Do Rib-Based Anchors Impair Chest Wall Motion in Early Onset Scoliosis (EOS)?

Introduction

Early onset scoliosis (EOS) is defined as a spine or chest wall deformity diagnosed prior to the age of 10 [1]. As severe deformity progresses in this patient population, children may go onto develop thoracic insufficiency syndrome (TIS), defined as the inability of the thoracic cavity to support normal respiration or lung growth [2,3]. Treatment for these children involves preservation of pulmonary function through restoration of a less deformed thoracic cavity, promoting thoracic growth, and limiting early spine fusion [4]. Those tasks may be accomplished through a variety of methods; primarily through distraction-based growth friendly instrumentation via traditional growing rods (TGR), vertical expandable prosthetic titanium ribs (VEPTR), or magnetically controlled growing rods (MCGR) [5–9]. Additional techniques such as growth guidance or the Shilla using the spine or the pelvis as anchor points also exist to manage these complex deformities [10,11].

An area of clinical equipoise in the EOS community is the type of proximal anchor used in a posterior distraction-based construct. Surgeons have various fixation options that can be broken up into either spine-based or rib-based anchor points. Biomechanical studies exist reporting on fixation strength for various proximal anchor constructs, but a large majority of the studies evaluate deformity correction percentage and complication rate which comprises mostly infections or proximal anchor failure [12–18].

There is a concern in pediatric spine deformity surgery that rib-based fixation may limit chest wall motion. This philosophy stems from the fusion mass created by device migration, the law of diminishing returns and auto fusion identified in EOS at the time of final fusion, and the complex thoracic cage osteotomies needed at time of graduate final fusion surgery needed to correct severe deformity [19–21].

However, no published literature exists that directly answers the question as to whether rib-based anchors impair chest wall motion in EOS. Unfortunately, until recently there was no practical methodology to evaluate the regional contributions of the various components of the respiratory cycle [22, 23]. In this study, we utilize a recently developed methodology, quantitative dynamic magnetic resonance imaging (QdMRI), to explore the answer to this clinical question. QdMRI [24, 25] is an imaging-based non-invasive methodology that provides quantitative information about regional respiratory function. Images are obtained during tidal breathing at rest, without the need for breath-holding, maximal expiratory effort, or other external monitoring devices, making this technique extremely practical for clinical use, even when patients are unable to follow specific breathing instructions [22, 26]. Other major strengths of QdMRI are that it can be performed in very young EOS patients [22] with high spatial and temporal resolution, affords large field of view that covers the entire thorax and abdomen, and offers sufficient image quality to utilize automatic artificial intelligence (AI)-based image segmentation of the lungs and other anatomical structures of interest [27–29].

We hypothesize that children with EOS managed with posterior distraction-based surgery and rib based proximal anchors will not have impairment of the chest wall contribution to the respiratory cycle post-operatively, and thus not experience a clinically significant chest stiffening.

Materials & Methods

Results

Tidal volume comparisons in EOS patients before and after surgery with and without age correction and without distinguishing between concave and convex sides of the spinal curve are shown in Table 2. Without age correction, all parameter tidal volumes increased after surgery (p<0.05). Both left and right chest wall tidal volumes significantly increased after surgery by 50.66% and 41.18% (without age correction), respectively, and by 17.19% and 24.71% (with age correction), respectively. Left and right Dtv (with age correction) did not significantly differ before and after surgery (p>0.05).

Table 3 shows the comparison of post-operative volumes with volumes of age-matched controls. All tidal volumes except LDtv (p<0.001) are close to that of normal children and do not differ significantly from those of control subjects (p>0.05).

Table 4 presents tidal volume comparisons before and after surgery when considering the directionality of the spinal deformity without age correction (Table 4a) and with age correction (Table 4b). In this analysis we did not identify major spinal curve types. CWtv significantly increased after surgery by 46.44% and 30.88% (without age correction) for concave and convex sides, respectively (Table 4a), significantly increased after surgery by 28.57% (with age correction) for the concave side (p < 0.002), and non-significantly increased after surgery by 14.82% (with age correction) for the convex side (p = 0.059) (Table 4b).

