Posterior open surgery is one of the most common surgical methods for the treatment of thoracolumbar fractures . There are several ineluctable drawbacks in conventional open posterior pedicle screw fixation. The muscles along the spine are stripped from the bone portion of the spine, exposing the facets and transverse processes to position of the screws accurately. This could lead to complications associated with the surgical approach, including excessive intraoperative bleeding, intramuscular loss of innervation, swelling and ischemia, even long-term muscle atrophy and scarring, which related to intractable back muscle pain and dysfunction after surgery . Anatomically, the posterior muscles and ligaments of the thoracolumbar spine play an important role in maintaining the stability of the corresponding segments. Therefore, these muscles and ligaments should be well preserved in the operation of thoracolumbar vertebral fractures .
In recent years, minimally invasive surgery has developed rapidly, broadening the surgical indications for thoracolumbar fractures without neurological symptoms. The placement of the pedicle screw by minimally invasive techniques does not require the stripping of paravertebral muscles and ligament tissue, thus reducing the incidence of approach-related complications. Multiple studies have demonstrated the advantages of minimally invasive nailing techniques, including reduced pain, less soft tissue injury and intraoperative bleeding, shorter postoperative hospitalization times, and rapid rehabilitation [3, 5–10].
In 1968, Wiltse et al.  first proposed the conception of the paraspinal muscle approach between multifidus and longissimus, which retained the integrity of posterior ligament complex, could produce less bleeding and surgical trauma compared with traditional open approach. Li et al.  found that the Wiltse approach had obvious advantages over the conventional open method in operative time, blood loss, postoperative drainage, postoperative hospitalization time, and postoperative improvement in VAS. Wu et al.  also found operation duration, blood loss, average length of incision and postoperative ODI of the paraspinal group were all obviously less than the traditional posterior approach group. Liu et al.  reported multifidus cross-sectional area decreased by only 7.6% in the Wiltse group compared to 35.4% in the posterior open group between pre-op and the last follow-up. This suggested that the Wiltse approach had a lower incidence of multifidus atrophy and fatty infiltration, so it was effective as a minimally invasive approach for thoracolumbar fracture.
In recent years, various fluoroscopic-based navigations have been introduced that provided the information of elaborate bony anatomy and experimented clinically [15–19]. The O-arm system is one of the intraoperative imaging platforms combined with Stealth Station navigation system that can be used to increase the accuracy of pedicle screw placement. Compared to traditional C-arm fluoroscope, the O-arm based navigation has several superiorities such as high quality of multi-dimensional images, lager filed of surgical view, and robotic positioning [17, 18]. Van et al.  performed a prospective multicenter clinical registry of thoracic, lumbar, and sacral pedicle screw placement using the O-arm navigation to assess the accuracy of screw placement. They evaluated a total of 1922 screws in 353 patients, and found only 2.5% of the screws were considered as misplaced. Silbermann et al.  assessed the accuracy of pedicle screw placement in lumbar-sacral spine between free-hand technique and O-arm based navigation method. The results noted that the accuracy rate was 99% in the O-arm group compared to 94.1% in the free-hand group.
Compared with the open surgery, the minimally invasive nailing with the Wiltse approach or O-arm 3D imaging for the treatment of thoracolumbar fractures have the advantages of less tissue trauma and bleeding, shorter operation and hospitalization time, and more accurate placement of pedicle screw [5–7, 9, 18, 19]. However, the comparison of clinical effect and radiological results between the two minimally invasive techniques has not been reported in the literature. In this study, there were no significant differences in intraoperative blood loss, length of incision, and postoperative hospital stay between the two groups. The VAS score and ODI score of the two minimally invasive techniques were significantly lower than the pre-op scores, and no distinction was drawn between the groups. Therefore, we believe that the two minimally invasive nailing methods can achieve the same therapeutic effect for thoracolumbar fractures. In aspects of operative time, the paravertebral approach group was obviously shorter than the O-arm navigation group, which possibly due to the fact that the placement and working of the O-arm imaging system is more time consuming and the surgeon is unfamiliar with the relevant special instruments. In addition, we also compared the surgical expenditure of the two groups, and showed expectable higher costs in OPSF group (59,035.4 ± 1,152.7 CNY) compared to WPSF group (48,142.1 ± 1,430.1 CNY). The principal reason for the cost difference was that more expensive implants and intraoperative neurophysiological monitoring were used in the O-arm navigation group.
The accuracy of screw position was 97.6% (164/168) in the WPSF group and 96.8% (151/156) in the OPSF group. There was no complication caused by the misplacement during follow-up. The results showed a better precision in the paravertebral approach group, though no significant difference was found. We reasoned that the placement of pedicle screws through Wiltse approach provide relatively intuitive vision. In this study, we used short-segment six pedicle screw fixation combined with intermediate screw fixation in both groups. Most authors reported that short-segment instrumentation with four pedicle screw was not adequate to achieve and maintain the reduction of thoracolumbar fractures and were associated with an unacceptable rate of failure [20, 21]. Compared to conventional 4-screw inter-segmental fixation, short-segmental fixation combined with intermediate screws enhanced the strength of fixation, which is helpful for maintaining reduction of the height and angle of the fractured vertebra, and allowed much earlier ambulation, which is important for recovery and avoiding complications . In the current study, the Cobb angle and VBA showed significant differences between pre-op and post-op in both groups. The average R values of the two groups were all restored to 99.9%, which means that both minimally invasive techniques can basically reset the fractured vertebra to the physiological height. Moreover, no distinct increase was presented in Cobb angle and VBA between post-op and final follow-up in both groups. The correction loss of R value was only 2.5% in the WPSF group and 2.7% in the OPSF group, respectively, with no clear discrepancy between groups. These indicate that the two minimally invasive techniques have satisfactory effects on the correction of kyphosis and preservation of segment height.
There also have some limitations in current research. First, the study was retrospective, and the treatment options of recruited patients mainly depend on their will, which implied non-randomized. Although there were no significant differences of preoperative clinical and X-ray data between the two groups. Second, the quantity of patients in this study was only 54 people, thus the conclusions drawn from statistics still not strong enough. Third, the follow-up duration of the study was relatively short. An extended observation should be needed for the evaluation of clinical efficacy, aggravated kyphosis, and the failure of fixation. In the future, the development of randomized controlled trials and more assessment methods will confirm the results of our study.