Artificial intelligence-assisted focused cardiac ultrasound training: A survey among undergraduate medical students

Introduction

Point-of-care ultrasound (POCUS), and specifically Focused Cardiac Ultrasound (FoCUS), is becoming increasingly recognised and widely adopted as an important tool for physicians in many fields, such as cardiology, emergency medicine, and critical care medicine.^1–3 Although not designed to replace formal sonographic examinations, it is a rapid, cost-effective, and accurate diagnostic tool, which provides clinicians with real-time visualisation and quick assessment of critical cardiac conditions requiring urgent medical attention.^3–5 Due to the trend of rapidly growing use of echocardiography, and the benefits of using FoCUS in a safe and efficient way, adequate education and training are mandatory.

There are important differences between available training and educational programmes for point-of-care echocardiography, and many physicians in training are only exposed to ultrasound late in their residency.^3,6–8 Early exposure to the fundamentals of FoCUS could help medical students understand anatomy and physiology and could provide a theoretical and practical foundation that can be built on during later clinical years. However, educational data evaluating early FoCUS training for undergraduate medical students is lacking.

Recent applications of artificial intelligence (AI) and deep learning (DL) have become an emerging technology in the field in various domains of medical imaging.^9–11 Such technologies integrated in point-of-care ultrasound devices provide real-time feedback to the user, increasing the chance of obtaining diagnostic images while also improving image interpretation.^12,13 These tools serve as expert guidance to the novice user, potentially facilitating and accelerating its learning curve.

This study aimed to survey and report undergraduate medical students’ learning and confidence level in their ultrasound skills before and after an artificial intelligence (AI)-assisted interactive FoCUS teaching session.

Survey

Before the start and immediately at the end of the AI-assisted FoCUS workshop, students self-reported their respective pre- and post-session confidence levels in ultrasound skills (image acquisition and anatomical recognition) on a 5-point Likert-type scale, with 1 as the lowest and 5 as the highest level of confidence (Table 1). Participants also completed questions regarding their background knowledge in ultrasound (pre-course survey) and overall satisfaction with the programme (post-course survey) (Table 1). Students were also encouraged to provide written feedback. The surveys were administered to the students in printed-paper form and all responses were kept anonymous.

OPEN IN VIEWER

Table 1.

Pre- and post-FoCUS teaching survey questions.

Pre- and post-teaching assessment of confidence level for the different learning objectives	Not confident at al	Not confident	Neither	Confident	Very confident
Adjusting image gain and depth	1	2	3	4	5
Obtaining basic cardiac views	1	2	3	4	5
Recognising left and right ventricle	1	2	3	4	5
Recognising left and right atrium	1	2	3	4	5
Recognising great vessels (aorta and IVC)	1	2	3	4	5
Recognising valves (mitral, tricuspid, and aortic valve)	1	2	3	4	5
Pre-teaching assessment of background knowledge in ultrasound	Free text
Experience in ultrasound
Ultrasound is part of medical curriculum
Post-teaching assessment of programme satisfaction	Strongly disagree	Disagree	Neutral	Agree	Strongly agree
The course was effective in achieving my learning objectives	1	2	3	4	5
POCUS can be useful to my learning journey in medical school	1	2	3	4	5
I plan to incorporate the content I learned into learning	1	2	3	4	5
Post-teaching written feedback	Free text
What could be improved in future sessions

POCUS: Point of care ultrasound; IVC: inferior vena cava.

Results

A total of 61 undergraduate students from different medical schools attended the workshop. Most of them were medical students from schools across the United Kingdom, two were students from outside the United Kingdom (one from Amsterdam UMC and one from Erasmus University Rotterdam).

All participants responded to the pre-course survey; response rate of the post-course survey was 85% (n = 52). There were 29 (48%) students who reported previous experience in ultrasound; however, this experience was mostly limited. For 35 (57%) students, ultrasound training was not an integrated part of the medical curriculum. Significant improvement in skills confidence was achieved in every domain (Table 2 and Figure 1), with the greatest improvement in obtaining the basic cardiac views (mean score pre-teaching 1.95 ± 0.94 versus mean score post-teaching 4.12 ± 0.55; p < 0.001; Figure 2). The vast majority (98%) considered the course as strongly effective in achieving the learning objectives and all students (100%) agreed that POCUS can be useful and should be integrated into the medical curriculum.

OPEN IN VIEWER

Table 2.

Confidence level of undergraduates before and after AI-assisted FoCUS teaching.

