Accuracy, Efficiency, and Patient‐ and Surgeon‐Reported Outcomes of Static Versus Robotic Computer‐Assisted Implant Surgery: A Randomized Clinical Trial

ABSTRACT

Objectives

To assess implant placement accuracy, efficiency, and patient‐ and surgeon‐reported outcomes of static computer‐assisted implant surgery (sCAIS) and robotic computer‐assisted implant surgery (rCAIS) in single‐tooth posterior edentulism.

Methods

In this randomized clinical trial, 36 patients requiring posterior single‐tooth implant restoration were allocated to sCAIS (n = 18) or rCAIS (n = 18). Preoperative planning for all patients was performed using cone‐beam computed tomography (CBCT) data. The DICOM datasets were imported into planning software to conduct virtual implant planning in accordance with prosthetic‐driven and bone‐supported principles. Implant placement accuracy was assessed by superimposing CBCT‐based preoperative planning with postoperative cone beam computed tomography, measuring linear (coronal, apical) and angular deviations. Surgery time and patient‐reported experiences (anxiety, intraoperative discomfort, postoperative pain) and surgeon‐reported feedback were recorded for each group.

Results

The rCAIS group demonstrated significantly superior accuracy, with lower coronal (0.70 ± 0.44 mm vs. 1.02 ± 0.37 mm, p = 0.018), apical (0.75 ± 0.41 mm vs. 1.11 ± 0.40 mm, p = 0.009), and angular deviations (1.20° ± 0.44° vs. 2.38° ± 1.03°, p < 0.001). However, rCAIS required significantly longer surgery time (49.53 ± 12.96 min vs. 24.38 ± 6.51 min, p < 0.001) and resulted in greater intraoperative patient discomfort (3.00 ± 1.60 vs. 1.67 ± 1.87, p = 0.019). Surgeons reported easier access to the surgery site with sCAIS (p < 0.01), whereas the complexity of the surgical procedure and primary implant stability of both groups were comparable.

Conclusion

The rCAIS provides superior implant placement accuracy compared to sCAIS but is associated with longer procedures and increased patient discomfort. The sCAIS offers reliable implant placement with simpler operation and greater adaptability. Clinical decision‐making should balance the benefits of accuracy with patient comfort, efficiency, and resource considerations.

Trial Registration

ClinicalTrials.gov (NCT06059573) and the Chinese Clinical Trial Registry (ChiCTR2300078795)

1. Introduction

The third European Association for Osseointegration (EAO) consensus conference highlighted increased success rates for dental implant restorations (10‐year rates reaching 94.9%). Nevertheless, prosthetic complications—often influenced by suboptimal implant positioning—remain a significant clinical concern (Albrektsson and Donos 2012). Prosthetically driven implant placement is crucial for optimal biomechanical load distribution, long‐term prosthetic maintenance, and aesthetic success (Belser et al. 1996). Computer‐assisted implant surgery (CAIS) is routinely performed in implant dentistry (D'Haese et al. 2017). The advent of technologies such as cone beam computed tomography (CBCT), intraoral scanning (IOS), rapid prototyping (RP), optical tracking, and haptic guidance has enabled the development of three major CAIS approaches: static computer‐assisted implant surgery (sCAIS), dynamic computer‐assisted implant surgery (dCAIS), and robotic computer‐assisted implant surgery (rCAIS) (Chen and Nikoyan 2021; Pimkhaokham et al. 2022).

Among these, sCAIS is currently the most widely used CAIS method. It implements preoperative planning software and computer‐aided design/computer‐aided manufacturing (CAD‐CAM) to fabricate three‐dimensional (3D)‐printed surgical guides for implant placement, significantly improving placement accuracy over freehand methods (Smitkarn et al. 2019; Van Assche et al. 2012; Tahmaseb et al. 2018; Guentsch et al. 2023). However, sCAIS suffers from several drawbacks, including prolonged guide fabrication time, restricted visibility and operative space, inadequate cooling, and a lack of intraoperative adaptability. Although dCAIS overcomes some of these issues by enabling real‐time navigation and obviating the need for surgical guides, it remains limited in compensating for hand tremors and providing tactile feedback (Wu et al. 2023; Wang et al. 2022). As a result, surgical accuracy in dCAIS continues to rely heavily on the surgeon's skill, and manual drilling errors persist.

More recently, with the rapid advancement of digital technology and the growing emphasis on accuracy in implant placement, rCAIS are being progressively integrated into clinical practice for oral implantology. In 2017, Zhao developed the world's first autonomous robotic assistant implant surgery, which was gradually introduced into clinical practice (BioMat'X FdOUdlASdCC et al. 2017; Bai et al. 2021). Like static (sCAIS) and dynamic (dCAIS) approaches, rCAIS begins with a preoperative planning phase using CBCT and intraoral scan data to determine the optimal implant position based on prosthetic and anatomical considerations. The distinction among modalities lies in the intraoperative execution: sCAIS uses prefabricated templates to guide drilling; dCAIS provides real‐time visual feedback via optical tracking systems; and rCAIS employs a robotic arm to execute the pre‐planned trajectory under continuous monitoring. The guidance modality is system‐dependent, utilizing optical tracking, haptic control, or a hybrid approach to achieve real‐time surgical navigation and calibration (Brief et al. 2002). The surgeon maintains active control in all phases: either by directing the robot via foot pedals or by manually guiding the handpiece, with the ability to intervene immediately if necessary. Although low‐evidence‐based preliminary studies (e.g., in vitro studies, case reports) report clinically acceptable outcomes for the rCAIS (Cao et al. 2020; Bolding and Uday 2022; Li et al. 2023; Chen et al. 2023) and suggest superior accuracy over sCAIS (Jia et al. 2023; Wang et al. 2023), high‐level evidence directly comparing them remains limited due to the relative novelty of clinical rCAIS system (Wang et al. 2026; Khaohoen et al. 2024). Extensive clinical data from randomized studies is required to support our understanding of the accuracy of rCAIS and sCAIS and explore the advantages and drawbacks of their clinical application. Furthermore, despite rapid digital advancements, we currently lack consensus on digital technology in implantology, which highlights the necessity for higher‐level evidence‐based research to establish clinical guidelines.

