Comparative analysis of dental implant placement accuracy: Semi‐active robotic versus free‐hand techniques: A randomized controlled clinical trial

Abstract

Background

Robot‐assisted implant surgery has emerged as a novel digital technology, and the accuracy need further assessment.

Purpose

This study aimed to compare the accuracy of single dental implant placement between a novel semi‐active robot‐assisted implant surgery (RAIS) method and the conventional free‐hand implant surgery (FHIS) method through a multicenter, randomized controlled clinical trial.

Materials and Methods

Patients requiring single dental implant placement were recruited and randomized into RAIS and FHIS group. Deviations at the platform, apex, and angle between the planned and final implant positions were assessed in both groups. Additionally, the evaluation of instrument and surgical complications was examined.

Results

A total of 140 patients (median age: 35.35 ± 12.55 years; 43 males, 97 females) with 140 implants from four different research centers were included, with 70 patients (70 implants) in the RAIS group and 70 patients (70 implants) in the FHIS group. In the RAIS and FHIS groups, the median platform deviations were 0.76 ± 0.36 mm and 1.48 ± 0.93 mm, respectively (p < 0.001); median apex deviations were 0.85 ± 0.48 mm and 2.14 ± 1.25 mm, respectively (p < 0.001); and median angular deviations were 2.05 ± 1.33° and 7.36 ± 4.67°, respectively (p < 0.001). Similar significant difference also presented between RAIS and FHIS group in platform vertical/horizontal deviation, apex vertical/horizontal deviation. Additionally, implants with self‐tapping characteristics exhibited significantly larger deviations compared with those without self‐tapping characteristics in the RAIS group. Both RAIS and FHIS methods demonstrated comparable morbidity and safety pre‐ and post‐operation.

Conclusions

The results indicated that the RAIS method demonstrated superior accuracy in single dental implant placement compared with the FHIS method. Specifically, RAIS exhibited significantly smaller deviations in platform, apex, and angular positions, as well as platform and apex vertical/horizontal deviations.

This clinical trial was not registered prior to participant recruitment and randomization. https://www.chictr.org.cn/showproj.html?proj=195045.

Summary Box.

What is known

• To date, there have been limited large sample comparative studies evaluating the accuracy of dental implant placement between robot‐assisted surgery and conventional free‐hand surgery methods.

What this study adds

• This multicenter randomized controlled clinical trial illustrated that RAIS offers superior accuracy in implant placement compared with FHIS. These results highlight the potential of robotic systems as a highly accurate and safe option in clinical practice.

1. INTRODUCTION

Accurate dental implant placement is crucial for long‐term success. Free‐hand placement might result in a high level of deviations from planned positions, thereby increasing the risk of prosthetic complications. ¹ Additionally, it poses a potential threat of damaging vital anatomical structures, resulting in complications such as severe hemorrhage and nerve injury. ¹ Static guides and dynamic navigation are used to achieve higher accuracy of implant position. ² However, static guides are limited by factors including poor visibility, big mouth opening, limited adjustability during surgery, poor cooling techniques and additional time for fabrication. ³ Dynamic navigation offers instant visual feedback but requires surgeons to keep an eye on the drill position in real time on the screen and in the surgical area, the outcome depends on the surgeon's hand‐eye coordination and surgical experience. Additionally, the high cost of equipment and complicated operation procedures have also made it difficult for this technology to be widely adopted. ⁴ , ⁵

Dental implant robotic‐assisted surgery systems have grown significantly and its applications are continuously expanding, and it combines the advantages of the physical restraints of static guides with the adjustability and visual capabilities of dynamic navigation, leading in increased accuracy and dependability. ⁶ , ⁷ For example, the first FDA‐authorized robotic system provides physical guidance by constraining the drill's position, orientation, and depth, thus eliminating deviations during the procedure. ⁸ , ⁹ Subsequently, Zhao et al introduced the world's first active robotic system, which autonomously adjusts in real‐time during intraoperative procedures and can perform surgical tasks directly on patients without explicit control by a surgeon. ¹⁰ Recent developments include the launch of an implant robot in 2022 and another system in 2023 that successfully placed highly accurate implants in patients with single tooth loss. ¹¹ , ¹²

As the application of robotic‐assisted technology continues to expand, the accuracy and efficiency of the robot is also one of the research concerns. Several studies compared the accuracy of dental implants placed using robots with other surgical techniques (free‐hand, static guides and dynamic navigation), and most studies indicate that various brands of robots demonstrate higher accuracy compared with free‐hand techniques, static guides, and dynamic navigation. ⁹ , ¹¹ , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ The accuracy of implant placement by robot‐assisted surgery indicated that the entrance, apical, and angle errors varied from 0.3 to 0.97 mm, 0.24 to 1.49 mm, and 0 to 7.13°, respectively. ²² However, oral implant robot surgery remains in the early phases of clinical research currently. Different robotic systems exhibit varying degrees of autonomy and comprise a range of components, including diverse visual recognition systems, design software, and robotic arms, which also differ across various programs.

