Introduction

Pedicle screws are widely implanted in the thoracolumbar area for a variety of pathological processes, traumatic and not [5]. Suboptimal placements of pedicle screws may lead to neurological and vascular complications, potentially requiring reoperation [1, 2, 6, 11, 15].

Pedicle screws are placed either free hand following anatomical landmarks, often augmented by intraoperative fluoroscopy, or using spine navigation based on intraoperative fluoroscopy or computerized tomography (CT) [1–3, 8, 10, 14, 17, 18].

The CT scans for spine navigation are most often acquired pre-operatively, even though recent reports have used intraoperative CT-based navigation [6, 7, 11].

The O-arm® Surgical Imaging System (Medtronic, Minneapolis, MN, USA) provides high-definition 3-D, multi-plane fluoroscopic images that can be automatically registered to a stealth neuronavigation station (Medtronic, Minneapolis, MN, USA) in a time-efficient process. As with any navigation system, it is imperative to assess the operational accuracy of the system while it is being used [12].

We describe our novel technique to continuously check the operational accuracy of the navigation system during thoracolumbar pedicle screw placement using O-arm-based neuronavigation.

Materials and methods

Patient population

We used the combination O-arm-stealth for intraoperative spine navigation during the placement of thoracolumbar pedicle screws in 12 patients in which a total of 82 screws were inserted. Table 1 details the levels of the spine that were instrumented. Six patients had traumatic injuries (four in the thoracic and two in the lumbar areas) while six patients had degenerative disease (all in the lumbosacral area). Nine patients underwent decompression and fusion and three patients underwent only fusion. All patients underwent post-procedural intraoperative O-arm imaging. All patients underwent post-operative thin slices control CT scan of the instrumented area (usually between 3 and 6 months). Follow-up ranged from 4 to 24 months (mean and median follow-up: 20 months).

Table 1 Pedicle screws distribution
Full size table

O-arm and stealth navigation

The O-arm (Medtronic, Minneapolis, MN, USA) is an imaging system that incorporates a flat panel detector and can provide standard fluoroscopic images or 3-D, volumetric high-definition fluoroscopic images. This dataset can be downloaded to a stealth workstation (Stealth Station, Medtronic, Minneapolis, MN, USA) where it is automatically registered and readied for navigation in about 55 s.

Operative technique

The following algorithm was used :

  1. Step 1: Preparation

    After the patient was positioned prone on the operating room table, the O-arm was used in a conventional X-ray mode to center the area of interest (spine segments to be navigated) in the anteroposterior and lateral plane.

    Next, small, 5-mm, linear incisions were made over the tips of the spinous processes of the spine segments to be navigated and 4-mm-self drilling micro-screws (Synthes, West Chester, PA ) were implanted into the exposed tip of the spinous processes. These micro-screws were to be used as reference points (internal fiducials) for verification and evaluation of the operational accuracy of the navigation system during every step of the surgery. In the lumbar spine, the navigation tracker was affixed in the iliac spine or on the cranial or caudal spinous process while in the thoracic area it was affixed on the cranial or caudal spinous process.

  2. Step 2. Image acquisition

    The spinal segment of interest was scanned using the O-arm and the images were automatically registered to the Stealth Station.

  3. Step 3. Accuracy verification

    The tip of the navigation pointer was placed on the head of the internal fiducials (spinous processes micro-screws) to assess operational accuracy. Operational accuracy was judged to be good when no difference was found between the navigation pointer tip position in patient’s space and the navigation pointer tip position in image space. (Fig. 1).

    Fig. 1
    figure 1

    Accuracy verification. The tip of the navigation pointer is placed on the head of the internal fiducial (spinous processes micro-screw) to assess operational accuracy

    Full size image

    If there is misalignment between these two positions we look for possible errors (smudged tracking balls, tracking instrument malfunction, etc.) and correct them.

  4. Step 4: Pedicle screws placement

    A navigated Pak needle was then introduced percutaneously using a short (1 cm) paramedian incision to make contact with the bone of the lateral elements of the spine. Once the Pak needle was in alignment with a suitable trajectory to enter the pedicle, we impacted it into the bone using a hammer (Fig. 2). We hammered the Pak needle well into the vertebral body. We then removed its stylet and placed a K wire through the needle. The K wire was slightly tapped into the bone advancing it a couple of millimeters from the tip of the needle. The Pak needle was then removed and a navigated tapper of appropriate size was used to tap into the vertebral body under continuous navigation guidance (Fig. 3). Next, the appropriate screw was selected and loaded on the percutaneous screw carrier. The insertion of the screw was again monitored using image guidance by navigating the screw carrier carrying on its tip the appropriate virtual screw (Fig. 4). Once the screw was inserted, the K wire was removed. Anytime we used a navigated instrument we double checked the operational accuracy according to step 3.