Although previous clinical studies on robotic‐assisted implant surgery have primarily focused on technical accuracy, operative time, and implant survival rate (Xu et al. 2023, 2024), there is a notable lack of high‐quality evidence regarding the patient's perspective. patient‐reported outcomes (PROs), which reflect patient perceptions of treatment, are essential indicators of success (Fu et al. 2023; McGrath et al. 2012). To the best of our knowledge, this randomized controlled trial is the first to systematically incorporate and evaluate both Patient‐Reported Outcome Measures (PROMs) and Patient‐Reported Experience Measures (PREMs) for this emerging technology. By doing so, this study provides novel insights that extend beyond technical performance. It aims to determine whether the perceived advantages of robotics translate into tangible benefits from the patient's viewpoint, such as reduced postoperative discomfort, improved functional outcomes, and a more positive overall experience compared to conventional guided surgery. This patient‐centered data are crucial for a comprehensive understanding of the value of rCAIS and for supporting shared clinical decision‐making (Youk et al. 2014).

Therefore, in this randomized clinical trial, we aimed to compare the accuracy of rCAIS with that of sCAIS. Additionally, we compared the patient‐ and surgeon‐reported outcomes of the two approaches.

2. Materials and Methods

This double‐arm randomized clinical trial was conducted at Tongji Hospital, Tongji Medical College of HUST, Hubei, Wuhan, China. The study protocol, approved by the Tongji Hospital Institutional Review Board (TJ‐IRB20231192), complied with the Declaration of Helsinki. The Consolidated Standards of Reporting Trials (CONSORT: Appendix S1) statement was followed (Schulz et al. 2010). After fully informing the participants of the study objectives and potential risks, we obtained written informed consent from them. All surgeries were performed at the Department of Prosthodontics and Implantology, Tongji Hospital, Tongji Medical College of HUST, between April 2024 and December 2024.

2.1. Sample Size Calculation

The sample size was calculated from the mean (standard deviation [SD]) of the angular deviation based on the results of a recent retrospective study (Jia et al. 2023) using G*Power (Version 3.1.9.7). The mean (SD) deviations used were 2.42 (1.55) in the fully guided implant surgery group and 1.48 (0.59) in the robotic system group, respectively. Based on pilot data (sCAIS group: 2.42 ± 1.55; rCAIS group: 1.48 ± 0.59), an effect size (Cohen's d) of 0.80 was estimated. For a two‐tailed t‐test with α = 0.05 and power = 0.80, a total sample size of 52 (26 per group) was required. Although our final sample size was 36 due to an inaccurate initial sample size calculation, a post hoc analysis confirmed that the achieved power to detect the large effect size observed in this study (Cohen's d = 1.49) was greater than 99%. Therefore, it is sufficient to draw meaningful conclusions.

2.2. Randomization and Allocation Concealment

An independent researcher uninvolved in the clinical procedures (LQ) generated the randomization sequence using SPSS software (IBM SPSS Statistics Inc., Chicago, IL) and prepared 36 opaque envelopes containing the group allocation. Consecutively enrolled patients who met the inclusion criteria were assigned a unique number. The researcher then placed slips indicating the corresponding group assignments (sCAIS or rCAIS) in sequentially numbered opaque envelopes. The randomization was performed at the patient level. For patients requiring two implants, both implants were placed using the modality (sCAIS or rCAIS) to which the patient was assigned. The surgeon (SK) opened the designated envelope immediately before virtual planning to determine the group assignment and designed the surgical plan accordingly. Blinding of the surgeon and surgical assistant is not feasible because of their active roles in performing distinct surgical procedures. Similarly, technicians analyzing postoperative implant positional deviations could not be blinded because the inherent characteristics of the sCAIS and rCAIS systems were readily discernible.

2.3. Study Population

Patients who required dental implant prostheses for a single missing tooth for at least 3 months and had an adequate bone volume for implant placement were invited to participate in this study. Each patient underwent cone‐beam computed tomography (CBCT) for initial screening and evaluation. The patients were allocated to each of the two groups using a computerized random number generator and screened according to the following inclusion criteria: (i) informed consent signed voluntarily, (ii) aged 18 years or older, (iii) one or two separate, single tooth losses in the posterior maxillary or mandibular region (premolar/M), (iv) sufficient bone volume for implant placement without any bone augmentation, (v) a minimum of seven residual teeth, and (vi) good communication with the physician and good compliance with the physician's advice regarding oral hygiene. The exclusion criteria were: (i) uncontrolled systemic diseases such as diabetes mellitus, hypertension, and bleeding disorders that would interfere with dental implant surgery, (ii) bisphosphonate treatment, (iii) untreated, uncontrolled periodontal disease, (iv) heavy smoker (more than 10 cigarettes per day), (v) poor oral hygiene, and (vi) unable to undergo CBCT due to pregnancy or planning for pregnancy.

Each patient underwent the insertion of one or two dental implants using the surgical modality to which they were assigned.

2.4. Preoperative Preparation and Virtual Planning

All patients in both groups underwent a preoperative CBCT and a digital intraoral scan (TRIOS, 3‐Shape, Copenhagen, Denmark). All virtual surgical plans for both the sCAIS and rCAIS groups were created by a single operator (Dr. SK), a qualified surgeon with over 10 years of experience in implantology.

Preoperative preparation of the sCAIS group involved the following: GuideMia software was used for implant planning and integration of CBCT and intraoral scan data (Figure 1). In accordance with the prosthetic‐driven principle, after selecting a virtual prosthesis from the software database, an implant was selected and a 3D implant placement was performed in the reconstructed CBCT/intraoral scan model. Implant parameters (length, diameter, and 3D position) were determined based on the virtual prosthesis position, occlusion, and biological principles. The surgical guides for the sCAIS group were digitally designed based on the fused 3D jaw model and the virtually planned implant positions. The design process, performed in the GuideMia software, involved defining the guide's anatomical coverage to achieve a stable and passive fit on the supporting teeth and/or mucosa. A virtual drill path report was generated and reviewed to confirm the clearance and accuracy of the planned trajectory. Upon final approval by the operating surgeon, the design was exported as a Standard Tessellation Language (STL) file. The guides were then additively manufactured using Stereolithography (SLA) technology. Post‐processing included washing in isopropanol to remove residual resin and post‐curing with UV light to achieve optimal mechanical properties.

FIGURE 1.

Surgical process of static computer‐assisted implant surgery.

For preoperative preparation of the rCAIS group, CBCT scans and intraoral scans were obtained using the same equipment and settings as in the sCAIS group. The data were then transferred into DentalNavi software (Beijing YaKebot Technology Co. Ltd., Beijing, China) for 3D reconstruction and implant planning. The use of two distinct software platforms was necessitated by a technical incompatibility, as the proprietary file format and closed architecture of the YakeRobot system did not allow for the import of surgical plans from the Guidemia software. The surgeon (SK) designed the implant parameters (length, diameter, and position) based on clinical, anatomical, and prosthetic considerations (Figure 2). The surgical steps for the robotic arm, including the drilling sequence, starting position, and depth, were planned.

