Evaluating a problem-based learning model integrated with 3D anatomy software and software-assisted annotation in undergraduate spinal surgery education: a randomized controlled trial

Abstract

Background

Traditional lecture-based learning (LBL) faces limitations in teaching complex spinal anatomy and surgical procedures. This study aimed to evaluate the efficacy of a novel Problem-Based Learning (PBL) model integrated with three-dimensional (3D) anatomy software and software-assisted annotation in spinal surgery education.

Methods

A randomized controlled trial included 120 fifth-year clinical medicine undergraduates, starting in August 2024. Participants were divided into an experimental group (n = 60, receiving 3D + PBL + annotation-assisted teaching) and a control group (n = 60, receiving traditional LBL). Outcomes were assessed via written tests (objective/subjective questions) and standardized questionnaires evaluating knowledge mastery, learning motivation, academic atmosphere, teacher-student interaction, and knowledge retention.

Results

The experimental group scored significantly higher on subjective questions (case analysis) than the control group (39.33 ± 5.38 vs. 32.08 ± 4.79, P < 0.001). Questionnaire results indicated that the experimental group reported significantly higher self-rated mastery in spinal endoscopic procedures, anatomy, Michigan State University (MSU) classification of lumbar disc herniation (LDH), and differential diagnosis (all P < 0.05). In addition, students in the experimental group expressed greater satisfaction with learning motivation, academic atmosphere, teacher–student interaction, and knowledge retention (all P < 0.05).

Conclusions

Integrating 3D anatomy visualization, software-assisted annotation, and PBL significantly enhances clinical reasoning, spatial understanding, and student engagement in spinal surgery education. This multimodal approach addresses the limitations of traditional methods and is recommended for broader application in orthopedic training.

Trial registration

Chinese Clinical Trial Registry (ChiCTR), ChiCTR2400082568. Registered on 01 April 2024.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12909-026-08652-7.

Introduction

Medical education is widely acknowledged as one of the most demanding disciplines, confronting inherent challenges stemming from rigorous academic requirements and professional standards [1]. The rapid accumulation of clinical knowledge has heightened focus on identifying pedagogical approaches to shorten the transition from medical student to physician, thus accelerating resident development [2]. Despite ongoing educational reforms in China over the past decade, Traditional lecture-based learning (LBL) remains dominant due to resource imbalances, faculty shortages, and constraints from conventional teaching philosophies [3].

PBL employs a student-centered approach featuring guided inquiry, self-directed learning, and interactive discussion. Enhanced motivation is observed in PBL students, who transition from passive knowledge recipients to active participants. Compared with LBL, PBL provides structured time for utilizing textbooks/digital resources to research solutions, exercise critical thinking, and conduct small-group discussions. This approach develops capacities for critical analysis and problem-solving [4–6]. PBL demonstrates higher student satisfaction and superior efficacy in developing communication skills, problem-solving abilities, and self-directed learning compared to traditional methods [7].

Conventional anatomy instruction relies primarily on two-dimensional media (e.g., textbooks, slides) supplemented by cadaver dissection. While pedagogically fundamental, this approach suffers from deficient spatial awareness, inadequate 3D contextualization, and limited clinical applicability. Integration of 3D digital technologies (e.g., dynamic anatomical visualization) has significantly enhanced structural comprehension. Student evaluations indicate these technologies markedly improve spatial cognition and surgical approach understanding, effectively bridging anatomical theory with clinical practice [8–10].

The application of QuPath software for spinal endoscopic annotation has transformed medical pedagogy. This computational tool develops physicians’ ability to integrate clinical reasoning with digital proficiency. Instructors can directly annotate structures and analyze images using intraoperative videos, facilitating the transition from anatomical models to clinical decision-making. This approach demonstrates stronger alignment with learner preferences and needs [11].

Consequently, this study investigates a PBL model integrating 3D anatomy with software-assisted annotation in spine surgery instruction, specifically focusing on lumbar disc herniation teaching. Results are presented below.

Materials and methods

General information

A cohort of 120 spine surgery interns who commenced training at our institution in August 2024 was enrolled. Participants were randomly allocated via random number table assignment into experimental (n = 60) and control (n = 60) groups. The experimental group comprised 31 males and 29 females aged 21–23 years (mean 22.5 ± 0.7 years). The control group included 30 males and 30 females aged 21–24 years (mean 22.5 ± 0.8 years). All participants were clinical interns with no statistically significant intergroup differences in age or sex distribution (P > 0.05).

