Vertebral bone quality score to predict cage subsidence following oblique lumbar interbody fusion

In the literature, many studies have reported favorable clinical outcomes of mini-open oblique lateral interbody fusion (OLIF) in patients with lumbar degenerative disk disease [1, 2]. Moreover, it is possible to perform OLIF through the physiological gap between the aorta and psoas major by making a small anterior–lateral skin incision [3]. This method avoids the invasion into the paraspinal musculature, is less invasive, and reduces the risk of injury to the lumbar plexus; in addition, it offers various advantages, such as minimal trauma and speedy recovery [4]. Indirect decompression of the neural elements is achieved using an enlarged cage to distract the narrow intervertebral space and the foramen [5]. Therefore, the degree of cage subsidence may have a greater impact on the clinical outcomes of OLIF than those of other posterior approach techniques where decompression is achieved directly.

Cage subsidence is one of the most commonly reported complications in OLIF, which leads to various compromised clinical outcomes, such as disk/foraminal height loss, column lordosis reduction, neuroforaminal stenosis, and recurrence of radiculopathy. In the relevant literature, low bone mineral density (BMD) has been considered as one of the main risk factors of cage subsidence, and it could be used as one of its predictors [6, 7]. The gold standard for BMD assessment and osteopenia or osteoporosis diagnosis is dual-energy X-ray absorptiometry (DEXA). However, previous studies have confirmed that DEXA cannot help in accurate BMD assessment of patients with lumbar degenerative diseases. Quantitative computed tomography (qCT) could be used as an alternative method for a more accurate assessment of bone quality. However, qCT is not routinely performed because it is expensive [8, 9]. Recently, a novel method of determining MRI-based vertebral bone quality (VBQ) score was proposed for evaluating the quality and fat infiltration of trabecular bone [10]. This score was found to significantly correlate with BMD assessed using the DEXA scan, which may help reduce investigation costs and radiation hazards [11, 12]. Moreover, the VBQ score has been recognized as an independent risk factor for fragility fractures, and it predicts the occurrence of cage subsidence and reoperation after transforaminal lumbar interbody fusion surgery [13, 14].

However, to the best of our knowledge, no studies have evaluated the relationship between the VBQ score and the occurrence of cage subsidence after OLIF surgery to date. Thus, this retrospective study aimed to assess whether the VBQ score can be used to predict the occurrence of cage subsidence after OLIF surgery.

In total, 102 patients were included in this study based on the inclusion and exclusion criteria. The demographic and clinical data of these patients are shown in Table 1. Overall, 39 (38.24%) patients with cage subsidence were included in the subsidence group, and their average subsidence was 3.59 ± 1.22 mm. The remaining 63 patients (61.76%) were included in the nonsubsidence group, and their average subsidence was 1.13 ± 1.56 mm. Comparison between the subsidence and nonsubsidence groups revealed that patients with subsidence were significantly older (62.10 ± 9.60 vs. 58.09 ± 8.82 years, p = 0.034) and had undergone more number of antiosteoporotic drug treatments [10 (25.64%) vs. 6 (9.52%), p = 0.030] than patients without subsidence. Additionally, patients with subsidence had a larger disk height change (3.51 ± 1.14 vs. 2.99 ± 1.05 mm, p = 0.022), more concave morphology of inferior endplate [1 (1.56%) vs. 14 (22.22%), p = 0.006] and superior endplate [2 (5.13%) vs. 16 (25.40%), p = 0.009] higher VBQ score (3.83 ± 0.82 vs. 2.98 ± 0.39 mm, p < 0.001), and a lower average lumbar DEXA T-score (− 1.76 ± 0.98 vs. − 1.02 ± 1.08 mm, p = 0.001) than patients without subsidence. In contrast, there were no significant differences in terms of other factors, including sex, BMI, age-adjusted CCI score, hypertension, history of smoking and drinking, chronic use of steroid, diagnosis, follow-up time, surgical level, cage height, cage length, cage position, preoperative and immediate postoperative disk height, and fusion status.

