Training and assessment of neonatal and infant cranial standard plane in ultrasound image for residents based on virtual ultrasound models

Li Zhihui; Yu Wei; Zhou Luyao; Zhou Jingran; Ding Tingting; Tang Zijian

doi:10.3969/j.issn.2096-4293.2025.11.003

Abstract：Objective To evaluate the effectiveness of virtual ultrasound models in standardized residency training of neonatal and infant cranial ultrasound and to establish a standardized assessment scheme based on virtual ultrasound models.Methods Eighteen resident physicians from the Department of Ultrasound Medicine at Shenzhen Children's Hospital,who were enrolled in standardized residency training,participated in this study.All of them received the theoretical instruction on cranial ultrasound first,and used the virtual ultrasound model as teaching aids for cranial ultrasound teaching.Two evaluators with over 10 years of experience in ultrasound independently performed objective scoring.Each resident underwent assessments before and after training,with statistical analysis performed on the results.The self-evaluation of residents before and after training was recorded by questionnaires.Results There was strong consistency between the two tests,and the intra-group correlation coefficient(ICC)of the scores before and after training were 0.955 and 0.955,respectively.The scores for various elements in the post-training assessment improved significantly compared with pre-training.Additionally,the scan time for cranial ultrasound decreased significantly after training.The self-evaluation of residents after training was higher than that before training.The above differences were statistically significant(all P<0.05).Conclusion Training and assessment using a virtual cranial ultrasound model can effectively enhance resident physicians'skills in pediatric cranial ultrasound and improve their confidence in performing the procedure.The virtual ultrasound model can be used for standardized training and assessment of cranial ultrasound of residents,and has high application value.