In 2023, a novel semi‐active robotic assistant implant surgery (THETA, Hangzhou Jianjia Medical Technology Co., Ltd, Hangzhou, China) was developed. Traditional emi‐active implant robots, is capable of independently entering and exiting the mouth, preparing the implant bed, and inserting the implant. The operator's responsibilities include replacing the drill, delivering directions, and monitoring the robot's activity. ²² Differently, this novel semi‐active robot requires the doctor to assist by manually guiding the drill downward during the drilling process. Previous studies have demonstrated its potential benefits. For instance, Chen et al evaluated the accuracy of this robot by in vitro model experiment compared with dynamic navigation, demonstrating higher accuracy of this novel robot. ¹⁵ Ding et al illustrated the clinical feasibility, accuracy, and safety of this robot through a series of case studies. ¹⁷ Shi et al compared the implant accuracy, safety and morbidity between this robot and free‐hand dental implant placement through a prospective study involving a limited sample size, the result enabled greater positional accuracy of the implant compared with free‐hand placement. ²³ To the best of the authors' knowledge, there have been no multicenter prospective randomized controlled clinical trials reported. To further enhance the persuasive assessment of this new semi‐active robot's accuracy and clinical usability, this study aims to further evaluate its clinical outcomes, by compare its deviations with the conventional free‐hand implant surgery (FHIS) method, with a multicenter, prospective, controlled, and randomized clinical trial.

2. MATERIALS AND METHODS

2.1. Study design

The present study was a multi‐center, randomized controlled clinical trial with simple randomization (1:1 allocation ratio), conducted at four centers, received approval from the Ethics Committee of center 1, Shanghai Ninth People's Hospital, School of Medicine, Shanghai JiaoTong University (Approval No. SH9H‐2022‐C27‐2), center 2, Zhejiang Provincial People's Hospital (Approval No. 2022013), center 3, Affiliated Stomatological Hospital of Nanchang University (Approval No. 2022032), and center 4, Wuhan University Stomatological Hospital (Approval No. 202207). To ensure data consistency and comparability across different research centers, a calibration meeting was held among all the clinical centers prior to the study to discuss and standardize operative protocols. Each clinician received written instructions regarding collection of experimental parameters to obtain acceptable inter‐examiner consistency. We also established a central monitoring team, centralized data collection and analysis and used double‐blind measurement and statistical analysis methods. The research adhered to the principles outlined in the 1975 Declaration of Helsinki, as revised in 2013. The study was conducted in accordance with CONSORT guidelines ²⁴ and was registered with the Chinese Clinical Trial Registry (ChiCTR2300071624). All participants were comprehensively informed about the surgical protocol and provided their written consent.

2.2. Participants selection

Patients with single missing tooth seeking dental implant rehabilitation were screened and recruited. Inclusion criteria were the following: (1) aged between 18 and 70 years; (2) individuals presenting with missing dentition necessitating implant restoration; (3) absence of soft and hard tissue defects or lesions within the designated implantation area that may impact the implant procedure; and (4) participate voluntarily and sign the informed consent form. Exclusion Criteria were the following: uncontrolled periodontitis, uncontrolled systemic diseases, mouth opening <35 mm, pregnancy or lactation, and any other conditions deemed inappropriate for trial participation by the investigator.

Two types of implants were used in this study: the cylindrical outer contour without self‐tapping Straumann® bone level (BL) or standard plus (SP) implant, and the tapered outer contour with self‐tapping NobelParallel conical connection (PCC) implant.

2.3. Sample size calculation and randomization

The sample size was determined based on several previous studies that reported a mean and standard deviation (SD) values of the platform/apex point of the implant. ²⁵ , ²⁶ , ²⁷ , ²⁸ Based on these parameters, the expected mean platform/apex deviation in the free‐hand group is 1.62/2.08 mm, with a SD of 1.0/1.24 mm. The RAIS group had an estimated mean and SD of 1.0 ± 0.3 mm for platform and 1.3 ± 0.5 mm for apex deviation. With α = 0.025, one‐sided testing, and a power of 1−β = 0.9, the sample size calculation is performed for a case ratio of 1:1 between the test and control groups, resulting in an estimated sample size of 55 patients per group. With 20% dropouts and randomization blocks, the sample size is increased to 70:70.

To ensure a balanced distribution of various uncontrollable non‐research factors among the comparison groups, the Interactive Web Response System (IWRS) is utilized for follow‐up allocation and randomization, with each center competitively enrolling participants.

