Minimally invasive posterior cervical microforaminotomy in the lower cervical spine and C-T junction assisted by O-arm-based navigation

Abstract

Navigated posterior cervical microforaminotomy (PCM) is a promising minimally invasive technique for treating radiculopathy caused by lateral disc herniations and foraminal stenosis. Between December 2009 and October 2010, 14 patients with unilateral foraminal disc herniations or foraminal stenosis at the C6-7 or C7-T1 level underwent PCM assisted by O-arm navigation using the METRx tubular retractor. The main symptoms were radicular arm pain with no significant neck pain. Successful relief of radicular pain was achieved in all 14 patients. Two of the patients were lost during follow-up, and three had to undergo further decompression due to remnant foraminal stenosis being discovered on intraoperative O-arm images. There were no cases of instability or recurrence, and the only complication observed was a dural tear in one patient, which was adequately treated with fibrin glue and bed rest. The duration of symptoms was 4.5 months on average. The mean operation time was 136 minutes, with the additional time required for the image guided surgery assisted by O-arm-based navigation being approximately 28 minutes on average. There were no other complications during the surgical procedure or in the immediate postoperative period. Posterior cervical microforaminotomy assisted by O-arm-based navigation is a safe, effective and minimally invasive procedure for the treatment of lateral disc herniations and foraminal stenosis of the lower cervical spine and C-T junction, offering the possibility of an accurate decompression with a reduced risk of segmental instability.

• Posterior cervical microforaminotomy
• O-arm
• navigation
• lower cervical spine
• tubular retractor

Introduction

Posterior foraminotomy for decompression of cervical nerve roots is a well established procedure among operative treatments for degenerative cervical spine diseases associated with radiculopathy. It was first described by Spurling and Scoville in 1944 [1] and Frykholm in 1947 [2], and was then modified by Scoville et al. in 1951 [3]. Its indications are for patients with posterolateral soft disc herniations or foraminal stenosis.

Although the number of these surgeries has decreased since the description of the anterior approach by Smith and Robinson [4] and Cloward [5] approximately 50 years ago, this technique has found specific applications in recent years with the advent of the keyhole foraminotomy and the description of ideal indications and long-term results [6–11]. Moreover, prevention of the approach- and graft-related complications found in anterior surgeries, as well as avoidance of fusion with preservation of the stability and mobility of the segment, have heightened the interest in posterior foraminotomy among spine surgeons.

The minimal invasiveness of the METRx system tubular retractor (Medtronic Sofamor Danek, Memphis, TN) enables muscle-splitting dissection without the traditional extensive subperiostal stripping of the paraspinal musculature, as used in open posterior approaches, which increases postoperative pain, blood loss, muscle spasm and dysfunction. In this way, the morbidity of the procedure is reduced, accelerating recovery and shortening hospital stay. However, minimally invasive PCM is not easily performed in obese patients who have a short neck, especially in the lower cervical spine or C-T junction.

Over the last decade, the use of image guidance systems has benefited surgeons seeking greater accuracy, especially for the placement of pedicle screws [12]. The complex three-dimensional (3D) anatomy of the spine, the proximity of neurovascular structures, and the demand for less aggressive procedures are some of the factors that have contributed to the growing use of navigation in spine surgeries. Previous articles have described successful image guided decompression of the cervical spine, either through a transcorporeal anterior microforaminotomy [13], or by anterior corpectomy and foraminotomy in cadaver models [14], [15].

This paper describes the surgical technique of minimally invasive posterior cervical microforaminotomy (PCM) assisted by an O-arm-based navigation system, as used in the treatment of unilateral radiculopathy due to foraminal disc herniation or foraminal stenosis in the lower cervical spine and C-T junction, and presents clinical results obtained in a group of patients.

Methods

Patients

Between March 2009 and October 2010, 14 patients underwent PCM assisted by O-arm navigation using the METRx tubular retractor at the C6-7 and/or C7-L1 level. The main symptoms were unilateral radicular arm pain without significant axial complaints, and some patients had slight motor weakness. Study images showed lateral herniated discs and/or foraminal stenosis in one or two segments (Figure 1). Patients with central disc herniations or spinal canal stenosis, or with symptoms of myelopathy or radiculomyelopathy, were excluded from the study. O-arm navigation was used in all the procedures and assisted the surgeons in performing full root decompression and achieving maximum preservation of facet joints.

Figure 1. T2-weighted MRI showing a right lateral soft disc herniation at C7-T1 (red arrow). Left: sagittal foraminal view; center: axial view; right: CT axial view.

Surgical technique

The patient is placed in a prone position under general anesthesia. Somatosensory evoked potentials and myotomal electromyography (EMG) are monitored throughout the procedure (NIM-Spine System, Medtronic, Inc., Memphis, TN). Before draping, a needle is inserted into the skin and a radiograph is obtained to locate the level that will be decompressed.

