Blog

TNBC tissue specimen

The study protocol and tissue specimen acquisition were approved by the Ethical Review Committee of Xiangya Hospital (2024121786). From 2024 to 2025, four pairs of samples from patients with histologically confirmed primary triple-negative breast cancer (TNBC) who underwent core needle biopsies at Xiangya Hospital (Hunan, China), received neoadjuvant chemotherapy with carboplatin and paclitaxel for the first time, and subsequently underwent surgery, without prior targeted or immunotherapy, were included in this study. Paraffin-embedded TNBC samples were collected for mass spectrometry analysis, with pre-treatment tissues obtained from core needle biopsies before NAT and post-treatment tissues from surgical resections. Ultimately, Eight TNBC tissue specimens were subjected to 4D-label free proteome analysis (Majorbio, China). To validate our findings, we further utilized paraffin-embedded primary triple-negative breast cancer (TNBC) surgical samples from 44 cases collected between 2023 and 2025. These samples were from patients who received neoadjuvant chemotherapy with carboplatin and paclitaxel for the first time, without prior targeted or immunotherapy, and underwent immunohistochemical analysis to investigate the relationship between THEM6 expression and prognosis. Tumor response to treatment was evaluated using RECIST 1.1, categorizing patients into responder (S) and non-responder (NS) groups. Additionally, human TNBC tissue arrays (WZ-TNBC1201) were purchased from Shanghai Outdo Biotech Co., Ltd.

Mass spectrometry

Paraffin-embedded samples were deparaffinized by incubating them in xylene at 37 °C with constant shaking, followed by centrifugation at 14,000 g to remove residual paraffin. This process was repeated until no wax remained. The samples were then sequentially rehydrated with ethanol and ultrapure water. Following rehydration, lysis was performed using 5% SDT buffer (SDS, 100 mM Tris-HCl, pH 8.5) supplemented with a protease inhibitor cocktail, followed by three rounds of homogenization. Ultrasonic fragmentation was carried out for 1 h, and the lysates were subsequently heated in a boiling water bath at 95 °C for 60 min. After centrifugation, the supernatant was collected for protein quantification using a BCA assay. SDS-PAGE was performed to assess protein integrity. Protein digestion was conducted overnight in TEAB buffer (100 mM) using trypsin, following reduction with TCEP and alkylation with IAM. The resulting peptides were desalted using HLB cartridges, vacuum-dried, and quantified. Finally, data-independent acquisition (DIA) mass spectrometry was performed on a timsTOF Pro2 instrument in DIA-PASEF mode, acquiring MS data across an m/z range of 400–1200 with 64 isolation windows for comprehensive proteomic analysis. Proteins with significant differential expression were identified based on criteria of P < 0.05 and an absolute fold change > 2. The association between THEM6 expression and overall survival (OS) in TNBC patients was analyzed using the KM-Plotter online tool (http://kmplot.com) based on the GSE96058 dataset [13].

Immunohistochemical staining

The tissue sections were dehydrated, subjected to citrate antigen retrieval, and blocked with 5% goat serum for 15 min at room temperature. They were then incubated overnight at 4 °C with a diluted primary anti-THEM6 antibody. Following this, the sections were treated with DAB, counterstained with hematoxylin, and dehydrated using ethanol. Images were captured and analyzed using ImageScope software (Leica Microsystems). The histological score was calculated as Total score = Proportion score × Intensity score, and samples were classified as high or low expression based on a median score of 4.

Cell lines and reagents

The human triple-negative breast cancer cell lines MDA-MB-231 and BT-549 were cultured in Dulbecco’s Modified Eagle Medium (DMEM) and RPMI 1640, respectively, each supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin. Cultures were maintained at 37 °C in a 5% CO2 atmosphere. Paclitaxel (HY-B0015) and Carboplatin (HY-17393) were obtained from MedChemExpress (MCE). The ferroptosis inhibitors ferrostatin (Fer-1, HY-100579) and liproxstatin-1 (Lip-1, HY-12726) were also purchased from MCE. Additionally, the pan-caspase inhibitor Z-VAD-FMK (HY-16658B), necrosis inhibitor necrostatin-1 (Nec-1, HY-15760), and autophagy inhibitor 3-methyladenine (3-MA, HY-19312) were sourced from MCE. Antibodies used in the study were acquired from the following suppliers: PGRMC1 (Zen-bio, 122868), THEM6 (Bioss, bs-15296R), FDFT1 (Proteintech, 13128-1-AP), GPX4 (Huabio, ET1706-45), SLC7A11 (Huabio, HA600098), ACSL4 (Huabio, ET7111-43), V5-tag antibody (Proteintech, 14440-1-AP), HA-tag antibody (Proteintech, 51064-2-AP), Flag-tag antibody (Proteintech, 20543-1-AP) and β-actin (Proteintech, 66009-1-Ig). lentiviral vectors including THEM6, PGRMC1 and the Negative Control, were purchased from GENERAL BIOL. Lentiviral particles containing shRNA targeting human FDFT1 (Locus ID 2222) were purchased from Origene.

