Introduction

An osteoid osteoma is a benign skeletal neoplasm composed of osteoid and woven bone that accounts for approximately 12% of benign skeletal neoplasms arising in the cortex of long bones. The majority of osteoid osteomas occur in children and adolescents, with spinal involvement in 10% to 20% of cases [13]. The most commonly affected part of the vertebrae is the neural arch in 75% of cases of spinal osteoid osteomas, with 33% involving the lamina, 20% involving the articular facets, and 15% involving the pedicles [43]. Osteoid osteomas located in the thoracic spine and particularly involving the vertebral body are even rarer [19]. Although most patients with adolescent scoliosis have little or no pain, osteoid osteomas are the most common cause in 2/3 of patients with painful scoliosis [17]. Patients with a spinal osteoid osteoma usually are treated successfully nonoperatively [8], and spontaneous healing of an osteoid osteoma may still occur within 3 to 4 years [24, 31]. Surgery generally is recommended for patients who do not respond to treatment with antiinflammatory drugs or when the tumor results in neural compression with deficits [8, 24, 33, 45].

However, anterior spinal surgery is a major and invasive option for treating osteoid osteomas of the vertebral bodies. The anterior approach to the thoracic spine traditionally has been gained by a posterolateral thoracotomy or a thoracolumbar incision, according to the affected level. Although the approach allows adequate exposure for curettage, there is the potential for substantial postoperative morbidity, including pain and compromised pulmonary function [7, 16]. Technologic advancements in endoscopic surgery have revolutionized traditional surgical approaches [3]. To date, video-assisted thoracoscopic surgery (VATS) has become a key to minimally invasive or access surgical approaches for a wide variety of anterior lesions in the thoracic spine [11, 25, 26, 44]. Video-assisted thoracic surgery has been used to perform a thoracic discectomy, resection of a spinal tumor, removal of an ossified posterior longitudinal ligament, spinal fusion, fractures, corpectomy, and spinal deformities [22, 26, 28, 44].

Even with video-assisted surgery, concerns regarding violation of the spinal canal leading to potential harm to vascular, neural, and other vital structures have resulted in techniques to improve the accuracy of spinal procedures [23, 25]. Several groups [4, 12, 15, 42] have reported a spinal navigation system (NAV) provides precise localization of the anatomy structures and instruments with minimal radiation exposure. These studies suggest accuracy of approximately 95% and reliability to reduce the margin of error in spinal surgery, such as pedicle screw insertion and resection of a spinal tumor with minimal bone removal. The demands concerning accuracy and less morbidity of spinal procedures have led surgeons to combine these newer technologies with their previous surgical arsenal [12, 39].

Preoperative image-guided localization associated with VATS has been reported as a useful therapeutic tool in the management of thoracic disease [9, 10, 36]. Although combining these techniques is not a new concept, its use in spine surgery has not been reported. We report a case of osteoid osteoma curettage in the thoracic spine performed with VATS combined with NAV (VATS-NAV).

Case Report

A 16-year-old girl was referred to our hospital with a 9-month history of mild left thoracic scoliosis (Fig. 1) and persisting dorsal thoracic pain that did not respond to NSAIDs. There was no history of injury. Her neurologic examinations and laboratory data revealed no abnormalities. CT and bone scintigraphy suggested an osteoid osteoma at the T9 level (Fig. 2).

Fig. 1
figure 1

An AP view radiograph of the thoracic spine shows mild left scoliosis.

Full size image
Fig. 2A–C
figure 2

(A) A bone scan shows focal increased tracer uptake (arrow) of the T9 vertebral body. (B) Axial and (C) sagittal CT scans of the spine show a nidus (12 mm × 10 mm × 8 mm) and its surrounding sclerosis (arrow).

Full size image

As nonoperative treatment was ineffective, we planned lesion curettage using VATS-NAV. The patient received general anesthesia and was placed in the left lateral decubitus position. Once the T9 level was identified under fluoroscopic control, entry points were drawn on the skin for performing a four-portal technique. The spinous process of the T9 vertebra was identified and marked on the skin for the transoperative attachment of the dynamic tracking device of the navigation-read instrument (Fig. 3). The navigation system we used was StealthStation® (Medtronic Sofamor Danek, Memphis, TN, USA).