The operating room set-up for an image-based navigation using the StealthStation Treon system (Medtronic Surgical Navigation Technologies, Louisville, CO) consists of an infrared camera positioned at the caudal end of the surgical table with its monitor placed on the opposite side from the principal surgeon to facilitate visualization during the procedure. The dynamic reference base (DRB) is attached over the cervicothoracic junction or over the upper thoracic levels, and contains passive markers that reflect light from the infrared light source integrated with the tracking camera system (Figure 2).

Figure 2. Operating room setting. (1) Intraoperative CT scan being performed with O-arm; (2) navigation monitor; (3) infrared camera system (IFS); (4) NIM (neural integrity monitor); and (5) dynamic reference base (DRB).

As soon as surgical preparation is complete, with the patient's position fixed on the operating table, an intraoperative CT image set is obtained with the O-arm equipment (Medtronic, Inc., Memphis, TN) and transferred to the StealthStation Treon image guidance workstation, where it is automatically registered (Table I).

Table I.  O-arm workflow.

Procedure	Time (min)
Draping	3
Moving O-arm into surgical field/  ensuring camera visibility	6
Scanning	1
Parking O-arm	3
Uncovering sterile drape	1
Ready to go	14
Total	28

The operative level is now again confirmed using the navigation system. With the aid of the sagittal CT and the fluoroscopy mode reconstructions aimed at the target facet joint and the axial CT views pointing over the laminofacet junction, the proper entry point is defined (Figure 3). A 1.6- to 2.0-cm incision is made in the skin as well as the cervical fascia. For two-level procedures, the incision is placed midway between the levels to be approached. Finger dissection is used to split the paravertebral muscles, then sequential dilators are serially inserted and the METRx tubular retractor is placed over the cortical bone. At this point, it is necessary to check the matching accuracy between the patient's anatomy and the images provided by the navigation system on the monitor. For this purpose, specific osseous landmarks inside the operative field are selected as references with the pointer probe, and are then compared to decide if safe navigation is possible. If the accuracy is not acceptable, the matching procedure must be repeated. Once the proper level has been reconfirmed with navigation, the METRx retractor is fixed in the selected position with a table-mounted flexible retractor arm.

Figure 3. Left: Preoperative radiograph showing the difficulty in finding the correct level (C6-7 level) for the surgical approach. Right: O-arm navigation images allowing proper localization of the level and laminofacet junction.

Remaining soft tissues are cleared from the field, so that the facet joint, both ipsilateral lateral masses and laminofacet junctions can be viewed satisfactorily. Subsequently, the pointer probe is used to determine the shape and size of the drill hole according to the preoperative plan. A 4-mm diamond burr with a high-speed drill (Black Max, Anspach, Palm Beach Gardens, FL) is preferred both to provide security for neural structures and for its hemostatic effect on the bone. In cases where only a small foraminotomy is necessary, a 3-mm diamond burr is used instead. Bone removal begins with the lateral part of the superior and inferior hemilaminae and then progresses to the medial portion of the descending facet. The drilling continues until the entire medial aspect of the ascending facet is exposed. The opened hole is then checked again with the navigation probe to determine whether any adjustment is needed. Sometimes, a 2- or 3-mm diamond burr is needed at this point to ensure safety and precision in drilling.

Subsequently, the ascending facet and the remaining hemilaminae are drilled until a thin layer of the deep cortical bone is visualized. Then, with a small 45° angled curette and a 1-mm Kerrison punch, the soft tissues covering the neural foramen and the lateral spinal canal are exposed. Using dissecting hooks and the same 1-mm Kerrison punch, the ligamentum flavum is removed, and now the lateral dural sac as well as the nerve root can be seen. In the event of epidural bleeding, the source can be filled with gelatin foam embedded with a thrombin activator component or Floseal (Baxter Healthcare Corporation, Deerfield, IL). We discourage the use of bipolar cauterization, and any such use should be limited, keeping it to a low-intensity mode. All surgical procedures following the positioning of the tubular retractor are performed under the operating microscope. After this, the features of the hole are reviewed visually, with a small probe to search for the rostral and caudal pedicles and with the navigation probe. To assure complete decompression of the root, O-arm scanning is always performed in all cases at this point in the procedure. We thereby expect to remove every remaining compression point inside the foramen, whenever indicated. After checking that the decompression is adequate, the operation wound is closed (Figure 4). During the follow-up period, a 3D CT scan is checked to evaluate the grade of laminofacet resection (Figure 5).

Figure 4. Left: A 1.6-cm skin incision for one-level decompression. Right: A 2.0-cm skin incision for two-level decompression.