Lentiviral transduction

Lentiviruses were packaged in HEK293T cells by co-transfection of the expression vector with packaging plasmids psPAX2 and pMD2.G using Lipofectamine 3000 (Invitrogen), following the manufacturer’s protocol. Viral supernatants were collected at 48 h post-transfection, filtered through a 0.45 μm membrane. Target cells (MDA-MB-231 or BT-549) were seeded in six-well plates and infected with lentivirus in the presence of 8 µg/mL polybrene. After 48 h, cells were selected with 2 µg/mL puromycin for one week. Stable expression was confirmed by Western blot analysis.

Cell viability assay

MDA-MB-231 and BT-549 cells were plated in 96-well plates at a density of 5 × 10^4 cells/ml. Following cell adherence, the cultures were treated with specified concentrations of carboplatin or paclitaxel for 48 h. Cell viability was evaluated using the Cell Counting Kit-8 (CCK-8) assay (MCE, HY-K0301), according to the manufacturer’s protocol. Absorbance at 450 nm was quantified using a microplate reader (PerkinElmer). The half-maximal inhibitory concentration (IC50) values were determined from dose-response curves generated using GraphPad Prism software.

Colony formation assay

Breast cancer cells were seeded into six-well plates and allowed to adhere overnight. They were then treated with 0, 0.5, or 1 µM carboplatin for 48 h. Following drug treatment, the medium was removed, and the cells were washed with PBS, trypsinized, and replated at a density of 2000 cells per well in new six-well plates. The cells were incubated for 14 days to allow colony formation. Colonies were subsequently stained with crystal violet, and the number of colonies in each condition was quantified.

Live/dead viability assay

Cells were seeded into 6-well plates at a density of 5 × 10^4 cells per well. After 48 h, cell viability was assessed using the Calcein AM/PI Live-Dead Cell Staining Kit I (APExBIO, K2247). Briefly, cells were incubated with 1 µM propidium iodide (PI) and 1 µM Calcein AM for 30 min at room temperature. Following staining, the cells were washed twice with PBS. Fluorescence microscopy was performed using an Axio Observer 3 fluorescence microscope (Carl Zeiss Microimaging) to visualize and differentiate live (green) and dead (red) cells.

Quantitative real-time PCR

Total RNA was isolated using TRIzol reagent (Invitrogen). cDNA was synthesized with SuperScript™ II reverse transcriptase (Invitrogen), and quantitative PCR was performed using Power SYBR Green PCR Master Mix (Takara). The primers used for the SYBR Green assays were as follows: THEM6-F:5′-GCAGCACTGGATCTCCTACAACG-3′; THEM-R: 5′- GGTCCTTGGTGACATCACTGAGC-3′; FDFT1-F:5′-GCAACGCAGTGTGCATATTTT-3′; FDFT1-R: 5′-CGCCAGTCTGGTTGGTAAAGG-3′; β-actin-F: 5′-CACCATTGGCAATGAGCGGTTC-3′; and β-actin-R: 5′- AGGTCTTTGCGGATGTCCACGT − 3′. Real-time amplification was carried out using an ABI Prism 7000 SDS (Applied Biosystems). Gene expression levels were quantified using the 2^−ΔΔCT method, with normalization to β-actin as the reference gene.

Western blot and ubiquitination assays

Cells were lysed using RIPA lysis buffer (MCE, HY-K1001) supplemented with a protease inhibitor cocktail (MCE, HY-K0010). The protein concentration in the cell lysate was determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific, 23225), following the manufacturer’s protocol. Samples were then denatured with SDS-PAGE protein loading buffer (Beyotime, D0071) at 100 °C for 5 min. Proteins were separated by SDS-PAGE and transferred to PVDF membranes (Millipore, ISEQ00010). After overnight incubation with the primary antibody, the membranes were probed with the secondary antibody at 37 °C for 1 h. Protein bands were visualized using Pierce™ ECL Western Blotting Substrate (Thermo Scientific, 32106) and captured using Image Lab software version 5.0 (Bio-Rad). For Ubiquitination assays, cells were lysed with 100 µL of NETN buffer (62.5 mM Tris-HCl pH 6.8, 2% SDS, 10% glycerol, 20 mM NEM, 1 mM iodoacetamide), boiled for 15 min, and then diluted 10-fold in NETN containing protease inhibitors, NEM, and iodoacetamide. After centrifugation, the supernatant was subjected to immunoprecipitation and analyzed by Western blotting [14].

