| Unique ID issued by UMIN | UMIN000057102 |
|---|---|
| Receipt number | R000065238 |
| Scientific Title | Impact of Real-Time AI Guidance on Image Acquisition Time in Resident Physicians Performing Point-of-Care Ultrasound: A Randomized Crossover Trial |
| Date of disclosure of the study information | 2025/02/21 |
| Last modified on | 2026/08/23 14:39:33 |
The impact of real-rime AI guidance on Point-of-Care Ultrasound examinations performed by residents
The impact of real-rime AI guidance on Point-of-Care Ultrasound examinations performed by residents
Impact of Real-Time AI Guidance on Image Acquisition Time in Resident Physicians Performing Point-of-Care Ultrasound: A Randomized Crossover Trial
Impact of Real-Time AI Guidance on Image Acquisition Time in Resident Physicians Performing Point-of-Care Ultrasound: A Randomized Crossover Trial
| Japan |
Healthy adult
| Cardiology | Emergency medicine | Adult |
Others
NO
To evaluate the impact of real-time AI assistance on image acquisition time and educational effectiveness in resident physicians performing point-of-care ultrasound (POCUS) examinations.
Efficacy
Confirmatory
Pragmatic
Not applicable
Time to image acquisition for five standard echocardiographic views recommended by the POCUS guideline in Japan: PLAX, PSAX-PM, AP4, SubC4, and SubC-IVC.
Time to image acquisition for all ten standard echocardiographic views, including the five views used for the primary outcome plus the following five views: PSAX-AV, PSAX-MV, AP5, AP3 and AP2.
The quality of acquired images.
Interventional
Cross-over
Randomized
Individual
Open -but assessor(s) are blinded
No treatment
NO
NO
Institution is not considered as adjustment factor.
YES
Central registration
2
Educational,Counseling,Training
| Maneuver |
Prior to the initiation of this study, all participating resident physicians will be administered a brief lecture by an emergency medicine specialist.
Resident physicians will perform echocardiographic examinations with AI assistance.
Prior to the initiation of the study, all participating resident physicians will be administered a brief lecture by an emergency medicine specialist.
Resident physicians will perform echocardiographic examinations without any AI assistance.
| 24 | years-old | <= |
| Not applicable |
Male and Female
Medical Qualification: a junior resident
A beginner in ultrasound with 1-2 years of experience
Have basic cardiac ultrasound knowledge
Prior formal training in echocardiography
20
| 1st name | Takashi |
| Middle name | |
| Last name | Shinato |
Sugita Genpaku Memorial Obama Municipal Hospital
Emergency and general medical department
917-8567
2-2 oote-cho, obama-shi, Fukui-ken 917-8567 Japan
0770-52-0990
doctor.s.u.t@icloud.com
| 1st name | Takashi |
| Middle name | |
| Last name | Shinato |
Sugita Genpaku Memorial Obama Municipal Hospital
Emergency and general medical department
917-8567
2-2 oote-cho, obama-shi, Fukui-ken 917-8567 Japan
0770-52-0990
doctor.s.u.t@icloud.com
Fukui prefectural hospital
none
Self funding
Ethics Committee of the Faculty of Fukui prefecural hospital
2-8-1 yotui, fukui-shi, Fukui-ken, Japan
0776-54-5151
doctor.s.u.t@icloud.com
NO
| 2025 | Year | 02 | Month | 21 | Day |
UMIN000057102
Unpublished
https://data.mendeley.com/preview/t7kn5hv44h?a=4038d8f9-8f0b-4c8a-9a91-7320a4190098
21
Total acquisition time for the five basic views did not differ significantly between the AI-assisted and unassisted conditions (192.9 vs. 125.0 seconds; mean difference [unassisted minus AI-assisted], -67.9 seconds; 95% CI, -138.7 to 2.9; P = 0.058). No significant differences were observed in acquisition time or success rate for any individual basic view.
| 2026 | Year | 08 | Month | 23 | Day |
Two residents were excluded from the analyses because their ultrasound clips were not saved according to the protocol, leaving 19 participants for analysis. Of these, 16 (84%) were male, 11 were postgraduate year (PGY) 1 residents, and eight were PGY-2 residents. None had previous experience with AI-assisted ultrasonography. Ten participants were assigned to the AI-assisted-first sequence and nine to the unassisted-first sequence; all received the allocated interventions.
A total of 21 residents were enrolled and assigned to either the AI-assisted-first sequence (n = 10) or the unassisted-first sequence (n = 11). Two participants in the unassisted-first sequence were excluded from the analyses because their ultrasound images were not saved according to the protocol. Thus, 19 participants were included in the final analyses: 10 in the AI-assisted-first sequence and nine in the unassisted-first sequence.
None.
The primary outcomes were the total acquisition time for five basic FoCUS views (SUM5: PLAX, PSAX-PM, A4C, SubC4, and SubC-IVC) and the acquisition success rate for each view. The secondary outcomes were the total acquisition time (SUM10) and acquisition success rates for 10 views, comprising the five basic views plus PSAX-AV, PSAX-MV, A5C, A3C, and A2C. Self-reported confidence in image acquisition was also assessed using pre- and post-trial questionnaires.
Completed
| 2025 | Year | 02 | Month | 23 | Day |
| 2025 | Year | 02 | Month | 01 | Day |
| 2025 | Year | 02 | Month | 23 | Day |
| 2025 | Year | 03 | Month | 31 | Day |
| 2025 | Year | 02 | Month | 21 | Day |
| 2026 | Year | 08 | Month | 23 | Day |
Value
https://center6.umin.ac.jp/cgi-open-bin/ctr_e/ctr_view.cgi?recptno=R000065238