​Inclusion criteria:​​

1. Full-time clinical undergraduates in their fourth academic year;

2. Capacity to articulate authentic feedback;.

3. Comprehension of study protocols with voluntary participation confirmed by written informed consent..

​Exclusion criteria:​​

1. Failure to meet any inclusion criterion;

2. Voluntary withdrawal due to personal circumstances;

3. Prior PBL teaching experience.

Method

Prepare 3D models and anatomical structure annotation pictures under spinal endoscopy

The Complete Anatomy 3D software was employed to generate 3D anatomical models, integrating 3D visualization technology with clinical anatomy education. This tool encompasses seven major systems, including skeletal, muscular, and nervous components, and is equipped with a dynamic annotation system, layer-by-layer dissection features, 360-degree rotation capabilities, and cross-sectional 3D reconstruction. Tissue staining techniques are used to highlight specific local structures, enabling students to thoroughly understand the anatomical relationships in spinal pathologies from multiple perspectives. Specifically, a lumbar spine model is constructed from bone tissue, intervertebral discs, and spinal nerves, clearly illustrating the correspondence between vertebral bodies, discs, and nerve roots at each segment. This assists instructors in effectively demonstrating compression patterns caused by various disc herniations, with emphasis on extraforaminal herniations and their impact on affected nerve roots, as well as clarifying the distinctions between exit roots and traversing roots. Additionally, cross-sectional models support student learning of lumbar MRI and CT imaging, while animated videos illustrate the progression of common conditions such as lumbar spinal stenosis and herniated disc disease, enhancing comprehension of disease evolution (Fig. 1). Appropriate permissions have been obtained for the use of images from the Complete Anatomy software for the purpose of this scientific publication.

Fig. 1.

Software-generated lumbar spine visualizations​. a Left 45° oblique view: Purple annotation indicates L5 ventral nerve root (left); L4-L5 vertebrae rendered transparent. b Left sagittal view: Purple annotation denotes L4 dorsal root ganglion (left); L4-L5 vertebrae transparent. c Posterior view: Green annotation highlights L4/L5 annulus fibrosus; L4-L5 vertebrae transparent. d Axial plane at L5 level: High-resolution anatomical plate. e Dynamic 3D visualization: Lumbar spinal stenosis (LSS) model. f Sequential animation: Lumbar disc herniation progression

The spine endoscopic anatomical structures were annotated using QuPath software (Open-source bioimage analysis software), with a total of 200 annotated images prepared for student instruction. These visual materials highlight common tissues encountered in spinal endoscopic surgery, including the ligamentum flavum, nerve roots, intervertebral discs, and osseous structures. Each tissue type is distinctly color-coded and labeled with its anatomical designation, enabling learners to effectively practice distinguishing and identifying these critical structures (Fig. 2). ​

Fig. 2.

QuPath-generated endoscopic annotation. Illustrative case demonstrating ligamentum flavum (yellow perimeter), nerve roots (red outline), and intervertebral disc (white boundary)

Research methods

The core instructional curriculum encompasses five principal modules: clinical manifestations and surgical indications for lumbar disc herniation; clinical presentations and classification systems pertaining to lumbar spondylolisthesis; the Michigan State University (MSU) classification framework with zonal distribution analysis for lumbar disc herniation; characteristic imaging features of lumbar disc herniation across radiography, computed tomography (CT), and magnetic resonance imaging (MRI); and critical differentiation between lumbar disc herniation and degenerative lumbar spinal stenosis.