Cycloheximide (CHX) Chase assay

To evaluate protein stability, MDA-MB-231 and BT-549 cells stably expressing either THEM6 or a negative control were seeded in six-well plates and cultured to approximately 70% confluence. CHX was added to the culture medium at a final concentration of 0.1 mg/mL. Cells were harvested at designated time points (0, 2, 4, 6, and 8 h) following CHX treatment. At each time point, cells were washed with ice-cold PBS and lysed in RIPA buffer supplemented with protease inhibitors. Lysates were clarified by centrifugation at 14,000 g for 15 min at 4 °C, and protein concentrations were quantified using a BCA assay. Equal amounts of protein were subjected to SDS-PAGE, followed by Western blot analysis to assess protein degradation kinetics.

ROS measurement

Intracellular ROS levels were measured using the fluorescent probe 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) (Sigma, 35845). This probe diffuses into cells and is enzymatically converted to the fluorescent 2′,7′-dichlorofluorescein (DCF). Triple-negative breast cancer (TNBC) cells were plated at a density of 6 × 10^5 cells per well in 6-well plates and incubated for 24 h. The cells were then exposed to 5 µM carboplatin for 48 h. After treatment, the cells were incubated with 10 µM DCFH-DA solution at 37 °C for 30 min, followed by three washes with phosphate-buffered saline (PBS). DCF fluorescence was visualized using an Axio Observer fluorescence microscope (Carl Zeiss Microimaging) with excitation at 485 nm and emission at 535 nm.

Ferrous iron measurement

Intracellular ferrous iron (Fe²⁺) content was quantified using the Iron Assay Kit (Abcam, ab83366). Briefly, samples were incubated at 25 °C for 30 min, followed by an additional 60-minute incubation with the iron probe at 25 °C to form a stable colored complex. Standard curve and reaction solutions were prepared according to the manufacturer’s instructions. The samples were then transferred to a microplate reader (PerkinElmer), and Fe²⁺ levels were determined by measuring absorbance at 539 nm.

Malondialdehyde measurement

Malondialdehyde (MDA) levels were measured in the lysates using the Lipid Peroxidation Assay Kit (Beyotime, S0131) according to the kit’s instructions. Briefly, 0.1 mL sample was mixed with 0.2 mL of the MDA detection working solution and incubated for 15 min at 100 °C. The samples were then allowed to cool to room temperature and centrifuged at 1000 × g for 10 min to collect the supernatant. Next, 200 µL of the supernatant was transferred to a 96-well plate, and absorbance was measured at 532 nm using a microplate reader (PerkinElmer). MDA levels were expressed as the ratio of the absorbance value to that of the control group.

Transmission electron microscope assay

Briefly, cells were seeded in 6-well plates at a density of 5 × 10^4 cells per well and treated with or without 5 µM carboplatin for 48 h. After treatment, cells were collected, washed with PBS, and fixed with 2.5% glutaraldehyde. The samples were then processed according to standard procedures, and images were obtained using a transmission electron microscope (Hitachi). The proportion of mitochondria with increased bilayer membrane formation was quantified using TEM images. To determine the size of mitochondria, TEM images were analyzed using ImageJ software to measure the mitochondrial area, and the relative mitochondrial size was provided.

Xenograft mouse model

The experiments were conducted with approval from the Animal Care and Use Committee of Central South University (China) (XY20240903005). Control or THEM6-overexpressing MDA-MB-231 cells (2 × 10^6) were orthotopically implanted into the mammary fat pads of 20 BALB/c nude mice (4 weeks old). Once the tumors reached approximately 50 mm³ in size, the mice were randomly assigned to four groups, with 5 mice per group: a carboplatin treatment group, in which carboplatin was administered via intraperitoneal (IP) injections at a dose of 50 mg/kg once weekly for three cycles, and a control group. Tumor volumes were measured every 3 days using calipers from the start of treatment and calculated using the formula (length × width2) × 1/2. Mice were sacrificed for tumor dissection on day 35 post-treatment initiation. H₂O₂ levels were measured using an H₂O₂ assay kit (Beyotime). Tumor tissue was homogenized, and absorbance at 560 nm was measured using a microplate reader (PerkinElmer).