2.4. Preoperative preparation and digital planning

The novel semi‐active dental implant robotic system (THETA, Hangzhou Jianjia Medical Technology Co., Ltd, Hangzhou, China) was used for the RAIS group. The system's architecture has been comprehensively detailed in Shi's report. ²³ The robot is positioned at the forehead of the dental chair, with the surgeon seated on the patient's right side and the assistant on the patient's left side. Figure 1 shows the functions of the individual components and the collaboration between them.

FIGURE 1.

Functions of the components and collaboration of the robot.

For the preoperative preparation process in the RAIS group, first, a U‐tubed registration device (Hangzhou Jianjia Medical Technology Co., Ltd., Hangzhou, China) was mounted on the patient's missing tooth site using a silicone impression material (Silagum, DMG, Hamburg, Germany). Then a cone‐beam computed tomography (CBCT) was taken with the same parameter (96 kV, 5 mA, voxel size: 0.15 mm; and field of view:12 × 9 cm). Later, the CBCT data were exported in the standard digital imaging and communications in medicine (DICOM) format. Second, a digital intra‐oral scan was performed by a scanner (TRIOS®, 3‐Shape, Copenhagen, Denmark) and a standard tessellation language (STL) file was exported for each patient. Third, CBCT data and the STL intra‐oral scanning data were imported into the surgical planning software (Cycad, version 1, Hangzhou Jianjia Medical Technology Co., Ltd, Hangzhou, China). The procedure of the planning software included maxillo‐mandibular segmentation, panoramic image reconstruction, mandibular canal depiction, and oral scan‐CBCT alignment, implant planning and fine‐adjustment.

For the preoperative preparation in the FHIS group, implant planning was conducted using DTX Studio™ (Nobel Biocare, Goteborg, Sweden) with combination of the CBCT data and intra‐oral scan data. The entire implant placement process was designed using the prosthetically driven implantation principle, with special attention paid to planning at the buccolingual center of the ridge, in the middle of the mesio‐distal width of the planned restoration, the axial load of the occlusal forces, and the equicrestal position of the endosteal portion of the implant. ²⁹ Hence, the implant type and location were visualized and confirmed preoperatively.

2.5. Surgical procedure

Following local anesthesia (articaine 4% with epinephrine 1:100000), flaps were minimally elevated. In the RAIS group, prior to the implant surgery, preparatory steps were taken as shown in Figure 2. First, calibration of the handpiece (position of the robotic arm) was executed using a standard plate used for tool calibration and a three‐plate reflective calibrated global probe. Subsequently, registration and calibration were performed to align the robotic arm accurately. The U‐tube was repositioned at the missing tooth site, and a patient‐tracking device was fixed to the contralateral teeth with self‐curing acrylic resin (Protemp™, 3M ESPE, Neuss, Germany). Calibrated probes marked registration points on the U‐tube, followed by accuracy verification. The U‐tube was then withdrawn from the patient's mouth to begin the implant procedure. The surgeon pressed the teach button (which allows free movement of the robotic arm when pressed) and moved the robotic arm to the implant site within 10 mm. Then released the teach button, the end of the robotic arm automatically orient itself approaching to the planned implant axis. After automatic positioning, the surgeon pressed and held the alignment button (which allows alignment movement of the robotic arm when pressed) to execute the alignment motion along the planned implant axis. The arm moved automatically in coordination with the patient's head, and the operator could only manipulate the handpiece in the axial direction of the designated implant after pressing the alignment button, ensuring precise drilling direction. Real‐time data on drilling parameters were monitored throughout on the screen. After finishing each step of cavity preparation, the surgeon released the alignment button, press the teach button, and retracted the robotic arm from the oral cavity, and repeated the above process until the implant bed preparation is completed. Ultimately, implant insertion was accomplished utilizing the robotic arm in the planned position. And a carrier can be inserted on the implant to further verify the actual depth of the implant immediately.

FIGURE 2.

Navigation and surgical procedure, including the calibration, registration and implant process on the navigation software Cypress and clinical practice.

In the FHIS group, based on the previous CBCT and intraoral scan data, an ideal implant position was designed. The implant surgeries at each center were performed by a senior physician, to make sure that the operation were precisely executed according to the plan.

2.6. Postoperative treatment

Flaps were sutured with 5‐0 Vicryl sutures. Post‐surgery CBCT scans were conducted with the same exposure parameters as pre‐surgery. The post‐operative regimen included 1 g of amoxicillin every 8 h for 3 days, 0.2 g of metronidazole every 8 h for 3 days, and rinsing with 0.12% chlorhexidine for 1 week. Postoperative complications, including secondary bleeding, swelling, infection, and nerve numbness, were noted. Sutures were removed 2 weeks after the surgery. The prosthesis was fabricated 3–6 months post‐surgery.

