Validation of single dynamic 3D structured light in precise pedicle puncture navigation

Abstract

Background

Spinal surgery navigation is a current research hotspot, with no reports on dynamic structured light navigation. This study explores a single 3D structured light system, utilizing one structured light camera to continuously emit light during surgery for dynamic guidance in pedicle screw placement.To investigate the methodology and principles governing minimally invasive puncture navigation utilizing single dynamic three-dimensional (3D) structured light, and evaluate its feasibility and accuracy.

Methods

Nineteen bovine lumbar spine bone experimental models were used. Performing percutaneous minimally invasive pedicle screw placement. Preoperative computed tomography (CT) scans were performed, 3D reconstructions were generated, and a screw pathway was designed. Intraoperative registration was then conducted, and the navigation channel was adjusted to align with the planned screw pathway. Kirschner wires were inserted, and their positions were graded using the Gertzbein–Robbins (G–R) scale. The alignment of the CT images and actual morphology was assessed. The entry and exit point 3D coordinates were collected before and after surgery to calculate offset values and angles.

Results

A total of 218 Kirschner wires were inserted, all achieving a grade of level A on the G–R scale, for a puncture accuracy of 100%. The intraoperative root mean squared was 0.442 ± 0.091 mm, and the point cloud-to-surface distance was 0.01 ± 0.007 mm. The preoperative and postoperative entry point offset was 1.044 ± 0.35 mm, respectively, and 1.439 ± 0.524 mm at the exit point. The screw path deviation was 1.12 ± 0.576º. A high correlation was observed in the distribution of entry and exit points both preoperatively and postoperatively. Trajectory deviations primarily occurred during wire insertion.

Conclusion

The single 3D structured light system enabled simultaneous real-time intraoperative scene scanning, static registration, and dynamic instrument tracking, facilitating high-precision navigation for minimally invasive puncture procedures.

Background

Spinal surgeries often require the reconstruction of spinal stability, inserting screws into the spine through the vertebral pedicle is the most commonly used technique [1]; however, improper placement can lead to severe complications [2]. Near-infrared light navigation technology has been increasingly applied in screw placement [3]. However, these screw placement systems typically comprise multiple components, with complex navigation processes, cumbersome coordinate transformations, inconvenient operation, and limited room for improvement in mechanical accuracy [4].

The innovative registration technique, which uses marker-free 3-dimensional (3D) structured light, simplifies operations, enhances efficiency, and ensures precision and safety, effectively meeting the demands of navigation-guided punctures [5]. 3D structured light is an active optical 3D measurement technology. This technology is widely used in various fields because of its high accuracy and real-time capabilities. Its core principle is a specialized projector that projects known, specific light spots, stripes, or coded patterns (known as structured light) onto the surface of the object being measured. Due to differences in the 3D shapes of the object’s surface, these pre-defined patterns undergo distortion and deformation. One or more cameras capture the deformed patterns from different angles. The precise 3D contour information of the object’s surface is obtained using chips and algorithms to compute the differences between the original and deformed patterns and construct a 3D model [6–8].

This study introduces dynamic 3D structured light in lumbar vertebral pedicle percutaneous puncture navigation. The process involved scanning the surface with structured light to generate a point cloud, aligning it with preoperative computed tomography (CT) scans, and dynamically tracking instrument placement for screw insertion using structured light. The approach facilitates minimally invasive puncture navigation within a single coordinate system using a single structured light system. The study findings provide theoretical insight and also serve as a practical reference for advancing innovative computer navigation methodologies.

Materials and methods

Materials

Equipment and software

Red-Crowned Crane System V1.0 (RCCS V1.0, independently developed) navigation software was used. A Sizector S028800 3D structured light camera and MPSizectorS SDK V2.21 (MEGA PHASE, China) software were used. Imaging was performed using Mimics 15.0 (Materialise Interactive Medical Image Control System 15.0, Materialise, Belgium) and spiral CT (INGENUITY CORE 128 CT, Philips, Netherlands).

Precision machining instruments

The navigation system included an aluminum alloy hollow cylindrical drilling tube (25 mm x 6.05 mm x 200 mm), and adjustment instruments consisting of bilateral support pillars, a lateral adjustment magnetic suction base plate, a magnetic base, vertical rods, an adapter plate, and a universal joint. The puncture tube comprised an outer sleeve (6 mm x 4.05 mm x 300 mm) and an inner sleeve (4 mm x 2.05 mm x 300 mm) (Fig. 1).

