Previously many studies have been conducted on the compensation of lumbar curvature after the selective thoracic fusion. Koller et al. (K. H et al., 2019) proposed an accurate prediction model for postoperative SLCC based on the analysis of many prospective STF cases. Danilo et al. (D & AO, 2016) conducted a retrospective cohort study of 42 Lenke 1 AIS patients and concluded that the main thoracic curve's overcorrection might result in less satisfactory results. Matthew et al. (MB et al., 2004) indicated that the preoperative push-prone is the best preoperative flexibility radiograph to predict the final lumbar curve measurement. Pasha et al. (P. S & JM, 2019) then developed a decision tree to define criteria for optimal lumbar curve correction following STF in Lenke 1 AIS. Schulz et al. (Schulz et al., 2014) pointed out that optimal postoperative outcomes for STF should include a lumbar Cobb angle less than 26 °, coronal balance 2 cm or less, deformity-flexibility quotient less than 4, lumbar correction more than 37%, and trunk shift less than 1.5 cm. However, these studies only focus on overall compensation behavior and aim to improve clinical strategies.
In this study, we calculated each disc wedge variation of distal unfused lumbar segments to further elucidate the characteristics of spontaneous compensation of the lumbar curve after STF. The results showed that the proximal two segments at level L1/2 and L2/3 accounted for most total compensation. The distal unfused lumbar segments provided the more distal the segment, the less compensation. Furthermore, we found that total disc compensation consisted of less than half of the total postoperative lumbar curve compensation. This phenomenon may indicate that the lumbar curvature is often affected and includes the thoracic vertebrae, such as T10, T11 and T12. However, since all of our cases chose L1 as LIV, our study did not further investigate the fused thoracic discs, we focused attention on the unfused lumbar segments. As shown in Fig. 2, the compensation ability of the lumbar segments showed a decreasing tendency, with a major role being played by the proximal adjacent lumbar curve. Moreover, our integral analysis indicated that the upper coronal lumbar curve was responsible for most of the compensation, which was consistent with the opinion of Na et al. (Na, Ha, Harms, & Choi, 2010). They were the first to divide the lumbar curve into the proximal and distal curves by their respective lumbar apex and concluded that looking at the proximal lumbar curve flexibility might be an alternative indicator for measuring the lumbar flexibility in MT-AIS patients treated by STF. We believe that the characteristics of residual lumbar curve after STF may be closely associated with the adding-on phenomenon and may provide evidence when choosing the correct LIV.
Then, what is the reason for the non-uniformity of unfused distal segment compensation? We believed that the flexibility of the distal unfused segments might be different. Zhao et al. (J et al., 2018) analyzed the characteristics of cobb angle distribution in the Lenke 5C AIS patients. They found that the disc angles had symmetric distribution in the main thoracolumbar/lumbar curve, while the distal segment is more flexible. The thoracolumbar/lumbar curve's apex was often L1 or L2 vertebrae, and its distal segments may correspond to the L1/2 and L2/3 segments that were consistent with our study results. Na et al. (Na et al., 2010) also found that the lumbar apex of 28 main thoracic curve patients was between L2 and L3, and concluded that the curve flexibility of the proximal lumbar area could be meaningful. Jansen et al. (RC, LW, E, & A, 2007) also concluded that in STF patients, the most correction was made in the upper part of the lumbar curve, while the distal lumbar curve seemed to be more rigid and less important in spontaneous curve correction. In addition, another reasonable hypothesis may be the mechanical effect of posterior fusion with the pedicle screw. This phenomenon was similar to complications on the sagittal plane, such as PJK(HJ et al., 2020; O. S et al., 2019) and DJK(P, M, M, F, & C, 2020), which we believed could result from stress concentration on the adjacent segments. Meric et al. (Enercan et al., 2015) have shown that facet joint degeneration is significant at the upper two levels adjacent to the LIV when performing STF. Furthermore, this may be explained by the principle of load-sharing, that when arthrodesis was applied, the posterior fixation conducted most of the forces to the lowest instrumented vertebrae. Furthermore, when this force is overloaded, the stress could be conducted to the most adjacent segments. This has related to coronal complications, which could lead to complications, including adding-on phenomenon and coronal imbalance.
Therefore, when planning surgical treatment strategies, the characteristics of unfused lumbar segments should be carefully considered. Inappropriate curve selection and excessive thoracic correction have been identified as the most common etiologies of coronal imbalance(J. H et al., 2018; W et al., 2018; Y et al., 2019). Meanwhile, numerous studies(MA & DM, 2011; Skaggs et al., 2016; W et al., 2018; X et al., 2016; Y et al., 2019) have demonstrated improper placement of the LIV is also an independent risk factor. It is also important to realize the heterogeneity of spontaneous compensation of unfused lumbar segments. When selectively fusing the thoracic curvature, the overall compensation ability of the lumbar curvature and the heterogeneity of compensation to avoid excessive compensation at the proximal end should be considered.
Even though our study focused on the residual lumbar curve segmental characteristics in Lenke 1 and 2 AIS patients who were performed STF, several limitations should be considered. First, we only included patients whose LIV was L1 vertebrae for the homogeneity analysis of disc compensation. Further researches on other LIV selection and comparison should be performed. Second, only coronal position data were studied in our research but not a sagittal plane, and there was no specific analysis of related complications. Finally, this was a single-center study, and multi-centric research should be conducted to further validate the results.