All statistically significant factors according to the univariable analysis were included in the multivariable logistic regression model. Detailed results of the effect size of each variable in the multivariable analysis are presented in Table 2. Notably, a higher VBQ score [odds ratio (OR): 23.158 ± 0.849, 95% confidence interval (CI) 4.381–122.399, p < 0.001] was the only variable that significantly predicted subsidence. We also conducted correlation analyses to evaluate the correlation among the VBQ score, average lumbar DEXA T-score, and amount of cage subsidence; the results of these analyses are shown in Table 3. A moderate correlation was found between the VBQ score and the amount of cage subsidence (r = 0.649, p < 0.001). Similarly, the VBQ score was moderately correlated with the average lumbar DEXA T-score (r = − 0.576, p < 0.001). However, a fair correlation was found between the average lumbar DEXA T-score and the amount of cage subsidence (r = − 0.418, p < 0.001).

Further, the ad-hoc analysis was performed to determine the cutoff values of the VBQ score and average lumbar DEXA T-score. Moreover, ROC curves were created for the preoperative VBQ score and average lumbar DEXA T-score as predictors for cage subsidence. Notably, the AUCs of the VBQ score and average lumbar DEXA T-score were 0.839 (95% CI 0.756–0.922) and 0.695 (95% CI 0.591–0.798), respectively. The cutoff value of the VBQ score was 3.435, with a sensitivity of 69.23% and a specificity of 88.89% (Fig. 2A). The cutoff value of the average lumbar DEXA T-score was − 0.75, with a sensitivity of 84.62% and a specificity of 44.44% (Fig. 2B).

In the present study, we retrospectively examined 102 patients who underwent OLIF surgery and identified 39 cases of cage subsidence. Some risk factors have been identified including old age, large number of antiosteoporotic drug treatments, increased disk height change, concave endplate morphology, low lumbar DEXA T-score, and high VBQ score. Notably, the novelty of this study is that we found that the VBQ score can independently predict postoperative cage subsidence in patients undergoing OLIF surgery. Cage subsidence is a clinical phenomenon that manifests as sinking of the cage into the vertebrae through an adjacent endplate. This phenomenon may lead to progressive disk height narrowing and poor clinical outcomes. Various subsidence rates have been reported in previous studies, and they range from 11 to 46.7% [3, 5, 17, 18]. In the present study, cage subsidence following OLIF was observed in 38.24% of the participants. This relatively higher subsidence rate in our study can be attributed to the use of a low threshold (≥ 2 mm of cage migration into adjacent endplates) to define subsidence. Moreover, the incidence of subsidence in the present study (38.24%) was much lower than that reported by Kotheeranurak et al. (46.7%), probably because we excluded patients with intraoperative endplate violations [18]. According to the reports in the relevant literature, the probability of occurrence of intraoperative endplate violations ranges from 16.8 to 33.1%, and these violations increase the risk of cage subsidence [19‐21]. Therefore, careful management of the surgical procedure is crucial to ensure that the endplate is completely undamaged.

Apart from the endplate violation, the risk factors that were thought to be associated with cage subsidence have been widely discussed in previous studies, including poor bone quality, over distraction, cage position and design, endplate sclerosis, age, and the presence or absence of the supplemental fixation [20, 22, 23]. In the present study, based on univariable analysis, patients with subsidence had older age, underwent greater number of antiosteoporotic drug treatments, had greater disk height change, and had a more concave endplate morphology compared to those without subsidence. Notably, bone quality and muscle mass tend to reduce with aging. According to a previous study, older patients are more likely to develop cage subsidence owing to the negative effects of age on bone microstructure [24]. Moreover, the use of antiosteoporotic drugs was recommended postoperatively for patients undergoing lumbar fusion surgery in order to reduce the subsidence rate [25]. However, in our study, the number of patients treated with antiosteoporosis therapy was more in the subsidence group, indicating that such patients have worse bone quality and are more prone to cage subsidence. Over the distraction is another important factor contributing to cage subsidence. Notably, it has been reported that disk space distraction may cause cage subsidence in anterior cervical and lumbar fusion cases [26, 27]. A large distraction leads to a large change in disk height, which further causes a significant increase in the compressive force between the cage–endplate interface. The selection of cage height has been suggested to be determined according to the disk height measured preoperatively. Furthermore, previous reports have proven that cage morphology affects cage subsidence [28, 29]. In particular, it has been believed that a flat endplate has better interface contact with an OLIF cage at the surface. A well-matched endplate-cage surface provides more even stress distribution and a larger area for endplate coverage, decreasing the incidence of cage subsidence. In contrast, a concave endplate provides a reduced contact area and leads to stress concentration, increasing the incidence of cage subsidence [30].

