The relationship between postoperative proximal junctional kyphosis and MRI-based pedicle bone quality scores in thoracic adolescent idiopathic scoliosis

Abstract

Objective: Proximal junctional kyphosis (PJK) is a major postoperative complication in patients with thoracic adolescent idiopathic scoliosis (AIS), and poorer bone quality condition is an important risk factor for the occurrence of PJK, meanwhile osteopenia in patients with AIS can persist into adulthood. The use of computed tomography Hounsfield units (CT-HU) and Quantitative computed tomography (QCT) to assess the quality of the pedicle bone condition can predict the occurrence of postoperative Proximal junctional problems (PJP). However, there is a lack of an MRI-based site-specific bone quality scoring system for assessing the condition of pedicle bone quality and its relationship with PJK. Therefore, the purpose of this study was to create an MRI-based site-specific bone quality scoring system for the pedicle and to evaluate its correlation with pedicle CT-HU values and its ability to predict thoracic AIS PJK, comparing the predictive ability of the Pedicle Bone Quality (PBQ) Score and the Vertebral Bone Quality (VBQ) Score.

Methods: This is a retrospective study to assess the relationship between postoperative PJK and the PBQ and VBQ scores in patients with Lenke1 and 2 AIS. Demographics, procedure-related data, and radiological data were collected, PBQ and VBQ values were measured on MRI T1-weighted unenhanced sagittal images, respectively. Simple linear regression and Pearson correlation analyses were used to determine the relationship of PBQ and VBQ score with pedicle HU values and Proximal junctional angle (PJA). Multivariate logistic regression analyses were used to identify risk factors associated with PJK, and in addition, univariate linear regression analyses were used to identify potential factors associated with PBQ score. Finally, Receiver operating characteristic (ROC) analysis and area-under-the-curve values were used to assess the predictive ability of PBQ and VBQ scores.

Results: A total of 181 patients (mean age 15.44 ± 3.22 years) were included in this study and the PJK occurred in 30 (16.6%) patients. and patients in the PJK group had significantly higher PBQ and VBQ values (3.25 ± 0.34; 3.69 ± 0.27) than those in the non-PJK group (2.65 ± 0.51; 2.86 ± 0.53; P < 0.001). Simple linear regression analyses showed a significant negative correlation between PBQ and VBQ with the pedicle HU values (r=-0.639, P < 0.001; r=-0.592, P < 0.001), and a significant positive correlation was found between PBQ (r = 0.642; P < 0.001), VBQ (r = 0.608; P < 0.001) with PJA. The results of multivariate logistic regression showed that the PBQ score was the risk factor for PJK (OR = 3.763, 95% CI = 4.324-12.844, P < 0.001), with 86.3% predictive accuracy, higher than VBQ (79.2%).

Conclusions: The novel PBQ score is a reliable tool for predicting postoperative PJK in patients with thoracic AIS. Compared to the VBQ score, it has a better correlation with pedicle CT-HU values and PJA, as well as better predictive ability for PJK.

Keywords: Adolescent idiopathic scoliosis; MRI; Pedicle bone quality score; Proximal junctional kyphosis; Site-specific bone quality score; Vertebral bone quality score.

Fig. 1

A: CT-HU values of the pedicles of UIV were measured on CT axial images; B: CT-HU values of the pedicles of UIV were measured on CT median sagittal images. CT-HU, Computed tomography-Hounsfield unit; UIV, upper instrumented vertebra

Fig. 2

A and B: Full-length frontal and lateral radiographs of the spine in a patient with typical Lenke 1AN AIS. The smaller unstructured lumbar curved makes it easy to measure the PBQ; C: PBQ measurement: The ROIs (rectangles) were placed in the area of cancellous bone of the pedicles of L1-L4; D: VBQ measurement: Measuring the signal intensity of the cancellous bone ROIs (circulars) in the vertebrae of L1-L4; and to measure signal intensity in the region of cerebrospinal fluid posterior to L3. AIS, Adolescent idiopathic scoliosis; PBQ, Pedicle bone quality; ROI, Range of interest; VBQ, Vertebral bone quality

Fig. 3

The linear regression results show the relationship between PBQ, VBQ and PJA. PBQ, Pedicle bone quality; VBQ, Vertebral bone quality; PJA, proximal junctional angle

Fig. 4

The linear regression pictures demonstrate the relationship between PBQ, VBQ and the pedicle CT-HU values. PBQ, Pedicle bone quality; VBQ, Vertebral bone quality; CT-HU, Computed tomography-Hounsfield unit

Fig. 5

ROC curve analysis for PBQ and VBQ prediction of PJK. PBQ, Pedicle bone quality; VBQ, Vertebral bone quality; PJK, proximal junctional kyphosis