Statistical analysis

Overall survival (OS) was analyzed using Kaplan–Meier curves and compared with the log-rank test. Data are presented as mean ± SD (n ≥ 3) and were analyzed using GraphPad Prism 6 software (GraphPad Software). Differences between groups with continuous data were evaluated using the Student’s t-test. p-values less than 0.05 were considered statistically significant.

The history of neurosurgery in Iraq is deeply rooted in ancient Mesopotamia, where early forms of surgical intervention, such as abscess drainage, were documented on cuneiform tablets​. The scientific structure of neurosurgery in Iraq took shape in the 1950s, with the first elective neurosurgical procedure being carried out by Dr. Najeeb Al-Yaaqubi and was further formalized in 1966 by Dr. Saad Al-Witry, considered the father of Iraqi neurosurgery. A milestone was passed when 1972 the Neurosurgery Teaching Hospital was inaugurated in Baghdad; this became a central focal point for neurosurgical training and practice [6]. The history of neurosurgery development took another course in the Kurdistan region of Iraq when international collaborations aimed at building local neurosurgical capacities in cities like Duhok started in 2012 [7].

The research findings are useful in understanding the role of the neurosurgical virtual laboratory at Neuroscience Hospital, Baghdad (Figs. 3, 4 and 5). The neuroscience hospital at Baghdad uses a stepwise, competency-based educational model in the neurosurgical lab. This was mainly developed through simulation-based training on clinically relevant scenarios. Iterative practice further reinforces learning, whereby immediate feedback from mentors during the process provides participants with opportunities to refine techniques and decision-making. It provided a very excellent avenue of professional growth between mentors and mentees. The mentors then took the mentees through iterative learning cycles of providing personalized feedback and clinical insights that could help fine-tune surgical precision and decision-making. Teaching in the neurosurgical lab was led by experienced neurosurgeons and senior residents with specialized expertise in simulation-based training.

Advanced Neurosurgical Training Equipment at Neuroscience Hospital, Baghdad. (A) High-resolution microscopes used for neurosurgical simulations, enabling residents to practice precision techniques in a controlled environment. Anatomical posters and reference books support theoretical learning alongside practical skills. (B) Endoscopic training setup featuring a head model and monitor, simulating real-time surgical procedures for enhancing skills in minimally invasive neurosurgery

Endoscopic and Microsurgical Simulation in Neurosurgical Training at Neuroscience Hospital, Baghdad. (A) Hands-on endoscopic training using a skull model, allowing residents to practice minimally invasive neurosurgical techniques. (B) Endoscopic view displaying the needed anatomy during a simulated procedure, enhancing visual-spatial understanding. (C) Microsurgical practice with high magnification, providing precision skill training in surgical anatomy and dissection. (D) Live demonstration of microsurgical techniques projected on a screen, used for educational purposes in ongoing resident workshops and seminars

A considerable proportion of participants (70%) reported weekly lab use, indicating that the facility is integral to their training and practice. Moreover, overall satisfaction with the lab experience was high, with 60% rating their satisfaction as 4 and 40% rating it as 5. This suggests that the lab is meeting the expectations of its users across various levels of experience. Notably, the chi-square test revealed a statistically significant association between satisfaction and the frequency of lab use (p = 0.010), highlighting that those who use the lab more frequently are more likely to be satisfied with their experience. Interestingly, satisfaction was not significantly influenced by gender (p = 0.519), indicating that both male and female participants had comparable satisfaction levels. However, years of neurosurgical experience significantly affected satisfaction (p = 0.009), with participants having more than 15 years of experience showing the highest levels of satisfaction. The gender distribution was skewed toward male participants, reflecting the broader gender imbalance currently observed in the neurosurgical field in our region.

The results demonstrate the high perceived educational value of various lab resources. The microscope and training models were consistently rated highly, with 40% of participants giving a maximum score of 5 for both tools. The impact of the 3D printers is rated lower: 20% gave a score of 5, and 40% gave a score of 4. This suggests variability in either utilization or perception of this resource. For medical textbooks, the rating was incredibly positive for 40% of the respondents.

