Exploring Full-cycle DeepSeek-assisted Case-based Learning in Undergraduate Radiology Education: A Respiratory System...

Area of Science:

• Medical Education Technology
• Artificial Intelligence in Radiology
• Case-Based Learning

Background:

• Traditional case-based learning (CBL) in respiratory radiology requires significant preparation time.
• Evaluating the full instructional cycle of CBL is crucial for optimizing medical education.

Purpose of the Study:

• To assess the effectiveness of DeepSeek-assisted case-based learning (CBL) in respiratory radiology education.
• To evaluate DeepSeek's role across preparation, implementation, and evaluation phases.

Main Methods:

• A prospective study involving 200 third-year medical undergraduates in 2025.
• Comparison of DeepSeek-generated cases versus Hospital Information System (HIS)-retrieved cases for preparation time.
• Assessment of DeepSeek-assisted group versus traditional CBL group during implementation and evaluation via test scores and questionnaires.

Main Results:

• DeepSeek-generated cases significantly reduced preparation time (p = 0.016).
• The DeepSeek-assisted group showed greater test score improvements (p < 0.05) and higher self-directed learning, interest, and efficiency.
• AI-generated personalized feedback was deemed educationally useful by the teaching department.

Conclusions:

• DeepSeek-assisted CBL is effective throughout the entire respiratory radiology instructional cycle.
• This AI approach enhances efficiency, student engagement, and learning outcomes.
• DeepSeek-R1 offers a valuable tool for personalized feedback and improved medical education.