Comparison of Accuracy and Time for Four Implant Placement Techniques Supporting Fixed-Partial Denture

Implant placement is based on the anatomy of the alveolar ridge, adjacent anatomical structures, and function and future restoration. To optimize implant placement, various guiding techniques have been introduced. Guides enable predictable implant placement based on preplanned restorations, reduced surgical time, less complications, and accurate placement with more favorable long-term success rate.^1–4 

Template-based guides (TBG) are conventional surgical guides based on a diagnostic wax-up and often serve first as a radiographic guide and are then transformed into a surgical guide. These guides are less accurate as they are usually fabricated on dental casts with no direct information of underlying bone and anatomical structures.⁵  Other limitations include time-consuming laboratory procedures and introduction of setting errors and positional artifacts. With cone beam computerized tomography (CBCT) and the digitalization of dentistry, computer-guided static and dynamic navigation systems were developed.^6–11  These systems incorporate three-dimensional (3D) radiographs and virtual planning of implants and prosthetic suprastructures. Static guides (SG) are fabricated using computer-aided design/computer-aided manufacturing (CAD/CAM) technology and require special surgical kits for implant placement.¹²  Static guides offer more accurate implant placement based on predetermined virtual plan and underlying bone anatomy but do not allow for intraoperative adjustments. Other disadvantages include long fabrication time and cost, and limited osteotomy visualization during the procedure. They can be difficult to accommodate in patients with limited mouth opening or limited space between teeth.^4–13 

Dynamic navigation (DN) is based on a real-time optical tracking of the jaw and the handpiece. The guidance system interactively displays with real-time feedback on the computer screen of information of the drill/implant position, depth and angulation throughout the procedure and according to the virtual treatment plan.^1–4  Dynamic navigation allows for intraoperative modifications during the osteotomy preparation without compromising the accuracy and offers optimal visualization of the surgical drill and implant site. It can be used on patients with limited mouth opening and limited space between adjacent teeth enabling improved ergonomics.^2,6,7 

Several studies have assessed and compared the accuracy of different guiding techniques. Though the accuracy of freehand (FH) implant placement may be sufficient for most clinical situations,¹⁴  guiding techniques provide significantly more accuracy.^2–7  The accuracy of dynamic navigation is comparable⁷  or even superior¹  to the accuracy of static guidance. Both static and dynamic guided methods are significantly more accurate than a TBG.¹⁵  Though experienced operators rely on their experiences when using less accurate methods such as FH or TBG, inexperienced practitioners entering into the field of implantology do not have that advantage. Implant placement without reliable guidance can result in large deviations from optimal implant positions, which can be particularly challenging to overcome when restoring multiple implants supporting a fixed partial denture (FPD).

This ex vivo pilot study aimed to evaluate efficiency and accuracy comparing 4 different implant placement techniques (DN, SG, TBG, and FH) in various locations on a simulation model for inexperienced operators. The hypothesis was that static and dynamic guidance techniques result in more accurate implant placement for multiple implants when used by novice operators. The secondary hypothesis was that the efficiency and accuracy would improve across consecutive attempts for all 4 techniques.

This study complied with the World Medical Association Declaration of Helsinki and the Code of Medical Ethics of Virginia Commonwealth University. The University Institutional Review Board approved the study protocol (IRB: HM20011878). Four senior dental students (2 females and 2 males) with no previous experience in dental implant placement participated in the study.

Operators were introduced to the study protocol and equipment including a review of implant placement and placement methods used in the study. Two implants for a 3-unit FPD were planned using 4 techniques: (1) Freehand (negative control) (FH); (2) Template-based guide (TBG); (3) Static guide (SG); and (4) Dynamic navigation (DN). Thirty-two maxillary and mandibular polymethylmethacrylate (PMMA) 3D were used. To provide balanced implant site distribution, maxillary model was missing left lateral incisor, canine, and first premolar while the mandibular model was missing both right premolars and a first molar (Figure 1). Treatment plan was based on the consultation with a board certified periodontist and oral and maxillofacial surgeon. Each operator placed implants in 2 maxillary and 2 mandibular models using each of the 4 methods totaling 32 implants into 8 maxillary and 8 mandibular models. All 4 operators followed the randomization schedule (computer software SAS EG version 6.1, SAS Institute, Cary, North Carolina, USA) for jaw model and placement method.

