Successful dental implant restoration depends on precise prosthetic-driven implant placement to minimize deviations from the preoperative plan. These deviations may exert unfavorable paraxial forces on the implant leading to fatigue, restoration failure, and peri‑implant bone loss [1,2]. Furthermore, implant malposition can often cause surgical complications such as damage to the inferior alveolar nerve and perforation of the Schneiderian membrane. Rehabilitating atrophic edentulous arches is particularly challenging and has a high risk of surgical complications. Implant placement accuracy can be influenced by surgeons’ inaccurate perceptions and hand tremors [3,4]. Strict demands are required of surgeons due to the long operative time and intense work of implant surgery for edentulous patients. In the digital era, computer-aided techniques that transfer preoperative planning information to the surgical phase can effectively solve these clinical problems.

Robotic systems have been investigated and applied to assist dental implant placement using haptic guidance and real-time visual navigation. For example, Yomi was the first commercial implant dentistry robot approved by the Food and Drug Administration (FDA) in 2017 [5]. As a semiactive robotic system, surgeons manually advance surgical drills along the planned path with rigid constraints to reduce operation error. For completely edentulous arches, proper implant position is essential to facilitate ideal biomechanical loading and obtain good esthetic results [6]. Bolding et al. [5] reported the first implant placement surgery under the guidance of Yomi in edentulous patients, with their case series showing that the procedure was safe and accurate. Furthermore, recent research has revealed that robots with an expected osteotomy deviation of less than 1 mm and angular deviation of less than 2° were more accurate than dynamic and static navigation techniques, suggesting promising applicability in the future [7], [8], [9].

Robotic systems require correct registration to match the coordinate system of patients to their radiological image. The procedure requires the alignment of fiducial markers or anatomical landmarks through rigid-body point-based registration [10]. In recent years, various registration strategies employing invasive or noninvasive markers have been extensively investigated for dental implant surgery. Compared with noninvasive markers such as splint templates, anatomical landmarks, or laser surface scans, invasive screw markers are more advantageous in terms of precision and stability in edentulous patients [11]. The number and distribution of fiducial makers can influence registration accuracy. West et al. [12] reported that widely spread fiducial markers, placing the centroid of markers near the target, and nonlinear arrangement of markers led to a relatively increased registration accuracy. Another in vitro study in 2014 showed that increasing the number of landmarks from four to eight did not significantly influence accuracy. Decreased registration errors were observed when the landmarks’ center was closer to the surgical region [13]. Fan et al. [14] recently explored the effects of the number and configuration of fiducial markers on registration accuracy during zygomatic implant placement. The registration errors decreased with an increase in the number of fiducial markers. Scattered distribution in the polygon span with at least five fiducial markers was recommended.

In general, previous studies disagreed on the optimal placement of fiducial markers. To our knowledge, no study has explored the effect of the number and distribution of fiducial markers on robot-guided implant placement accuracy in completely edentulous arches. Therefore, this in vitro study aimed to evaluate whether the number and distribution patterns (dispersed or localized) of fiducial markers influenced the accuracy of robot-aided implant placement in edentulous mandibular arches, determined by measuring discrepancies between planned and actual implant positions. The null hypothesis was that neither the number nor the distribution of fiducial markers would influence robot-guided implant placement accuracy in completely edentulous arches.