In the article about the surgical skills in neurosurgical residency training, Liu et al. [8] focused on the especially important role of neurosurgical residency training in the context of surgical skills laboratories. They sought to enhance residents’ technical skills in complex skull base operations through a cadaveric-based, structured dissection curriculum and modern equipment at Cleveland Clinic. Furthermore, three-dimensional printing has evolved to be a robust neurosurgical education and anatomy training tool. Thiong’o et al. (2021) [9] describe the role that 3D printing plays in neurosurgical simulation, including skull base surgery and vascular procedures, to practice complex surgical skills outside the operating theater. It has proved beneficial in decreasing the learning curve for difficult procedures. Also, Baskaran et al. (2016) [10] point out that 3D printing precision in generating anatomical models from patient-specific data has dual benefits including improved surgical training and preoperative planning. Realistic simulation of neurosurgical tasks can be developed using additive manufacturing processes such that skill acquisition is improved, and patient outcomes are positive. Innovative problem-solving using 3D printing received a wider range of responses, topping at 30%, rating it 2. This would suggest that while 3D printing is recognized as important, it is not yet integral to every participant’s training or practice, probably because of its recent introduction or unfamiliarity with the technology.

Research has often been central to the goals that decide career choices and build surgical skills in medical students. Awad et al. (2016) [11] note that this trend is notably reflected in the number of medical students who receive research grants, of which more than 50% go on to pursue a residency in neurosurgery. Additional support for this view is inferred from the fact that 40% of the participants rated the lab’s role in research skills as 4, and 30% rated it as 5. Likewise, leadership and decision-making skills promoted by the non-medical resources of the lab are also rated to be four by 50% of the participants and rated to be five by 20%. This shows that the role of the laboratory goes beyond technical skills to include research competencies regarding professional development. 60% of all the participants highly rated the lab’s contribution to fostering innovation. The entire neurosurgical residency is six years in length, during which time the critical emphasis in each year has been tailored, almost in a pillar-like fashion. Lab training is organized in such a way that it complements this pillar development and progressively increases in intensity from foundational-level lab skills to more advanced surgical-based decision-making as the years progress.

The integration of neurosurgery into the curricula of medical schools remains an important underdeveloped feature in the medical education system worldwide. Lee et al. (2020) [12] study indicated that the level of neurosurgical exposure varied grossly across different regions, with only 39.7% of students reporting any form of neurosurgical experience during their education. The idea is that regular use of the lab enhances participants’ ability to innovate in their practice. Kato et al. (2020) [1] review global disparities in neurosurgical education between developed and developing countries. Despite advances in surgical techniques or diagnostic tools, many developing regions still face huge barriers, including access to limited resources, training, and modern technology. They also advocate for international collaboration to close these gaps and support programs. Kanmounye et al. (2020) [13] discuss how the role of the Foundation for International Education in Neurological Surgery has transformed to decrease global neurosurgical disparity through education. Until recently, FIENS, founded in 1969, focused on brief mission trips but, since then, has transformed into a more sustainable model through the education of local neurosurgeons and the establishment of residency programs in LMICs. This model, labeled “service through education,” has enhanced the development of neurosurgical systems in LMICs and has led to a sustainable effect due to local ownership and international cooperation.

Limitations

This study has several limitations. The sample size is relatively small, limiting the generalizability of findings to a broader population of neurosurgeons and trainees. Additionally, the study relies on self-reported data, which may introduce response bias. The cross-sectional design does not allow for the assessment of the long-term impacts of the neurosurgical lab on clinical outcomes. The survey instrument was developed specifically for this study and has not been previously validated, which may influence the reliability and interpretability of the results. Moreover, all participants were active users of the lab, which may introduce selection bias and lead to an overestimation of satisfaction and perceived benefit. Furthermore, while the study highlights the effectiveness of simulation-based training, it does not compare outcomes with traditional training methods. Future research should incorporate larger cohorts, objective skill assessments, and longitudinal follow-up to validate these findings.

This study was conducted in three phases, in accordance with the Design and Development Research (DDR) framework [16]. Phase 1 (not described in this article) involved a need analysis via literature review to identify gaps in students’ learning interest, understanding, and engagement of the occupational health subject as well as the need of a novel programme. Phase 2 involved the design and development of a metaverse programme known as MATOSH. In Phase 3, the MATOSH programme was implemented among medical students to empirically evaluate its effectiveness in improving their interest, understanding, and engagement in the occupational health subject. The study flowchart is illustrated in Fig. 1 below.

Phase 2 (Design and Development)

Phase 2 involved two components, which were the design and development of the MATOSH programme. This phase was conducted over a two-month period from June 2024 to July 2024.

Study design, study location, and study population

The content design of the MATOSH programme adopted a nominal group technique (NGT). Third-year medical students from the Faculty of Medicine at the National University of Malaysia served as the study population. This cohort was selected because the occupational health subject was introduced to them for the first time during the public health posting. Meanwhile, the development of MATOSH involved both academicians and third-year medical students from the same university, all of whom participated exclusively through online meetings.