A diagnostic wax-up was performed on jaw models simulating 3-unit FPD restorations on 2 implants (Figure 1). A thermoplastic template was trimmed to the marginal gingival level providing sufficient support on the adjacent teeth (Figure 2a). A hole was drilled through the occlusal surface of the thermoplastic template according to the ideal position of the access hole of the planned implant crown. Implant position was planned under the proposed restoration and marked on the duplicated jaw model (Figure 2b). An optimal implant insertion line was determined according to the marked position, access hole, and axis of the adjacent tooth and a planned implant using a laboratory surveyor (Figure 2c). Insertion lines for both implants supporting the FPD were paralleled and gutta-percha bars were secured inside the thermoplastic template (Figure 2d). Proposed implant crowns were filled with acrylic resin (Biocryl ICE, Great Lakes Dental Technologies, Tonawanda, New York), polymerized, adopted to the PMMA jaw model and polished (Figure 2e and f).

All PMMA models were tagged with 4 fiducial markers: 2 on the buccal aspect apically to the canines and 2 on the posterior aspect of the model (Figure 2g). They ensured the 3-dimensional orientation of the CBCT scan and enabled accurate superimposition for analysis. A preoperative CBCT scan (iCAT FLX V10, Imaging Sciences International LLC, Kavo, Germany) of the model with a radiographic guide was taken using settings: voxel size of 0.2 mm, 7.4-second exposure time, 120 kVp, 5 mA, and 611.6 mGy/cm². One radiographic guide was used for 4 identical jaw models' CBCT scans. All CBCT scans were imported into the Navident (ClaroNav, Ontario, Canada) dynamic guidance system software. Appropriate implant sizes (diameter and length) were selected and virtually planned in optimal positions according to the radiographic guides. Jaw models were marked according to the guidance method and randomization schedule and secured in the manikin head on a dental chair to simulate clinical settings (Figure 2h).

Virtual plans of implants on CBCT were available throughout the FH attempts for visual reference and guidance. The necessary measurements were taken on the computer screen and the implants were placed without using any physical or 3D guidance.

The radiographic guides were transformed into the TBGs by removing gutta-percha, widening the access hole to the 3-mm diameter, and opening access to the buccal surface of the mandibular and the palatal surface of the maxillary guides. Template-based guides enabled good visual control of the depth, angulation, and access for narrow pilot drills to the desired depth of the osteotomy (Figure 2g). Due to increasing drill diameter, the enlargement of the osteotomy and implant placement was performed without the guide.

Models for this group were scanned with an intraoral scanner (Trios, 3Shape, Copenhagen, Denmark). Model scan and CBCT scan with radiographic guide were imported into a computer software for guided surgery and implant planning (Implant Studio, 3Shape). Virtual implant planning was performed according to the radiographic guides. Surgical guides were virtually designed and exported as standard tessellation language (STL) format files into 3D printing computer software (PreForm, Formlabs, Somerville, Massachusetts, USA) for print layout and adding the support structure to the guides. Fabrication was done using a 3D printer (Form 2, Formlabs) and surgical guide resin (Formlabs) (Figure 3a). Guides were washed in 99% isopropyl alcohol for 10 minutes (Form Wash, Formlabs), dried and post-cured for 30 minutes at 60°C (Form Cure, Formlabs), the support structures were then removed, guides were polished and metal sleeves (Zimmer Biomet, Warsaw, Indiana, USA) were inserted into the guide to enable rigid drill stabilization (Figure 3b). The guides were autoclaved for 20 minutes at 121°C. For implant placement, a guided surgical kit and tube adapters (Zimmer Biomet) were used according to the manufacturer's guidelines (Figure 3c and d).

All models were tagged with 4 fiducial markers, and a preoperative CBCT scan with a radiographic guide was taken (Figure 4a). CBCT files were imported into dynamic navigation system software and implants were virtually planned (Figure 4b). Handpiece and jaw were equipped with tags for tracking (Figure 4c) and the computer-guided dynamic navigation device was placed in front of the operator, enabling good visualization of the computer screen displaying a real-time drill position and angulation in relation to the planned implant position (Figure 4d). A tracking stereoscopic camera was secured above the operating field tracking a spatial relationship between the jaw-tag secured on a model, drill-tag on the handpiece, and the tag on the tracing tool (Figure 4e). Registration of the jaw model to its on-screen CBCT scan was done by selecting 6 landmarks, locating them on the model and tracing around them using a tracer tool. The registration ensures accurate positioning and continuous tracking of the jaw during the navigated procedure (Figure 4f). Handpiece and each drill were calibrated prior to use using a calibrator. Accuracy of registration was checked by placing a tip of the calibrated drill on the cusp tip of the neighboring tooth and validating its on-screen location. In case of inaccuracy, the registration can be repeated on the model. During the implant site preparation, real-time video feedback in relation to planned implant position guides the drill positioning. The position, depth, and angulation of the drill is continuously tracked and related to the planned implant position. The interactive display on the screen provides accurate visual feedback through color-coded graphic software alerting the operator of position and accuracy (Figure 4c–f).

