Multiple challenges of deformity correction have been well documented in treating the bizarre and complicated DS-NF1, including the management of dislocated rib head penetrating into spinal canal. The neurologic status and the severity of intra-spinal occupation by rib head are the pivotal factors that determine the treatment strategy for this condition [8]. No reports have been addressed the effectiveness of apical fixation on treating RHD. The present study, for the first time, comprehensively assessed and confirmed the benefits that screw/hook placement at vertebra with RHD could improve the extraction degree of the penetrated rib head from the spinal canal in DS-NF1.
Previously, resecting the dislocated rib head should be aggressively performed when the following two aspects were indicated [5, 7, 8]. First, the spinal cord being compressed by rib head was detected by T2-weighted MR images[8]. Second, painful rib hump by palpation induced symptoms of radicular neuralgia or limb weakness [7]. After posterior laminectomy and decompression, penetrated rib head could be directly visualized and resected [6]. However, this intra-spinal procedure increased bleeding and operative time, which may bring new risk of neurological complications [4]. Moreover, adhesion between the rib head and the dural sac of the spinal cord should be carefully evaluated, so as to avoid aggravating the nerve damage by violent traction. Accordingly, Mukhtar et al. proposed a ‘left in situ’ strategy as an alternative when the excision of rib head could not be accomplished directly via posterior approach [6]. This concept involved preservation of rib head and removal of a 5-cm segment of rib’s lever arm, which could be helpful for avoiding the interference and traction of the spinal cord and meanwhile eliminating the injury of the ‘leverage’ effect of the rib tip on the spinal cord. Besides, removal of periosteum completely was a necessary preventative method for the regeneration of ribs and reconstruction of the continuity between the osteotomy ends [6]. The remained rib head might get reshaped over time [6]. However, no validated follow-up data was reported currently.
The above-mentioned management strategies were apposite for patients with existing or impending neurological impairment, while for patients who were neurologically intact, whether or not to resect the dislocated rib head was still in controversial. The current mainstream perspective was more inclined to retain the dislocated rib head to avoid interfering the spinal cord, tearing the dural sac, and reducing the bleeding and operative time [5, 10]. Mao et al. suggested that more correction of vertebrae translation and restoration of normal RVA was beneficial for reducing the invasion of the rib head dislocated into the spinal canal [5]. Cai et al. indicated that this phenomenon could be attributed to the laxity or dislocation of the costovertebral articulations, which allowed withdrawal and pullout of the penetrated rib being possible when the corresponding vertebrae was dragged to the concavity following corrective maneuvers [16]. This implied a passive mechanism that superior deformity correction, which was more or less correlated with higher implant density [17], was beneficial for extracting the dislocated rib head from spinal canal. On the other hand, the degree of retraction of dislocated rib head could theoretically be further improved if the vertebrae with RHD were anchored and dragged back to midline with well derotation, serving as an active mechanism. Therefore, screw/hook insertion at vertebra with RHD was supposed to be an ideal fashion retracting the penetrated intracanal rib head in DS-NF1.
Technically, screw insertion at the pedicle was relatively challenging compared with hook placement at the lamina. The dystrophic and thin pedicles brought about significant challenge of reliable pedicle screw insertion, high frequency of screw malposition [18] and high risk of cord injury. Aside from mutable screw trajectory, low bone mineral density in NF-1 might also contribute to the poor biomechanical properties of the pedicle-screw at these levels [18]. All these technical difficulties and the attendant risks rendered the apical dystrophic spinal segments a screw-forbidden zone in most circumstance. Currently, no quantitative data was available supporting that the screw/hook placement at dystrophic pedicle was beneficial and should be tried with utmost effort in case of RHD with impending neurological deficits.
The result of the present study showed a rate of 50.9% (27/53) for pedicle screw placement in the corresponding dystrophic vertebrae with dislocated rib head. And the correction rates of Cobb angle, IRL, VT and the RVA in the screw/hook group were significantly higher than those in the non-screw/hook group. Further, more corrections of Cobb angle, VT and the RVA contributed significantly to the correction of IRL through multiple linear regression analysis. The above results confirmed that after screw/hook placement combining with higher corrections of Cobb angle, VT and the RVA at vertebrae with RHD, the three-dimensional spatial relationship between rib head and spinal canal could be changed via translational traction and derotational withdrawal. This was helpful for maximizing the degree of spontaneous withdrawal of dislocated intracanal rib head. Accordingly, if possible, the screw/hook placement at vertebrae with RHD should be aggressively advocated. With the extensive clinical application of O-arm navigation, the precise screw placement will be further increased [18]. This was confirmed by Jin et al. that the accuracy of pedicle screw insertion in apical region was increased from 67% with free-hand technique to 79% with O-arm-based navigation technique [18]. Aside from the immediate postoperative withdrawal, the increased spinal stability provided by rigid screw/hook constructs at apical area was also beneficial for preventing progression of RHD and the attendant risks of cord injury [10], all of which were supportive of rib head retaining strategy in neurologically intact DS-NF1 patients.
The clinical relevance of the current study is that despite being risky and challenging, screw/hook placement at dystrophic vertebrae with RHD was beneficial for spontaneous retracions of the intracanal rib head. For patients without neurological symptoms, preoperative pedicle, lamina and vertebral body morphology at the corresponding level of RHD should be carefully evaluated, which is crucial for assessing the possibility and risk in apical fixation. Navigation-based technique can be used to improve the accuracy of pedicle screw placement. If the pedicle at corresponding level of RHD is not a good recipient to well contain the pedicle screw, the ‘in-out-in’ technique is applicable [19]. Otherwise, the lamina hook should be considered as an alternative when pedicle screw insertion was failed. For extreme cases, if the lamina is too thin and osteoporotic to anchor the hook, the segment corresponding to dislocated rib head can be left without apical fixation. In this situation, increasing the implant density in the adjacent segments and improving the curve flexibility using Grade 2 osteotomy are indirect ways to improve the withdrawal of rib head by increasing the corrections of Cobb angle, vertebral translation and rotation. During all these procedures, intraoperative neurophysiological monitoring is indispensable. Lastly, we would like to emphasize that the outcome of this management strategy was highly correlated with the surgeon’s experience and operative technique, and should be carried out meticulously. These aforementioned knowledges are essentially instructional and would assist in treating patients with this condition.
The limitations of this study should be addressed. First, CT evaluations at the long-term follow-up were not available currently. Thus, whether or not the correction of the extracted rib head would be lost due to the loss of main curve correction, formation of pseudarthrosis or failure of internal fixation could not be precisely elucidated. Second, it was impossible to tell the difference between screw and hook insertion on degree of withdrawal of rib head due to the limited sample size.