Cervical spinal pathologies, encompassing degenerative conditions, trauma, infections, and tumors, are significant causes of myelopathy and radiculopathy, often necessitating surgical intervention to alleviate neural compression and stabilize the spinal column.1–3 Among the various surgical techniques, anterior cervical corpectomy has become a widely utilized and effective approach, particularly when the pathology involves the vertebral body, allowing for extensive decompression of the spinal cord and nerve roots.

Corpectomy is a surgical technique that provides extensive decompression and fusion at the cervical level and is also used for certain thoracic pathologies. It is considered more complex compared to traditional anterior cervical discectomy and fusion,1 as is thoracic corpectomy, which can be performed through various approaches such as anterior (supra-sternal or trans-sternal), dorsolateral, or costotransverse.2 The surgeon's experience plays a crucial role in each case. Thoracic pathologies are often traumatic or neoplastic, whereas cervical pathologies are primarily degenerative.3 Corpectomy is indicated for a wide range of conditions including degenerative disc disease, traumatic injuries, tumors,4 infectious lesions, and posterior longitudinal ligament calcification.5,6 It offers diverse radiological and clinical outcomes,7,8 and in some series, is comparable to anterior cervical discectomy and fusion in terms of effectiveness.9,10

The procedure is associated with complications such as dysphagia (5.3%), esophageal perforation (0.2%), recurrent laryngeal nerve paralysis (1.3%), infection (1.2%), adjacent segment disease (8.1%), pseudoarthrosis (2.0%), graft or hardware failure (2.1%), cerebrospinal fluid leakage (0.5%), hematoma (1.0%), and Horner's syndrome (0.4%).11,12 The global hospital mortality rate is 1.6%, with a complication rate of 18.4% and a median hospital stay of 6 days. Having three or more spinal levels treated and a surgery duration exceeding 6h are significant predictors of postoperative complications.13 Controversies exist regarding surgical techniques due to these complications, though some series suggest that the complications are similar to other anterior approaches.14 Advantages of corpectomy in multi-level disease, particularly cervical, include shorter surgical times but increased intraoperative blood loss, which is proportional to the number of levels treated. Comparing corpectomy to anterior cervical discectomy and fusion, there are almost no differences in multi-level disease, making it a safe procedure.15

This procedure advantages include direct access to the pathology and the ability to achieve broad decompression and subsequent spinal fusion, leading to improved neurological outcomes. Corpectomy is the treatment of choice for certain cervical and thoracic conditions, offering excellent decompression, though its stability for multi-level disease still warrants debate. Each case should be individually assessed to determine the best surgical approach and material.16

Historically, autologous bone grafts, typically harvested from the iliac crest or fibula, were the gold standard for reconstructing the spinal column after corpectomy due to their osteoconductive, osteoinductive, and osteogenic properties.16 However, their use is associated with considerable drawbacks, including donor site morbidity (e.g., pain, hematoma, infection, fracture), limited graft availability, and potential for graft collapse or resorption. These limitations have driven the development and adoption of alternative reconstructive materials, such as allografts and, more recently, various types of metallic cages.16–18

Over the years, new prosthetic materials such as titanium cages have been implemented, showing safety and good results in various studies.17–20 While bone grafts are still used, innovative techniques are being explored.21 Expandable titanium cages have emerged as a promising alternative, offering several theoretical advantages. They can be precisely expanded in situ to achieve optimal distraction and restoration of spinal alignment, potentially reducing the risk of subsidence and providing immediate biomechanical stability. Moreover, their use eliminates the need for a second surgical site, thereby mitigating donor-site complications. While studies have demonstrated the safety and efficacy of these prosthetic implants, and their non-inferiority to traditional methods in many aspects, a comprehensive understanding of their comparative neurological outcomes against autologous bone grafts remains an area of ongoing investigation.

The aim of this study was to retrospectively compare the neurological outcomes, clinical improvement, complications, and technical parameters between autologous bone grafts and an advanced expandable titanium cage system for vertebral body reconstruction following cervical corpectomy.

A total of 48 patients were obtained. Following a thorough screening process, 40 patients were enrolled in the study, with 20 undergoing corpectomy with artificial implants (ADD plus™/Te-Corp™) and 20 receiving reconstructions with autologous bone grafts (fibula or iliac crest), Fig. 1. The two groups were compared across multiple demographic and clinical variables (Table 1 and Fig. 2).

Fig. 1.