FIGURE 2.

Surgical process of robotic computer‐assisted implant surgery.

The same prosthetic‐driven planning parameters and safety criteria (e.g., distance to vital structures, optimal implant size and position for the final restoration) were applied uniformly to all patients during the virtual planning phase, irrespective of their allocation to the sCAIS or rCAIS group. All implants and abutments used in this study were OsseoSpeed EV (Astra Tech Implant System, Dentsply Sirona, Hanau, Germany), HealDesign EV (Astra Tech Implant System, Dentsply Sirona, Hanau, Germany), and TiDesign EV (Astra Tech Implant System, Dentsply Sirona, Hanau, Germany). Each patient underwent a case‐specific surgical protocol, which provided detailed intraoperative guidance to the surgeon.

2.5. Surgical Procedures

2.5.1. Implant Placement With sCAIS

The position and stability of the surgical guide were verified before implantation to avoid placement errors. The patient was then administered local anesthesia (articaine 4% with epinephrine 1:100,000), and a full‐thickness soft tissue flap was elevated. A fully guided surgical protocol was utilized. Osteotomy preparation and implant placement were performed using the digital surgical guide in accordance with the guided surgery protocol of the Astra system (Figure 1).

2.5.2. Implant Placement With rCAIS

No fiducial markers were required during the preoperative CBCT acquisition. Instead, a patient‐specific reference template, fabricated preoperatively, was securely fixed to the adjacent healthy teeth to the implant site at the beginning of the surgery. Following local anesthesia, incision, and flap elevation, the patient‐specific intraoral reference template was positioned. To ensure its stable and precise seating, a gun‐type hybrid injection technique was used to apply temporary crown resin uniformly at the connection interface. This securely attached template, which contains fiducial markers, served as the dynamic spatial benchmark for the robotic system. The spatial position of the robotic arm was recorded using an end‐effector marker. The handpiece was then fitted with a calibration plate for drill‐tip calibration. Using a calibrated registration probe tracked by the navigation system, the surgeon sequentially and precisely touched the center of each fiducial marker on the physical template. This process established a spatial transformation between the preoperative CBCT dataset (virtual space) and the patient's actual anatomy (physical space), enabling the robot to align the planned implant trajectory with the surgical site.

Next, the robotic arm autonomously reached the implant site and performed the implant osteotomy according to the predefined surgical plan. After each step of osteotomy preparation, the robotic arm was automatically moved outside the mouth for drill changes. The surgeons were only required to control the robot via the foot pedals during the entire procedure; they monitored real‐time drilling parameters (position, depth, orientation, and force) and adjusted the plans as necessary. Finally, the robotic arm placed the implant automatically (Figure 2).

2.5.3. After Implant Placement

Both groups, following implant placement, the implant stability quotient (ISQ) were measured by performing resonance frequency analysis (RFA) and insertion torque value (ITV) were meatured by a wrench (Astra Tech Implant System, Dentsply Sirona, Hanau, Germany). Next, a cover screw was placed for submerged healing. The flap was sutured tension‐free using 5‐0 non‐absorbable sutures (Prolene, Ethicon). Postoperative medication was standardized. All patients were prescribed a prophylactic antibiotic (amoxicillin 500 mg three times daily for 5 days). Postoperative analgesics were not part of the routine prescription protocol and patients were instructed not to take additional over‐the‐counter pain medications without consulting the study team. Patients were instructed to rinse with 0.12% chlorhexidine gluconate mouthwash twice daily for 2 weeks. Sutures were removed (e.g., 7–10 days) postoperatively. Second‐stage surgery to expose the implant and place a healing abutment was performed (e.g., 3–4 months) after the initial implant placement.

2.6. Outcome Measurement

2.6.1. Accuracy Outcomes

All patients underwent immediate postoperative CBCT using preoperative parameters. The preoperative Computer‐Aided Design (CAD) surgical plan and the postoperative DICOM data were imported into the same DentalNavi software. This integrated workflow eliminated the need for file format conversions or data transfer between disparate software programs, thereby minimizing a potential source of error. The software's built‐in reference point registration function that matching anatomical landmarks (cusps, fossae, foramina, etc.) was utilized to perform a 3D spatial matching, achieving precise superimposition of the virtual implant from the preoperative plan and the actual implant from the postoperative CT scan. The software automatically segmented the implant from the CBCT data to generate an initial position. Due to software interoperability limitations, the surgical plan from GuideMia (sCAIS) could not be directly imported into Dental Navi (used for accuracy measurement). Therefore, for the sCAIS group, the postoperative DICOM data (with the placed implant) and the original planning DICOM data were both imported into Dental Navi. The planned implant position was then manually re‐created within Dental Navi by aligning the dataset with the preoperative plan as the reference, allowing for subsequent deviation analysis. This automated output was then reviewed by an operator. If any discrepancy between the generated model and the actual implant imagery was observed, manual adjustments were performed to ensure the virtual model's position and angulation precisely matched the clinical reality in all anatomical planes. Following this co‐registration, the software's postoperative analysis module automatically calculated the distance deviations (coronal/apical: global, buccolingual, mesiodistal, and apicocoronal) and angular deviations between the planned and actual implant positions (Figures 3 and 4). The postoperative analysis to determine the deviation between the planned and actual implant positions was performed independently three times for each implant by a single researcher (LZL). The mean value of these three independent measurements was subsequently calculated and used as the definitive value for all statistical analyses.

FIGURE 3.

Illustration of the parameters of deviation between the planned and actual placed implants. ① Coronal global deviation; ② Apical global deviation; ③ Angular deviation; ④ Coronal deviation in the buccolingual direction; ⑤ Coronal deviation in the mesiodistal direction; ⑥ Coronal deviation in apicocoronal direction; ⑦ Apical deviation in the buccolingual direction; ⑧ Apical deviation in the mesiodistal direction; ⑨ Apical deviation in the apicocoronal direction; ⑩ Buccolingual angular deviation; and ⑪ Mesiodistal angular deviation.

FIGURE 4.

Postoperative evaluation of the implant position. The green profile indicates the planned implant, and the red profile indicates the actual placed implant.

2.6.2. ITV and ISQ

After installing the implant, ITV was evaluated using the wrench (Astra, D. sinroma). Next, a SmartPeg was screwed into each implant and RFA was performed using Osstell Mentor (Osstell ISQ, Osstell AB, Göteborg, Sweden). The ISQ‐1 was then measured twice in both the buccolingual and mesiodistal directions and the minimum value was recorded. The second implant stability measurement (ISQ‐2) was performed seven days after the second surgery. Both ISQ‐1 and ISQ‐2 were measured using the same SmartPeg device.

