OC07.05: Integrating artificial intelligence as a virtual assessor for ISUOG's Basic Training trainees
Artificial intelligence (AI) is being massively integrated in healthcare in general, and in imaging in particular. The ISUOG Basic Training (BT) program is a 4-step program with a practical component necessitating evaluation of required images from the participating trainees by ISUOG faculty. Given the challenges and the required time commitment with the expansion of the program, we sought to evaluate the role of integrating AI as a virtual assessor for the trainees in the ISUOG BT program. Retrospective study where images obtained by ISUOG BT trainees and scored in the traditional way by an ISUOG faculty member were subsequently scored by AI. The images included obstetrical images obtained during live scans in accordance with the requirements of BT's 20+2 planes, and gynecological images obtained using various simulators. The images were of varied quality as some were saved directly from the machine (high quality) and others were captured using a smart phone (low quality). The scoring of the images was compared between the ISUOG faculty and AI being mindful of the image quality. Pearson correlation test (r) was used to evaluate the correlation between the assessors. A p-value <0.05 was considered significant. There were 965 images submitted by 35 trainees where 338 (34%) were of high quality and 637 (66%) were of low quality. The scoring between the ISUOG faculty and AI was statistically comparable with good correlation (r = 0.78, p < 0.01) on the high quality images. However, it was lower for the low quality images while maintaining a good correlation (r = 0.71, p < 0.01). AI is a promising tool in ultrasound education where it can have a significant role in throughput and provide quality image assessment comparable to ISUOG faculty. In addition, it has the potential to serve as a tutor to the sonologist with direct feedback on how to enhance image quality. This will aid in standardization of assessment and feedback to trainees on a large scale to ensure safety in accordance with ISUOG BT mission and vision.
Authors
- Nicole El Helou
- Catalina Valencia (ORCID: https://orcid.org/0000-0003-4476-2042)
- G. Rameh
- Reem Abu-Rustum (ORCID: https://orcid.org/0000-0002-8427-7351)
- I. Aboujaoudé
- G. E. Chalouhi (ORCID: https://orcid.org/0000-0002-3302-416X)
- Suresh Seshadri (ORCID: https://orcid.org/0000-0001-5236-8112)
Institutions
- University of Florida (US)
- Institut Necker Enfants Malades (FR)
- Centre Hospitalier de Bigorre (FR)
- Lebanese Hospital Geitaoui-University Medical Center (LB)
- Fetal Care Research Foundation (IN)
Publication Details
- Journal
- Ultrasound in Obstetrics and Gynecology
- Published
- 2024-09-01
- DOI
- https://doi.org/10.1002/uog.27766
- Primary Topic
- Medical Imaging and Analysis
- Type
- article
- Field-Weighted Citation Impact
- 0.00