​The experimental group implemented an integrated PBL teaching methodology augmented by 3D anatomical modeling and digital annotation tools. This approach encompassed five structured phases: 1) Pre-session preparation involved distributing representative clinical cases (lumbar spinal stenosis, disc herniation, and vertebral compression fractures) and spine endoscopy videos one week prior, with instructions to download the Complete Anatomy software to review 3D anatomical models;2) Session initiation commenced 10 min before instruction, delineating PBL workflows and facilitating self-organized groups of 8–10 students. Each group designated a leader to coordinate autonomous case analysis, collaborative discussion, knowledge documentation, and synthesis of systematic findings; 3) Instructor-guided exploration utilized Complete Anatomy to demonstrate spinal structures through interactive 3D manipulation (zooming, rotation, tissue opacity manipulation), enabling students to investigate pathological progression and surgical principles. Screenshots of the surgical video were annotated in QuPath with color-coded structural demarcations (e.g., yellow for ligamentum flavum, red for nerve roots, white for intervertebral discs) to enhance spatial recognition of anatomical relationships; 4) Facilitated discussion featured moderated group presentations addressing predefined clinical questions, followed by cross-group interrogation and peer-assisted clarification, with instructor documentation of conceptual understanding; 5) Conclusion entailed faculty synthesis of key insights, resolution of contentious concepts, and strategic guidance for methodological competence development, facilitating student acclimatization to the PBL pedagogical framework.

​The control cohort received conventional educator-centered pedagogy. Prior to each session, participants were required to review designated educational materials. During lectures, instructors systematically disseminated knowledge using multimodal teaching tools, including blackboard diagrams, multimedia presentations, video demonstrations, and anatomical specimen models. Instruction strictly followed standardized textbook sequences, with educators emphasizing fundamental concepts and clinically challenging topics through syllabus-aligned supplementary explanations. Concluding each session, the instructor synthesized core knowledge points and guided students in consolidating understanding through memorization and conceptual reinforcement.

Comparison of teaching effects

Theoretical examination

The summative assessment employed a closed-book format comprising two equally weighted sections, each accounting for 50% of the total 100-point score. Examination content was rigorously derived from didactic materials, with objective items (primarily multiple-choice questions) curated from historical medical practitioner databases to evaluate factual knowledge acquisition. Subjective components consisted of clinical case analyses authored by orthopedics faculty and independently graded by non-participating instructors to ensure impartiality. This integrated testing framework simultaneously assessed discrete knowledge retention through standardized items while evaluating systematic disease comprehension and analytical capabilities via scenario-based applications.

Questionnaire survey

At the conclusion of the instructional period, a structured nine-item questionnaire was administered across two domains (Tables 2 and 3), with responses scored on a 10-point Likert scale (1 = lowest satisfaction, 10 = highest satisfaction). Table 2 specifically evaluated students’ perceived efficacy of the PBL teaching model combined with 3D anatomy and software-assisted annotation in mastering complex spinal surgery concepts, while Table 3 comparatively assessed satisfaction levels with both teaching methodologies. These instruments were designed to quantify educational outcomes beyond standard examinations, with higher aggregate scores indicating stronger endorsement of each pedagogical approach.

Table 2.

Self-assessment of medical knowledge by cohorts

Items	Experiment	Control	T	p
Basic steps of spinal endoscopic surgery	7.32 ± 1.12	5.56 ± 1.78	6.43	< 0.001
Spinal anatomy	6.88 ± 1.32	4.72 ± 2.23	6.44	< 0.001
MSU classification of LDH	7.89 ± 1.49	7.14 ± 1.48	2.73	0.007
Clinical manifestations of the involved nerve roots	6.93 ± 1.89	5.19 ± 1.88	5.07	< 0.001
Differential diagnosis of LDH	7.25 ± 1.56	5.23 ± 1.69	6.80	< 0.001

The self-assessment scores are presented as mean (± s). Statistical significance between the experimental and control groups was assessed using a t-test. Items were rated on a 10-point scale

Table 3.

Comparison of evaluations and teaching satisfaction between groups

Items	Experiment	Control	T	p
Learning motivation	7.45 ± 1.16	5.62 ± 1.57	7.25	< 0.001
Academic atmosphere	7.08 ± 1.21	5.73 ± 1.52	5.42	< 0.001
Teacher-student interaction	7.22 ± 1.28	5.30 ± 1.28	8.22	< 0.001
Knowledge absorption score	8.21 ± 1.43	7.30 ± 1.72	3.15	0.002

Learning motivation, academic atmosphere, teacher-student interaction, and knowledge absorption scores are reported as mean (± s). Significant differences were determined using a t-test

 A P-value of less than 0.05 was considered significant

The structured nine-item questionnaire was used to assess students’ mastery of spinal surgery concepts and satisfaction with the integrated 3D-PBL-annotation pedagogy. This questionnaire was developed specifically for this study to measure perceived mastery (Domain A) and learning experience/satisfaction (Domain B). The English version of the questionnaire is provided as Supplementary File S1.