2.7. Outcomes

The primary outcomes included global deviation at the implant platform, global deviation at the implant apex, and angle deviation between planned and placed implants. The secondary outcomes included the horizontal and vertical deviations at the implant platform and apex, as well as the instrument evaluation for software functionality and machine performance, safety assessment, and evaluation of surgical morbidity and complications.

2.8. Accuracy measurement

The preoperative and postoperative CBCT data were superimposed using the 3D Slicer software (version 4.13; Harvard, Boston, USA), following the method reported by Talmazov et al. ³⁰ First, the preoperative CBCT scans containing the coordinates of the planned implant position and the postoperative CBCT scans were imported into the 3D Slicer software. Then, the preoperative and postoperative 3D reconstructed models were superimposed by placing of three alignment points of the same position. Precise alignment is achieved by point cloud acquisition algorithms of 3D Slicer and adjusted manually if necessary. After the automatic recognition of the actual implant position, the coordinates of planned and actual implant were compared. The long axis of the planned and the placed implant positions were compared and measured for angular deviation. For platform and apex points, global, horizontal and vertical deviations were analyzed separately (Figure 3). Two medical professionals from Center 1 and Center 2, who was not involved in the study performed all the assessments twice dependability. They received professional training prior to measurements and passed consistency tests. The average of their collected data was utilized for final statistical analysis.

FIGURE 3.

Implant positional accuracy measurements. (A), Pre‐ and post‐surgery cone‐beam computed tomography superimposition, automatically defining and manually checking the actual implant position. (B), ① Angular deviation. ② Platform global deviation. ③ Platform vertical deviation. ④ Platform horizontal deviation. ⑤ Apex global deviation. ⑥ Apex vertical deviation. ⑦ Apex horizontal deviation.

2.9. Instruments, safety and surgical morbidity evaluation

Instruments evaluation were conducted throughout preoperative preparations and the surgical procedure, focusing on software functionality, safety, operational stability and device defects. The software functionality was assessed in areas such as data management, image display during surgery, implant planning, step design and robot control, with satisfaction rated as satisfied, neutral, or unsatisfied. Safety evaluation covered mechanical and electrical aspects. Mechanical safety was confirmed if there were no unintended movements or damage, whereas electrical safety was verified by the absence of leakage. Operational stability was deemed satisfactory if the system operated continuously. Device defects were any risks to health and safety from normal use.

In addition, surgical morbidity associated with implant surgery were recorded, including intraoperative bleeding, potential damage to adjacent structures, nerve or neighboring tooth injuries, implant excessive deviation, unacceptable deviation/displacement, postoperative infections, bleeding, wound dehiscence, and short‐term implant failure.

2.10. Statistical analysis

Statistical analyses were performed using SPSS 22.0 (IBM Corp., Armonk, NY, USA). The age distribution between the two groups was evaluated by independent t‐test, whereas the distributions of other factors were evaluated by χ ² test. Mann–Whitney U test was performed to compare the accuracy of platform, apex, and angular deviation among the RAIS and FHIS group, and the differences of the deviations associated with implant position, length, type, and bone augmentation in RAIS. The Kruskal–Wallis test was utilized to compare the differences of the deviations among bone density and implant diameters. The statistical significance was set at p < 0.05.

3. RESULTS

3.1. Study population

A total of 163 patients was screened and 140 satisfied the eligibility criteria were randomized and received the allocated treatment between January 2023 and September 2023 (23 were excluded). 140 patients were randomly assigned to two treatment groups (n = 70 each group). The study was conducted at facilities located in eastern and southern China, with all participants being of Asian ethnicity. The distribution of participants among the four centers is as follows: Centers 1 and 4 each had 28 participants (14 in both RAIS and FHIS groups), Center 2 had 59 participants (29 in RAIS and 30 in FHIS groups), and Center 3 had 25 participants (13 in RAIS and 12 in FHIS groups). Bone density was measured by the method described by Putra et al. ³¹ Figure 4 shows the CONSORT diagram. Patient and implant demographics are shown in Table 1. No obvious intergroup differences were observed in terms of the RAIS and FHIS group.

FIGURE 4.

CONSORT diagram.

TABLE 1.

Demographic and clinical characteristics of the included patients.