Fig. 1.

Illustration of the precision machining instruments, including the design diagram (A) and physical diagram (B) of the adjusting instruments, as well as the design diagram (C) and physical diagram (D) of the puncture instruments

Materials

Bovine lumbar spine bone and pig skin are common products available in the market. Kirschner wires are commonly sold items by medical device companies, while the polymeric polyurethane bandage is sourced from Jingyi Medical Technology Co., Ltd. in Yangzhou, Jiangsu, China.

Methods

Preparation of bovine lumbar spine bone experimental model

The experimental protocol was developed in accordance with the ethical guidelines of the Helsinki Declaration and was approved by the Human Ethics Committee of Lecong Hospital, Shunde (2023) No. 86. The spinous and transverse processes of the bovine lumbar vertebrae 1–6 (L1–L6) were trimmed and contoured by wrapping them in cloth filled with sawdust (Figs. 2A-C). Processed vertebrae was then covered with pig skin, encased in a polymeric polyurethane bandage, and secured to a wooden base using φ1.5 mm Kirschner wires and screws (Fig. 2D).

Fig. 2.

Illustration of the process of creating the bovine lumbar spine bone experimental model. (A) Bovine lumbar spine bone. (B) Pig skin covering the bovine bone with trimmed transverse and spinous processes. (C) Contoured bovine bone. (D) Wrapping with a polyurethane polymer bandage

Preoperative design

CT scanning was performed on the experimental model. Preoperative CT scanning was performed, followed by 3D reconstruction of the skeleton and body surface using Mimics 15.0. The generated stereolithography (STL) files, which is a file format that describes the geometric surface shapes of three-dimensional objects, were then imported into the left window of the RCCS.

Segmentation of the vertebrae

The “Segment” button was clicked to initiate segmentation in RCCS. The lateral surfaces of each vertebral body were outlined by marking along the superior endplate of the lumbar vertebrae, the upper edge of the facet joints, and the inferior endplate of the lumbar vertebrae. Once the outlines were complete, the process continued by selecting “Segment In” and then “Confirm Segmentation” to sequentially segment the independent L1–L6 vertebrae.

Preoperative puncture channel design

Next “Primitive Factory” was clicked and “Cylinder” was selected, followed by “Translate/Rotate” to position the cylinder along the long axis of the vertebral pedicle. A puncture channel (φ2 mm) was designed for each vertebra from L1 to L6. Finally, all information related to the body surface, skeleton, individual vertebral bodies, and the preoperatively planned screw pathway was consolidated into a folder labeled “Preoperative Design Plan.”

Structure light scanning intraoperative scene and registration

Structured light debugging. At the initiation of structured light scanning, the placement of both the experimental model and the navigation drilling tube was confirmed to be within the structured light scanning range (Fig. 3A).

Fig. 3.

Positioning and fine registering. (A) Positioning of the experimental model and structured light camera. (B) Fine registration of the point cloud with the body surface. (C) Obtaining the root mean square (RMS) value. (D) Measurement of the distance between the point cloud and the body surface

The body surface point cloud for initial alignment and cropping was obtained, followed by denoising. “Start,” then “Stop,” and finally “Capture” were clicked in the right RCSS window to capture a frame of the current intraoperative scene’s point cloud, which was then transferred to the left window. Next, the “Align two clouds by picking (at least 4) equivalent point pairs” option was used to roughly align the point cloud with the body surface. Then, the point cloud was trimmed using the “Segment” tool and any noise points were removed with " Statistical Outlier Removal”.

Fine registration of the body surface point cloud with the CT body surface, and measurement of registration accuracy. Afterward, “Finely register already (roughly) aligned entities (clouds or meshes)” was selected to achieve a more precise alignment with the body surface and obtain the RMS value (Fig. 3B-C). Finally, “Compute cloud/mesh distance” was clicked to measure the distance between the point cloud and the body surface (Fig. 3D).

The preoperative design plan was moved to the original surface point cloud position for subsequent pedicle screw puncture navigation. After point cloud registration, the matrix was copied, “Apply Transformation” was clicked, and the matrix was pasted. The “Apply Inverse Transformation” option was chosen, and the Preoperative Design Plan folder was moved to align with the position of the intraoperative scene point cloud. This allowed the CT images to be positioned in the surgical space, enabling navigation based on the CT images displayed on the screen.