Statistical methods

All datasets underwent comprehensive statistical analysis using SPSS version 25. Examination scores are presented as mean ± standard deviation (SD). Quantitative variables were analyzed employing one-way ANOVA, while categorical data underwent Chi-square testing with Bonferroni correction applied for multiple comparisons. Statistical significance was established at P < 0.05, revealing significant inter-group differences.

Result

Examination results

Statistical analyses revealed no significant difference in objective test scores (multiple-choice items) between the experimental and control groups (P > 0.05). Conversely, the experimental group demonstrated significantly higher performance on subjective assessments (case analyses) compared to controls (P < 0.01) (Table 1).

Table 1.

Comparison of objective and subjective scores between groups

Group	Experiment	Control	T	p
Objective topic	37.58 ± 5.89	39.61 ± 5.67	1.92	> 0.05
Subjective topic	39.33 ± 5.38	32.08 ± 4.79	7.79	< 0.001

The objective topic and subjective topic scores are reported as mean (± s) for both groups. A t-test was used for statistical comparisons

A P-value of less than 0.05 was considered statistically significant

Results of the questionnaire survey

Post-intervention distribution of 120 anonymous questionnaires yielded a 100% response rate with complete data integrity, revealing statistically significant intergroup differences (P < 0.05) in both knowledge acquisition and instructional satisfaction. The experimental group demonstrated superior performance across four key knowledge domains: (1) Basic steps of spinal endoscopic surgery (2), spinal anatomy (3), MSU classification of LDH (4), Clinical manifestations of the involved nerve roots (5), Differential diagnosis of LDH (Table 2). Correspondingly, satisfaction metrics showed significantly higher scores (P < 0.05) for the experimental cohort in learning motivation, interactive learning climate, faculty-student engagement, and knowledge assimilation (Table 3), confirming the integrated methodology’s efficacy in enhancing both cognitive competencies and educational experience.

Discussion

The convergence of technological advancement and interdisciplinary integration has amplified the multidimensional complexity of clinical practice since the 21st century. To enhance medical students’ clinical competency and achieve innovation-driven excellence in medical education, traditional pedagogy—characterized by unidirectional knowledge transfer and isolated skill training—demonstrates significant limitations in cultivating dynamic problem-solving capabilities essential for clinical contexts. While conventional medical education exhibits efficacy in foundational knowledge dissemination through systematized delivery frameworks and modular curricula [12], contemporary spinal surgery instruction remains constrained by didactic “instructor-recipient” methodologies that prioritize rote memorization over critical engagement, thereby suppressing innovative cognition and perpetuating theory-practice disparities.

Spinal surgery education presents unique pedagogical challenges due to its inherently integrative nature, requiring synthesis of complex knowledge networks spanning deformity correction, infection control, trauma reconstruction, and oncologic intervention—all intricately interconnected with anatomical, pathological, radiological, and biomechanical principles [13, 14]. Traditional two-dimensional instructional materials fail to convey three-dimensional spatial relationships among vertebral components, neural structures, and vascular systems, creating substantial cognitive barriers to understanding pathological biomechanical alterations. Core learning obstacles include: (1) Novices’ difficulty correlating anatomical structures with clinical presentations, particularly in distinguishing intraspinal versus extraspinal pathologies; (2) Theory-practice disjunction exacerbated by dependence on planar atlases; and (3) Inability to contextualize pathophysiological changes within biomechanical frameworks. Educational reform must prioritize ​three-dimensional visualization​ and ​clinical reasoning integration​ to overcome these discipline-specific challenges [15].