2.6.3. Surgical Time

The total surgical time (TST) was calculated as duration from the first incision to suturing completion. In the rCAIS group, the TST included the preparation time (PT) for calibration and registration and the operative time (OT).

2.6.4. Patient‐Reported Outcomes

Patients were interviewed for patient‐reported results of preoperative anxiety, intraoperative experience, and postoperative complications.

1. Before implementing any clinical intervention, participants were assessed for dental anxiety using a modified dental anxiety score (MDAS) questionnaire supplemented with two questions focusing on dental implant surgery. Therefore, two questions were added inquiring about patients' feelings toward implant surgery and rCAIS. Patients completed the MDAS in the waiting room before surgery.

2. Intraoperative experience was immediately assessed using a postoperative questionnaire with a 10‐point numeric rating scale (NRS). Surgery duration, discomfort (specific causes), and overall satisfaction were evaluated, with “0” meaning no feelings for the item or not satisfied at all and “10” meaning very strong feelings for the item or completely satisfied.

3. Participants were assessed for postoperative pain intensity during the first week (sixth hour, first, second, third, fifth, and seventh days). Data were recorded using patients' self‐assessment on a NRS 0–10.

2.6.5. Surgeon‐Reported Outcomes

A self‐developed scale was used to assess the surgeon‐reported outcomes. The assessment content of the scale focused on the ease and convenience of the operation and the primary stability of the implant.

2.6.6. Intraoperative and Postoperative Complications

Intraoperative and postoperative complications were also analyzed, including bleeding (intraoperative/postoperative), adjacent nerve/tooth injury, excessive implant deviation, infection, wound dehiscence, and early implant failure (≤ 3 months).

2.7. Statistical Analysis

Measurements were imported into Statistical Package for Social Sciences (SPSS) version 25.0 (IBM SPSS Statistics Inc., Chicago, IL). The chi‐square test was used to compare demographic data and implant characteristics between the sCAIS and rCAIS groups. Normality of the data distribution of resulting data were calculated using the Shapiro–Wilk test. Generalized estimating equation (GEE) with unstructured working correlation structures was used to compare the difference of continuous variables between the two groups. The GEE method was used to account for the fact that repeated observations (several implants) were available for a single patient. The relationship between ITV and ISQ was assessed using Pearson's correlation analysis. The strength of the correlation was interpreted based on the correlation coefficient (r) as follows: |r| < 0.3 (negligible), 0.3 ≤ |r| < 0.5 (low), 0.5 ≤ |r| < 0.7 (moderate), and |r| ≥ 0.7 (strong). A p‐value of < 0.05 was considered statistically significant.

3. Result

3.1. Basic Characteristics

Thirty‐six patients (40 implants) who underwent dental implant surgery between April 2024 and December 2024 were included (Figure 5). The patients were randomly assigned to either sCAIS or rCAIS groups. The sCAIS group comprised three men and 15 women, and their mean age was 40.94 ± 13.35 years. The rCAIS group comprised nine men and nine women, and their mean age was 34.16 ± 9.60 years. The sCAIS and rCAIS groups underwent 21 and 19 implant placements, respectively. Table 1 summarizes the descriptive characteristics of the patients and implants.

FIGURE 5.

Consolidated Standards of Reporting Trials (CONSORT) flowchart.

TABLE 1.

Demographics of patients and implants included in the study.

Group	sCAIS group	rCAIS group	p (chi‐square‐test)
No. of patients	18	18
Age
Mean ± SD (years)	40.94 ± 13.35	34.16 ± 9.60	0.08
Gender (%)
Female	15 (83.3%)	9 (50%)	0.075
Male	3 (16.6%)	9 (50%)
No. of implants	21	19
Implant diameter (%)
4.2	6 (28.57%)	4 (21.05%)	0.855
4.8	15 (71.42%)	15 (78.94%)
Implant length (%)
8	7 (33.3%)	9 (47.4%)	0.605
9	11 (52.4%)	9 (47.4%)
11	3 (14.3%)	1 (5.3%)
Implant position (%)
Premolar	4 (19.0%)	2 (10.5%)	0.756
Molar	17 (81.0%)	17 (89.5%)
Jaw (%)
Mandible	6 (28.6%)	4 (21.1%)	0.855
Maxilla	15 (71.4%)	15 (78.9%)

Abbreviations: rCAIS, robotic computer‐assisted implant surgery; sCAIS, static computer‐assisted implant surgery; SD, standard deviation.

3.2. Implant Placement Accuracy

The rCAIS group demonstrated significantly lower values of three accuracy outcome variables compared with the sCAIS group (angular deviation, 1.20° ± 0.44° and 2.38° ± 1.03°; coronal global deviation, 0.70 ± 0.44 mm and 1.02 ± 0.37 mm; and apical global deviation, 0.75 ± 0.41 mm and 1.11 ± 0.40 mm for the rCAIS and sCAIS groups, respectively) (p < 0.001, p < 0.05, and p < 0.01, respectively) (Table 2).

TABLE 2.

Comparison of the angular deviation, 3D deviation at the platform, and 3D deviation at the apex between the sCAIS and rCAIS groups.

	sCAIS group	rCAIS group	MD (95% CI)	p
	Mean (SD)	Mean (SD)
Coronal global deviation (mm)	1.02 (0.37)	0.70 (0.44)	0.37 (0.13 to 0.62)	0.002**
Apical global deviation (mm)	1.11 (0.40)	0.75 (0.41)	0.39 (0.15 to 0.64)	0.002**
Angular deviation (°)	2.38 (1.03)	1.20 (0.44)	1.195 (0.69 to 1.69)	0.000***
Coronal deviation in the buccolingual direction (mm)	0.31 (0.25)	0.25 (0.17)	0.07 (−0.03 to 1.90)	0.184
Coronal deviation in the mesiodistal direction (mm)	0.26 (0.20)	0.13 (0.11)	0.07 (−0.03 to 0.17)	0.191
Coronal deviation in apicocoronal direction (mm)	0.84 (0.46)	0.58 (0.48)	0.23 (−0.05 to 0.52)	0.118
Apical deviation in the buccolingual direction (mm)	0.42 (0.31)	0.25 (0.20)	0.109 (−0.32 to 0.10)	0.311
Apical deviation in the mesiodistal direction (mm)	0.37 (0.25)	0.19 (0.17)	0.176 (0.04 to 0.30)	0.008**
Apical deviation in the apicocoronal direction (mm)	0.84 (0.47)	0.58 (0.48)	0.232 (0.06 to 0.52)	0.125
Buccolingual angular deviation (°)	1.54 (1.07)	0.84 (0.52)	0.28 (−0.04 to 0.61)	0.093
Mesiodistal angular deviation (°)	1.61 (1.20)	0.62 (0.53)	1.26 (0.50 to 2.025)	0.001**
Depth deviation	0.84 (0.46)	0.58 (0.48)	0.23 (−0.05 to 0.52)	0.118

Note: MD adjusted according to the generalized estimating equations (GEE).