Figure 5. Postoperative posterior view of a 3D CT scan showing the small keyhole for decompression of a C6-7 left foraminal stenosis and demonstrating adequate preservation of facet joints.

Results

The clinical results are summarized in Table II. From December 2009 to October 2010, 14 patients and 16 levels were operated. The series comprised 13 men and 1 woman whose mean age was 49.8 years (range: 37–67 years). Two patients needed a two-level decompressive surgery while 13 had single-level disease. Two patients have subsequently been lost to follow-up; however, their clinical results were excellent up until 3 months after surgery. The mean symptom duration was 4.5 months, and the mean follow-up duration was 8.6 months (range: 6–13 months). Successful relief of radicular pain was achieved in all patients. The mean preoperative VAS (Visual Analog Scale) score for arm pain was 7.7, which improved to 1.8 (p < 0.001), and the patients’ mean score on the Oswestry Disability Index (ODI) had improved from 47.4% to 10.3% (p < 0.001) by the time of the last follow-up. There were no cases of instability or recurrence; however, one patient underwent fluoroscopy-guided selective nerve root block (SNRB) at the affected level due to residual pain and a tingling sensation, and showed a marked improvement in arm discomfort. A further three out of 12 patients required medication (Pregabalin, Lyrica, Pfizer) for 1–3 months due to residual tingling sensation. No patients have suffered from moderate to severe axial neck pain (VAS 4∼10).

Table II.  Demographics for posterior cervical microforaminotomy assisted by O-arm-based navigation.

Case number	Sex	Age	Operated levels	Side	Intraoperative bleeding (cc)	Operation time (min)
1	M	49	C6-7/C7-T1	Left	200	190
2	M	53	C6-7	Left	50	120
3	F	67	C6-7	Left	130	140
4	M	48	C6-7	Left	170	85
5	M	45	C6-7	Left	120	115
6	M	52	C6-7	Left	50	140
7	M	57	C6-7	Left	120	95
8	M	43	C6-7	Left	150	195
9	M	52	C6-7	Right	130	100
10	M	54	C6-7	Right	80	115
11	M	40	C6-7/C7-T1	Left	250	200
12	M	37	C6-7	Right	90	140

The only complication observed was a dural tear in one patient, which was adequately treated with fibrin glue and bed rest. The operation time was 136 minutes on average, with the additional time for the image guided surgery assisted by O-arm-based navigation being approximately 28 minutes on average. There were no other complications during the surgical procedure or the immediate postoperative period.

Discussion

Posterior cervical foraminotomy is a proven and established technique in the spine surgeon's arsenal for treating radicular symptoms originated from lateral disc herniations and foraminal stenosis. Its safety and effectiveness have already been demonstrated in previous studies, with good-to-excellent results being achieved in more than 90% of cases [6–11]. Among its advantages, fusion is avoided in the operated level, stability and mobility are preserved, and there are no complications related to the approach, such as are found with anterior cervical techniques.

Another important benefit relates to the tubular retraction system, which was previously described by Foley and Smith for lumbar microdiscectomy and was then adapted for cervical spine surgeries [16]. Extensive muscle-stripping dissection is not necessary, reducing postoperative pain and discomfort as well as muscular dysfunction and blood loss. Some studies have demonstrated equivalent clinical results between open and tubular retractor-assisted posterior decompressions; however, the minimally invasively operated cases had less blood loss, shorter hospitalizations, lower analgesic requirements, and less postoperative neck pain [17], [18]. Nonetheless, it is important to remember the classic contra-indications for posterior foraminotomy, which are related to poor results: pure axial neck pain, gross cervical instability, symptomatic central disc herniations and kyphotic deformity. All patients with such findings were excluded from the study and underwent other forms of treatment.

In this study we performed the same technique for two-level ipsilateral foraminotomy in two patients. Postoperatively, the patients had the same satisfactory results as the single-level treated patients. Holly et al. already demonstrated that two-level minimally invasive posterior cervical foraminotomy has rates of success and complications equivalent to those of conventional foraminotomy [19]. Gala, O’Toole and colleagues have also described the technique using the METRx system and included two-level procedures, in which a same-length incision was placed between the levels of interest [20], [21]. We used this same approach whenever necessary.

We have found that the main pathological processes that are conducive to nerve root compression are acute posterolateral soft disc protrusions and/or osteophytes originating anteriorly from the uncovertebral joint or posteriorly from the facet joint. We believe that the preoperative planning is one of the most important steps in the procedure and thus recommend analyzing both MRI and CT scans, on the grounds that the former are more diagnostic for acute disc protrusions whereas the latter are better for evaluation of osteophytes and arthrosis.

