Accuracy of thoracic pedicle screw placement in adolescent idiopathic scoliosis patients using the entry point identified by new landmarks: a computed tomography study

Nowadays, TP screws placement is still a challenging procedure since the maximal permissible translation error is less than 1 mm and the permissible rotation error is less than 5° in thoracic spine[18]. Screw misplacement not only posed a significant threat to adjacent vital neurovascular or visceral structures but also decreased the pull-out strength of the screw, which consequently increased the possibility of instrumentation-related failures [19, 20]. Such risks substantially increased in scoliosis patients with abnormal vertebral morphology and changed anatomic relationship between the vertebrae and surrounding vital structures [21, 22].

To improve the accuracy rate of TP screw insertion in AIS patients, several image-guided techniques had been described, such as fluoroscopy, computer-assisted technique or intraoperative navigation technique, and so on [10‐13]. However, these image-guided techniques increased both the radiation exposure and surgery time. The free hand technique was simple and convenient and also exhibit good results of TP screw placements. Since the margin of error for TP screw is small, accurate TP screw insertion with free-hand technique depends on accurate entry point. Many previous studies had described different entry points for TP screws. The Roy-Camille technique selected the intersection between the mid-lines of the facet joint and transverse process as the entry point [23]. Vaccaro reported the entry point to be the junction of the superior border of transverse process and the middle of the superior facet for T4–T9 levels [17]. However, these points had been considered to be inappropriate due to the high incidences of pedicle perforation. Parker identified the entry point in the middle of the triangle formed by the pars interarticularis, the lower edge of the superior articular facet and the medial border of the transverse process. The entry points were more medial and cephalad from T12 to T7 and were more lateral and caudal above T7 [24]. Kim also suggested different entry points at different thoracic regions. The entry point of the 12th thoracic vertebra was located at junction of the bisected transverse process and lamina at the lateral border of the pars. The entry point moved medially and cephalad in the middle thoracic region and moved laterally and caudally in the proximal thoracic region [15]. Although these entry points had been proven to be practicable, they were not quantified, which made the young surgeons without much experience on screw placement confused and increased the risks of pedicle perforations.

To qualitatively identify the ideal entry point for TP screws in AIS patients, Qi used a new anatomic landmark as the points of reference, which was the most concave point at the junction of the transverse process and the folding part of the lamina [25]. Surgeons can select the entry point according to distances between landmark and entry point in both axial plane and sagittal plane suggested by Qi at each level. However, it is not easy for surgeons to remember the different distances from the landmark to entry point at different levels. Modi introduced the concept of an ideal pedicle entry point for TP screw with the freehand technique which is located at the base of the superior facet at the junction of the lateral one third and medial two thirds [18]. However, about 5 mm of inferior articular process overhangs the base of the superior articular process should be removed to expose the entire base of superior facet, which might lead to more bleeding and more operation time.

In this study, we propose a constant entry point which can be easily identified without removing the inferior facet. We noticed that in axial plane, projection of lateral border of the superior facet on the pedicle was always located in direct line with lateral wall of the pedicle (Fig. 3). Therefore, the lateral border of the superior facet can be used as a landmark for identifying entry point. Modi selected the entry point at the base of superior facet and no screws showed inferior violation [18]. Therefore, we think the intersection of lateral border of the superior facet and the base of superior facet is a proper entry point for TP screw insertion. However, the lateral edge of the superior facet was often covered by inferior facet, especially on the concave side of thoracic vertebra. Only a curved border connecting lateral border of the superior facet and the transverse process can be seen after exposure of posterior elements. However, removing inferior facet is time-consuming and may increase the blood loss. We thought the intersection of the curved border and inferior facet (Point a in Fig. 1) was approximately the distal end of lateral border of superior facet. O’Brien’s study suggested that the vertebra rotated and shifted as a whole in AIS patients [26]. We presumed that the lateral border of superior facet should be parallel to the supraspinous ligaments at the same level. Therefore, we draw a line from the intersection of the curved border and inferior facet (Point a in Fig. 1), paralleling to the supraspinous ligaments. The intersection of this line and the base of superior facet was selected as the ideal entry point for TP screw insertion in AIS patients. Using this entry point, the overall perforation rate of screw placement was only 6.9% and no screw-related complication happened. Kwan evaluated the accuracy of 2020 pedicle screws from 140 AIS patients and found the total perforation rate was 20.3% with majority of the perforations occurred in thoracic region [9]. In Modi’s study, a total of 152 screws were displaced either medially or laterally in 448 TP screws with the total perforation rate of 33.9% [18]. Qi reported the total perforation rate of 7.8% in 306 TP screws from 21 AIS patients using a landmark determined by CT scan [25]. In the literature, the reported intraoperative blood loss for AIS surgery with pedicle screw instrumentation ranged from 500 to 1212 ml [27‐32] and the mean operation time ranged from 146 to 378 min [28, 32‐34]. Both the average blood loss and average operation time of our cases were comparable with previous studies. Hence, our method of selecting entry point was both convenient and efficient for TP screw placement in AIS patients.

In the current study, we noticed that all the anterior perforation occurred in the upper thoracic spine, which might explain the highest perforation rate in this region (15.7%). Previous studies had demonstrated that the morphology of proximal thoracic spine was smaller than that of distal thoracic spine and the smaller morphology of proximal thoracic spine was most likely associated with an increased risk of anterior cortical perforation [35, 36]. Jiang’s study reported that the both esophagus and trachea were located in proximity to anterior cortex of proximal thoracic spine in AIS patients [21, 22]. Spine surgeons should choose appropriate screw length to avoid anterior cortex perforation in TP region. As we know, screw placement on the concave side is more difficult than that on the convex side due to the larger converge of the pedicle screw and smaller pedicle diameter. Smorgick reported higher incidence of pedicle perforation on the concave side when compared with the convex side [37]. However, in our study, the perforation rate is similar between the convex and concave side. This result is similar to Min’s study [38]. It is because all the TP screws were placed by the surgeon (J.J) who was experienced on screw placement in AIS patients and was very cautious when inserting screw on the concave side.

There were several limitations which should be mentioned in this study. Firstly, although our new method for selecting the entry point achieved high accuracy of screw placement, the perforation rate in the upper thoracic spine was as high as 15.7%, indicating that this method might be less useful in upper thoracic region. Secondly, the sample size was relatively small. Further study with large sample size is need to verify the accuracy of TPS insertion with our new method. Thirdly, as we know, the accurate TP screw placement depends not only the selection of entry point but also other important steps, such as proper use of awl, the convergence of the screw, and so on. All the TP screws were placed by an attending surgeon (J.J) who was experienced in screw placement. Whether the residents can achieve the same accuracy of screw placement with this method needs further study to evaluate.