Sample size

Fifteen third-year medical students and an occupational health lecturer from the Faculty of Medicine at the National University of Malaysia (UKM) were invited to the NGT to identify the top five important occupational health topics to be included in the MATOSH programme. The development of MATOSH involved three academicians (i.e., one occupational health and two computer science lecturers) and three third-year medical students.

Sampling method

The 15 medical students in NGT were recruited through purposive sampling. The sampling frame, consisting of a list of medical students, was obtained from the posting coordinator for the public health posting. Students (the sampling units) were selected based on the following inclusion criteria: (i) completion of the public health posting; (ii) attendance of at least 75% of the occupational health lectures (i.e., three out of four lectures); and (iii) access to a metaverse platform, regardless of prior experience with its use. However, students were excluded if they: (i) had recently failed the public health posting; or (ii) declined to provide informed consent for participation in the study. An experienced occupational health lecturer who currently teaching the subject at the same university was also invited to the NGT. For the development of MATOSH, the developer team (consisted of three academicians and three medical students) were recruited through purposive sampling method from the researchers’ network.

Study instruments

To determine and include the top five important occupational health topics for the content design of the MATOSH programme, an interview schedule was developed. It consisted of a central question of “what topics should be included in the MATOSH programme to effectively teach occupational health to medical students”. Meanwhile, to develop the MATOSH programme, a password secured laptop (Microsoft Surface Go with 16GB RAM) was utilised. Besides that, online software such as Spatial.io and Sketfab.com were used to design the virtual environment and virtual avatars, respectively. Both of these software were used under free account plans.

Study procedure

For the NGT (for MATOSH programme content design), 15 third-year medical students who met the inclusion criteria and an occupational health lecturer were invited to a meeting room at the faculty. As an introduction, the participants were brief regarding the aim of the NGT, which was to determine the top five important occupational health topics to be included in the MATOSH programme. Initially, participants engaged in silent idea generation, where they independently wrote down as many relevant occupational health topics as they could within 10 min. This was followed by a round-robin sharing session, where each participant took turns presenting one idea at a time. All ideas were recorded verbatim by the facilitator on a shared screen. This continued until all participants had contributed all their ideas. No discussion or evaluation occurred during this stage to ensure equal contribution from all participants. In the clarification and consolidation phase, the participants reviewed the compiled list of occupational health topics. Any overlapping or similar ideas were grouped or rephrased with the consensus of the participants. Participants could ask questions for clarification, but detailed debates were discouraged to maintain focus and efficiency. Next, during the voting and ranking phase, each participant was asked to provide scores for each occupational health topic based on importance (“1” = least important until “5” = most important). Rankings were submitted anonymously using Google form. In the final step, results were tallied, and the five topics with the highest percentages were identified and displayed. A group discussion followed to confirm consensus and ensure all participants were satisfied with the outcome. This process resulted in a ranked list of the top five important occupational health topics, which served as the foundation for content design of the MATOSH programme.

Once the top five important occupational health topics were identified, the development of a metaverse programme began by following a structured seven-step process [17]. The first step involved defining the metaverse concept by aligning content with the existing occupational health subject, which included four lectures on occupational hazards, HIRARC, SOCSO roles, relevant legislation, and universal precautions in hospital. Content development was guided by consultations with one occupational health lecturer and two computer science lecturers. The second step established the technological requirements, which identify the necessary hardware and software components. MATOSH was developed using Spatial.io [18], a free web-based platform that supported metaverse creation without the need for advanced devices such as VR or AR goggles, and without requiring personal cloud storage, as all data resided on the platform itself. The third step focused on designing virtual entities and environments. The virtual space was modelled as a hospital using Spatial.io templates, while virtual three dimensional hospital staff were generated from Sketchfab.com [19]. The fourth step integrated social interaction dynamics to enhance student engagement. As such, students were able to customise avatars, navigate through the virtual hospital using “W””, “A”, “S”, “D” buttons, interact with other users, and dicuss occupational health scenarios virtually. The fifth step addressed the development of a virtual economy, in which MATOSH was decided as a non-profit educational tool with no commercial or NFT elements. The sixth step involved testing and optimisation through internal evaluation and feedback from three medical students, which allowed for improvements in usability and content. The final step was the launch and publication of MATOSH on Spatial.io, with access restricted to educational and research use, and no online promotional efforts.