Low BMD has been proven to be a critical risk factor for cage subsidence. Proper preoperative assessment of bone quality helps in developing a reasonable surgical plan and taking relevant preventive measures to improve surgical outcomes and reduce associated complications. Although DEXA is recognized as the gold standard for bone density assessment, it is associated with many issues that can lead to inaccurate evaluations, such as spinal deformities, overlying soft tissue, previous compression fractures, bowel content, osteophytes, and aortic atherosclerosis [31‐33]. In the present study, the average lumbar DEXA T-score was fairly correlated with the amount of cage subsidence; this finding was similar to the findings of previous studies, which reported that BMD assessed by DEXA scanning had a significantly weak correlation with the amount of cage subsidence. Therefore, this score is not a good predictor for cage subsidence in OLIF surgery and demonstrates low accuracy (0.695). Previous studies have reported that Hounsfield units calculated by qCT may help assess bone quality more accurately than the DEXA T-score [9, 34]. Notably, low Hounsfield units indicate low BMD, which further indicates a higher incidence of developing cage subsidence [7, 35]. However, qCT is more expensive and associated with a high risk of radiological hazards; hence, it is not suitable for use in routine examinations.

Previous studies have examined the MRI-based VBQ score to assess bone quality [10, 36]. In particular, the use of this score is inexpensive and completely radiation-free. Osteoporotic bone is usually characterized by trabecular atrophy and local adipocyte replacement on histological analysis [37]. The use of the VBQ score is based on the theory that more atrophy and fatty infiltration in osteoporotic trabecular bone causes higher SI on T1-weighted images. Owing to the easy-to-teach methodology and rapid results unaffected by confounding factors, the measurement of this score has shown good reproducibility and reliability, demonstrating that this is a practical tool for clinical use [14, 38]. In previous studies, the use of the VBQ score was verified against the use of DEXA to identify osteoporotic bone and to independently predict the risk of fragility fractures, cage subsidence, and vertebral compression fracture in patients with spine metastases [13, 39, 40]. In the present study, the VBQ score was moderately correlated with the average lumbar DEXA T-score and amount of cage subsidence. This score is known to be an accurate indicator of bone quality; in addition, it is known to facilitate the identification of osteoporotic bone. More importantly, the VBQ score was found to be a significant independent predictor of cage subsidence after OLIF surgery according to the multivariable logistic regression analysis in our study. Patients with VBQ scores higher than the threshold are approximately 23 times more likely to experience subsidence than those with lower scores. Moreover, the VBQ score predicts cage subsidence better than the average lumbar DEXA T-score, with an accuracy of 0.839. This difference may be attributed to the fact that the VBQ score is a site-specific bone density measure that directly reflects vertebral bone quality at the weight-bearing site while excluding factors affecting whole-region standard DEXA measurement, making its assessment more precise than that of the DEXA T-score. Notably, endplate bone quality—a novel site-specific MRI-based bone quality assessment proposed by Jone et al. [41] could predict severe cage subsidence after standalone lateral lumbar interbody fusion even better than the VBQ score. These findings suggest that the VBQ score should be adopted in clinical practice for patients with OLIF surgery in order to improve the efficacy and safety of treatment.

This study has several limitations. First, the retrospective design was subject to certain biases, such as the selection bias in the data collection and the potential observer-expectancy bias. Second, the relatively small cohort from a single health center may explain the lack of statistical significance of some previously identified risk factors for cage subsidence in our multivariable logistic regression analysis. Finally, some important laboratory examination parameters were not assessed in the present study, including calcium, phosphorus, lipids, and cholesterol levels. A previous study reported that patient physiology affects the VBQ score and found a higher VBQ score in patients with hyperlipidemia and healthy bones [42].

The VBQ scoring method can be used to evaluate bone quality with good effectiveness and accuracy and to independently predict postoperative cage subsidence in patients undergoing OLIF surgery. Preoperative measurement of the VBQ score should be considered to identify high-risk patients and take timely precautions to minimize complications.