**1. Introduction**

Spinal instability is commonly treated through surgical fixation involving vertebral screw insertion. Conditions that frequently result in spinal instability are fractures after trauma, spinal deformities, tumors, and degenerative diseases. Spinal fixation is aimed at gaining stability and preventing subsequent neurological deficit. Generally, the bilateral positioning of screws in the vertebrae around the level(s) of instability is needed to achieve the immobilization of the unstable segments. The accurate insertion of screws is essential for achieving safe and optimal spinal fixation surgery. Conversely, malpositioned screws can induce damage of vital structures or result in the failure of fixation [1,2].

**Citation:** Pijpker, P.A.J.; Kuijlen, J.M.A.; Tamási, K.; Oterdoom, D.L.M.; Vergeer, R.A.; Rijtema, G.; Coppes, M.H.; Kraeima, J.; Groen, R.J.M. The Accuracy of Patient-Specific Spinal Drill Guides Is Non-Inferior to Computer-Assisted Surgery: The Results of a Split-Spine Randomized Controlled Trial. *J. Pers. Med.* **2022**, *12*, 1084. https://doi.org/ 10.3390/jpm12071084

Academic Editor: Morohoshi Takao

Received: 18 May 2022 Accepted: 27 June 2022 Published: 30 June 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Traditionally, freehand insertion of screws is performed according to anatomical landmarks and through fluoroscopy control. However, because freehand screw insertion does not account for structural variations, including anatomical variance or severe deformations, there is an increased risk of malpositioned screws and related neurovascular complications [3–6]. Advances in spinal surgery have led to the development of computer-assisted surgery (CAS) navigation systems, also often mentioned as image-guided surgery, or intraoperative navigation. Initially, CAS systems relied on preoperative CT, which had to be manually re-registered for each individual vertebra and was associated with substantial registration errors. With the advent of modern CAS systems that rely on intraoperative acquired CT, this time-consuming repetitive calibration for individual vertebrae became redundant, leading to the increased use of CAS in spinal surgery. The accuracy of screw insertion has been substantially improved with current CAS systems, with a significantly reduced misplacement rate compared with the misplacement rate associated with freehand screw insertion [7,8]. Although spine surgeons with specific subspecialities (e.g., minimally invasive surgery and oncology) have widely incorporated CAS into their practices, the technology has not been adopted in all hospitals. In addition, the accuracy of screw insertion in cases with a highly mobile cervical spine reportedly fails to meet the high accuracy that is achieved in other spinal areas, which is most likely induced by the intraoperative shifting of segments. Moreover, despite these modern techniques, screw malpositioning does still occur, with the reported prevalence of malpositions being as high as 19% [9,10]. For these reasons, there is an ongoing demand for alternative navigational technologies that can be used to facilitate accurate spinal screw insertion.

Recent developments in medical computer-aided design and manufacturing techniques have given rise to completely new methods of surgical planning, commonly referred to as 3D virtual surgical planning (VSP) and patient-specific instrumentation (PSI) technology. The PSI technology comprises 3D-printed instrumentation that allows for the translation of the VSP to surgery. PSI are currently widely applied within different specialties [11], and the use of 3D-printed drill guides (3DPGs) has been attracting increasing attention as a promising navigational tool for spinal screw insertion. Recent studies have shown that 3DPGs are feasible for cervical and thoracic spine instrumentation, as demonstrated in cadaveric studies [12–17] and in clinical trials [18–23]. Moreover, their specific applications have been described for scoliosis surgery [24,25], C2 lamina screw insertion [26], and C1–C2 transarticular screw insertion [27]. Compared with CAS, the 3DPGs have several advantages. First, it provides the benefit of having a preformulated screw plan that includes the direction, length, and thickness of the screws. Second, it eliminates the need for intraoperative fluoroscopy. Third, it is applied to individual vertebra and is thus unaffected by intervertebral motion, which, in the case of CAS, can lead to workflow interruptions resulting from re-registration. Fourth, 3DPGs do not induce a surgical line of sight interference as in the case with CAS, which requires to constant looking back and forth between the patient and the screen. Last, 3DPGs do not require costly hospital investments.

The findings of various studies reported in the literature that have directly compared the accuracy of freehand and 3DPG screw insertion have all demonstrated the superiority of guides over the freehand technique [28–30]. However, comparisons of CAS and 3DPG are sparse within the literature; only one study has compared groups of patients instrumented with 3DPG and CAS [31]. To the best of our knowledge, a randomized comparative study of 3DPG and CAS has not been previously conducted.

Although 3DPGs appear to have several advantages, to become accepted as a viable alternative navigational tool its accuracy must first be shown to be comparable to that of CAS. Consequently, the objective of this randomized trial was to demonstrate the noninferiority of 3DPG relative to CAS in terms of the accuracy of pedicle and lateral mass screw insertion.