Our study demonstrates that integrating 3D anatomy visualization, software-assisted annotation, and PBL effectively addresses limitations of traditional pedagogy by enabling intuitive observation of complex anatomical relationships and accelerating identification of critical landmarks, thereby enhancing instructional efficiency. This multimodal approach stimulates learner engagement, develops clinical application proficiency, and strengthens collaborative competencies in simulated environments. Quantitative outcomes confirm the experimental group’s superior performance in clinical case analyses (P < 0.05) and significantly higher satisfaction (P < 0.05) across knowledge mastery, content comprehension, learning motivation, and self-directed learning capacity—indicating substantial improvements in knowledge translation, clinical reasoning, and learning initiative. Given orthopedics’ unique educational demands, this 3D-PBL integration elevates theoretical performance, enhances complex concept assimilation, and improves interdisciplinary communication while yielding superior learner satisfaction [16]. Complementary evidence suggests incorporating 3D-printing with PBL and case-based learning may further augment knowledge acquisition and self-assessment in clinical education [17], prompting development of scalable pedagogical frameworks across medical disciplines to advance educational innovation through synergistic technology-pedagogy integration.

Our findings align with existing scholarship in affirming PBL’s efficacy in enhancing learner engagement, clinical reasoning, and collaborative competencies [18, 19]. The integration of three-dimensional visualization technology with PBL demonstrates significant pedagogical advantages within spinal surgery education, evidenced by improved knowledge retention, heightened disciplinary interest, and strengthened clinical decision-making capabilities [14]. This integrated approach fosters comprehensive understanding of spinal pathologies, augments problem-solving skills, and cultivates interactive learning environments. Implementation of PBL in standardized residency programs enhances theoretical knowledge acquisition, diagnostic reasoning, and collaborative practice [20]. Crucially, medical students’ PBL experiences positively predict post-graduate development of interdependent competencies in social interaction and cognitive processing [21]. Substantial evidence confirms PBL’s superiority over traditional lecture-based pedagogy in developing self-directed learning strategies and problem-solving aptitude [22–25], with Zhang et al.‘s meta-analysis (n = 6,180 students) establishing PBL as among the most effective strategies for improving medical education quality and efficiency [26].Mitrousias et al. experimentally demonstrated that 3D software-assisted instruction yields significantly higher spatial cognition test scores than cadaveric dissection [27]. The 3D methodology not only enhanced comprehension of complex anatomical relationships but also improved three-dimensional spatial judgment accuracy – advantages notably absent from traditional cadaveric approaches despite their tactile authenticity. 3D models provide dynamic visualization of anatomical structures with superior clarity [28], while offering logistical flexibility and customizable case-based teaching unconstrained by temporal or spatial limitations. This eliminates ethical and preservation challenges inherent to pathological specimen utilization [29]. Distinctively, our model synthesizes PBL methodology with 3D anatomical software and endoscopic annotation technologies, creating a comprehensive pedagogical framework. This integrative approach surpasses standalone PBL or 3D visualization by delivering authentic anatomical representation through enriched multimodal resources and intuitive spatial awareness development. The synthesized pedagogy provides an optimized experiential platform for spinal surgery education that substantially enhances learning effectiveness beyond conventional methods.

The pedagogical innovation of this study lies in establishing an integrated 3D-endoscopic annotation system, where intraoperative spinal endoscopy frames are structurally annotated to achieve precise spatial mapping of 3D anatomical models onto clinical scenarios. This framework leverages high-resolution 3D reconstructions coupled with endoscopic visualization, enabling students to observe nuanced spinal structural dynamics and critical surgical sequences—enhancing both learning efficiency and comprehension of complex anatomical relationships. Given the elevated complication risks inherent in spinal procedures (including neurological deficits, infections, and substantial hemorrhage, particularly in neuromuscular scoliosis, spondylolisthesis, and adult spinal deformity cases [30]—which may precipitate chronic pain, functional impairment, and protracted rehabilitation), surgeons’ ability to visualize intricate anatomy through this system proves critical for navigating complex pathologies [31]. When addressing PBL scenarios, students utilize 3D anatomy platforms with endoscopic annotations to investigate the anatomical foundations underlying clinical problems, adopting a surgeon’s perspective while simultaneously mastering contemporary spinal instrumentation and advanced surgical techniques. This integrated approach demonstrates pedagogical superiority in developing clinical problem-solving competencies. Within spinal surgery education, the triad of 3D visualization, annotation-guided learning, and PBL methodology elevates students’ clinical competency while reinforcing PBL’s centrality in medical curricula. This multimodal pedagogy cultivates future physicians who possess not only comprehensive knowledge but also demonstrably enhanced communication, problem-solving, and self-directed learning capabilities—essential attributes for modern medical practice.