Baseline Characteristics	RAIS group	FHIS group	p‐value
Age (year)	18–67	19–68	0.664 a
Mean ± SD	35.6 ± 12.3	35.1 ± 12.8
Gender			0.855 b
Male	21	22
Female	49	48
Total number of implants	70	70
Implant position
Maxillary			0.737 b
Anterior teeth	2	5
Premolar teeth	6	8
Posterior teeth	11	23
Mandible			0.167 b
Anterior teeth	1	0
Premolar teeth	6	9
Posterior teeth	44	25
Bone Augmentation			0.237 b
Yes	14	20
No	56	50
Implant system			0.864 b
Straumann	41	40
Nobel PCC	29	30
Bone Density			0.638 b
D1	0	0
D2	7	4
D3	54	57
D4	9	9
Self‐tapping			0.864 b
Yes	29	30
No	41	40
Implant diameter (mm)			0.732 b
3.3	2	5
3.75	1	2
4.1	25	20
4.3	23	25
4.8	14	15
5	5	3
Implant Length (mm)			0.448 b
8.0	8	5
8.5	10	5
10	50	56
11.5	1	3
12	1	1

Abbreviations: FHIS, free hand implant surgery; RAIS, robotic assisted implant surgery.

^a Independent t‐test.

^b Chi‐Square p‐value.

3.2. Implant positional accuracy

The angular, coronal, apical and depth deviations were calculated for each group. Table 2 shows the implant accuracy results. For the RAIS group, the mean and SD of the platform global deviation, apex global deviation and angular deviation is 0.76 ± 0.36 mm, 0.85 ± 0.48 mm and 2.05 ± 1.33°, respectively. The vertical and horizontal deviation of the platform is 0.39 ± 0.30 mm, 0.61 ± 0.39 mm, whereas the vertical and horizontal deviation of the apex is 0.40 ± 0.36 mm, 0.71 ± 0.39 mm. For the FHIS group, the mean and SD of the platform global deviation, apex global deviation and angular deviation is 1.48 ± 0.93 mm, 2.14 ± 1.25 mm, 7.33 ± 4.67°, respectively. The vertical and horizontal deviation of the platform is 0.74 ± 0.82 mm, 1.13 ± 0.69 mm, whereas the vertical and horizontal deviation of the apex is 0.73 ± 0.76 mm, 1.89 ± 1.19 mm. Compared with the free‐hand group, the RAIS group exhibited significantly lower platform, apex, and angular deviations, indicating a notably higher degree of accuracy (Figure 5).

TABLE 2.

Platform, apex, and angular deviation between different characteristics in RAIS group.

	Number of implants	Platform deviation (mm)			Apex deviation (mm)			Angular deviation (°)
		Global	Vertical	Horizontal	Global	Vertical	Horizontal
Jaw
Maxilla	19	0.77 ± 0.35	0.44 ± 0.29	0.57 ± 0.33	0.84 ± 0.41	0.44 ± 0.31	0.67 ± 0.36	1.93 ± 0.97
Mandible	51	0.76 ± 0.37	0.37 ± 0.30	0.62 ± 0.31	0.86 ± 0.51	0.39 ± 0.38	0.72 ± 0.41	2.09 ± 1.45
p‐value a		0.995	0.291	0.517	0.874	0.306	0.634	0.726
Position
Anterior + Premolar	15	0.70 ± 0.49	0.42 ± 0.44	0.49 ± 0.35	0.88 ± 0.79	0.48 ± 0.63	0.70 ± 0.55	2.06 ± 1.45
Molar	55	0.78 ± 0.32	0.38 ± 0.25	0.64 ± 0.30	0.85 ± 0.37	0.38 ± 0.25	0.71 ± 0.35	2.04 ± 1.31
p‐value a		0.054	0.543	0.039**	0.136	0.496	0.183	0.897
Implant length
<10 mm	18	0.83 ± 0.45	0.48 ± 0.39	0.62 ± 0.36	0.99 ± 0.70	0.52 ± 0.55	0.77 ± 0.53	2.51 ± 1.87
≥10 mm	52	0.74 ± 0.32	0.35 ± 0.26	0.61 ± 0.30	0.81 ± 0.38	0.36 ± 0.27	0.68 ± 0.34	1.88 ± 1.06
p‐value a		0.614	0.301	0.877	0.448	0.301	0.824	0.190
Implant diameter
3.5 mm 3.75 mm	3	1.04 ± 0.99	0.63 ± 0.79	0.81 ± 0.64	1.49 ± 1.67	0.92 ± 1.30	1.13 ± 1.11	3.21 ± 3.10
4.1 mm 4.3 mm	48	0.72 ± 0.33	0.37 ± 0.27	0.56 ± 0.30	0.80 ± 0.39	0.37 ± 0.28	0.68 ± 0.35	2.12 ± 1.27
4.8 mm 5.0 mm	19	0.83 ± 0.29	0.38 ± 0.25	0.69 ± 0.28	0.87 ± 0.32	0.40 ± 0.25	0.71 ± 0.33	1.68 ± 1.05
p‐value b		0.368	0.941	0.249	0.565	0.765	0.830	0.565
Implant system
Straumann	41	0.67 ± 0.24	0.32 ± 0.21	0.54 ± 0.24	0.72 ± 0.32	0.32 ± 0.21	0.59 ± 0.30	1.44 ± 0.72
Nobel	29	0.90 ± 0.45	0.47 ± 0.38	0.71 ± 0.38	1.04 ± 0.61	0.51 ± 0.50	0.87 ± 0.45	2.90 ± 1.53
p‐value a		0.024**	0.168	0.082	0.003**	0.245	0.001**	<0.001**
Bone augmentation
Yes	14	0.74 ± 0.30	0.39 ± 0.28	0.55 ± 0.28	0.82 ± 0.47	0.41 ± 0.26	0.62 ± 0.44	1.76 ± 1.02
No	56	0.77 ± 0.37	0.38 ± 0.31	0.62 ± 0.32	0.86 ± 0.49	0.40 ± 0.39	0.73 ± 0.38	2.12 ± 1.40
p‐value a		0.883	0.786	0.378	0.607	0.542	0.082	0.481
Bone density
D2	7	0.73 ± 0.66	0.48 ± 0.49	0.54 ± 0.46	1.00 ± 1.08	0.61 ± 0.81	0.77 ± 0.73	2.31 ± 2.14
D3	54	0.77 ± 0.32	0.36 ± 0.26	0.62 ± 0.31	0.83 ± 0.38	0.37 ± 0.26	0.69 ± 0.36	2.06 ± 1.28
D4	9	0.76 ± 0.31	0.44 ± 0.36	0.56 ± 0.15	0.88 ± 0.41	0.43 ± 0.40	0.73 ± 0.26	1.72 ± 0.91
p‐value b		0.355	0.880	0.292	0.512	0.890	0.440	0.892