Patient distribution by surgical technique. Flowchart showing the selection of 48 patients initially considered for inclusion. After applying exclusion criteria, 40 patients were retained for final analysis: 20 underwent reconstruction with autologous bone grafts (fibula or iliac crest) and 20 received expandable titanium cages (ADD plus™/Te-Corp™).

Fig. 2.

Comparative summary of surgical and clinical parameters. Bar chart illustrating key variables between the two groups, including ΔmJOA score, surgical time, intraoperative blood loss, hospital stay, and levels treated. Note the significantly greater neurological improvement in the ADD plus™/Te-Corp™ group.

The mean age of patients in the ADD plus™/Te-Corp™ group was 58.5±11.3 years, while the autologous graft group had a mean age of 65.1±9.8 years (p=0.055). Although the difference was not statistically significant, a trend toward older age was observed in the autologous graft group. The sex distribution showed a higher male-to-female ratio in the ADD plus™/Te-Corp™ group (2.33:1) compared to the autologous graft group (1.22:1) (Table 2).

Neurological outcomes

Patients treated with expandible titanium cages had significantly lower preoperative mJOA scores (8.95±3.03) compared to those in the autologous graft group (11.25±3.5; p=0.0323), suggesting greater initial neurological impairment. However, the postoperative mJOA scores were similar between groups (13.75±2.15 vs. 12.9±3.78; p=0.3875). The mean improvement in mJOA was significantly greater in the ADD plus™/Te-Corp™ group (4.8±1.85 vs. 2.2±0.87; p=0.0001), indicating superior neurological recovery (Fig. 3).

Fig. 3.

Neurological outcomes pre- and postoperatively. Bar chart comparing preoperative and postoperative modified Japanese Orthopaedic Association (mJOA) scores in both groups. The expandible titanium cages group showed a significantly greater improvement, despite lower baseline scores.

Baseline neurological function, measured with the mJOA scale, differed significantly between treatment groups, introducing a potential source of bias. To account for this, we performed adjusted analyses.

An ANCOVA model, controlling for preoperative mJOA and age, showed that patients treated with titanium cages had significantly higher postoperative mJOA scores compared to those treated with autologous grafts (adjusted coefficient=+2.50, p=0.0007).

To further reduce baseline imbalances, we applied inverse probability of treatment weighting (IPTW) using propensity scores estimated from age and preoperative mJOA, Fig. 4. After weighting, covariate balance improved substantially, with standardized mean differences falling below conventional thresholds. The IPTW-weighted regression indicated a trend toward better neurological outcomes in the ETC group (coefficient=+1.75, p=0.092), although this result did not reach conventional statistical significance.

Fig. 4.

(A) Standardized mean difference (SMD) before and after applying inverse probability of treatment weighting (IPTW). The SMD measures the balance of patient characteristics between the two treatment groups. An SMD value less than 0.1, indicated by the dashed line, suggests good balance. The black bars show the unweighted data, which had a significant imbalance in the baseline mJOA pre score. The light gray bars show the IPTW-weighted data, which are much closer to zero, indicating that the groups are now more balanced. (B) This histogram displays the distribution of propensity scores (PS) for the two treatment groups: expandible titanium cages (ETC) and autologous graft (AG). The graphs’ purpose is to check for adequate overlap between the two distributions, which is a key assumption for IPTW analysis. The graph shows sufficient overlap between the two groups. (C) The boxplot shows the distribution of Hirabayashi recovery rates. ETC demonstrate higher and more consistent recovery rates compared with autologous grafts. (D) Bar chart of mean Hirabayashi recovery rate with standard deviations. ETC group shows higher mean recovery (65.0%) compared to autologous graft (45.1%).

Regarding clinically meaningful improvement, the responder rate (≥2-point increase in mJOA) was 100% in the ETC group versus 75% in the autologous group (p=0.064).

Taken together, these results suggest that, after adjustment for baseline neurological status, titanium cages are associated with greater neurological recovery compared to autologous grafts. The IPTW analysis supports the robustness of this association, although weighted models indicate that the effect size may be more modest than suggested by the unadjusted comparison.

Also, the mean Hirabayashi recovery rate was 65.0±27.9% in the titanium group and 45.1±65.2% in the autologous graft group. Median recovery rates were 65.2% and 36.4%, respectively. The distribution of recovery rates is illustrated in Fig. 4.

This study provides a comparative evaluation of expandible titanium cages (ADD plus™/Te-Corp™) versus autologous grafts (fibula or iliac crest) in anterior spinal reconstruction.