Intraoperative navigation adjustment for puncture

Cylindrical fitting, simultaneously obtaining the navigation pathway. In the left window, “Settings” was clicked to configure the navigation channel with a diameter of 2 mm, a length of 800 mm, and an instrument diameter of 25 mm. “Start” was then clicked, followed by “Navigation, “ to immediately display the navigation drilling tube point cloud and the generated cylinder in the right window. At the same time, the left window displayed the navigation channel fitted by the cylinder.

Each individual vertebra and its designed pathway were displayed. The options for CT surface and point cloud were unchecked, and the vertebra and pathway designed for the puncture were selected. This way, only the individual vertebra, the designed pathway, and the navigation pathway were shown in the left panel.

Intraoperative adjustment of the puncture direction was performed. First, the universal joint was used to adjust the direction, ensuring that the navigation pathway was aligned parallel to the designed pathway. Then, the position was fine-tuned using the longitudinal rod and magnetic suction base to align the navigation channel with the planned pathway.

A puncture was made along the navigation channel. Both the navigation channel and the vertebra were rotated to confirm satisfactory alignment between the navigation channel and the designed screw path, following the direction of the navigation drill sleeve at this moment. The nested puncture tubes and φ2.0 mm stainless steel Kirschner wires were manually inserted into the navigation drilling tube. A small skin incision was made, and the Kirschner wire was guided to the bone surface by inserting the inner tube up to the bone surface while advancing the outer tube slightly through the skin. At this point, the electric drill was connected to the Kirschner wire. The Kirschner wire was advanced until it exited the anterior edge of the vertebral body, and then the wire was trimmed (Fig. 4A-B).

Fig. 4.

Navigation-guided puncture. (A) Puncture tube and Kirschner wires. (B) Intraoperative puncture

Gertzbein–robbins scale grading and observation of CT images and actual morphology

Postoperative CT scans were performed on the experimental model containing the Kirschner wires. Then, the Kirschner wires were removed from the model. Three-dimensional modeling and the reconstruction of bone structures were performed using Mimics 15.0 software. The resulting STL files were then exported and imported into RCCS. The preoperative and postoperative bone images were initially coarsely registered, followed by fine registration using the aforementioned methods. The position of the Kirschner wires was assessed by reviewing CT images of the wires and the postoperative screw pathway. Then, they were graded using the Gertzbein–Robbins (G–R) scale [ 9]. Disassemble the model and extract the bovine bones, and φ1.5 mm stainless steel Kirschner wires were threaded through various entry and exit holes in the bones to observe the actual positional morphology of the Kirschner wires within the bones.

Experiment data collection and statistical analysis

The average intraoperative registration accuracy was calculated and is expressed as mean ± standard deviation (M ± SD). In the same coordinate system, the center point where each planned screw pathway intersected the preoperative bone at the entry and exit points on the bone surface was identified as the preoperative entry and exit points. After wire extraction, the center points of the entry and exit holes in the bone were considered to be the postoperative entry and exit points. The 3D output coordinates of each point were used to calculate the deviations between the preoperative and postoperative entry and exit points, as well as angular deviations in the screw pathway.

All statistical analyses were performed using SPSS 26.0 software (IBM, USA) and Origin 2018 (OriginLab, USA). A correlation analysis was conducted on the axis coordinate values (X, Y, and Z) of the preoperative and postoperative entry and exit points. A Pearson correlation coefficient (R²) of ≥ 0.8 was considered to indicate a strong correlation between two variables. Quantitative data following a normal distribution are expressed as mean ± standard deviation (M ± SD), and a t-test was used for comparisons. The results of data not following a normal distribution are presented as median (M) and interquartile range (Q), e.g., M (Q1, Q3), and the rank sum test was applied. The significance level was set at α = 0.05, with p-values>0.05 indicating no significant difference. The coordinate values of each axis (X, Y, and Z) were compared before and after surgery.

Results

G–R scale classification

A total of 218 Kirschner wires were inserted using puncture technique. All Kirschner wire positions were rated as grade A on the G–R scale, resulting in a puncture accuracy of 100% (Fig. 5A).

Fig. 5.