Abbreviations: RAIS, robotic assisted implant surgery, FHIS, free hand implant surgery. *p < 0.05, **p < 0.01

^a Mann‐Whitney U test.

^b Kruskal‐Wallis test.

FIGURE 5.

The results of implant accuracy between the robot‐assisted implant surgery (RAIS) and free‐hand implant surgery (FHIS) group.

3.3. The effect of different characteristics on the accuracy of the RAIS and FHIS group

Table 2 illustrates the positional deviations observed among patients with different characteristics and implant types. In the aspect of implant type, for non‐tapered BL and SP implants, the platform, apex, and angular deviations were 0.67 ± 0.24 mm, 0.72 ± 0.32 mm, and 1.44 ± 0.73°, respectively. In contrast, for tapered PCC implants, these deviations were measured at 0.90 ± 0.45 mm, 1.04 ± 0.61 mm, and 2.90 ± 1.53°, respectively. Tapered PCC implants displayed significantly greater errors in global platform, apex, and angular deviations compared with those non‐taped characteristics (SP and BL) within the RAIS group (p < 0.05). However, no significant differences were observed regarding implant length, diameter, jaw location (maxilla, mandible), tooth position (anterior, premolar, molar), bone augmentation, or bone density within the RAIS group.

The Table S1 presents the platform, apex, and angular deviations between different characteristics in FHIS group. The results indicate that there are no significant differences among the evaluated parameters.

3.4. Instrument evaluation

The software functionality across all criteria was found to be satisfactory. Both mechanical and electrical safety were confirmed, with no unintended movements, toppling, part detachment, mechanical breakage, injuries, or electrical leakage occurring during the trial. Operational stability was also deemed satisfactory, as the system functioned continuously throughout the trial. For the device defects, in one case of the RAIS group, there was a trouble of removing the temporary adhesive resin fixing the patient‐tracking reflective device after the implantation process was finished due to the significant concavity of the adjacent tooth, and a high‐speed handpiece was then used to remove the resin. Neither the RAIS nor the FHIS group displayed any additional safety concerns. The surgeon was pleased with the software's functionality and the machine's general operational efficiency, and there were no instrument flaws, work stability issues, or unfavorable occurrences connected to the robotic system. For the patient, the RAIS did not pose a safety risk.

3.5. Surgical morbidity and complications

None of the RAIS or FHIS groups reported any surgical implant problems, such as implant displacement above the apical anatomical limit, bleeding during surgery, or damage to nearby teeth or nerves. There was no presence of wound dehiscence, infection, or post‐operative hemorrhage as the wound healed. However, one implant in the FHIS group failed to osseointegrate and was removed at the three‐month postoperative follow‐up, whereas all the other implants osseointegrated evenly.

4. DISCUSSION

The present study evaluated the accuracy of a novel semi‐active oral implantation robotic system through a multicenter prospective randomized controlled trial, and the results demonstrated outstanding accuracy of the implant (average global platform deviation, global apex deviation, and angular deviation were 0.76 mm, 0.85 mm, and 2.05°, respectively) compared with free‐hand surgery, indicating a high level of accuracy.