Among the 98 studies acquired in 49 paired EOS patients, there were 8 with PTC, 52 with MTC and 20 with TLC and 18 with LC. EOS patients with the MTC type of major spinal curve type formed the largest subgroup with 26 paired EOS patients before and after surgery. Table 5 shows the results of CWtv in this MTC group of EOS patients while considering the concave and convex sides of the spinal curve without age correction (Table 5a) and with age correction (Table 5b). The results of CWtv for the MTC subgroup of EOS patients without distinguishing concave and convex sides of the spinal curve before and after surgery are listed in supplemental Table S1, where CWtv (left or right) significantly increased after surgery (with or without age correction). CWtv significantly increased after surgery for both concave and convex sides without and with age correction (p < 0.05). CWtv increased after surgery by 64.43% (without age correction) and 50.24% (with age correction) on the concave side of the spinal curve, and increased after surgery by 45.35% (without age correction) and 35.17% (with age correction) on the convex side of the spinal curve. Table 4b shows after age correction, the average ratio of tidal volumes Dtv/CWtv in MTC EOS patients after surgery became close to and not significantly different from that of normal children on the convex side of the spinal deformity (1.44±1.10 in MTC EOS patients vs. 1.39±0.67 in normal children (p = 0.78)) compared to that on the concave side of the spinal curve (1.06±0.52 in MTC EOS after surgery vs. 1.39±0.67 in normal children (p = 0.02)). The sub-analysis of CWtv for the 10 patients in the TLC subgroup revealed no significant changes after surgery. TLC and LC subgroups considered together showed no significant changes on CWtv after surgery; see Table S2 for more details.

Figure 1 depicts a male EOS patient with thoracic spinal dextroscoliosis on the AP radiograph, which also displays 3D MRI surface renditions of his lungs at EE and EI pre-operatively (top row) and post-operatively (bottom row). In this patient, CWtv and Dtv increased after surgery, with bilateral chest wall tidal volumes increasing from 56.71 cc to 127.39 cc (a 124.63% increase), and bilateral hemi-diaphragm tidal volumes combined increased from 91.22 cc to 109.94 cc (a 20.52% increase).

• Download figure
• Open in new tab

Figure 1.

Frontal chest radiographs, and 3D surface rendition results for lungs at EE and EI derived from dMRI for one male EOS patient with leftward concavity of thoracic spine before surgery (the first row) and after surgery (the second row).

Discussion

Pediatric spine deformity surgeons have a variety of operative techniques to control early onset scoliosis to prevent the development of thoracic insufficiency syndrome. Both Redding and Wang recently concluded that it is important to monitor the impact of different surgical strategies on pulmonary development including the use of various surgical devices and anchor points [32, 33]. However, little data exists, and hence information regarding the long-term changes in lung function during spinal surgical and non-surgical treatments for EOS must be expanded upon [33]. In the current study, we investigated how rib-based fixation affects chest wall motion and thus its contribution to respiratory tidal volume in pediatric patients with EOS by utilizing the recently developed QdMRI methodology [25,26]. We obtained measurements from both the concave and convex sides of the spinal deformity to assess the contribution of chest wall to respiration on its own, relative to the diaphragm, and in comparison, to normal children, before and after surgical intervention.

We observed that globally all tidal volumes of the chest walls and hemi-diaphragms significantly increased after surgery (Table 2). When comparing the post-operative volume contributions from the chest wall to age matched controls we do not see a statistically significant difference (p=0.51 and p=0.41, for left and right chest wall tidal volumes, respectively) (Table 3). This data indicated that the operative intervention utilizing rib-based anchors for EOS in this cohort can return the chest wall contributions of lung volume in the respiratory cycle to near normal, despite the anchor points being attached to the ribs. These results are important to the pediatric spine deformity surgeon as it provides support for proximal anchor attachments to the ribs given the known problematic migration of spine-based pedicle screw anchors in growing constructs [34, 35].

