Cone beam CT-datasets of the skulls of five fresh-frozen edentulous cadavers were obtained (Planmeca, ProMax 3D Max, Stockholm, Finland; 576 slices, voxel size 0.3 mm, FOV: 11 × 16 cm). The settings were in accordance with the clinical settings used for implant planning. A 3D model of the zygomatic bone and maxillae was created using ProPlan CMF 3.0 (Materialise, Leuven, Belgium) software.
Virtual surgical planning
To mimic the clinical problem of a maxillary defect, typical examples of maxillary tumor resection surgeries were planned virtually (Fig. 1a, b). 3D VSPs are created that included partial resection of the maxilla, leaving a maxillary defect, based around an assumed tumor volume that would be suitable for obturator prostheses supported by zygoma implants. The defects created in this experiment were classified as low-level Brown Class 2b maxillectomies [13].
Based on the 3D VSP, surgical cutting guides were designed and printed to transfer the resection plan to the cadavers. Next, an obturator prosthesis was designed in the software matching the virtually created defect. Pre-existent dentures were not available for any of the cadavers. The maxillary soft tissues were segmented in order to design digital maxillary dentures as base templates for the final obturator prostheses. Implementing the digital obturator prostheses completed the VSP and enabled the digital planning of the prosthetic implant platform positions.
The position of the zygomatic implants was planned backward from the position of the prostheses. The zygomatic implant heads, to support and fixate the prostheses to the zygomatic implants, were placed in the most ideal prosthodontic positions, slightly palatal from the occlusional plane (Fig. 1c, d). In a vertical dimension, enough space for a future bar superstructure and acrylic was taken into account (Fig. 1e). Horizontally, the spacing between the prosthetic implant platforms was carefully chosen in order to fit a clip retention system and to enhance any necessary cleaning of the implants (Fig. 1f).
The positioning of the zygomatic implant was planned with the tip of the implant placed in the lateral cortical bone of the zygomatic complex. The assumption was that placing the apical part of the implant in the cortical bone provides optimal primary stability and will cover the bone on the lateral side of the implant.
The preferred apical and abutment positions of the zygomatic implants, implant lengths, and obturator prosthesis were designed virtually. Subsequently, patient specific implant drill and placement guides were designed based on the final virtual set-ups (3Matic Medical, Materialise, Leuven, Belgium) (Fig. 1g, h). The drilling/placement guides were developed to fit the following bone structures: alveolar ridge, nasal floor, and zygomatic arch. The guides were printed from polyamide, produced according to the ISO 13485 standards for medical devices, at Oceanz (Ede, the Netherlands). The study ultimately resulted in an advanced implant guide design. The addition of centered channels in the drill-guide enables angled cuts and the length of the channels form an integral depth stop for the drill. The insertion of stainless steel (316 L) milled drill sleeves in the channels should minimize deviation of the drill trajectories. The maxillary bone-supported part included an extension to the nasal aperture to verify good positioning of the guide [14] and was connected with crosslink arms to the zygomatic bone-supported part. In addition, the guide was supplied with holes for temporary fixation with mini screws.
Surgical procedure
The cadaver surgery was split into two series to evaluate the findings and, if necessary, to adjust the guides and/or obturator prosthesis before the second test. The surgery was performed by OMF surgeons involved in the planning process, and the supportive visual documentation of the planned guide position was always present in the operating room. Two cadaver heads were thawed before surgery for the first session, and the other three were thawed later for the next session.
To create the Class IIb maxillary defects, the cutting guides were placed on the denuded bone of the maxilla and the zygoma (Fig. 2a). The stability and fit of the bone-supported cutting guide was verified. The left-sided maxillectomies were guided by the surgical templates. The resected specimen, mimicking a tumor resection, was removed, resulting in a Class IIb defect. The cutting guides were removed and subsequently the implant drilling guide was fitted and placed.
The precise alignment to the underlying bone structures was verified. The drill-guide was fixed to the bone on two anatomical locations (zygoma and premaxilla) using 2.0 mm cortical locking screws (KLS Martin, Tuttlingen, Germany) (Figs. 2c and 3a). After drilling, the implant beds through the guide, according to the drill sequence for oncological zygomatic implants (Southern implants, South Africa), the metal sleeves were removed (Fig. 3b). The guide was designed to direct the angle and depth of the implant placement (Fig. 3c). The VSP planned implant lengths were placed and the final prosthetic platform position was checked by the maxillofacial prosthodontist and the guide was removed (Figs. 2b and 3d). The obturator prosthesis was then fitted, which provided the surgical team with a visual check as to whether the emergence of the zygomatic prosthetic platforms was favorable or not, in relation to the pre-planned slots in the obturator prosthesis. The surgical procedure was finalized by fixating the obturator prosthesis. Non-engaging prosthetic cylinders (Southern implants, South Africa) were fixed to the obturator prosthesis with light cured resin to fix the prosthesis firmly on the zygomatic implant abutments (Fig. 2d). The obturator prosthesis was checked for balance support on the contralateral side of the residual maxilla. After the surgical procedure, the obturator prostheses were removed and the heads underwent a post-operative cone beam CT scan to analyze implant accuracy.
In preparation for the second cadaver operation session, two alterations were made to the working method. The first alteration was the use of more rigid and solid crosslink arms on the drilling guide, to minimize guide movement due to vibrations during drilling. Secondly, longer mini screws were used to fixate the guides to the bone. The longer 8 mm screws were better for retention in the slightly porous cadaver bone.
Outcome measures
The primary outcome of this study was the fit of the prosthetic cylinders connected to the placed zygomatic implants in the preoperative positioned slots of the obturator prostheses. It was noted if the obturator prostheses needed adjusting to fit the cylinders. Both surgical and prosthetic steps were based on one virtual surgical plan and had to tally with the final positions of the prosthetic implant platform above the designed screw access holes in the dental arch of the obturator prosthesis. In all five cadavers, the support for the obturator prostheses had to be on the remaining maxilla and should match the surgical resection. A placement accuracy of within 3 mm of the prosthetic cylinders in the slots were considered to be successful for a prosthesis, resulting in a passive fit.
A secondary outcome measure for 3D planned series was the zygomatic implant placement accuracy. The post-operative CBCT-data were obtained in a similar fashion as for the pre-operative CBCT. The post-operative maxillae were segmented and, the implant positions were matched with the 3D VSP. The post-operative implant positions are determined by two observers. The most distal part of the long axis of the implant was used as the abutment position (Fig. 4a), so that the results were not dependent on a rotation along the long axis. The entry and exit positions in the zygomatic bone were defined by the intersection of this long axis with the virtual maxilla.
Two coordinate systems were defined:
-
1)
The Implant’s Coordinate System (ICoS); the z-axis runs along the long axis of each planned implant.
-
2)
The Occlusion Coordinate System (OCoS); congruent with the axial, saggital, and coronal planes, where the axial plane is defined by the occlusion plane of the virtual obturator prostheses.
The planes perpendicular to the z-axis, and running through the planning’s abutment point, defined the entry and exit points for the ICoS measurements. The intersections of the implant long-axes with these planes were defined as the corresponding points of the outcome. Then, the distance between the abutment, the entry and exit points, and their corresponding points were defined. Also, the 3D angular deviations between the planning’s and the outcome’s long axes were determined (Fig. 4a–c). Unpaired t tests were conducted as well as the intraclass correlation (ICC) of the implant reconstruction between the two observers.