Abbreviations: 95% CI, 95% confidence interval; rCAIS, robotic computer‐assisted implant surgery; sCAIS, static computer‐assisted implant surgery; SD, standard deviation.

*p < 0.05; **p < 0.01; ***p < 0.001.

No significant differences were observed in coronal deviation in the buccolingual direction, coronal deviation in the apicocoronal direction, and apical deviation in the apicocoronal direction between the two groups (p > 0.05). However, the majority of linear and angular deviations: coronal deviation in the mesiodistal direction; apical deviation in the buccolingual direction; apical deviation in the mesiodistal direction; buccolingual angular deviation; and mesiodistal angular deviation were significantly lower in the rCAIS group than in the sCAIS group (p < 0.05) (Table 2). We further analyzed the directions of the platform and apical deviation, and a scatter plot illustrating the deviations of all 40 implants in each direction revealed that almost all implants were placed further coronal to the planned position (Figure 6).

FIGURE 6.

Scatter plots illustrating implant deviations in three dimensions. (a) Coronal deviation in the buccolingual and mesiodistal direction. (b) Coronal deviation in the buccolingual and apicocoronal direction. (c) Coronal deviation in the mesiodistal and apicocoronal direction. (d) Apical deviation in the buccolingual and mesiodistal direction. (e) Apical deviation in the buccolingual and apicocoronal direction. (f) Apical deviation in the mesiodistal and apicocoronal direction. Red dots represent the sCAIS group (n = 21); green dots represent the rCAIS group (n = 19). The red shaded circle indicates the range within which deviations in both the x‐ and y‐axis directions are ≤ 1 mm; the green shaded circle indicates the range within which deviations in both directions are ≤ 2 mm. These circles visually illustrate the precision and consistency of implant placement within each group.

3.3. ITV and ISQ

No significant differences were observed in ITV and ISQ scores between the two groups (p > 0.05) (Table 3). Both the sCAIS and rCAIS groups demonstrated a significant difference between ISQ‐1 and ISQ‐2 (p = 0.002 for sCAIS, p < 0.0001 for rCAIS) (Table 4). Considering all implants, a positive linear association was observed between the ITV and ISQ at implant placement (Pearson correlation, 0.510; p < 0.001) (Figure 7).

TABLE 3.

Comparison of the ITV and ISQ scores between the sCAIS and rCAIS groups.

	sCAIS group	rCAIS group	MD (95% CI)	p
	Mean (SD)	Mean (SD)
ITV (Ncm)	26.67 (11.10)	29.11 (11.46)	−1.14 (−8.47, 6.19)	0.093
ISQ	75.00 (4.81)	76.21 (4.21)	−0.553 (−3.30, 2.20)	0.694

Note: MD adjusted according to the generalized estimating equations (GEE).

Abbreviations: rCAIS, robotic computer‐assisted implant surgery; sCAIS, static computer‐assisted implant surgery; SD, standard deviation.

TABLE 4.

Comparison of ISQ‐1 and ISQ‐2 scores of sCAIS and rCAIS groups.

	Mean ± SD	p
sCAIS
ISQ‐1	75.00 (4.81)	0.002**
ISQ‐2	81.00 (2.62)
rCAIS
ISQ‐1	76.21 (4.21)	0.000***
ISQ‐2	81.39 (4.66)

Note: The ISQ‐1 and ISQ‐2 were measured immediately after implantation and 7 days after the second surgery.

Abbreviations: rCAIS, robotic computer‐assisted implant surgery; sCAIS, static computer‐assisted implant surgery; SD, standard deviation.

**p < 0.01; ***p < 0.001.

FIGURE 7.

Correlation between implant stability quotient (ISQ) and implant torsion value (ITV) in sCAIS (red) and rCAIS (green) groups. The solid red line represents the linear regression line fitted to all data points. The dark red shaded area around the regression line represents the 95% confidence interval for the regression estimate, indicating the precision of the fitted line. The light red shaded area represents the 95% prediction interval for individual observations, indicating the expected range of ISQ values for a given ITV. Pearson's correlation analysis revealed a moderate positive correlation (r = 0.510, p < 0.001), demonstrating that higher insertion torque values are associated with greater primary implant stability.

3.4. Surgical Duration

The surgical duration for implant placement was significantly shorter in the sCAIS group than in the rCAIS group (24.38 ± 6.52 min vs. 49.53 ± 12.96 min, p < 0.000) (Table 5).

TABLE 5.

Comparison of the total surgical time and operative time between the sCAIS and rCAIS groups.

	sCAIS group	rCAIS group	MD (95% CI)	p
	Mean (SD)	Mean ± SD
TST (min)	24.38 (6.51)	49.53 (12.96)	−1.14 (−8.47, 6.19)	0.000***
PT (min)	—	12.06 (4.28)	—	—
OT (min)	24.38 (6.51)	37.47 (10.70)	−13.97 (−19.6, −8.32)	0.000***

Note: MD adjusted according to the generalized estimating equations (GEE).

Abbreviations: OT, operative time; PT, preparation time (calibration and registration); rCAIS, robotic computer‐assisted implant surgery; sCAIS, static computer‐assisted implant surgery; SD, standard deviation; TST, total surgery time.

***p < 0.001.

3.5. Patient‐Centered Results

Tables 7 and 8 detail the PROs of the two groups.

1. Modified dental anxiety score (MDAS)

TABLE 7.

Patient‐reported intraoperative experience of the clinical treatment procedures.

	p
	Friedmann test	Pairwise test versus 6 h with Bonferroni correction
		1st	2nd	3rd	5th	7th
sCAIS	< 0.001***	0.750	0.002**	< 0.001***	< 0.001***	< 0.001***
rCAIS	< 0.001***	1.000	0.008**	< 0.001***	< 0.001***	< 0.001***

Abbreviations: rCAIS, robotic computer‐assisted implant surgery; sCAIS, static computer‐assisted implant surgery.

***p < 0.001.

TABLE 8.