Efficacy was evaluated as the time required for:

• Preoperative time: presurgical procedures according to placement method used:

  • FH: CBCT, diagnostic wax-up, radiographic guide preparation, virtual implant planning

  • TBG: CBCT, diagnostic wax-up, radiographic guide preparation and transformation, virtual implant planning

  • SG: CBCT, model scanning, virtual prosthetic and implant planning

  • DN: CBCT, virtual prosthetic and implant planning

• •

  Operative time: time spent on the actual implant placement. Additionally, calibration time was measured during operative time for DN as this method requires registration and tracking of the jaw and instruments as part of implant placement procedure.

Time was measured for every attempt using all 4 methods. The average time per implant and per model was calculated for each placement method separately and compared among methods and operators. The methods for measuring time were objective and did not require inter/intra-rater reliability measures. Time was measured independently by the third person, using stopwatch application of Apple smartphone. For each operative and preoperative time starting point and end point were standardized and objectively used through all measurements. Each procedure and time measurement started after the start sign by the person recording time and finished when the procedure was completed and instrument placed back on the table.

Accuracy was assessed by comparing the positions of virtually planned to the actually placed implants by superimposing a preoperative CBCT scan (Figure 4b) and postoperative CBCT scan of the same model and calculating deviations (Figure 4g) Preoperative and postoperative CBCT scans were taken using the same settings and superimpositions were performed by aligning the fiduciary markers on the CBCT scans. Four deviations between planned and placed implant positions (Figure 4h) were calculated digitally using EvaluNav software (ClaroNav, Ontario, Canada):

• Entry deviation (lateral 2D)—lateral deviation of the cervical part of implant

• Apical vertical deviation (V)—height deviation between apex of planned and placed implant

• Apical 3D deviation—deviation between apexes of planned and placed implant taking into account entry, apical vertical, and angular deviation

• Angle deviation—deviation of the long axis of placed implant according to planed position

Average deviations were calculated for each implant, each placement method, and each jaw model and were compared among implant sites, methods, and operators. Deviation measurements were objective since they were calculated by the software and did not require assessment of inter/intra-rater reliability.

Procedure times and deviations were summarized with mean and standard deviation. Differences in total procedure time, drilling time, and the 4 deviation measures were compared using repeated-measures analysis of variance. The predictor variables were method and the jaw. Post hoc pairwise comparisons were adjusted for using Tukey's HSD. SAS EG v.6.1 (SAS Institute, Cary, NC, USA) was used for all analyses. The assumptions for these models include normality of residuals and equality of the variance. Due to evidence of unequal variance (Levene's test) among the 4 methods, the models allowed for unequal variance across the 4 methods. The assumption of normality was not violated for any of the models based on graphical investigation of normal quantile–quantile plots. Statistical methods and results were reviewed and approved by an independent statistical reviewer.

A summary of the average operation time and deviations can be found in Table 1.

The total procedure time (including preoperative and operative time) was significantly associated with the method (P value < .0001) but not the jaw (P value = .9958). The DN method was the fastest, with an average of 32.1 minutes, which was significantly faster than any of the other methods (adjusted P value < .0001). Static guidance was the second fastest method with an average of 143 minutes and faster than the FH and TBG methods (adjusted P values < .0001). The difference between TBG and FH was on average 1.5 minutes (180.0 vs 181.5) and was not statistically significant (adjusted P value = .7172).

The operative time (implant placement) was significantly associated with method (P value = .0001). The fastest placement time was with SG with an estimated average time of 5.0 minutes, which was significantly faster than FH (adjusted P value = .0029) and DN (P value < .0001) but not TBG (adjusted P value = .0965). The TBG took on average 7.7 minutes and was significantly faster than DN (adjusted P value = .0014) but not FH (adjusted P value = .1408). The FH method averaged 10.2 minutes. The longest drilling time was the DN method, which averaged 13.5 minutes. Although not statistically significant, the mandibular implants took longer on average (9.8 vs 8.5 minutes, P value = .1626).