Robots can be classified as active, semi‐active, and passive based on their level of autonomy. ³² , ³³ Passive robots require the operators to manually guide the robotic for all procedures, including entering and exiting the mouth, preparing the implant bed, and placing the implant. A study found that a passive robot showed good implant position accuracy (global platform deviation, 1.04 mm; global apex deviation, 0.95 mm; angular deviation, 2.56°) when used in edentulous arch. ⁹ Active robots can autonomously enter and exit the mouth, prepare the implant bed, and insert the implant. The operator's only responsibility is replacing the drill, issuing instructions, and monitoring the robot's operation. ³⁴ Semi‐active robots can prepare the implant bed and install it independently. However, the robotic arm requires operator traction to enter and exit the mouth. ²² Different semi‐active robots also have slightly different operations, for example, the semi‐active robot used in the present study, the robotic arm possesses functions for only depth and alignment control. Before automated positioning begins, the operator must hold the teach button and manually advance the drill within 10 mm to the implant site. After automated positioning is done, the operator must hold the alignment button while gently exerting downward force along the line's direction during drilling. After finishing the drilling, the operator pulls the drill out of the patient's mouth by holding the teach button. The operator can use the teach and alignment buttons to operate the robot arm. This approach enhances surgical safety by allowing the robot arm to respond promptly to sudden significant changes in the patient's head position during surgery, preventing potential emergency situations where the arm fails to withdraw in time. This structure is one of the major difference in operation between these two semi‐active robots, thereby increasing the safety of the procedure. ¹⁵ , ¹⁷ , ²³ In contrast, another semi‐active robot does not require the physician to exert a downward force during the procedure, the drill bit autonomously descends until the specified depth is achieved. ³⁵ So far, only one research has compared the accuracy among different types of robots, Xu et al compared the surgical efficiency and accuracy of implant placement among robots with three robots, including semi‐active robot, active robot and passive robot in a vitro study, the preparation times for the active, passive, and semi‐active groups were 3.85 ± 0.17 min, 2.14 ± 0.06 min, and 1.65 ± 0.19 min, respectively. The mean operation time for the passive group (3.76 ± 0.59 min) was shorter than that of the active (4.89 ± 0.70 min) and semi‐active (4.59 ± 0.56 min) groups (all p < 0.001). For the implant accuracy, the results indicate that active and semi‐active robots exhibit similar implant accuracy, whereas the passive robot shows higher deviations, highlighting the impact of different human–robot interactions have variable surgical efficiencies. ¹⁴ Although implant robotic systems on the market now are reported to have very high surgical accuracy, differences in machine architecture, surgical processes, and technical features result in varying procedures, efficiency, and accuracy among manufacturers.

For the comparison between robot and traditional static guides and dynamic navigation, most previous researches have reported higher accuracy in robots. For static guides, He et al assessed the accuracy of an active robotic and fully guided static computer‐assisted implant surgery by in vitro models and in vivo patients research, found that the active robot showed significantly higher accuracy than the fully static guides. ²⁰ For edentulous implant placement, Wang et al compared the accuracy of an active robot with fully guided static computer‐assisted implant surgery (CAIS) template in edentulous, the result showed that the mean and SD deviation of robot at the implant platform, apex, depth, and angle was 0.65 ± 0.25 mm, 0.65 ± 0.22 mm, 0.49 ± 0.24 mm, and 1.43 ± 1.18°, respectively, and demonstrated superior accuracy compared with CAIS templates. ³⁶ Another retrospective clinical study also demonstrated higher accuracy of implants placed with by an active robotic compared with static computer assisted implant surgery. In addition, they found that implant regions had no significant influence on the accuracy of implant placement. ¹⁹ However, a recent in vitro study demonstrated that static‐guided surgery resulted in more accurate and precise implant placements, yielding superior outcomes compared with robotic‐guided surgery under controlled laboratory conditions. ³⁷ For dynamic navigation, Chen et al evaluated the accuracy of the robot used in this study by in vitro model experiment compared with dynamic navigation, demonstrating higher accuracy of this novel robot. ¹⁵ Tao et al compared the accuracy of the robotic system with the dynamic navigation system in a clinical study, the results showed that a robotic system exhibits lower deviations both in partially and completely edentulous jaws. ¹³ In contrast, Mozer et al compared RAIS with static computer‐assisted implant surgery (sCAIS), the overall angular deviation was 2.66 ± 1.83° for the robotic system and 0.68 ± 0.38° for guided surgery using static guides (p < 0.001), the 3D‐deviation of the implant platform at crest level was for sCAIS 0.79 ± 0.28 mm and RAIS 1.51 ± 0.53 mm (p < 0.001) and at the apex for sCAIS 0.82 ± 0.26 mm and for RAIS 1.97 ± 0.79 mm (p < 0.001), respectively, robotically guided implant surgery was less accurate in terms of trueness and accuracy than traditional sCAIS in this in vitro study. ³⁷