Fig. 3
figure 3

The intraoperative setup of the VATS-NAV procedure is shown in this diagram. The dynamic tracking device of the navigation-read instrument is attached to the T9 spinous process (*). The dots represent the location of the portals.

Full size image

We then checked the location of the nidus and planned the entry point and projection of a hole toward the lesion (Fig. 4). To access the spinal canal and vertebral body, the rib head and proximal 3 cm of the ribs of the target lesion were removed. Using a 3-mm high-speed diamond burr, total nidus curettage was completed (Fig. 5). The surgical time was 60 minutes and the total time in the operating room was 90 minutes.

Fig. 4A–B
figure 4

(A) The NAV monitor shows the pointer (arrow) on the osteoid osteoma of the T9 vertebral body. (B) A VATS view shows a congruent image with NAV. 1 = working channel; 2 retractor channel; 3 = suction/irrigation channel; L = lung; d = diaphragm; r = rib; OO = osteoid osteoma; * = segmental vessels.

Full size image
Fig. 5
figure 5

A postoperative VATS view of the osteoid osteoma after total nidus curettage is shown. C = vascular clip; P = pointer of NAV; OO = osteoid osteoma.

Full size image

The patient reported immediate complete relief of preoperative pain and required only mild analgesics for postoperative pain related to the chest drain. On the second postoperative day, the chest drain was removed and the patient was discharged from the intensive care unit after review of a chest radiograph. She was allowed early mobilization without walking aids and received respiratory physiotherapy during hospitalization. The patient was discharged on the fifth postoperative day. She had improvement of scoliosis and was free of symptoms 5 months after treatment. CT scans at that time showed complete removal of the nidus without spinal instability (Fig. 6). Histologic examination revealed the typical structures for an osteoid osteoma.

Fig. 6A–B
figure 6

(A) Axial and (B) sagittal CT scans of the T9 vertebral body show the nidus curettage (arrow). C = vascular clip; * = removed rib.

Full size image

Discussion

Patients with an osteoid osteoma can be treated successfully nonoperatively, and spontaneous healing of an osteoid osteoma may occur. However, some patients cannot tolerate such long-term nonoperative treatment. If nonoperative treatment fails, the subsequent alternative would be surgery. Currently percutaneous radiofrequency ablation is the minimally invasive surgery recommended if the site of the osteoid osteoma permits [29], as in an osteoid osteoma located in the extremities. However, an osteoid osteoma located in the posterior arch, pedicle, or posterior wall of the vertebral body carries a high surgical risk of neurovascular damage (thermal damage) because the target is close to the spinal cord, nerve roots, and arteries [34]. For this reason, we selected surgical resection of the nidus using VATS-NAV for our patient.

VATS allows for access to anterior spinal lesions using minimally invasive principles. The literature suggests VATS can be performed with the same accuracy and completeness as is possible with the conventional open approach but through much smaller skin and muscle incisions [6]. This procedure is associated with less postoperative pain, better cosmesis, earlier return to normal activity, and lower perioperative morbidity such as blood loss and pulmonary complications [3]. The disadvantages of the thoracoscopic approach are the steep learning curve [25], keeping up with evolving surgical technique and instrumentation, and higher costs [41].

Numerous studies have described the use of video-assisted surgery in the management of spinal tumors [1, 18, 23, 27, 37], however we found few studies regarding primary tumors of the spine. Mori et al. [28] in 2011 described en bloc extirpation for an osteoid osteoma of a thoracic vertebral body through a thoracoscopic approach (without NAV). Our experience with endoscopic excision of spinal osteoid osteomas was reported [2, 14], but we had not yet used the NAV system.