	Pre-teaching (n = 61)	Post-teaching (n = 52)	Difference	P-value
Adjusting gain and depth of image	1.88 ± 1.07	3.75 ± 0.86	1.86	< 0.001
Obtaining basic cardiac views	1.95 ± 0.94	4.12 ± 0.55	2.16	< 0.001
Recognising ventricles	2.75 ± 1.26	4.15 ± 0.72	1.40	< 0.001
Recognising atria	2.67 ± 1.18	4.06 ± 0.72	1.38	< 0.001
Recognising great vessels	2.58 ± 1.09	3.69 ± 0.92	1.11	< 0.001
Recognising valves	2.49 ± 1.12	4.02 ± 0.78	1.53	< 0.001

Data are presented as mean ± standard deviation. Likert-type scale used to score the different questions: 1 = not confident et al; 2 = not confident; 3 = neither; 4 = confident; 5 = very confident.

OPEN IN VIEWER

Figure 1.

Mean survey response score pre- and post-FoCUS teaching.

OPEN IN VIEWER

Figure 2.

Pre- and post-teaching Likert-type scores for students’ confidence level in obtaining basic cardiac views.

Discussion

This study aimed to evaluate the usefulness of an AI-assisted interactive FoCUS teaching session to undergraduate medical students. We found that the workshop was highly effective, significantly improving students’ knowledge and confidence levels in every tested domain. Students strongly supported the idea of integrating ultrasound training into the medical curriculum and reported that more exposure to ultrasound would be helpful in achieving learning objectives.

Despite the rapidly growing use of point-of-care ultrasound, our survey demonstrates a still limited role of ultrasound training into the medical curriculum. For approximately 60% of students, ultrasound training was not integrated into the curriculum, and only half of the participants had any experience with ultrasound. Moreover, this experience was usually reported as very little and consisted mainly of observational exposure without practical training.

Although the workshop was only short, the benefit reported by all students was evident. The earlier students are exposed to POCUS, the more likely that these skills will become durable and remain in the long term. Therefore, an integrated point-of-care ultrasound curriculum and more systematic and early exposure of medical students to ultrasound holds promise to be very beneficial.

In this study, highly innovative AI-integrated point-of-care ultrasound devices were used to support medical students in their learning process. AI has the potential to accelerate the students’ learning curve, and to increase the effectiveness of POCUS. AI integrated in point-of-care devices can give the user immediate actionable feedback to improve the image, and to increasing the user’s confidence level in image interpretation. This was demonstrated in our study as all students reported a tremendous increase in confidence level for both obtaining and interpreting the cardiac ultrasound views. Therefore, with all this in mind, focused cardiac ultrasound training should ideally begin as early as possible in one’s career, and be aided by modern technology such as AI. AI-assisted ultrasound served as a valuable complement to expert faculty in teaching medical students by enhancing the learning process. It provided students with real-time feedback and guidance, helping them to acquire technical skills more efficiently. AI tools can highlight key anatomical features, suggest optimal imaging techniques, and even identify potential errors, allowing students to learn from their mistakes in a supportive environment. This technology also enabled faculty to focus on higher-level instruction and personalised mentoring, as AI can handle routine assessments and reinforce foundational concepts. Together, AI-assisted ultrasound and expert faculty create a more effective and comprehensive educational experience.

Several limitations to our survey data needs to be taken into account. Students who registered for and attended the workshop may have been those with a greater interest and aptitude for ultrasound, which may have introduced selection bias, possibly affecting the generalisability of the results. The data was handled entirely anonymously with no mechanism to statistically pair the changes in confidence scores. The second questionnaire was undertaken immediately after the teaching session and therefore represents short-term effects without information on the sustainability of the effects on confidence level over time. Finally, student confidence level was self-reported, and there was no objective way of assessing this improvement.

Despite its limitations, this small-scale study is to our knowledge one of the first to look at the footprint that AI could leave on POCUS training of undergraduate students. This is probably a good example versatile medical education must be in order to adapt to fast changing times.

Further studies could compare novice inexperienced AI users with experienced POCUS operators in terms of diagnostic efficiency in time-critical settings: in acute and emergency medicine, training a bigger number of operators to a basic level might prove to be a true life -saving innovation.

In conclusion, this study has demonstrated that AI-assisted interactive FoCUS training can be an effective and powerful tool to increase ultrasound skills and confidence levels of undergraduate medical students. Therefore, integration of such ultrasound courses into the medical undergraduate curriculum should be considered.