    Fig. 2
    figure 2

    The navigated Pak needle is at the entry point

    Full size image
    Fig. 3
    figure 3

    The navigated tapper is shown entering the vertebral body

    Full size image
    Fig. 4
    figure 4

    The insertion of the screw is again monitored using image guidance by navigating the screw carrier that has on its tip the appropriate virtual screw

    Full size image
  5. Step 5: Assessment of the position of the placed screws

    After all the pedicle screws were inserted, we obtained another O-arm high-definition scan to check on the pedicle screws position (Fig. 5 left). We used the Mirza et al. scoring system to evaluate and assess the position of the inserted screw [10].

    Fig. 5
    figure 5

    Post-procedural intraoperative O-arm scan showing good position of pedicle screws(left) and thin-slice post-operative CT of the same level (right). Both images yield comparable information

    Full size image
  6. Step 6

    Once we were done with pedicle screws placement, we proceeded with the rest of the operation that included posterior fusion with or without decompression.

Results

Overall, 82 screws (40 in the thoracic region and 42 in the lumbosacral area ) were placed in 12 patients (Table 1).

No patient experienced any complication related to pedicle screws placement. We repositioned two screws during surgery because the screws had minimal (less than 2 mm) lateral pedicle cortex breech.

None of the screws needed to be repositioned after surgery and hence no patient was brought back to the OR for screws repositioning. The post-operative thin slices CT did not yield any different information than the post-procedural intraoperative O-arm (Fig. 5).

All of our screws were in excellent position according to Mirza’s evaluation system [10].

The implanted internal fiducials were visualized very well and very easily in all the 3-D images. The operational navigation accuracy assessed using the internal fiducials was found to be excellent. The number of trajectories available to place pedicle screws using our technique was plentiful and we were able to take best advantage of each vertebra different anatomy.

Discussion

Free hand or conventional fluoroscopy-based pedicle screw placement is associated with suboptimal pedicle screw placement in 5–55 % of cases [1, 4, 6, 11, 13],while the incidence of anatomical injuries and reoperation is of 6.9–8 % [2, 5] and 4–8.8 %, respectively [1, 15, 16]. With the advent of navigation-based pedicle screws placement, the incidence of these events has decreased [18].

It is clear that navigation helps in improving the accuracy of pedicle screws placement. A recent meta-analysis report gives a median accuracy of 93.7 % (range 72 to 100 %) with navigation versus a median accuracy of 86.6 % (range 27 to 100 %) when navigation is not used [9].

One advantage of navigation based on intraoperative image data set acquisition (3-D fluoro or intraoperative CT) versus navigation based on pre-operative image acquisition is that the former is performed with the image acquired with the patient in the surgical position, therefore decreasing spine segments motion. In addition, the former may allow for multiple level registrations at once and does not need presence of the posterior elements of the spine, making navigation doable in revision cases [11, 19].

When one looks at the overall cost associated with placement of pedicle screws using navigation based on intraoperative image acquisition, the significant capital equipment investment of an intraoperative imaging machine, about $600,000 in the O-arm case, must be balanced by the total cost (expenses and loss of professional revenues) linked to the need of bringing a patient back to the operating room to reposition a pedicle screw. This total cost may be as high as $15,000–20,000. This is in addition to the extra cost that may be associated with treatment of injuries linked to malpositioned pedicle screws. Moreover, there is an extra value provided to the patient by avoiding neurological injuries and a second surgical procedure.

Our technique is novel because it allows for at will check of the operational accuracy, using our implanted internal fiducials, therefore controlling for errors due to instrument malfunction or system failure. As the internal fiducials were placed on each segment to be instrumented and as the spinal segment represent a volumetric unit, we felt that excellent accuracy at the tip of the spinous process would translate into excellent operational accuracy throughout that spine segment; our results seem to validate this assumption [5]. Our technique may represent an useful addition to existing technique to use neuronavigation to place pedicle screws.

Conclusions

The accuracy of pedicle screw placement in the thoracolumbar area is enhanced by using our proposed algorithmic step wise technique that includes at will check of the operational accuracy of the navigation system while utilizing the O-arm and Stealth Station. Our proposed algorithmic technique allows for efficient check of intraoperative navigational accuracy and for use of a wide variety of pedicle screws entry points.