Kirschner wire, postoperative screw pathway, preoperatively planned screw pathway, lumbar spine, and postoperative morphology. (A) Kirschner wire with the vertebral horizontal plane. (B) Postoperative screw pathway with the vertebral horizontal plane. (C) Rear view of the preoperatively designed screw pathway and Kirschner wire (the white arrow indicates the preoperatively planned screw pathway, and the green arrow indicates the Kirschner wire). (D) Side view of the preoperatively planned screw pathway and the Kirschner wire. (E) Side view of the lumbar spine, the preoperatively planned screw pathway, and the Kirschner wire. (E1) Amplified side view of L3. (E2) Amplified rear view of L3. (F) Posterior view of bovine Kirschner wires. (G) Anterior view of bovine Kirschner wires

Observation of CT images and actual morphology

CT images and morphological observation indicated that the postoperative screw pathway and Kirschner wire positions at the vertebral pedicle were consistent with the preoperative design (Fig. 5B-G).

Intraoperative registration accuracy

The mean RMS value and the average distance between the point cloud and the body surface were 0.442 ± 0.091 mm and 0.01 ± 0.007 mm, respectively, indicating minimal registration errors and high precision (Table 1).

Table 1.

Image registration (RMS)and the average distances between the point cloud and the body surface

Subject	RMS	the average distances between the point cloud and the body(mm)
Model 1	0.52705	0.00903
Model 2	0.494479	0.00445
Model 3	0.382186	0.01450
Model 4	0.433367	0.01830
Model 5	0.676392	0.00055
Model 6	0.325441	0.00653
Model 7	0.488297	0.01791
Model 8	0.337745	0.01525
Model 9	0.475575	0.00787
Model 10	0.530171	0.01851
Model 11	0.395644	0.01146
Model 12	0.436667	0.01221
Model 13	0.442627	0.00508
Model 14	0.343609	0.00225
Model 15	0.369532	0.00793
Model 16	0.349672	0.01049
Model 17	0.395812	0.00386
Model 18	0.574068	0.00238
Model 19	0.421586	0.02479
M ± SD	0.442 ± 0.091	0.010 ± 0.007

Preoperative and postoperative wire entry and exit point deviations and screw pathway deviation angles

The mean entry point deviation was 1.044 ± 0.35 mm, the mean exit point deviation was 1.439 ± 0.524 mm, and the mean nail pathway deviation was 1.12 ± 0.576º. These values demonstrate minimal deviation and high precision.

Preoperative and postoperative correlation of entry and exit wire points

The R² value of the entry wire point on the X-axis was 0.99935, 0.99857 on the Y-axis, and 1 on the Z-axis. The R² value of the exit wire point on the X-axis was 0.99951, 0.99709 on the Y-axis, and 1 on the Z-axis (Fig. 6). The displayed distribution of entry and exit wire point positions demonstrated a high degree of correlation.

Fig. 6.

Correlation between preoperative and postoperative entry and exit wire points. (A) The R² value of the X-axis entry wire points. (B) The R² value of the Y-axis entry wire points. (C) The R² value of the Z-axis entry wire points. (D) The R² value of the X-axis exit wire points. (E) The R² value of the Y-axis exit wire points. (F) The R² value of the Z-axis exit wire points

The median difference in X-axis coordinates between preoperative and postoperative entry points was 0.07 (−0.585, 0.686), with no significant difference (Z = −1.393, p = 0.164). The average difference in Y-axis coordinates was 0.01 ± 0.362, also without a significant difference (t = 0.461, p = 0.646). The median difference in Z-axis coordinates was 0.098 (−0.473, 0.676), which was significantly different (Z = −2.027, p = 0.043). The median difference in X-axis coordinates between preoperative and postoperative exit points was 0.121 (−0.711, 0.754), with no significant difference (Z = −0.631, p = 0.528). The median Y-axis coordinate difference was 0.644 (−5.64, 1.19), which was significantly different (Z = −9.299, p < 0.001). The mean difference in Z-axis coordinates was significantly different at 0.138 ± 0.618 (Z = 3.291, p = 0.001). These findings suggest that trajectory deviation occurred during wire insertion.

Discussion

The precision of a navigation system’s final positioning is closely tied to the accuracy of the registration process. The structure light-based navigation method, known for its non-radiative nature, low learning curve, and markerless operation, incorporates a wide range of point clouds in the registration process to enhance precision. This aspect warrants further investigation. Therefore, the application of structured light in navigation registration, instrument tracking methods, procedures, and effects, particularly in minimally invasive navigational surgeries, warrants further investigation.