Granular analysis looking at concave and convex sides of the spinal deformity showed that the concave side of the deformity exhibited a greater increase in CWt, Dtv, & the ratio of Dtv/CWtv contributions post-operatively than pre-operatively (Table 4a & 4b) when compared to the convex side. The sub-cohort of MTC, which represents the majority of our cases and is the dominant phenotype by the Lenke classification for adolescent idiopathic scoliosis, exhibited greater increase for CWtv, Dtv, and Dtv/CWtv (Table 5).

Other methods exist for assessing chest wall motion in humans, but no study reached definitive conclusion regarding the effect of posterior-based distraction instrumentation on the respiratory cycle when anchored to the ribs [36–40]. Seddon et al suggested using physical devices placed on the chest wall in conjunction with imaging devices, such as MRI, to measure lung volume, and requiring breathing controls during image acquisition [41]. Such techniques would not be practical for use in EOS patients given their young age and often difficulty with following instructions. Additional studies in the literature compared scoliosis patients via 2D cross-sectional lung displacements or 3D volumes with normal subjects, and concluded that chest wall motion in scoliosis patients is worse than in normal subjects, but this is the first report documenting restoration of normal function post-operatively and regain of function compared to normative control in a patient population that is extraordinarily difficult to assess pulmonary function. Early reports highlighting this novel technology rib-based anchors & the respiratory cycle in children with thoracic insufficiency provides the pediatric spine community an opportunity to answer many pulmonary based questions in a heterogenous patient population that is historically difficult to study. The additional value of this technology lies in the glossary of normative subjects available for comparison, providing surgeons with a baseline to compare to when discussing operative intervention and outcomes with parents.

This study is not without limitation. One limitation of this study is the small number of EOS patients with major spinal curve types beyond MTC such as proximal thoracic, thoracolumbar, or lumbar curves, as not all EOS patients have the same curve type. Another limitation is the small numbers of male and female patients, which precluded gender specific comparisons with curve types, although our normal cohort had roughly equal and large number of male and female subjects. The greatest limitation is the fact that in EOS the most common implant used today is the MCGR, which would preclude a child from getting the QdMRI during treatment, and thus would only allow for post-fusion imaging [42, 43]. Additional limitations include that all data set stems from a single center. Lastly, while the results may demonstrate no detriment to chest wall mobility post-implant, concerns about the possibility of rib-based anchors driving chest wall stiffness cannot be completely dismissed until this cohort is followed to maturity or the time of definitive fusion as stiffness may appear with time. Despite these limitations, in this first study, our results suggest that chest wall motion is not impaired in pediatric patients with early onset scoliosis after rib-based fixation.

Conclusions

The novel technique of quantitative dynamic magnetic resonance imaging demonstrates that rib-based fixation does not appear to decrease the contribution of chest wall function to the respiratory cycle with chest wall tidal volumes significantly increasing post-operatively for both concave and convex sides of EOS patients with a main thoracic scoliosis and even approaching that of normative age matched controls.

Supplemental materials

Table S1: Global comparisons of tidal volumes (in cc) of MTC subgroup of EOS patients before and after surgery without distinguishing concave and convex sides of spinal curve. Each cell shows mean (upper number) and standard deviation (SD) (lower number) values. P values (pre-surgical vs. post-surgical) are shown using paired t-testing. Ratio r = ((post-pre)/pre) × 100%.

Table S2: Global comparisons of tidal volumes (in cc) of TLC+LC subgroups of EOS patients before and after surgery without distinguishing concave and convex sides of spinal curve. Each cell shows mean (upper number) and standard deviation (SD) (lower number) values. P values (pre-surgical vs. post-surgical) are shown using paired t-testing. Ratio r = ((post-pre)/pre) × 100%.