12.1 Introduction

Over 600,000 children in the United States have pediatric scoliosis, with the vast majority diagnosed with adolescent idiopathic scoliosis (AIS). ¹ While many of these patients are managed with nonoperative care, a significant number undergo an operation, which is often a major undertaking that carries substantial risk. When the decision is made to operate, it is often not for current symptoms, as most patients are asymptomatic. ² Thus, avoiding complications, both short-term and long-term, is one of the most important aspects of surgical care of this pathology.

12.2 Surgical Approach to Pediatric Scoliosis

Surgical treatment of AIS is generally recommended when curvature is greater than 45° and progressing or when the terminal curvature exceeds 50 degrees. ³ The principal aims of surgery are primarily to halt curve progression with spinal fusion and secondarily correct existing curvature with rod-and-screw constructs. The surgical approach to spinal fusion can be through open anterior, posterior, or combined approaches. There are also posterior percutaneous or “minimally invasive” approaches to AIS correction. ^{4 ,​ 5 ,​ 6}

While there are a variety of options, operative treatment of pediatric scoliosis is most often performed through a posterior approach. ² Through the posterior approach, deformity correction can be achieved through rod contouring and derotation maneuvers, as well as through various osteotomies. ^{7 ,​ 8 ,​ 9} Some of the more involved osteotomies, such as pedicle subtraction osteotomies and vertebral column resections carry significantly higher risk of complications. ^{9 ,​ 10} In addition, the current surgical trends for pediatric scoliosis have shifted away from hook-and-wire fixation, toward pedicle screw instrumentation. ^{11 ,​ 12 ,​ 13} In the pediatric spine, pedicle screw placement has been shown to be safe in the vast majority of cases, but there remains a risk for malpositioned screws that may result in neurologic injury. ¹⁴ The deformities seen in congenital scoliosis and neuromuscular scoliosis can be even more challenging as the rotation of the vertebral bodies can make pedicle screw placement difficult and sometimes impossible.

12.3 Pedicle Screw Instrumentation

Posterior instrumentation in AIS includes screws inserted into pedicles of the vertebra that are connected by a rod to fix the spine while fusion occurs. ¹⁵ The placement of these pedicle screws can be guided by either anatomical landmarks or computer-guided navigation. ¹⁶ Inappropriate pedicle screw placement may result in intraoperative pedicle fractures, wound infection, and encroachment on the spinal canal causing neurological complications. ¹⁷ Studies of non-navigated, freehand screw placement have established baseline levels of inappropriate screw placement. ^{18 ,​ 19 ,​ 20 ,​ 21}

Pedicle screws can be placed with “freehand” technique, meaning with the use of bony landmarks and palpation of the bony canal through the pedicle. But imaging can be used to assist the instrumentation process with several technological advances. The most recent advancement is the use of the intraoperative CT for navigation of pedicle screw placement. While navigation can assist with osteotomies, including hemivertebral resections for cases of congenital scoliosis, ²² the most natural use of navigation is to increase the accuracy of pedicle screw placement. ^{23 ,​ 24} Moreover, navigation can allow for placement of pedicle screws in situations previously deemed unsafe, such as severe rotational deformity or narrow pedicles. ^{25 ,​ 26 ,​ 27 ,​ 28 ,​ 29} Additionally, in cases of revision surgery for pediatric scoliosis, the normal landmarks for freehand pedicle screw placement may be distorted. Navigation can be useful in these situations, as well as if an osteotomy through the prior fusion mass is needed. ³⁰ Lastly navigation may also have a role in minimally invasive techniques for pediatric scoliosis. ³¹

12.4 Freehand Technique

Freehand technique has been used widely in orthopedics to place implants. Before more widespread use of intraoperative fluoroscopy, percutaneous pinning of the hip was performed freehand with external landmarks and tactile feedback. Radiographs were performed to confirm proper positioning after placement of implants. However, in the age where imaging is readily available intraoperatively, most surgeons use a combination of visual landmarks, tactile feedback, and fluoroscopic imaging to place screws. Freehand placement of pedicle screws can be fast and safe, but it is important to evaluate the results of this technique. Many surgeons do not get postoperative CT and do not look critically at screw placement other than plain films, but some centers have tried to answer this question.