Figure 5 displays total procedure times along with operative and preoperative times for each of the 4 methods.

Results from the repeated-measures analysis to determine associations for method and jaw with the 4 deviation measures are given in Table 2 and Figure 6. The differences in vertical apex deviations were not significantly different for the 4 methods (P value = .3937) but were marginally higher for the maxilla than the mandible (0.48 mm vs 0.31 mm, P value = .0651). The 2-dimensional (2D) entry deviation was significantly related to the method and jaw combination (P value = .0254). The 2D entry deviation was significantly lower for SG than any of the other 3 methods for the mandible (0.35 mm vs 0.83–1.23 mm). None of the pairwise comparisons were statistically significant on the maxilla. Entry 2D deviation differed between the jaws for the FH method only (adjusted P value = .0256). The 3D apex deviation significantly depended on the method and jaw combination (P value = .0431). On the mandible, the FH method had significantly higher 3D apex deviations than DN (adjusted P value = .0100) and SG (adjusted P value = .0042). Similarly, the TBG method had significantly higher apex 3D deviations than DN (adjusted P value = .0086) and SG (adjusted P value = .0032). For the maxilla, the only significant difference was that the FH method was associated with significantly higher apex 3D deviations than SG (adjusted P value = .0359). TBG resulted in marginally higher deviations on the mandible than maxilla (adjusted P value = .0664). The overall angle deviation also depended significantly on the method and jaw combination (P value = .0154). On the mandible, the FH method had significantly higher angle deviation than DN (adjusted P value = .0118) and SG (adjusted P value = .0036). Similarly, the TBG method had significantly higher overall angle deviations than DN (adjusted P value = .0029) and SG (adjusted P value = .0006). For the maxilla, the FH method had significantly higher overall angle deviations than SG (adjusted P value = .0316) and marginally higher than DN (adjusted P value = .0537). The only method that differed significantly between the 2 jaws was TBG, which had higher deviations on the mandible than maxilla (adjusted P value = .0310).

Guiding methods for dental implants, their clinical use, accuracy, and efficacy have been studied by different authors.^1,7,16–19  Expert practitioners participated in most of these studies^6,7  and only a few included novice operators and their early implant placement experiences.^17,20  While experienced practitioners use all techniques including SG and DN, inexperienced operators might benefit more from using guided methods to improve accuracy and efficiency during their initial implant placement attempts. Limited evidence exists as to which method can help optimize accuracy and efficiency of implant placement for inexperienced operators and can be used predictably in early stages of skill development and spatial orientation awareness. This study compared 3 guided methods to the freehand implant placement for novice operators.

All methods offer acceptable outcomes; however, static and dynamic guidance resulted in higher accuracy and efficiency than TBG and FH implant placement on a simulation model. In entry deviation, SG was superior to the other 3 techniques, especially on the mandible where the implant sites were posterior vs anterior in the maxilla. Advantages of SG are likely contributable to the precision and rigidity of the guide providing great stability to the implant drills not allowing any positional deviations. Dynamic navigation offers accurate visual feedback but no physical support to the drill, possibly presenting a challenge and introducing deviations in the entry of the osteotomy if the tip of the drill is unstable and operator hesitant due to lack of experience.

In apical and angle deviations, SG and DN were significantly more accurate than TBG and FH placement, again especially on mandible with posterior implant sites. Having accurate and continuous dynamic or rigid static guidance throughout osteotomy preparation and implant placement facilitates the precise apical and angular position of the implant and aligns it with the virtually planned one. Novice operator can rely on SG for fast, reproducible, and accurate implant placement with very little room for introducing deviations. It is important that SG is accurately planned and fabricated with intimate fit as intraoperative guide adaptations are not feasible and the guide limits the visual feedback of the osteotomy during the procedure. Static guides can be bulky and require special drill sets with longer drills rendering them difficult to use in patients with limited mouth opening. The size, bulk, and rigidity of SG can limit the ability for adequate cooling saline delivery to the osteotomy site possibly leading to thermal damage. Planning and fabrication of SG requires computer expertise and costly equipment. The advantage of dynamically navigated implant placement is visual feedback and guidance on the computer screen through the entire procedure. Though the planning and calibration takes time and equipment is costly, the placement is efficient and accurate. No guide is needed and osteotomy can be visualized and corrected throughout the procedure. The drills are normal size and the mouth opening does not present a limitation with this technique. Placement of multiple implants proved predictable and reliable with DN for inexperienced operators aligning with previously reported observations that single implant placement with DN can be learned quickly and accurately.^17,20 