A study by Tu et al. reported that preoperative CT navigation significantly reduced both radiation exposure and screw preparation time compared to intraoperative 3D C-arm imaging [10]. However, It has been reported that O-arm navigation does not guarantee enhanced accuracy in screw placement [11]. Fatima et al. employed various robotic surgical techniques, which resulted in longer durations compared to the control group [12].

Most procedures utilizing near-infrared light navigation require preoperative tool registration and the placement of multiple trackers, thereby extending the surgical duration [13]. In the case of extensive segment surgeries, multiple image scans are required during the procedure, resulting in notable radiation exposure [14, 15]. The system’s accuracy can be substantially undermined by loose, misplaced, contaminated, or distantly placed trackers [16, 17].

He et al. employed structured light for bone scanning, producing thousands of registered point clouds that emphasized the benefits of cost-effectiveness, speed, non-radiation, and high precision [18]. The 7D Surgical System (FLASH System; Machine-vision Image Guided Surgery (MvIGS) System) uses structured light to scan the surface of the spine, capturing registered point clouds. By integrating near-infrared light tracking for screw placement instruments, FLASH was reported to achieve a screw placement accuracy of 98.7%, while MvIGS reached 94.2%, thereby improving the efficiency of intraoperative navigation workflows. Structured light navigation eliminates the need for tracker placement; this reduces additional trauma and its associated drawbacks [19–23]. Structured light registration has a short processing time, allowing for multiple rapid and repeated registrations without the need to rescan images. Each “FLASH Fix” averages only 16.54 s, enabling image drift correction in real-time [24]. It presents functionalities similar to dynamic machine vision. However, the 7D Surgical System cannot yet be applied to minimally invasive spinal surgery, as instrument tracking still relies on near-infrared light navigation technology [25]. With the future integration of dynamic sensors, video-vision transformation models, and sensor-based training, dynamic machine vision can be realized for real-time navigation in surgical procedures. If there can be an artificial intelligence-guided percutaneous puncture navigation robot that integrates 3D structured light and dynamic sensing localization technology, which can real-time track surface displacement and ensure the accuracy and stability of the puncture path during the procedure, it would represent a significant breakthrough in medical positioning and navigation. This study on dynamic structured light navigation aligns with the emergence and development of dynamic machine vision.

The results of this study demonstrate that the Kirschner wires consistently achieved G–R grade A, with an RMS value of 0.442 mm. The entry point exhibited a deviation of 1.044 mm, the exit point showed a deviation of 1.439 mm, and the angular deviation of the screw pathway was 1.12º. Vardiman et al. reported exit point deviations in robot-assisted lumbar pedicle screw placement ranging from 1.7 to 1.9 mm at the exit point, 1.8 to 2.3 mm at the entry point, and angular deviations between 2.0 and 2.8° [26]. Zhu et al. pioneered the application of single static structured light navigation. They reported an entry point error of 0.28 ± 0.16 mm and an angular error of 0.49 ± 0.24º. Previous studies on various types of navigation have achieved varying degrees of satisfaction. In comparison, the qualitative assessments in this study demonstrated excellent puncture efficacy, and the quantitative analyses showed high accuracy [27].

Image-guided navigation and preoperative surgical planning are important elements defining digital surgery [28]. The present study aligns with the principles of digital surgery and achieves high-precision punctures, primarily due to the following features. Dynamic 3D structured light has advantages that include a camera fixed atop a post 2 m high, with a spatial resolution of 3 megapixels. Its working distance ranges from 0.9 to 1.5 m, with a field of view of 800 mm x 605.8 mm at 1.5 m. Static structured light systems are limited to single-frame imaging and lower precision; however, the camera used in dynamic 3D structured light systems rapidly projects a series of structured light patterns. This process facilitates surgical location tracking and the real-time tracking of screw placement instruments (navigation drilling tube) by capturing and aligning point clouds with the body surface (static registration). The streamlined navigation process, simplified coordinate transformation chain, minimal cumulative errors, and enhanced precision collectively improve overall accuracy.

