Innovative Learning in Anatomy Education: Assessing the Impact of Low-Cost 3D Deep Learning Anatomical Models in Museum-Based Instruction.

BACKGROUND: Anatomy education traditionally relies on cadaveric dissection and static models, which can be costly, ethically complex, and limited in interactivity, hindering students' ability to grasp clinical correlations and spatial relationships. This study evaluates the effectiveness of a virtual anatomy museum approach, leveraging low-cost, interactive 3D models and deep learning, to enhance clinical reasoning and anatomical understanding among medical undergraduate students.

METHODOLOGY: In this non-experimental, observational study, designed as a proof of concept, 40 undergraduate students from the Bachelor of Applied Medical Sciences program at King Khalid University are divided into Target (n = 20) and Control (n = 20) groups based on pre-assigned enrolment numbers. The Target Group engages with a web-based virtual anatomy museum platform featuring interactive 3D anatomical models (e.g., brain, spine, knee) in a museum-like environment over a 15-week semester, fostering exploration and clinical understanding. The platform, powered by a PointNeXt deep learning backbone and AnatoVision Block, provided interactive visualization, annotations, and quizzes. The platform provides interactive features, such as rotating models to examine structures like the cerebellum for neurological assessments or the femur for fracture diagnosis, along with clear labels and quizzes. The Control Group receives traditional lecture-based instruction supplemented with 3D-printed anatomical models, serving as a baseline to contextualize the Target Group's experience rather than a direct comparator, given the qualitative differences between interactive digital and lecture-based approaches with physical models. Post-intervention, students' clinical reasoning and anatomical proficiency are assessed using case-based questionnaires targeting 5 dimensions: acceptance, clinical problem-solving creativity, motivation, knowledge gain, and communication skills. Students also rate their experience via a 5-point Likert scale questionnaire, providing insights into the platform's potential to enhance anatomy education in resource-limited settings.

RESULTS: The Target Group scored significantly higher (mean 4.48 ± 0.66) than the Control Group (mean 3.45 ± 0.70) in overall performance (p < 0.01). Notable improvements were observed in anatomical understanding (95% vs. 65%, p = 0.01), clinical scenario interpretation (90% vs. 70%, p = 0.02), and motivation (85% vs. 50%, p = 0.05). Most Target Group students (92%, n = 18) agreed that the virtual 3D models enhanced clinical relevance, spatial orientation, and interdisciplinary communication, with 100% recommending its curriculum integration. Technical difficulties were reported by 25% of the Target Group (p = 0.49).

CONCLUSION: The virtual anatomy museum approach, utilizing low-cost, interactive 3D models, demonstrates strong potential to enhance clinical reasoning and anatomical proficiency among medical students at King Khalid University. The platform, explored as a proof of concept, fosters skills critical for clinical practice, such as identifying cranial sutures for neuroimaging or planning surgical corridors for hip replacements, offering a scalable, engaging supplement to medical education in resource-limited settings. The Target Group's experience, contextualized against a Control Group receiving lecture-based instruction with 3D-printed models, highlights the platform's ability to support hands-on learning.