Surgeon‐reported intraoperative experience of the clinical treatment procedures.

	sCAIS group	rCAIS group	MD (95% CI)	p
	Median (range)	Median (range)
The complexity of the treatment procedure	3 (3, 3)	3.5 (2.75, 4)	−0.12 (−0.27, 0.01)	0.088
The degree of mouth opening interference	3 (2.75, 4)	4.5 (3.75, 6)	−0.33 (−0.52, −0.14)	0.000***
The stability of implant placement	8 (6, 8.25)	6.5 (5, 7.25)	0.13 (0.01, 0.25)	0.24

Note: MD adjusted according to the generalized estimating equations (GEE).

Abbreviations: rCAIS, robotic computer‐assisted implant surgery; sCAIS, static computer‐assisted implant surgery.

***p < 0.001.

All patients completed the MDAS questionnaire; of the five original MDAS questions, those with the highest scores for anxiety were tooth drilling (5.15 ± 3.16), scaling (4.83 ± 3.08), injection (4.52 ± 2.83), waiting for the surgery in the waiting room (4.22 ± 2.77), and the appointment for the next day (4.16 ± 2.69). Regarding dental implant surgery, the anxiety score was 4.72 ± 2.80, with the highest score (5.33 ± 2.80) reported for procedures involving robot‐assisted surgery. No significant differences in anxiety scores were observed between the groups for any question (p > 0.05).

• 2Intraoperative experience

The majority of patients reported relatively high satisfaction with implant surgery, and no statistically significant difference was observed in overall satisfaction between the two groups (p > 0.05). However, when evaluating procedural discomfort, patients in the rCAIS group experienced significantly higher levels of discomfort than those in the sCAIS group (p < 0.05). Analysis of the frequency distribution of the four discomfort factors revealed that the primary sources of discomfort were mouth‐opening fatigue and instrument‐pulling sensations (Table 6).

• 3Postoperative pain questionnaire 1 week after surgery

TABLE 6.

Patient‐reported intraoperative experience of the clinical treatment procedures (NRS: 0–10).

Questions	sCAIS group	rCAIS group	MD (95% CI)	p
Degree of intraoperative discomfort	1.67 (1.87)	3.00 (1.60)	−1.13 (−8.47, 6.19)	0.001 a
Uncomfortable factors	—	—	—	0.570
Keep mouth open	7	17	—	—
Instrumental traction	6	5	—	—
Saliva in the mouth	3	5	—	—
Drill vibration	4	3	—	—
Overall satisfied scores	8.61 (2.68)	8.72 (2.014)	0.11 (−0.08, 0.31)	0.248

Note: MD adjusted according to the generalized estimating equations (GEE).

Abbreviations: rCAIS, robotic computer‐assisted implant surgery; sCAIS, static computer‐assisted implant surgery.

^a p < 0.05.

According to the Mann–Whitney U test results, no significant difference was observed in postoperative pain between the groups at every time point (p > 0.05) (Figure 8). Both groups demonstrated a consistent decrease in pain scores as the recovery period progressed. Notably, we observed a marked reduction in pain scores in both groups from the second postoperative day (p < 0.001) (Table 7).

FIGURE 8.

Postoperative pain scores over the first 7 days following implant surgery. Box plots show NRS pain score distributions for sCAIS (red) and rCAIS (green) groups at 6 h, 1, 2, 3, 5, and 7 days. p‐values indicate between‐group comparisons (all p > 0.05).

3.6. Surgeon‐Reported Outcomes

Upon evaluating the complexity of the surgical procedures and the stability of implant placements in both groups, we observed no statistically significant differences (p > 0.05) in surgeon‐reported outcomes between the two groups. However, a statistically significant difference was observed in the impact of patients' mouth opening on surgery (p < 0.01) (Table 8).

3.7. Surgical Morbidity and Complications

No patient in the sCAIS group reported adverse events. However, one implant in the rCAIS group failed to osseointegrate and was removed at the three‐month postoperative follow‐up. All other implants osseointegrated evenly. Additionally, one case in the rCAIS group required an intraoperative modification of the pre‐operative plan due to significant difference between the actual bone quality and that anticipated on the CBCT.

4. Discussion

In this randomized controlled trial, the implant accuracy, surgical duration, and patient‐ and surgeon‐reported outcomes were analyzed and compared between the sCAIS and rCAIS. The results indicate that both technologies provide accurate placement of a single implant in relation to the planned position, and that rCAIS facilitates a more accurate implant position than sCAIS. However, the rCAIS approach requires a longer surgical duration and results in worse intraoperative experience in patients than the sCAIS.

In this study, implant deviations in the sCAIS group (coronal: 1.02 ± 0.37 mm; apical: 1.11 ± 0.40 mm; angular: 2.38° ± 1.03°) were significantly greater than those in the rCAIS group (coronal: 0.70 ± 0.44 mm; apical: 0.75 ± 0.41 mm; angular: 1.20° ± 0.44°), which is consistent with existing literature (Bai et al. 2021; Jia et al. 2023; Khaohoen et al. 2024). In an in vivo and in vitro study involving the same robotic systems, the rCAIS demonstrated mean angular, coronal, and apical deviations of 1.01° ± 0.87°, 0.58 ± 0.60 mm, and 0.58 ± 0.60 mm (in vitro) and 0.95° ± 0.50°, 0.45 ± 0.28 mm, and 0.47 ± 0.28 mm (in vivo), respectively, which were significantly lower than those in the sCAIS group (He et al. 2024). The majority of linear and angular deviations were significantly lower in the rCAIS group than in the sCAIS group. Compared with rCAIS, sCAIS resulted in a greater deviation at the implant apex than at the platform, both buccolingually and mesiodistally. This discrepancy likely stems from the angular deviation of the sCAIS being nearly twice that of the rCAIS, which amplifies the apical deviation from the coronal site. Scatter plots confirmed more dispersed apical deviations on the sCAIS, which is consistent with other studies (Arısan et al. 2010; Koop et al. 2013). This phenomenon might be attributed to the diminishing restrictive effect of the surgical guide as the drill advances after penetrating the tissue. In contrast, the rCAIS utilizes real‐time navigation for dynamic drill path correction. Additionally, in the sCAIS group, the relatively large tolerance between the metal guide sleeve and drill may have contributed to greater horizontal deviations at the implant apex than at the corona. As the tolerance between the drill tip and guide sleeve increases, the horizontal deviations also tend to increase (Koop et al. 2013). Although we observed no significant differences in the distribution of deviation in direction between the two groups, the scatter plots of the rCAIS group were more concentrated, indicating superior performance in angular control.

