Upper and middle thoracic fractures are rare among spinal fractures. Approximately 10–20% of general spinal traumas are observed in this region (12). Upper thoracic vertebral injuries often result in axial stress and bending with rotation and dislocation. This type of injury is generally seen at T4-T6 levels in motorcycle riders (13).
With the development of spine surgery, posterior thoracic interpedicular screwing has become more important. The key to this operation is the correct placement of the pedicle screw in one step, but it has been difficult due to safety concerns relating to the upper thoracic pedicle properties. The upper thoracic spine pedicles are thinner, and there are different angles for each spine. Thoracic pedicles are thin, short, and narrow, and their cortex is thin and fragile; therefore, thoracic pedicles are easily broken during screwing (14). Additionally, the angles of the thoracic pedicles are different from each other, which has made the rate of error in placing thoracic pedicle screws at one time very high, causing serious consequences by damaging the surrounding tissues (15).
The anatomy of the thoracic pedicles is more complex, and screw insertion is more difficult in complex thoracic fractures and vertebral malformations. The penetration rate can reach 30–40% in the insertion of the thoracic pedicle screw with the freehand technique (16). CT-based navigation systems are used to guide the placement of pedicle screws on the spine. However, the accuracy of the systems is questioned, and the failure rate is high (8.5–11%). Its use is not common due to its disadvantages, such as intraoperative position changes and spinal instability, lack of real-time navigation, and high cost (17).
With the use of 3D printing in spine surgeries, the production of guide plates, and provision of preoperative simulation, the accuracy of operations has increased. The fact that the accuracy is not affected by the intraoperative position and the higher reliability of guide plates provides superiority to navigation systems. Providing preoperative simulation and using the model as a guide during surgery reduce the surgeon's margin of error and operation time (18).
Controls performed with 2D fluoroscopy in upper thoracic interpedicular screwing show high error rates. Feyza Karagöz et al. retrospectively examined 113 pedicle screws between T2-T8 in 24 patients without coronal deformity. The control of the pedicle screws was checked during the operation by C arm fluoroscopy and postoperative CT. The faulty pedicle screw insertion rate was found to be 20.3%, 27.4% between T2-T5 and 14.5% between T6-T8 (19). Pedicular screws were applied to T4-T12 levels by 5 experienced surgeons on 5 fresh cadavers - Vaccaro et al. In postoperative CTs, a faulty screw placement rate was found at a rate of approximately 41%. Of these, 21 screws were observed to be in the vertebral canal by preparing the medial wall of the pedicle (20). In our study, in patients who underwent surgery using the freehand technique, a total of 18 (15.7%) incorrect pedicular screw placements were observed, 16 screw grade 2 (14%) and 2 screw grade 3 (1.7%).
With the application of 3D compression in spine surgeries, personalized production of guide plates, and preoperative simulation of the operation on the model have increased the accuracy of operations. The relatively easier and cheaper operating processes and the high reliability of the guide plates have removed the limitations of navigation methods. Lu et al. used 3D modeling as an aid to cervical pedicle or vertebral plate screw placement and proved that it can provide correct placement of the screws (21). Customized 3D spine models and screw insertion guide plates can be used to aid screw insertion and ensure the correct insertion of screws.
Mizutani et al. designed 3D models to apply cervical pedicle screws and achieved good results with guide plates in placing cervical pedicle screws (22). Sugawara et al. created personal 3D navigation models for thoracic pedicle screws and applied pedicle screws under their guidance simply and safely. In 103 patients, 813 screws were placed with 3D guides. In postoperative CT scans, 801 screws (98.5%) were placed without cortical violation, and no injury to the vessels and nerves was observed (23). Wei Hu et al. placed 56 pedicle screws in 7 patients with upper and middle thoracic trauma using the 3D printing-supported preoperative plan method. Regarding the placement of 56 screws according to postoperative CT images, 33 were grade 0, 18 were grade 1, 4 were grade 2 (perforated sidewall), and 1 was grade 3 (perforated sidewall, no vascular nerve injury). The accuracy rate was 91% (24). In our study, screw placement was performed according to postoperative CT images of 116 pedicle screws placed in the upper thoracic spine in 15 patients with preoperative 3D printing support and guidance, and 104 (89.6%) were grade 0, 8 (7.0%) were grade 1, and 4 (3.4%) were grade 2. Grade 3 positioning was not observed in any screw, and the pedicular screw placement accuracy rate was 96.5%. Comparing the pedicle screw placement accuracy of the upper thoracic vertebrae (96.5%) and the pedicle screw placement accuracy (84.2%) of the freehand technique in the 3D printing-supported group, the difference was statistically significant (p < 0.05).
In the study by Yue Pan et al., 37 patients with spinal deformities were operated on, with group 1 (20 patients, 396 screws) supported by 3D printing and group 2 (17 patients, 312 screws) supported by the freehand method. The operation time in group 1 was 283 ± 22.7 minutes. In group 2, it was 285 ± 25.8 minutes. The operation time was found to be shorter in group 1, although the difference was not statistically significant (p = 0.89) (25). In our study, whereas the operation time was 134 ± 22 minutes for group 1, it was 152 ± 38 minutes for group 2. The difference in operation times was statistically significant (p < 0.05).
In the study by William Clifton et al., for 40 C7, 40 T6, and 40 L5 pedicle screws, the rate of agreement between the pedicle positions studied on preoperative models and the postoperative pedicle screw positions was found to be 100% for C7, 100% for T6, and 93% for L5 (26). In our study, 2 (13.3%) of the 15 upper thoracic fracture patients (group 1) were T3, 5 (33.3%) were T4, and 8 (53.3%) were T6 fractures, which were operated on by preoperative planning using 3D modeling. For these 15 patients, the concordance rate between pedicle positions studied on preoperative models and postoperative pedicle screw positions was 93.8% for T3 fractures, 94.7% for T4 fractures, and 98.4% for T6 fractures.
For spinal surgeons, it takes a long time to experience an upper thoracic pedicle screw. Additionally, vertebral canal violations and vascular injuries are common in this region (27). Thanks to the preoperative surgical simulation of the 3D printing-supported model, the application of the upper thoracic pedicle screw will become more efficient and easier. In this study, the accuracy rate obtained in the 3D printing-supported group was 96.5%, which was higher than that of the freehand technique group. We think that the 3D printing-supported method in upper thoracic pedicle screw application will shorten learning time, provide easier learning on the model, and increase pedicular screw placement accuracy.