This study acknowledges limitations including (1) constrained sample size and demographic homogeneity (n = 120 from a single institution), potentially compromising external validity and generalizability; (2) methodological considerations such as possible confirmation bias in participant surveys due to awareness of the novel intervention, alongside unvalidated applicability beyond spinal surgery to other orthopedic subspecialties like arthroplasty or trauma surgery; and (3) exclusive focus on short-term outcomes without assessing long-term knowledge/skill retention. We recognize that the sample size of 120 participants is relatively small and restricted to a single institution, which may affect the external validity and generalizability of the results. Despite this, the study provides valuable insights into the effectiveness of integrating 3D anatomy visualization and PBL in spinal surgery education. Future research should consider expanding the sample size and including participants from multiple institutions to enhance the generalizability of the findings. Additionally, longitudinal studies could help assess the long-term retention of skills and knowledge acquired through this teaching model. Despite these constraints, the research offers meaningful guidance for optimizing orthopedic pedagogy. Future investigations should expand multi-institutional cohorts to enhance representativeness, validate the integrated 3D-annotation-PBL methodology across orthopedic subspecialties, and conduct longitudinal studies evaluating sustained clinical competency development—thereby advancing evidence-based implementation of innovative educational frameworks in surgical training.

Conclusion

This study suggests that integrating PBL with interactive 3D anatomy software and software-assisted endoscopic annotation may enhance spinal surgery education compared to traditional lectures. The integrated approach was associated with better performance in clinical case analysis, indicating stronger clinical reasoning. Students also reported greater mastery of complex spinal concepts and higher satisfaction with learning motivation, interaction, and knowledge retention. By bridging anatomical theory, spatial visualization, and clinical application, this multimodal PBL model presents a promising strategy for advancing orthopedic training, though further validation across larger and more diverse cohorts is required.

Abbreviations

LBL

Lecture–Based Learning

3D

Three–Dimensional

PBL

Problem–Based Learning

MSU

Michigan State University classification

LDH

Lumbar Disc Herniation

LSS

Lumbar Spinal Stenosis

CT

Computed Tomography

MRI

Magnetic Resonance Imaging

Authors’ contributions

Feng Yuan and Maji Sun conceived and designed the study. Wenbo Li, Ziyao Ding, and Shuo Feng contributed equally as co-first authors. Wenbo Li, Ziyao Ding, Wenkang Xu, and Haixu Qi were responsible for material preparation, data collection, and analysis. Qirui Zhu and Bingxu Xiao assisted in data interpretation and results presentation. Shaoyu Zhu contributed to statistical analysis. The initial draft of the manuscript was written by Wenbo Li and Qirui Zhu. Shuo Feng critically reviewed and revised the manuscript. All authors read and approved the final version of the manuscript. Feng Yuan and Maji Sun supervised the project and acquired funding.

Funding

This research was supported by the Paired Assistance Scientific Research Project of the Affiliated Hospital of Xuzhou Medical University (Grant No. FXJDBF2024204), the Construction Project of High-Level Hospital of Jiangsu Province (Grant No. GSPJS202515，GSPJS202516，GSPJS202506), the High-Level Research Project Cultivation Program of the Affiliated Hospital of Xuzhou Medical University (Grant No. PYJH2025033), and the Special Project on New Medical Technologies for the Construction of High-level Hospitals of the Affiliated Hospital of Xuzhou Medical University (Grant No. GSPJS202403), the Jiangsu Provincial Department of Science and Technology (Grant No. BE2022708), and the 2025 High-Level Scientific Research Project Cultivation Program of Xuzhou Medical University Affiliated Hospital Approved Projects (Grant No. PYJH2025104).

Data availability

The datasets used and/or analyzed in this study are available from the corresponding author upon reasonable request. We do not have ethical permission to upload the datasets to a repository. Please note that all research data have been anonymized for confidentiality purposes.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of the Affiliated Hospital of Xuzhou Medical University (XYFY2023-KL146-01). All participants had informed consent prior to the study’s investigation, the subjects were healthy adults, and the Helsinki Declaration was not violated in the study. It is to confirm that all methods were carried out in accordance with relevant guidelines and regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.