An interesting finding of this study was the systematic tendency for the majority of the implants to be placed slightly shallower than planned. This observation warrants further discussion. We hypothesize that this coronal shift could be attributed to several technical factors. In sCAIS, surgeons may exhibit a natural caution to avoid over‐preparation, particularly when approaching vital anatomical structures like the inferior alveolar nerve or maxillary sinus. This caution could manifest as a subconscious tendency to under‐prepare the osteotomy depth. Additionally, the tactile feedback during the final threads of implant insertion might lead the surgeon to stop slightly short of the planned depth to avoid excessive insertion torque, especially in denser bone qualities. In rCAIS, the observed shallower placement may be attributed to the rigid arm of the robotic system lacking the tactile feedback that allows human surgeons to make minute compensatory adjustments during final implant placing (Wu et al. 2023; Singthong et al. 2024).

Primary stability can be affected by implant characteristics, bone quantity and quality, and surgical techniques (Blume et al. 2021; Jenner et al. 2025; Farronato et al. 2020). A dual‐parameter evaluation system was utilized to systematically record IT and ISQ during surgery, thereby enabling a comprehensive assessment of initial implant stability. After a waiting period of 3 months, both the sCAIS and rCAIS groups demonstrated sufficient stability at the second assessment. The results confirmed that both techniques reliably helped achieve initial stability in the posterior region. The implants used in this study were Astra Tech EV implants with a small thread pitch design, enhancing primary stability through increased bone‐implant contact area, which is consistent with in vitro findings comparing EV to NA and Straumann BLT implants (Karl and Irastorza‐Landa 2017). Furthermore, Pearson's correlation analysis revealed a strong positive correlation between the ITV and ISQ, which is consistent with the findings of other studies (Farronato et al. 2020; Bergamo et al. 2021).

Although most studies emphasized implant positioning accuracy, the surgical duration was also quantified. The operative time (from first drilling to last suture) was significantly longer in the rCAIS group than in the sCAIS group. This difference may be due to minor intraoperative head movements. Although the robotic system is equipped with a “real‐time tracking” system that dynamically detects and corrects the patient's head movement trajectories, this process requires additional time. Currently, many dental implant‐guided systems consist of surgical guides, drill handles, and drills. In this study, the ASTRA guide system used in sCAIS integrates the drill sleeve directly, eliminating drill handle adjustments and enabling nondominant hand stabilization of the mandible. This enhances the efficiency and reduces systematic errors between the drill and guide sleeve. However, excessive head movement during robotic procedures necessitates re‐registration, underscoring the critical need for effective head and mandible stabilization. Loose reference markers or significant deviations requiring recalibration. Additionally, the rCAIS is an emerging technology that effectively utilizes the rCAIS and relies on the acquisition of appropriate foot‐pedal‐hand‐eye coordination during surgery, which entails a substantial learning curve. The operative duration is anticipated to decrease with the proficiency of surgeons utilizing rCAIS.

Average operative times for both the sCAIS and rCAIS groups were slightly longer than those indicated by other studies (sCAIS group: 14.5 ± 2.48 min; rCAIS group: 21.5 ± 9.3 min) (Chen et al. 2023; Wu et al. 2021). The prolonged operative time may be due to the following factors: (i) the flap approach used in this study required additional time for incision, flap elevation, and suturing; and (ii) intraoperative imaging and recordings were performed to obtain comprehensive clinical data, which also contributed to the extended operative time. Furthermore, one rCAIS case required an intraoperative plan modification, which increased the operative time. Nevertheless, this also demonstrates the high flexibility of the rCAIS. The real‐time feedback and dynamic control of rCAIS allowed for immediate adjustment of the implant position and drilling parameters, enabling the procedure to continue as planned. In contrast, such an intraoperative discrepancy in the sCAIS group would likely necessitate a deviation from the static guide, potentially converting the procedure to a freehand approach, which could be considered a protocol deviation or dropout. This highlights a significant clinical advantage of rCAIS: its adaptability to unanticipated anatomical variations, thereby enhancing procedural safety and potentially reducing the need for salvage interventions.

Furthermore, the integration of Patient‐Reported Outcome Measures (PROMs) and Patient‐Reported Experience Measures (PREMs) is increasingly recognized as an essential component of a comprehensive clinical trial design (Casaca et al. 2023). These data offer a holistic view of treatment effectiveness and are critical for aligning clinical practice with patient values and priorities, ultimately guiding improvements in healthcare delivery and shared decision‐making. So patient‐centered results were assessed, including preoperative anxiety, intraoperative experience, and postoperative pain, obtained from questionnaires. Regarding to preoperative anxiety, no significant difference in preoperative anxiety was observed between the two groups. Notably, patients reported significantly higher anxiety scores for the rCAIS, potentially attributable to cognitive biases arising from limited public awareness of this novel technology. Therefore, targeted patient education in robot‐assisted implantology is clinically significant in reducing anxiety. According to the intraoperative experience questionnaires, the discomfort level was significantly lower in the sCAIS group than that in the rCAIS group. This is consistent with the results of another clinical randomized controlled trial (Shi et al. 2025). Patient feedback revealed that prolonged mouth opening (mean duration > 45 min) and perceived mechanical traction as the primary sources of discomfort. This difference may stem from the requirement of sustained maximum mouth opening during posterior rCAIS procedures to accommodate the trajectory of the robotic arm. Consequently, robotic‐arm designs should prioritize miniaturization and enhanced flexibility to reduce intraoral spatial demands and minimize patient discomfort. Although no differences in postoperative pain was observed between the two groups, both groups demonstrated a significant decrease in pain over time. Patients reported significant pain relief on 2nd days postoperatively compared with 6 h postoperatively, which aligns with the normal healing process of the peri‐implant soft tissues. However, at the sixth hour and first day postoperatively, the pain scores tended to be slightly higher in the rCAIS group than in the sCAIS group, which might be associated with longer operative times. The flap surgical technique used in this study prolonged the wound exposure time, thereby potentially increasing the risk of local tissue damage. Considering that we included single implant surgery without complex procedures only (e.g., bone augmentation and sinus lift), the overall postoperative pain was mild, with minimal analgesic requirements.

Regarding the surgeon‐reported outcomes, both techniques offer similar operational convenience and satisfaction. However, posterior anatomical constraints limit robotic‐arm positioning, necessitating maximum mouth opening for proper implant placement. Shi et al. (2025) reported limited accessibility of the rCAIS in posterior surgeries. Fatigue‐induced mandibular movement or reduced opening increases the risk of collision with adjacent teeth, requiring intraoperative robotic trajectory recalibration. Conversely, sCAIS allowed for manual mandible stabilization and reduced continuous mouth opening time (by 15–20 min), demonstrating superior clinical adaptability.