In addition to the inherent factors of the robotic system, accuracy may also be influenced by individual patient characteristics, as well as various aspects such as differences in implant morphology, diameter, length, bone density and other related factors. In this study, tapered implants presented significant deviation than non‐tapered implants in the RAIS group, this result is different with a previous study which compare the accuracy of immediate implant placement of cylindrical implants (CI) and tapered implants (TI) of different lengths using a robotic dental implant system RemebotDent (Remebot, Beijing Ruiyibo Technology Co., Ltd.), showed no significant effects of implant length or shape on implant accuracy. ³⁸ Wei et al also demonstrated that implant macrodesign (tapered and straight implants) did not affect accuracy and primary stability in immediate implant using dynamic navigation. ³⁹ This discrepancy in our results may be attributed to the unique design of the THETA, as surgeons need to exert a gentle downward force along the alignment direction during drilling and implantation, wherein self‐tapping conical implants may be more prone to deviation compared with non‐self‐tapping cylindrical implants. As for the bone density, no significant different was found between D2, D3 and D4 bone in this study, this result is also different with a previous study which indicated that low bone density might be risk factors influencing the accuracy of implant placement with computer‐guided surgery. ³¹ In addition, no significant differences were observed regarding implant length, diameter, jaw location (maxilla, mandible), tooth position (anterior, premolar, molar), and bone augmentation or not within the RAIS group in our study. Further research with larger sample sizes is required to address these issues.

Moreover, marker invisibility could sometimes happen during surgery when the camera was interfered by the surgeon's head or under other improper using. Different robots have unique configurations, the trolley and binocular camera of the robot in this study are integrated and positioned at the patient's head. One benefit of this configuration is that the binocular camera can detect markers throughout the surgery with more accuracy, reducing the chance of invisibility. In contrast, other kinds of robots usually have a modular design, the camera is normally oriented toward the patient's feet, marker invisibility is prone to occur during surgery. Furthermore, to handle the robotic arm and prevent unintentional movements during drilling processes, manual button control (alignment and instruction buttons) is designed.

Compared with static guides, robotic implant surgery presents several limitations. The device's high cost is a significant drawback, and preoperative registration adds to preparation time. Surgeries on both sides of the jaw require re‐registration, and visual obstruction issues may occur during the procedure. ³⁷ When compared with navigation technology, robotic surgery also involves high costs and maintenance. The current robotic arms lack tactile feedback, leading to a loss of haptic sensation concerning the bone's density during drilling. This absence of tactile feedback prevents the surgeon from assessing the resistance to drilling in bone or models, creating a complete disconnect. As a result, it limits the surgeon's ability to determine the actual bone density relative to the drilling resistance and adjust their drilling protocols accordingly to achieve primary stability during implant placement. ³⁷

Despite the positive findings of this investigation, several limitations must be addressed. First, the study focused primarily on the accuracy of the RAIS compared with the FHIS, necessitating additional observations on prospective studies with large samples to compare with dynamic navigation and static guides for accuracy, surgical efficiency, and postoperative evaluation, which will be more convincing. Second, preoperative preparation of robotic surgery requires implant design, preoperative calibration, registration, verification, which greatly increases the time of preoperative preparation, other technologies such as artificial intelligence can be combined to simplify preoperative procedures and improve efficiency in the future. Third, in measuring the accuracy of implant position, it is essential to consider the tolerance of the alignment between preoperative and postoperative CBCT scans. Misalignment between these scans can lead to errors in assessing the precise location of the implant, thereby affecting the overall accuracy measurement. Moreover, the robot used in the present clinical trial is limited to a single quadrant of implant placement. Its application in cross‐quadrant and full‐arch implant surgery needs further exploration in future.

5. CONCLUSION

Within the limits of this study, it is proposed that the novel semi‐active robot‐assisted implant surgery gives more accuracy in implant placement than free‐hand implant surgery. These findings highlight the promise of robotic systems as a highly precise and safe choice in clinical practice.

FUNDING INFORMATION

This work was supported by grants from Major Scientific and Technological Project of Zhejiang Province (No. WKJ‐ZJ‐2328).

CONFLICT OF INTEREST STATEMENT

The authors declare that there is no conflict of interest.

Supporting information

Table S1. Platform, apex, and angular deviation between different characteristics in FHIS group.

Yang F, Chen J, Cao R, et al. Comparative analysis of dental implant placement accuracy: Semi‐active robotic versus free‐hand techniques: A randomized controlled clinical trial. Clin Implant Dent Relat Res. 2024;26(6):1149‐1161. doi: 10.1111/cid.13375

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.