RCCS requires only partial scanning of the navigation drilling tube surface with structured light, allowing the algorithm to accurately fit the approximately 25 mm diameter drilling tube without effects from prolonged exposure or the partial obstruction of the light source. Single-step registration enables extensive, long-segment, multi-directional navigation-guided punctures. The inherently 3D nature of surgical design and navigation-guided punctures allows for the meticulous design of the screw pathway from the surface to the interior in a way that closely resembles human visual observation. Additionally, the left and right windows provide distinct viewing perspectives within the same coordinate system. The left window is dedicated to screw pathway design and navigation-guided punctures, while the right window facilitates the real-time monitoring of screw placement instrument tracking and cylinder fitting effects.

The instrument maintains a high degree of coaxiality between the navigation drilling template, the puncture cannula, and the Kirschner wire. The center of the navigation drilling tube functions as the working channel, facilitating minimally invasive punctures.

This study exclusively used model of skin-covered bovine bone. and lacked characteristics such as respiratory movement, bleeding, and anatomical variability. Thus, the models did not reproduce the complexity of the human anatomy or surgical conditions. This limitation should be addressed by using more realistic models, for example, cadaveric or clinical settings. Although 218 punctures were performed in this study, they were derived from only 19 models. Additionally, the study was performed under controlled laboratory conditions, with direct experimental comparisons of existing navigation systems and testing restricted to the lumbar puncture region (not involving the thoracic or cervical levels), limiting generalizability. This research is currently in the technological development and prototype validation phase. Future studies should address workflow integration, ergonomics, and compatibility with other operating room equipment to investigate practical surgical implications.

Before medical devices can transition from the laboratory to hospitals, they must move through several developmental stages, including regulatory approval, clinical validation, product implementation, market introduction, and application promotion.

Many manual operations are primarily intended to demonstrate the underlying principles, appearing somewhat cumbersome compared to the automated processes of established navigation systems. The development of RCCS modules, such as three-dimensional reconstruction, computer-assisted design, automatic registration, and real-time depth sensing, is ongoing. Robotic arms will be used for assisted puncture, and the associated software will be embedded. The next step involves customizing and developing polymeric polyurethane grids to facilitate rapid preoperative skin surface preparation.

Animal puncture experiments and cadaver specimen studies for the thoracic and cervical regions will be conducted in the future. Conduct direct experimental comparisons with existing navigation systems and gradually transition to clinical use, with the aim of achieving convenient navigation for punctures while preserving the surgical experience and habits of practitioners. Once RCCS is perfected, finalized, and commercialized, operations, such as image reconstruction, surgical design, registration, adjustment, puncture, and depth measurements, will be automated. At that point, the learning curve will be similar to that for existing surgical robotic techniques, making it convenient for surgeons to use.

Currently, navigation surgery is transitioning toward digital and radiation-free technologies, with image-guided navigation facilitating for the integration of upcoming cutting-edge innovations [29]. The ongoing work aims to produce advanced software, devices, and instruments with expanded functionalities through the fusion of chips, precision machining, automation, artificial intelligence, and other advancements. The evolution of navigation surgery based on structured light will enable enhanced simplicity, speed, minimally invasive procedures, and heightened precision.

Conclusion

The single 3D structured light system was shown to efficiently perform intraoperative scene scanning, static registration, and the dynamic tracking of screw placement instruments. This integrated approach was designed to facilitate precise navigation for minimally invasive procedures.

Abbreviations

3D structured light

3-dimensional structured light

CT

Computed tomography

RCCS

Red-crowned Crane System

Mimics

Materialise Interactive Medical Image Control System

STL

Stereolithography

RMS

Root mean square

G–R

Gertzbein–Robbins

MvIGS System

Machine-vision Image Guided Surgery System

Authors’ contributions

XHC: Conducted the study. The data were collected, analyzed, and interpreted. Wrote the manuscript.HXX: Edited the manuscript and planned the project.GDZ: Designed the study, interpreted the data, and edited the manuscript.BWY: Planned the project.ZPH: Edited the manuscript and reviewed the manuscript.WHX: Edited the manuscript.All the authors have read and approved the final manuscript.

Funding

No Funding.

Data availability

All data generated during this study are available from the corresponding author on reasonable request. There was no data published previously.

Declarations

Ethics approval and consent to participate

The study was approved by the “Medical Ethics Committee of Lecong Hospital, Shunde District, Foshan City”. The reference number was “(2023) No. 86”. All procedures adhered to the relevant guidelines and principles of the Declaration of Helsinki. Informed consent has been obtained from all participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.