A seminal study was conducted to critically evaluate freehand screws using a postoperative CT. In 112 screws that were assessed, 12.5% screws were misplaced and two screws were on the aorta. ³² Another report of over a thousand freehand screws placed in 60 pediatric deformity patients evaluated with postoperative CT showed about 10% had significant medial or lateral pedicle wall violations. ³³ The T4–T9 screws placed in the concavity were seen to be the highest risk after evaluating freehand thoracic screws. ³⁴ Overall it appears that about 1 in 10 freehand screws is in a suboptimal location, but they may not have negative clinical effects. Medial and lateral breach is often well tolerated and even anterior penetration can have no effect in many cases as aortic and esophageal injury is exceedingly rare.

12.5 Freehand Technique with Image Guidance

Two-dimensional (2D) imaging can be performed readily at most hospitals. When freehand technique is used, plain radiography can demonstrate the cascade of the screws in two planes. This can be time consuming as films need to be printed and repeated if images are inadequate. These films do not assess the axial plane, where most errors are usually detected.

Another technique involves the mobile C-arm. The image intensifier can be draped into the surgical field and used flexibly to image in any plane but only takes one image at a time. Two C-arms can be used to take simultaneous biplanar images to save time. Still, axial plan information is difficult to interpret and surgery must be performed wearing lead. The patient and staff are all at risk for radiation exposure.

Limitations in these 2D techniques led to the development of intraoperative CT scans (e.g., O-arm), which can be used with or without navigation. Without navigation, the CT can be performed after implant placement to critically assess pedicle screw location after freehand technique, using the axial plane, which is more ideal. Using the CT with navigation allows real-time confirmation of the surgeon’s visual and tactile information showing the containment in the pedicle and the length and diameter of the screw. This maximizes the size of the screw that can be placed safely. The drawbacks include the expense of the machinery and the radiation exposure to the patient.

12.6 Navigation Technique

Navigation is the synthesis of physical landmarks on the surgical field with previously obtained or intraoperatively obtained imaging to form a visual guide for the surgeon when the target is deep to the surgical exposure. In spine surgery, navigation techniques elucidate pedicle morphology since the pedicle is usually not directly palpable or viewable during surgery unless a decompression or osteotomy is performed. In most modern systems, a posterior exposure is performed normally. Then a CT scan is performed with an O-arm (Fig. 12‑1) after a 3D array is attached to the spinous process (Fig. 12‑2) usually using the most cephalad spinous process exposed. When draping is complete (Fig. 12‑3 a), the entire team is evacuated from the room and the CT is run (Fig. 12‑3 b, c) and is uploaded to the navigation station. The surgeon registers the probe after the CT scan (Fig. 12‑4). Previous navigation systems used registration of anatomic landmarks after this step, but more recent CT-based systems have obviated this time-consuming step. These newer systems register the probe immediately and the probe becomes “live,” meaning it can be seen on the monitor superimposed on the anatomic structures in the axial plane with the invisible structures beneath the probe in line with the axis of the probe (Fig. 12‑5 a).

The image translation to the flat panel display screen shows the tip of the probe placed at a proposed starting point or bony landmark and the projected trajectory of any length-and-diameter screw before it is placed. These can be color coded for better visualization (Fig. 12‑5 b). One can then move the hand in space in order to optimize angulation in any plane. The awl, probe, tap, and screw can also be done under navigation or by freehand depending on surgeon preference.

Since the ideal way to assess the fidelity of screw placement is to use postoperative low-dose CT examination to visualize the screws in the pedicle, these systems can be compared to freehand techniques. ^{28 ,​ 35} Compared to freehand screw placement, navigated techniques have been associated in these studies with more optimally placed screws, fewer unacceptably placed screws, and fewer screw removals. ²⁸ Moreover, 2D navigation performs favorably compared with non-navigated techniques and 3D navigation performs favorably compared with 2D navigation. ^{26 ,​ 36 ,​ 37 ,​ 38} In combination with the navigation, a robotic arm can be attached to the bed and the trajectories confirmed by the surgeon can be instrumented by the robot. Robot-assisted implantation of pedicle screws has not yet been shown to outperform other navigation techniques. ^{39 ,​ 40} Notably, navigation setup times and cost increase with the complexity of navigation.