Although computer‐assisted technologies including rCAIS have helped achieve automation, have advanced implant placement accuracy, efficiency, safety, and visualization, clinician expertise remains paramount for comprehensive diagnosis and treatment planning. This requires a multidimensional analysis of anatomical (bone volume and neurovascular anatomy), biomechanical (occlusal forces), aesthetic (gingival phenotype), and patient‐specific factors (systemic health and compliance). Building on advances in AI and deep learning for data‐driven treatment optimization (Chen and Chen 2024), in this study, we aimed to develop a fully digital, evidence‐based decision‐support framework for implantology. To synthesize the key outcome measures and contextual factors relevant to our study, a conceptual framework was developed (Figure 9). The construction of this framework was based on a review of core outcome sets in implant dentistry and clinical trial methodology literature (Tonetti et al. 2023). This system integrates multimodal data including patient demographics (age and missing tooth type), clinical measurements (ISQ, deviation values, and operative time), and PROs (Figure 9). Leveraging machine learning algorithms to analyze this integrated data, the system aims to recommend the optimal computer‐assisted implant surgery (CAIS) modality (sCAIS, dCAIS, or rCAIS) for specific clinical scenarios (e.g., single posterior tooth loss, aesthetic zones, insufficient bone volume) and provide personalized surgical parameters. This facilitates a shift from experience‐driven towards evidence‐based clinical decision making.

FIGURE 9.

A comprehensive digital decision‐making and evaluation system for oral implantology. This schematic illustrates the multidimensional factors considered in evaluating the success of computer‐assisted implant surgery. The framework encompasses both Surgeon‐reported outcomes (e.g., accuracy, implant position, initial primary stability) and Patient‐reported outcomes (e.g., satisfaction, oral health‐related quality of life), whereas also accounting for influential patient‐level variables (e.g., number of tooth loss, time duration, age). The diagram was created to visually synthesize the key domains identified as critical to comprehensive outcome assessment in our study and the broader literature. It was developed using Adobe Illustrator CC 2023.

This study had some limitations. First, the small sample size (n = 36) may have limited the statistical power. Second, the accuracy assessment relied on postoperative CBCT registration rather than intraoral scan data, presenting two concerns: (i) CBCT quality impacts data superimposition and coordinate extraction, with IOS demonstrating superior accuracy in prior studies (Derksen et al. 2019); and (ii) additional CBCT exposure may conflict with ALARA (as low as reasonably achievable) principles (Hämmerle et al. 2015). Third, due to the technical constraints of the integrated robotic system, which did not allow for the import of external surgical plans, we were required to use two different software platforms for preoperative planning. Although this introduces a potential confounding variable, as differences in software algorithms and user interfaces could theoretically influence the planning process itself, we implemented a blinded, independent assessment of all primary outcomes to minimize its impact on the results. Consequently, the reported differences in surgical accuracy are attributable to the execution of the surgical technique rather than to the measurement method. Future technological developments enabling cross‐platform compatibility would be valuable to eliminate this variable in subsequent comparative studies. Additionally, the assessment of postoperative pain may be subject to confounding, as patients were not prescribed routine analgesics but were not actively monitored for unreported over‐the‐counter pain medication use. Although no patients reported taking analgesics, the possibility of undisclosed self‐medication cannot be ruled out. Finally, long‐term outcomes (e.g., ≥ 5‐year implant survival and complication rates) are warranted to fully evaluate implant survival rates between sCAIS and rCAIS.

Although the rCAIS demonstrates superior implant position accuracy over the sCAIS, its clinical implementation faces challenges. The complex robotic workflow necessitates ≥ 3 personnel, increasing procedural complexity and costs, and not yet standardized. Furthermore, one implant in the rCAIS group failed to osseointegrate. The precise cause of this failure remains undetermined. The role of haptic feedback (or its absence) in such failures warrants careful consideration. Although advanced robotic systems incorporate engineered safeguards against thermal injury in vitro (Zhao et al. 2024; Liu et al. 2024), the translation of these protections to the clinical setting, where bone density is heterogeneous and surgical conditions are variable, remains an area of active investigation. It is plausible that the inability to perceive subtle tactile cues in vivo could, in certain edge cases, contribute to suboptimal surgical execution that is not solely related to heat generation but could also affect primary stability. This hypothesis, however, requires direct validation in future clinical studies designed to correlate intraoperative robotic parameters with histological and clinical outcomes. This event underscores that both techniques, despite their overall high accuracy and success, are not without risk. Future studies with larger cohorts are necessary to fully understand the factors influencing long‐term survival in guided implant surgery.

This study focused on a relatively simple clinical scenario (single tooth loss without bone defects). Future research should investigate complex cases (e.g., multiple missing teeth, bone augmentation, and aesthetic zones) to strengthen the clinical evidence. rCAIS may prolong the duration of surgery and increase costs, thereby potentially affecting patient acceptance. Multicenter studies should emphasize patient‐centered outcomes and socioeconomic analyses, including cost‐effectiveness analyses, to fully evaluate the clinical value and broader applicability of the rCAIS.

5. Conclusion

In this study, both the sCAIS and rCAIS were found to achieve clinically acceptable accuracy in guiding 3D implant placement, with the rCAIS demonstrating superior accuracy. Although both systems achieved satisfactory primary stability, the rCAIS requires longer operative times. The sCAIS offers better surgical accessibility, reduces procedure duration, and enhances patient experience. Future research should focus on streamlining rCAIS workflows to optimize treatment outcomes and patient experience.

Author Contributions

Zhilin Luo: writing – original draft, writing – review and editing, conceptualization, investigation. Wantong Zhou: investigation, software. Min Wang: investigation; software. Lianyi Xu: conceptualization, methodology, formal analysis. Xijin Du: conceptualization, methodology, formal analysis. Yingguang Cao: conceptualization, methodology, investigation. Ke Song: conceptualization, methodology, investigation, funding acquisition, supervision.

Funding

This was part of an “investigator‐initiated study,” and dental implants were delivered free of charge by DENTSPLY SIRONA Implants (Mondal, Sweden).

Disclosure

We wish to disclose that a conference abstract related to this work was presented as a poster at the Digital Dentistry Society (DDS) Annual Meeting (2025) and published in the Journal of Dentistry: Clinical Applications (Supplement #48). The current manuscript represents the complete study with full methodology, results, and discussion, which has not been published elsewhere.

Ethics Statement

The study approved by the Tongji Hospital Institutional Review Board (TJ‐IRB20231192) complied with the Declaration of Helsinki.

Consent

Investigators obtained informed consent before enrolling participants in clinical trials.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Appendix S1: clr70138‐sup‐0001‐AppendixS1.pdf.

Data Availability Statement

The data are available from the corresponding author upon reasonable request.