The use of expandable titanium cages in cervical corpectomy is well-supported in the literature due to their advantages over autogenous bone grafts, such as eliminating donor site morbidity, providing immediate structural support, and facilitating precise height restoration and maintenance of sagittal alignment.

Studies by Zaïri et al.27 and Awad et al.28 generally affirm the favorable clinical and radiological outcomes associated with expandable cages for cervical spondylotic myelopathy. Prakash et al.,29 focusing on winged expandable titanium cages, also reported considerable clinical improvement, aligning with the overall positive perception of these implants. Despite the relatively small sample size, our findings contribute to the growing body of evidence suggesting that expandable titanium cages are a viable and potentially superior alternative to traditional grafting techniques, particularly in terms of early neurological recovery.

Regarding patient demographics, the mean age of our study population was 61.75±10.95 years (ranging from 28 to 82 years). This age profile is consistent with other studies in the field, such as Das et al.30 (59.5–60.9 years), Pojskic et al.31 (61.3 years), and Zaïri et al.27 (60 years), indicating that our cohort represents an older patient population often affected by complex cervical pathologies. However, when compared to other cohorts like Patil et al.32 (45.2±16.6 years), Burkett et al.33 (48 years), and Prakash et al.29 (39.74 years), the mean age in our study was considerably higher. This difference, potentially ranging from 13 to 22 years, highlights variations in patient populations across different institutions and regions, which might influence surgical indications, outcomes, and complication profiles. Consistent with the broader literature, there was a generalized male predominance observed in our study population.

However, our findings present a notable point of divergence when compared to some existing literature. Specifically, Patil et al.,32 in a retrospective descriptive study comparing expandable cylindrical cages to iliac crest autografts a comparison directly analogous to ours found no significant differences in clinical improvement (including neurological presentation) between the two groups. This stands in direct contrast to our observed statistically significant superior mJOA improvement in the expandable cage group. Possible reasons for this discrepancy could include differences in patient populations, specific expandable cage designs (our study utilized the ADD plus™/Te-Corp™ system), surgical techniques, variations in postoperative rehabilitation protocols, or differences in the assessment methodology of clinical improvement.

Patients treated with ETC implants exhibited significantly greater improvement in mJOA scores at three months postoperatively (ΔmJOA=4.8) compared to those who underwent reconstruction with autologous bone grafts (ΔmJOA=2.2; p=0.0001). This superior neurological recovery occurred despite a worse baseline neurological status in the ETC group, suggesting a more effective decompression and stabilization of the spinal cord when using expandable cage systems. These results align with those of Awad et al.,26 who reported an average mJOA increase of 3.3 points and a 97% fusion rate in a review of 143 cases of cervical spondylotic myelopathy treated with expandable cages.

Keshavarzi et al.34 also reported excellent outcomes with expandable cages in thoracolumbar reconstructions performed for neoplastic, infectious, or traumatic etiologies, with no postoperative neurological deterioration and high rates of sagittal balance restoration. While their study focused on a broader anatomical and pathological spectrum, their conclusions support the biomechanical and clinical advantages of expandable cage systems in anterior column reconstruction.

Furthermore, a systematic review and meta-analysis by Das et al.30, which compared expandable and non-expandable cages (rather than autografts), also concluded that there was no significant difference in JOA score improvement between these two types of cages. While this comparison is distinct from ours (cage vs. cage versus cage vs. autograft), it introduces nuance by suggesting that even within the realm of prosthetic implants, a clear superiority in neurological outcome is not universally established across all designs. Our study focusses on the comparison with autografts, where a significant difference was found, highlights the continued relevance of evaluating these older versus newer reconstructive methods. The larger effect size observed in our study (4.8 vs. 2.2 delta mJOA) might reflect a more pronounced benefit in our specific cohort or with the particular implant system used.

Surgical parameters such as operative time and hospital stay did not differ significantly between groups in our study, consistent with prior literature. However, a trend toward greater intraoperative blood loss was observed in the ADD plus™/Te-Corp™ group (520±263.78mL vs. 372.5±227.75mL), likely attributable to the higher number of levels treated in this group (2.2 vs. 1.75). Similar findings have been reported by Keshavarzi et al.,34 who noted increased blood loss in multilevel reconstructions but found the overall complication rates to be low and manageable.