Statistical analysis

In the NGT, the total scores for each occupational health topic were calculated and converted to percentages (i.e., total score of each topic divided by 75 and then convert to 100%), and the five occupational health topics with the highest percentages were selected. Descriptive statistics, including frequency counts and total scores, were used to summarise the ranking data. No inferential statistical tests were applied, as the purpose of the NGT was to reach group consensus rather than to test hypotheses.

Phase 3 (Implementation and Evaluation)

Phase 3 consisted of two components, namely the implementation and evaluation of the effectiveness of MATOSH in improving medical students’ interest, understanding, and engagement in the occupational health subject. This phase was conducted over a 9-month period, from August 2024 to April 2025.

Study design, study location, and study population

The implementation component of this phase of the study involved a cognitive debriefing to validate the MATOSH programme, while the evaluation component of the MATOSH programme employed a quasi-experimental design. Third-year medical students from the Faculty of Medicine at the National University of Malaysia served as the study population. This cohort was selected because they were introduced to the occupational health subject for the first time during the public health posting.

Sample size Estimation

Given the skill engagement score for students with conventional learning = 3.78 ± 0.55 points [20], skill engagement score for students with e-learning = 3.56 ± 0.81 points [20], significance level = 0.05, power of study = 80%, the required sample size was 208 students (i.e., 104 students for intervention and 104 students for control group). This was calculated using Pocock’s formula for two means. Although 208 students were recruited, 36 did not return their responses, resulted only 88 and 84 students in the intervention and control groups, respectively.

Sampling method

In Phase 3, medical students were recruited through purposive sampling. The sampling frame, which was the list of medical students, was obtained from the posting coordinator for the public health posting. Students (the sampling units) were selected based on the following inclusion criteria: (i) enrolment in the public health posting at the time of recruitment; (ii) attendance of at least 75% of the occupational health lectures (i.e., three out of four lectures); and (iii) access to a metaverse platform, regardless of prior experience with its use. Exclusion criteria included: (i) students who were repeating the public health posting at the time of recruitment; and (ii) students who declined to provide informed consent for participation in the study.

Study instruments

In the implementation component of Phase 3, a semi-structured interview schedule was developed for the cognitive debriefing. This interview schedule included six questions pertaining to the content clarity, content relevance, visual appearance, usability, education value, technical functionality.

Meanwhile, in the evaluation component of Phase 3, three tools were utilised to evaluate the effectiveness of the MATOSH: the Study Interest Questionnaire (SIQ), Occupational Health End of Module (OH-EOM) test paper, and the Student Course Engagement Questionnaire (SCEQ). These instruments were employed to evaluate students’ interest, understanding, and engagement in the occupational health subject, respectively.

The Study Interest Questionnaire (SIQ) is an 18-item, unidimensional questionnaire developed by Schiefele et al. in 1987 [21]. It was intended to measure students’ interest in a specific field of study, based on the Educational Interest Theory. The 18 items in the SIQ were ranked on a Likert scale from 1 (“not at all true”) to 4 (“completely true”), with scores ranging from 18 to 72 points. Higher total scores were associated with more interest in the student’s field of study. Items 1, 2, 5, 7, 9, 10, and 11 were reverse coded (see Appendix C). Cronbach’s alpha was 0.90 and test-retest reliability was 0.67 over a 2-year period. Convergent, discriminant, and concurrent validity of the SIQ was demonstrated by correlations between SIQ and intrinsic motivation (r = 0.46, p < 0.001), extraversion (r = 0.01, p > 0.05), and use of deep learning strategies (r = 0.45, p < 0.001). Students’ exam performance over 2 years correlated with SIQ scores as evidence of predictive validity (r = 0.33, p < 0.05).

The Occupational Health End of Module (OH-EOM) test paper was created by the lecturers responsible for teaching the occupational health subject in the public health posting. This test was initially developed to assess students’ understanding of the occupational health subject at its conclusion. It comprised 10 multiple-choice questions (with options A to D), and students were required to select the best answer. Each correct response was awarded one mark, resulting in a total score range of 0 to 10. Higher marks indicated a higher level of understanding towards the occupational health subject. This test paper was vetted by two occupational health lecturers at the Department of Public Health Medicine, Faculty of Medicine, UKM.