#### **2. Materials and Methods**

#### *2.1. Study Design*

The SpineGuides study was a single-center, prospective, investigator-driven study that randomly allocated screws to either 3DPG or CAS-assisted instrumentation in patients undergoing clinically indicated fixation surgery. All consecutive patients scheduled for cervical and/or thoracic spinal fixation surgery were eligible for this study. Exclusion criteria were: (1) patients aged below 16 years, (2) scoliosis surgery, (3) previous surgical history entailing laminectomy or the application of osteosynthesis material at the target levels, (4) urgent cases, and (5) unilateral instrumentation. This trial was undertaken after obtaining approval from the ethical board of the local medical institution (ref no. M19.229543). Written informed consent was obtained from each patient prior to their enrollment in the study. The trial has been registered on euclinicaltrials.eu with a registration number of 2022-500880-11-00.

The study was designed to determine whether the accuracy of screw insertion in spinal fixation surgery performed with the 3DPG navigational technique is non-inferior relative to that of screw insertion performed using the CAS technique. Because spinal bone geometry, density, and microstructure can vary widely among subjects, a "split-spine" design was selected for the study. The two navigational techniques were randomly assigned to either the left or the right side of the spinal column. The split-spine design removed the influence of interindividual variability in the study arms and also ensured that vertebral levels and screw insertion techniques (mass lateral vs. pedicle) were evenly distributed.

The two techniques were randomly assigned to the right or left side of the spinal column by generating balanced permutations via computer randomization. The randomization was constrained by blocks such that an equal number of techniques per side are obtained. The inclusion was limited to bilateral screw insertion, which resulted in a consistent number of screws at each vertebral level within both study groups. The randomization scheme was created using the online tool at randomization.com (http: //www.randomization.com, accessed on 5 January 2019) prior to commencing the study, and randomization codes were enclosed in sealed envelopes.

#### *2.2. Virtual Surgical Planning*

Virtual surgical planning (VSP) was carried out in accordance with a previously developed workflow, and 3DPGs were fabricated according to our previously published blueprint [17,23]. The brief description of the VSP and PSI steps are as follows. First, the preoperative CT was imported into medical image data segmentation software (Mimics, Materialise, Leuven, Belgium). Threshold-based segmentation was performed for each vertebral level to obtain 3D anatomical models, and optimal trajectories were manually defined. Then, 3DPGs were designed and manufactured in polyamide in accordance with ISO 13,485 standards and sterilized for intraoperative use by autoclave steam sterilization. Knowledge of the allocation was concealed from the 3D-specialist (PP) in charge of planning screw trajectories and designing the guides.

#### *2.3. Surgical Technique*

The standard surgical procedure was applied, starting cranially and continuing caudally placing screws sequentially at each level. Because randomization was concealed from the 3D specialist, the 3DPGs were designed with bilateral drill holes. Therefore, they would only fit if screws had not yet been inserted at the level of interest. Accordingly, for each level to be instrumented, the protocol stated to start with the 3DPG-assigned side followed by the contralateral CAS-assigned side. The 3DPGs were positioned after performing meticulous removal of soft tissue. Pilot holes (2 mm) were drilled at high-speed using appropriate drill stops (Figure 1). In the case of thick screws, the trajectories were enlarged/expanded using a straight pedicle probe. The study protocol ensured that the 3DPG burr hole checking was not performed using the CAS system, in order to keep the study arms separate. For the CAS study arm, the screw trajectories were created according to the

standard CAS guiding procedure applied at our neurosurgical department, which consists of several steps. The steps applied in the CAS-assisted screw placement were as follows: (1) entry point identification using the CAS pointer and its marking with a ball-tipped burr, (2) definition of an optimal trajectory using the pointer, and (3) creation of a burr hole through alternated probing (or drilling for lateral mass) and pointer-based checking until the desired trajectory was achieved. For the purposes of this study, the ultimate CAS trajectory was saved intraoperatively by positioning the CAS pointer in the drill hole and storing the trajectory's coordinates within the system. By opening the saved data using the CAS cranial module (instead of the regular spine module) postoperatively, we were able to retrieve the trajectory's coordinates in the digital imaging and communications in medicine (DICOM) format (Figure 2). The cone-shaped pointer tip ensured concentric positioning within the drill hole. During the procedure, no navigated drill or screwdriver was used, as these tools were not part of our center's standard procedure, nor it was available within the collection of instruments. The CAS setting comprised a mobile Arcadis Orbic 3D fluoroscopy C-arm (Siemens Medical Solutions, Erlangen, Germany) combined with a Brainlab optical navigation system (BrainLab Curve, BrainLAB, Munich, Germany).

**Figure 1.** 3DPG positioning and high-speed pilot hole drilling. A drill stop was slid over the drill bit to prevent penetration of the anterior cortex.

**Figure 2.** Retrieval of the DICOM coordinates of the intraoperatively stored trajectories accomplished by opening the spine planning file inside the cranial module of the planning software.