In general, the intent of NAV is to improve the accuracy of spine surgery. Studies comparing conventional and computer navigation techniques have shown the superior accuracy of this technology [39]. However, few studies [4, 12, 15, 42] document the accuracy of NAV for spinal tumors. Publications concerning primary tumors are even rarer; to date, three papers have reported a total of 12 patients with spinal osteoid osteomas treated with the assistance of NAV [20, 32, 40]. Assaker et al. [5] reported two cases of image-guided endoscopic spine surgery but not for tumors.

Traditionally curative treatment of a spinal osteoid osteoma could be performed using VATS after initial fluoroscopic guidance. However intraoperative localization of the nidus is often difficult. A wide surgical resection of the bony structure is required to ensure removal of the nidus [35, 38]. Therefore, the NAV system may guide intraoperative nidus resection with good accuracy, resulting in complete intralesional excision without sacrificing more bone than necessary thus resulting in instability of the spine [30]. Intraoperative CT guidance or iso-C three-dimensional intraoperative spinal navigation in our patient could have provided the ability to resect a lesion [21, 32], but with much more radiation exposure than with the NAV system that required only initial fluoroscopic localization of the vertebral target. The surgery also could have been performed using VATS alone. However, we do not believe the endoscopic view from VATS alone is adequate to delineate safely the tumor margins and its relationship with some adjacent structures. We found NAV provided adequate orientation to the surgeon with sufficient accuracy on the anatomic landmarks that were identified preoperatively with the NAV system workstation (Fig. 4). It facilitated precise planning of the curettage and the surgical vector to the targeted small and subcortical lesion at the T9 level. It also helped to define the tumor margins, limits of curettage, and the adjacent spinal canal. Moreover, the NAV system has other advantages over conventional fluoroscopic guidance such as three-dimensional (3-D) visualization of structures, the spatial relationship between instruments and anatomy in real-time, and especially, no radiation exposure.

Curative treatment of a spinal osteoid osteoma can be performed with VATS after initial fluoroscopic guidance. However intraoperative localization of the nidus is often difficult. A wide surgical resection of the bony structure is mandatory to ensure removal of the nidus. Therefore, the NAV system may guide intraoperative nidus resection with good accuracy resulting in complete intralesional excision without sacrificing more bone than necessary. An intraoperative CT in this case could present the same result, but with much more radiation exposure than the NAV system with only initial fluoroscopic localization of the vertebral target.

However, caution is needed concerning NAV. First, accurate and reliable navigation requires a low system error and high rate of congruence between the patient’s preoperative 3-D images and the surgical anatomy. The frame of reference has to be attached to the vertebra in such a manner that the vertebra and the frame became one rigid body. The site of attachment may be the pedicle or spinous process of the vertebra. Obviously, attachment to the pedicle is more rigid than attachment to the spinous process, as reported previously [5]. We chose the spinous process because the osteoid osteoma covered the pedicle area on the right side and because of the lateral decubitus position of the patient. Moreover, we usually prefer to attach the frame of reference to the spinous process because usually a pedicle attachment takes time to transfer the patient from the radiology department to the operating room, as reported previously [5]. The time needed for performing VATS-NAV in the anterior thoracic spine in our patient was approximately 2 hours, which is comparable to the time for the open procedure when performed by the same surgical team.

The second problem is that tracking of optical array devices can be obscured or displaced by surgeons and/or instruments during the procedure. If any displacement occurs during image acquisition or surgery, the whole procedure can become useless and even dangerous if the surgeon is not aware of this displacement. This also could lengthen the time of surgical procedures. During our patient’s surgery, displacement of the frame of reference attached to the T9 spinous process occurred twice. However, we did not spend much surgical time adjusting the images for the NAV. Moreover, this surgery is not totally blinded, as seen in pedicle screw insertion, because of the real-time video endoscopic view.

Progressive advances in the technology of spinal surgery have evolved to offer fewer adverse events and less morbidity for patients. We believe VATS-NAV allows spinal operations to be performed in less time with less morbidity, and minimizes the hazards of radiation for the surgeon, patient, and operating room staff. VATS-NAV provided real and virtual images of the targeted lesion and positioning of the instruments in real time. Further prospective and comprehensive studies may confirm our initial impressions regarding the VATS-NAV combination.