The Student Course Engagement Questionnaire (SCEQ) was developed by Handelsman in 2005 to assess engagement among university students [22]. It consisted of 23 items with four domains, including skills engagement (9 items), participation/interaction engagement (6 items), emotional engagement (5 items), and performance engagement (3 items) (Handelsman et al. 2005). Participants responded using a 5-point Likert scale (i.e., 1 = “much less like me” to 5 = “much more like me”) (see Appendix E). The total scores on the SCEQ ranged from 23 to 115. Although there was no specific cut-off point in this questionnaire, higher total scores indicated a higher level of engagement. The SCEQ demonstrated good internal consistency across the four engagement domains (Cronbach’s alpha ranged from 0.76 to 0.82). It also demonstrated good convergent validity, in which the SCEQ scores were significantly associated with absolute and relative engagements (β = 0.16–0.38).

Study procedure

For the implementation component of Phase 3, 15 third-year medical students who met the inclusion criteria (similar to those described in Phase 2) were invited to a lecture hall for cognitive debriefing. They were first requested to independently explore the MATOSH programme using their personal devices (e.g., smartphones, tablets or laptops). Following this, each student participated in a cognitive debriefing interview. The feedback obtained from these interviews informed essential modifications to enhance the design, functionality, and usability of the metaverse. All cognitive debriefing interviews were audio-recorded to serve as a reference for the research team.

For the evaluation component of Phase 3, medical students who were assigned to the control group (n = 104) received only conventional lectures. Meanwhile, the medical students allocated to the intervention group (n = 104) received both conventional lectures and the MATOSH programme.

For the control group, four occupational health lectures were conducted, with one lecture per week in a lecture hall. The four lectures covered the following topics: types of occupational hazards, HIRARC, the roles of SOCSO, occupational safety and health legislations, and universal precautions in hospital. At the conclusion of the posting, the researchers distributed the study instruments (SIQ, OH-EOM test paper, and SCEQ) for in-class assessment. Prior to participation, students were briefed on the research project and asked to complete an online informed consent form. Those who declined participation were instructed to remain seated quietly in the lecture hall. Medical students who consented to participate received three Google Forms, namely the SIQ, OH-EOM, and SCEQ. They were instructed to complete all tasks within 20 min without engaging in peer discussion.

Medical students in the intervention group also attended four weekly occupational health lectures, covering the same topics as the control group: types of occupational hazards, HIRARC, the roles of SOCSO, occupational safety and health legislations, and universal precautions in hospital. Additionally, the intervention group had an extra session dedicated to introducing them to the MATOSH metaverse programme. During this fifth session, students received instructions on how to navigate the MATOSH platform and were given one hour to explore its content using their own mobile devices (e.g., smartphones, tablets or laptop). At the end of the posting, the researcher distributed the study instruments (SIQ, OH-EOM test paper, and SCEQ) for in-class assessment. Students were briefed on the research project and asked to complete an online informed consent form before participating. Those who chose not to participate were instructed to remain quietly in the lecture hall. Students who consented to participate received three Google Forms, namely the SIQ, OH-EOM, and SCEQ, and were asked to complete all tasks within 20 min, without engaging in peer discussion.

Statistical analysis

Qualitative data from the implementation component (cognitive debriefing) was analysed thematically using inductive coding. Meanwhile, in the evaluation component, descriptive analysis was conducted to outline the sociodemographic characteristics of the medical students who participated in the study. Continuous data were reported as mean and standard deviation if data were normally distributed, or as median and interquartile range if data were not normally distributed. Categorical data were presented as frequencies and percentages.

Since all data was normally distributed (ascertained using Kolmogorov-Smirnov test), a paired t-test was conducted to assess any significant difference between pre- and post-intervention SIQ, OH-EOM, and SCEQ scores for both intervention and control groups (i.e., within group comparison). Meanwhile, an independent t-test was performed to assess the significant differences in the SIQ, OH-EOM, and SCEQ scores between the intervention and control groups (i.e., between groups comparison). A significance level of p < 0.05 was set. All statistical analyses were conducted using SPSS version 28.

Ethical considerations

The current study received ethical approval from the Ethics Committee of the National University of Malaysia (JEP-2025-219). Online information booklets outlining the study’s objectives and procedures were provided to the medical students via Google Forms. Online informed consent was obtained from the medical students prior to their participation. Medical students were not penalised if they refused to participate in the study. They were also taught with the same level of commitment and the same amount of academic content, regardless of their group allocation. The investigators adhered to the principles outlined in the Declaration of Helsinki as well as the Malaysian Good Clinical Practice Guidelines.

No personal identifiers were captured in the Google Form. Only authorised research team members had access to the research data, and their access was monitored and restricted based on necessity. The collected data were securely stored in a password-protected, encrypted online database hosted on a secured server. Study data were retained for a period of five years following the publication of the study results. At the end of the five-year retention period, all study data were permanently deleted or destroyed.