Discussion

Transpedicular CII fixation is a reliable method. However, high incidence of complications associated with screw insertion along the wrong trajectories indicates the need for new methods of safe implantation [14]. Bone perforations were observed in most studies although some authors report high safety of “free hand” method [15, 16]. Thus, P. Punyarat et al. reported bone perforations by 12 (23%) screws during transpedicular CII fixation (grade 1 — 10 (19%) screws, grade 2 — 1 (2%) screw, grade 4 – 1 (2%) screw) [17]. R.J. Bransford et al. reported that 83.3% of screws were completely surrounded by the bone. “Free hand” installation of 0.3% of screws into the axis resulted vertebral artery injury [18].

Spinal navigation based on intraoperative CT or polypositional radiography does not exclude the risk of bone perforation with possible complications. J.W. Hur et al. reported that transpedicular CII fixation using O-arm was followed by cortical layer perforation with screw exit for more than 2 mm in 7.6% (10 out of 92) of cases. Two screws perforated vertebral artery [19]. J. Randall et al. reported perforation-free implantation of screws into the axis in 67% of cases that was significantly lower compared with “free hand” group [15]. M. Uehara et al. used intraoperative CT navigation and reported perforation of the axis in 12.5% of cases (degree 3 in 5% of cases) [20].

3D navigation matrix technology is an inexpensive and effective method ensuring reduced risk of intraoperative complications in patients with complex trajectories of screw system implantation. This approach also increases accuracy of screw placement, reduces the number of repeated introductions of the screw and irradiation of a patient. According to literature data, this method demonstrates high efficiency and availability. The last characteristics are similar to those of intraoperative CT navigation [21—23]. In 2014, S. Kaneyama et al. analyzed an effectiveness of navigation matrices for screw implantation into the axis (23 patients, 48 screws). The authors used 3 types of matrices: for determining the insertion point, for forming the passage by a needle, for introduction of the screw. Mean axial deviation was 0.36±0.62 mm, sagittal deviation — 0.30±0.24 mm. Class 1 of implantation was obtained for 46 screws, class 3 – for 2 screws. Class 2 and 3 was observed for 2 screws. No complications were recorded [13].

In 2016, S. Guo et al. reported the results of cadaveric experiment. Implantation with navigation matrices was significantly more accurate (95.8% — grade 1, 4.2% — grade 2) compare with “free hand” technique (grade 1 — 72.7%, grade 2 — 17, 3%, grade 3 and 4 — 4.5% for each class) [24].

The Japanese authors reported high accuracy of implantation during Harms surgery in 2017. Forty-eight screws (24 CI, 20 transpedicular screws into CII, 4 translaminar into CII) were implanted with mean deviation of 0.70±0.42 mm in frontal plane at the level of the middle of the pedicle. All screws did not extend beyond the bone [25].

In 2017, specialists from China implanted 74 screws into СI and СII (37 screws in each group). In the experimental group, the screws were installed using navigation matrices. Thirty-two (86.5%) screws were completely inside the bone, 3 screws (8.1%) perforated the bone for less than 1 mm, reinstallation was not required. Two screws (5.4%) perforated the bone for over 1 mm. In the control group, 23 (62.2%) out of 37 screws were inside the bone, 3 (8.1%) screws perforated the bone for less than 1 mm, 11 screws – over 1 mm (29.73%). Thus, acceptable installation accuracy was 94.6% for navigation matrices and 70.3% for standard fluoroscopy [26].

It should be noted that the described technology is a fundamentally new solution in spinal navigation due to some features. Unlike intraoperative CT navigation, screws are inserted along scheduled trajectories. Prolonged preoperative planning for designing and manufacturing matrices is offset by reduced time of surgery and incidence of re-implantations. Another feature is available designing, manufacturing and application of navigation devices in various institutions that was demonstrated in our study (the matrices designed and manufactured in the same laboratory were used in two different centers). Thus, complex screw implantation may be performed with high accuracy in medical institutions not equipped with CT navigation.

Considering our experience, we concluded the shortcomings of this technology. The first main aspect is long learning curve including the study of 3D designing and printing, development of optimal matrix design, determination of suitable materials and parameters of printing. The second drawback is prolonged preoperative planning including designing, manufacturing and sterilization of the matrix. Intraoperative disadvantages are the need for careful skeletization of the vertebral structures for tight contact with the matrix, possible breakdown of the matrix and its deformation.