There are certain issues that make the posterior cervical microforaminotomy a difficult procedure for specialized surgeons. Of these, the complex 3D anatomic characteristics of the spine, especially the cervical vertebrae, and the limited visibility provided by a tubular retractor only 2 cm in diameter are the most noteworthy. However, other important features, e.g., the proximity of important neural and vascular structures and the requirement for a satisfactory decompression without destabilizing the segment, have motivated the search for a safe and effective technique which preserves the minimally invasive character.

Although the current practice of posterior microforaminotomy and endoscopic foraminotomy has demonstrated their effectiveness in several studies, Fessler and Khoo related that 8% of cases of radiculopathy symptoms remained the same after surgery [17], while Kim and Kim presented success rates of 86.4% in 22 patients undergoing tubular retractor-assisted decompression [18]. In our opinion, in cases where the radiating pain is caused by true foraminal stenosis or lateral disc herniations, a complete decompression can reduce the percentage of non-satisfactory results.

The segment stability issue must also be taken into account. Previous reports have highlighted the importance of preserving at least 50% of the facet joint at the risk of compromising shear strength or giving rise to segmental hypermobility [22], [23]. With careful and accurate preoperative planning, based on CT or MR images, it is possible to determine the proper amount of bone that must be removed from the ascending and descending facets without risking eventual instability. In this regard, O-arm-based navigation provides the best quality reference images prior to surgical programming and ultimately directs the surgeon to accomplish an adequate and secure decompression. Before commencing the drilling and during the procedure itself, the pointer probe is used to check the shape and size of the keyhole, as well as to determine whether any adjustment in trajectory or depth is necessary. For this purpose, axial and sagittal reconstructions are often verified on the navigation monitor.

Image guided surgery has been used to assist in a variety of spinal procedures which demand accuracy. Several reports have described its use in pedicle screw placement, as well as for ventral approaches, disc replacement and tumor surgeries [24–28]. Albert, Klein and colleagues have demonstrated in two different studies the feasibility of performing anterior cervical foraminotomies and corpectomies on an image guided basis, and have achieved improved precision and additional safety with respect to the integrity of the vertebral artery [14], [15]. In a recent report, Kim et al. highlighted the invaluable assistance provided by O-arm-based spine navigation in defining the orientation for a satisfactory decompression via anterior cervical microforaminotomy [13]. As in our present study, the technique provided valuable assistance in circumstances where anatomic localization is difficult due to the limited field of view.

Another notable benefit of the O-arm-assisted approach derives from the ease of locating the correct surgical level, thereby facilitating the selection of the appropriate entry point. Patients having short and thick necks with lower-level disease present some difficulty if traditional preoperative radiography is to be used. With the O-arm-based approach, navigation can help to minimize the size of the skin incision.

As a result of our experience with cases operated with O-arm assistance, we came to recognize the remarkable importance of the superior articular process (SAP) in the pathology of cervical foraminal stenosis, which is easily visible in a foraminal O-arm view (Figure 6). In three cases we extended the foraminotomy because remaining compression due to a SAP osteophyte was visualized upon checking intraoperative O-arm scans. Consequently, we consider this examination to be indicated after all decompressions to determine whether the performed microforaminotomy is adequate, in order to avoid the need for revision surgery.

Figure 6. Intraoperative O-arm images used to check the adequacy of decompression. A foraminal view (lower right) can be obtained to evaluate the remnant foraminal stenosis.

All of the 12 patients improved with respect to their radicular pain, and in one case a small dural tear was found by the end of the procedure. There has been no recurrence so far, though most of the patients have not yet completed follow-up.

In these cases, the DRB was attached over the upper thoracic spine or over the mid-thoracic levels, because much of the accuracy was lost when the DRB was fixed to the operating table. We did not attach the DRB to the patient's spinous process because we consider this to be inappropriate in minimally invasive surgery. The DRB was instead attached to the patient's dorsal surface, in a position close to the level to be approached.

We strongly believe that the longer operating time with O-arm-based navigation, compared to that for conventional surgeries, is compensated for by all the benefits mentioned above. We also believe that the surgeon's knowledge of local anatomy and proficiency in performing the conventional techniques are essential for the successful and safe use of the minimally invasive PCM assisted by O-arm-based navigation.

The shortcomings in our study are the limited number of cases and the short follow-up period.

Conclusion

Posterior cervical microforaminotomy assisted by O-arm-based navigation is a safe, effective and minimally invasive procedure for the treatment of lateral disc herniations and foraminal stenosis of the lower cervical spine and C-T junction, offering the advantage of an accurate decompression together with a reduced risk of segmental instability.

Declaration of interest: This study was sponsored by a fund from the Wooridul Spine Foundation.