Studies comparing static and dynamic guidance observed similar accuracy for both techniques and reported similar deviations from the study findings.^1,7,19  Despite the lack of experience of operators participating in this study, results are consistent with the literature. Interestingly, there were no differences in deviations with DN when comparing our results to those performed by experienced surgeons.⁷  The influence of experiences obtained through implant placement with dynamic navigation guidance can provide a reliable technique for novice and experienced operators.²⁰ 

This study observed similar accuracy obtained with TBG and FH techniques. One possible explanation is lack of rigidity of TBG and inadequate stability of the drill during initial osteotomy. The drills inside TBGs tend to deviate in mesiodistal and buccolingual dimension, and require prior experience to make intraoperative adjustments. These guides also can reduce visibility and mislead the drill in suboptimal position. However, TBGs remain popular due to their relatively easy and inexpensive fabrication. Experienced operators can overcome TBG's limitations by correcting for positional errors, which was not observed for inexperienced operators in this study. Our observations are consistent with a previous study that reported even greater angular deviations for TBG than for the freehand method observed in this study.¹⁵ 

Accurate single implant placement is easier to achieve than multiple implants to support a multiunit FPD, which presents a spatially more challenging task. Freehand placement for multiple implants may be feasible for an experienced operator envisioning the distribution of available bone, space, and occlusal alignment for future prosthesis. Observations in this study suggest that on the other hand, the increased complexity related to spatial positioning of multiple implants can be challenging for an operator without prior experience. Even in the presence of a careful virtual implant plan and visual display of target implant positions, the FH and TBG techniques do not result in the same accuracy as SG and DN when used by inexperienced operators. Use of precise guidance methods is therefore indispensable for accurate placement in absence of experience and during training.

Time efficiency should evaluate all stages involved in implant placement including planning and guide fabrication. The present study assessed efficiency for preoperative (CBCT, presurgical planning, guide fabrication) and operative (surgical planning, calibration for DN, and implant placement) phase. Guide fabrication and setup for guided methods required additional preparation time and costly equipment but in return resulted in decreased operative time and increased accuracy. Dynamic navigation only requires CBCT for preoperative planning, making it the most time efficient method when total time is considered. Conventional radiographic guides were used for TBG and FH methods. Due to time required to fabricate these radiographic guides, the longest preoperative times were observed for TBG and FH methods. Modifying the radiographic guide into a surgical guide for TBG method does not significantly affect the preoperative time. When assessing efficiency for implant placement (operative time), the SG technique is significantly faster and more accurate than other techniques. Consistent with previous reports, operative time for novice operators using DN quickly improves across attempts.²⁰  When placing multiple implants with no or limited prior experience, dynamically navigated or statically guided placement offers higher accuracy and more favorable efficiency than FH and TBG-aided placement.

The main limitation of this study was its ex vivo design on PMMA models, which does not have the same properties as natural bone but in turn offered better standardization among different operators for assessing accuracy and efficacy of different implant placement methods. Similar patterns of efficiency and accuracy are observed in clinical applications of these 4 methods; however, the reproducibility in an ex vivo model is more predictable than an in vivo model due to individual patient characteristics that may affect the measured variables inconsistently. The ex vivo study on PMMA models offers consistency and reproducibility and reduces confounding variables. Future research efforts should explore application of these trends and validate these observations in in vivo clinical studies.

Despite the study limitations, we can conclude that dynamic navigation was the most time-saving considering presurgical and surgical time combined for the novice operators in the study. Using a static guide offers a significantly more time-saving surgical procedure than other techniques. Considering accuracy, static guide technique and dynamic navigation are superior to and significantly more accurate than using template-based acrylic guide or freehand implant placement.

The authors would like to thank Dr Leroy Thacker, PhD for providing an independent review of the methodology and results of the manuscript. Dr. Thacker is an Associate Professor in the Department of Biostatistics in the School of Medicine at Virginia Commonwealth University, Richmond, Virginia, USA and Director of Statistical Services at VCU Wright Center for Clinical and Translational Research.