Our study also reaffirms that expandable cages facilitate intraoperative adaptability and ease of placement. Unlike mesh cages or structural grafts, which require precise sizing and forceful impaction, expandable devices allow for in situ adjustment to the defect, minimizing the risk of endplate violation or over distraction a concern echoed in the literature by Awad et al.28 and other biomechanical studies. Moreover, the broad surface contact of the expandable cage enhances endplate load distribution, potentially improving fusion conditions despite a smaller central graft cavity.

With respect to complications, no statistically significant differences were observed between groups. The ETC group had a slightly lower complication rate (2 vs. 4 patients), although the difference was not statistically significant. This aligns with prior data showing complication rates below 10% for expandable cages used in cervical and thoracolumbar regions. Importantly, none of our patients experienced cage migration, neurological deterioration, or need for reoperation within the follow-up period.

It is noteworthy that the preoperative mJOA scores were significantly different between the groups (p=0.0323), with the autologous graft group having a higher mean mJOA score (indicative of better baseline neurological function). Despite this, the expandable cage group demonstrated a more substantial change in mJOA, further emphasizing its efficacy in promoting neurological recovery, even from a potentially more severe baseline, or demonstrating its ability to achieve greater functional gains. Furthermore, a trend towards older age was observed in the autologous graft group, although this difference did not reach statistical significance (p=0.055).

Besides, the Hirabayashi recovery rate,24,25 a standardized metric derived from the Japanese Orthopedic Association (JOA) score,22 is an essential tool for objectively evaluating postoperative neurological improvement in patients with myelopathy. In our pilot study, although the use of expandable titanium cages showed a superior recovery rate (65.0%) compared to autologous bone grafts (45.1%), our findings suggest that the degree of neurological recovery is more strongly influenced by the preoperative severity of the myelopathy than by the specific surgical system used.

This result is consistent with existing literature, which has shown that decompressive interventions, such as laminoplasty and corpectomy, can achieve significant regression of neurological symptoms and, in some cases, complete recovery, as noted by Dmitrievich et al.35

Varying recovery rates have been reported in different studies, highlighting the diversity of outcomes depending on the pathology and technique. For example, Tumturk et al.36 reported a recovery rate of 83.3% after hemilaminectomy for thoracic spinal stenosis, while a study on cervical corpectomy for ossification of the posterior longitudinal ligament (OPLL) recorded a rate of 67.01%. Similarly, Lu et al.37 found no statistical difference in the 58.1% recovery rate in patients over 70 years old compared to a younger control group. Additionally, Du et al.38 concluded that both anterior cervical discectomy and fusion (ACDF) and anterior cervical corpectomy and fusion (ACCF) can result in satisfactory neurological outcomes with no significant difference in the Hirabayashi recovery rate between the two techniques.

The results of our study, although preliminary, reinforce the idea that effective neural decompression is the primary factor in neurological improvement, rather than the choice of reconstruction technique. The consistency with the literature suggests that both expandable titanium cages and autologous grafts are viable options and that the baseline severity of the myelopathy is a key predictor of recovery potential.

These findings support the use of artificial implants as a viable alternative to autologous grafts, especially in patients with multiple-level pathology or when donor site morbidity is a concern.

The findings of this pilot study suggest that both autologous bone grafts and expandable titanium cages are safe and effective surgical options for spinal decompression and stabilization via corpectomy. A key finding from our analysis, adjusted for baseline differences in mJOA scores, is that we found no statistically significant difference in final neurological outcomes between the two systems.

This result is clinically relevant as it indicates that postoperative neurological recovery is primarily driven by the severity of the patient myelopathy at the time of surgery, rather than by the type of implant material used. The IPTS analysis confirmed this relationship, demonstrating that patients with more advanced myelopathy (lower mJOA scores) have a greater potential for neurological improvement.

While this study is not large enough to establish the superiority of one system over another, it does validate both as viable options in clinical practice.

The results provide crucial preliminary evidence that can guide decision-making, suggesting that the choice between an autologous graft and an expandable cage could be based on other factors, such as cost, implant availability, surgical time (as discussed in the results), or surgeon preference, without compromising the patient's short-term neurological outcome.

Despite the limited follow-up, this study lays the groundwork for future research, underscoring the need for larger, multi-center studies to not only validate these findings but also to examine long-term outcomes, including fusion rates and the incidence of late complications.

Conceptualization: AMPG, JCLV, UGG. Formal analysis: AMPG, JCLV. Investigation: AMPG, DAVM, JCLV. Methodology: AMPG, JCLV, UGG. Project administration: OMC, AIDT, UGG. Writing – original draft: AMPG, DAVM, JCLV. Writing – review & editing: AMPG, JCLV, UGG.