC, United States; Jeffrey Holloway, University of South Carolina School of Medicine Columbia, Columbia, SC, United States

Background: Point-of-care ultrasound can provide a rapid, noninvasive, and inexpensive hemodynamic assessment of pediatric patients by evaluating the diameter and collapsibility index (CI) of the inferior vena cava (IVC). GE has incorporated artificial intelligence (AI) software, including an Auto-IVC feature, that automatically calculates IVC diameter and CI.

Objective: This study aims to evaluate the performance of the Auto-IVC feature of GE Venue Go 4 in pediatric patients by comparing AI-generated measurements of IVC diameter and CI with manual calculations.

Design/Methods: This study included healthy patients ranging from ages 0 months to 18 years presenting for wellness checks at Prisma Health Children's Hospital Outpatient Clinic. Patients with acute or chronic illness were excluded. Images were acquired in accordance with the American Society of Echocardiography (ASE) guidelines for IVC measurement. POCUS exams were performed using Motion Mode of the GE Venue GO 4 ultrasound machine with a 1.4-5.7-MHz curvilinear probe to obtain a longitudinal time-lapsed view of the IVC over the course of three tidal breaths. Auto-IVC was used to calculate the average minimum and maximum diameter and CI of the IVC. The average of three manually measured maximum and minimum IVC diameters was used to calculate the average diameter and collapsibility index (CI) for each patient. AI and manual measurements were analyzed with paired t-tests.

Results: A total of 135 patients were enrolled in the study and divided into age groups: < 6 months (n=11), 6≤age < 12 months (n=9), 12≤age < 36 months (n=12), 3≤age < 5 years (n=19), 5≤age < 9 years (n=25), 9≤age < 13 years (n=37) and 13≤age < 18 years (n=22). The GE Venue GO 4 was unable to calculate Auto-IVC data for pediatric patients less than 12 months. The level of statistical significance for all tests was p≤0.05. No statistically significant difference between auto-IVC and manual measurements was found except for in the 5≤age < 9 years age group.

Conclusion(s): No significant difference was identified between manual and Auto-IVC measurements across most pediatric age groups, indicating good agreement between the two methods. However, a statistically significant difference was observed in the 5≤age < 9 years age group for minimum and maximum diameter, suggesting potential variability in Auto-IVC performance in this group. The principal limitation of this study is the relatively small sample size. Overall, these findings suggest that the Auto-IVC feature may be a reliable substitute for manual measurement of IVC diameter and collapsibility index in pediatric patients over a year of age.

P-value Table
P-values comparing manually calculated measurements to Auto-IVC generated values for maximum and minimum IVC diameters and collapsibility index broken down into ages groups.

Ultrasound Image of Auto-IVC Function
Auto-IVC function obtaining maximum diameter, minimum diameter, and collapsibility index.