Response to Letter to the Editor Regarding the Article Entitled “Trabecular Bone Remodeling After Posterior Lumbar Interbody Fusion: Comparison of the Osseointegration in Three-Dimensional Porous Titanium Cages and Polyether-Ether-Ketone Cages” by Segi et al

To the Editor:

We appreciate the insightful comments made by the authors of the letter regarding our article entitled “Trabecular Bone Remodeling after Posterior Lumbar Interbody Fusion: Comparison of the Osseointegration in Three-Dimensional Porous Titanium Cages and Polyether-Ether-Ketone Cages.” ¹ We are grateful that our discussion was stimulating enough to prompt letters from readers and that the editors allowed us to present this meaningful discussion in print. Below, we discuss our opinions and limitations in response to the feedback we received, as well as our vision for the future.

Sample Size and Statistical Power

First, we acknowledge the importance of the discussion on sample size. Although posterior lumbar interbody fusion (PLIF) is a widely used procedure, its technical details vary slightly among institutions. For this reason, we preferred to analyze patients who underwent PLIF using a certain technique rather than recruit more patients from multiple institutions. One limitation of this policy was the low statistical power. Getting closer to the truth may be possible by estimating the required sample size after evaluating the sensitivity and specificity of postoperative computed tomography (CT) imaging.

Discrepancy in the Number of Participants

We thank the authors of the letter for identifying the error in the presentation of the number of patients. Initially, the reported number of excluded patients was higher than the actual, because those who met multiple exclusion criteria were counted in duplicate. After reviewing the dataset, we confirmed that the correct number of patients in the analysis was 101, after excluding from the 151 patients those who had specific medical reasons (n = 16), required reoperation because of postoperative infection (n = 1), and lacked postoperative CT (n = 33).

Evaluation of Osteoporosis

In addition, we acknowledge the lack of osteoporosis evaluation. Undoubtedly, osteoporosis is an important factor that affects implant-related complications after spinal fusion surgery, including PLIF.^2,3 Unfortunately, the inability to assess osteoporosis was an unavoidable limitation, because quantitative evaluation was lacking in many patients in this cohort. Furthermore, the importance of optimizing bone health to improve spinal fusion surgery outcomes has been gaining recognition. ⁴ Considering their availability for osteoporosis, various medications may have affected the findings in the fusion segment. Therefore, osteoporosis evaluation is an important point in this field. Notably, in terms of imaging findings, a previous study has suggested the absence of correlation between TBR occurrence and hip T-score. ⁵ Therefore, we deemed that an integrated view of osteoporosis, TBR, and mechanical complications remains unexplored.

Subjective Interpretation of Imaging Findings

We agree that interpretation of imaging findings may skew the results of this study. However, we recognize that in studies in which diagnostic imaging is the primary outcome, the impact of interobserver variability in the interpretation of results is inevitable. The TBR and other diagnostic imaging items used in this study were previously reported to have relatively small interobserver variability. ⁵ Therefore, in this study, imaging diagnoses were determined by a collegial system. Nevertheless, we would like to establish a scoring system that will enable stable interrater agreement. Previous scoring systems designed for X-ray findings, including but not limited to the Burkus classification ⁶ proposed by the authors of this letter, are surprisingly simple. Currently, there is a need for a scoring system designed for complex imaging technologies.

Long-Term Follow-Up

We completely agree that investigating the long-term outcomes of TBRs and segment stability is interesting. Indeed, we have already reported that TBRs may resolve after 5 years in segments that have established solid stability. ⁷ However, large-scale and long-term follow-up data would be difficult to obtain, because the cost and radiation exposure that repeated CT imaging entails may be unreasonable for patients who are doing well. We are currently exploring approaches to address this challenge.

Mechanism of VEC Formation in 3DTi Cages

Our opinion on bone response to the elastic modulus of the cage was entirely hypothetical. As the authors of the letter suggested, biomechanical studies and finite element analysis are important for understanding segmental stability after spinal fusion surgery. However, the intervertebral changes we tried to understand in-depth in our studies were not simple mechanical questions but involved biological responses. Living bone tissue must respond to both the implant and the continuously changing stress transfer. Therefore, far, we have only observed piecemeal results at specific time points; therefore, a mechanical approach may play an important role in filling the gap.

Patient-Reported Outcomes

Investigation of the relationship between patient-reported outcomes and imaging findings, including TBR, was essential and arguably the ultimate goal of the research. However, the reported outcomes of PLIF are clinically stable, ⁸ and worse clinical outcomes may not be the direct effects of unfavorable imaging findings, such as cage subsidence and screw loosening. ⁹ At present, we are focusing on collecting homogeneous specimens as much as possible through retrospective studies to gain an integrated understanding of imaging findings. We hope to link these results with actual patient perspectives in the future.

The imaging findings after intervertebral fusion are altered by unexplained laws. One may argue that a randomized controlled trial is necessary to accurately answer the question raised by the authors of this letter and to understand the general theory of this law. However, given the significant results in the previously reported relatively small retrospective studies, it can be said that from a clinical standpoint, the consideration has already entered the practical stage. We believe that another approach, such as basic research and machine learning for diagnosis, may be needed.

Again, we thank the authors of the letter for their interest and anticipation of TBR and their valuable suggestions. We hope that our answers will satisfy their questions and those of the other readers of this prestigious journal.