- Email: [email protected]
Influence of Magnetic Resonance Imaging Features on Surgical Decision-Making in Degenerative Cervical Myelopathy: Results from a Global Survey of AOSpine International Members
Accepted Manuscript The Influence of MRI Features on Surgical-Decision Making in Degenerative Cervical Myelopathy: Results From a Global Survey of AOSpine International Members Aria Nouri, MD MSc, Allan R. Martin, MD, Anick Nater, MD, Christopher D. Witiw, MD MS, So Kato, MD, Lindsay Tetreault, PhD, Hamed Reihani-Kermani, MD, Carlo Santaguida, MD, Michael G. Fehlings, MD PhD PII:
S1878-8750(17)30917-8
DOI:
10.1016/j.wneu.2017.06.025
Reference:
WNEU 5896
To appear in:
World Neurosurgery
Received Date: 4 May 2017 Revised Date:
2 June 2017
Accepted Date: 5 June 2017
Please cite this article as: Nouri A, Martin AR, Nater A, Witiw CD, Kato S, Tetreault L, ReihaniKermani H, Santaguida C, Fehlings MG, The Influence of MRI Features on Surgical-Decision Making in Degenerative Cervical Myelopathy: Results From a Global Survey of AOSpine International Members, World Neurosurgery (2017), doi: 10.1016/j.wneu.2017.06.025. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
SC
RI PT
The Influence of MRI Features on SurgicalDecision Making in Degenerative Cervical Myelopathy: Results From a Global Survey of AOSpine International Members
M AN U
Aria Nouri MD MSc1, Allan R. Martin MD1, Anick Nater MD1, Christopher D. Witiw MD MS1, So Kato MD1, Lindsay Tetreault PhD1, Hamed Reihani-Kermani MD2, Carlo Santaguida MD3, Michael G. Fehlings MD PhD1,4
TE D
Affiliations 1Division of Neurosurgery, Toronto Western Hospital, University Health Network, Toronto, Ontario, Canada 2Department of Neurosurgery, Kerman University of Medical Sciences, Kerman, Iran 3McGill University Health Centre, Montreal, Quebec, Canada 4Department of Surgery, University of Toronto, Toronto, Ontario, Canada
EP
Corresponding Author:
AC C
Michael G. Fehlings MD PhD FRCSC FACS Professor of Neurosurgery Halbert Chair in Neural Repair and Regeneration University of Toronto 399 Bathurst St. Toronto Western Hospital, Toronto, Ontario Canada Email: [email protected] Phone: 416-603-5229 Fax: (416) 603-5627
ACCEPTED MANUSCRIPT
RI PT
Survey on MRI features and DCM Surgical-Decision Making
ABSTRACT
SC
Objective: We conducted a survey to understand how specific pathologic features on MRI influence surgeons toward an anterior or posterior surgical approach in Degenerative Cervical Myelopathy (DCM) .
M AN U
Methods: A questionnaire was sent out to 6,179 AOSpine International members via email. This included 18 questions on a 7-point Likert scale regarding how MRI features influence the respondent’s decision to perform an anterior or posterior surgical approach. Influence was classified based on the mean and mode. Variations in responses were assessed by region and training.
AC C
EP
TE D
Results: Of 513 respondents, 51.7% were orthopedic surgeons, 36.8% neurosurgeons and the remainder were fellows, residents or “other”. In ascending order, multilevel bulging discs, cervical kyphosis and a high degree of anterior cord compression had a moderate to strong influence toward an anterior approach. A high degree of posterior cord compression had a strong-moderate influence, while multilevel compression, OPLL, ligamentum flavum enlargement, and congenital stenosis had a moderate influence toward a posterior approach. Neurosurgeons chose anterior approaches more and posterior approaches less, in comparison to orthopedic surgeons (p<0.01). Of note, 59.8% of respondents were equally comfortable performing multilevel (≥3 levels) anterior and posterior procedures, while 61.5% did not feel comfortable in determining the surgical approach based MRI alone. Conclusion: Specific DCM pathology influences the choice for anterior or posterior surgical approach. These data highlight factors based on surgeon experience, training and region of practice. They will be helpful in defining future areas of investigation in an effort to provide individualized surgical strategies and optimize patient outcomes.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
AC C
EP
TE D
M AN U
SC
RI PT
KEYWORDS: Cervical Spondylotic Myelopathy (CSM), Ossification of the posterior longitudinal ligament (OPLL), Degenerative Cervical Myelopathy (DCM), surgical approach, anterior, posterior, imaging
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making INTRODUCTION
METHODS
TE D
M AN U
SC
RI PT
Degenerative Cervical Myelopathy (DCM) encompasses a set of age-related or genetically determined degenerative changes of the cervical spine involving multiple spinal structures that result in progressive spinal cord dysfunction due to cord compression and dynamic injury.1 Patients with DCM typically present with a wide range of neurological deficits including numbness, weakness and gait impairment. While the diagnosis of DCM is based on clinical presentation, MRI is necessary for both confirmation of a clinical diagnosis and surgical planning. Although radiographs and CT may be acquired as well, the clear depiction of the cord and spinal structures makes MRI a powerful tool. Once the decision is made to treat DCM with surgery, surgeons may opt for an anterior and/or posterior approach to achieve spinal cord decompression and cervical spine stabilization depending on the underlying pathology, their personal preference and various other factors2. As a consequence, there has been considerable interest to determine which approach may be more effective when both can be used.3-8 Indeed, this topic is currently being investigated in a multicenter study (CSM-S),1 which is randomizing patients into anterior or posterior surgical treatment when there is equipoise for undertaking either procedure.9 Given that anatomical factors have an important role in the surgical-decision making process, we conducted an international survey of AOSpine International members to evaluate which specific MRI features influenced surgeons’ choice between an anterior or posterior approach and whether type of specialty and region of practice influenced surgical decision making.
AC C
EP
A group of nine spine researchers, neurosurgery residents, orthopedic surgeons and neurosurgeons discussed the various MRI features that could potentially influence the choice for an anterior or posterior surgical approach for treatment of DCM (Figure 1). The group composed 18 questions based on a seven-point Likert scale to indicate no preference or either a strong, moderate or weak preferences for an anterior or posterior surgical approach (See Appendix A). Respondents were also asked to provide additional answers about their surgical preferences, as well as their specialty, region of practice and their highest level of training. An option was provided to allow for additional comment or suggestions. The English language questionnaire was constructed using Survey Monkey (www.surveymonkey.net, Palo Alto, USA) and a link asking to participate in the survey was sent out to the AOSpine International community (6,179 members) via email in August 2016. Two additional reminders were sent, and the survey was closed in October 2016.
Statistical Analysis Likert scale responses were approximated by a linear scale and attributed a value of 3, 2 and 1 corresponding to strong, moderate and weak preference for anterior 1
https://clinicaltrials.gov/ct2/show/NCT02076113
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making surgery, respectively. Conversely, responses were attributed a value of -3, -2 and -1, corresponding to strong, moderate and weak preference for posterior surgery, respectively. A response of no preference for either anterior or posterior was valued as 0.
SC
RI PT
Degree of influence Overall influence of a MRI factor was estimated based on the mode and mean. The mode was identified by the most frequent response. The mean (m) was computed in the following manner: No influence was determined if the mean approximated 0 (0.50≤m≤0.50); Increasing degree of influence were apportioned equal ranges of 0.83 away from the no influence range in the following manner: (1) Strong, moderate and weak influence for anterior surgery was considered for values of m>2.16, 1.33m≥-2.16, -0.50>m≥-1.33, respectively.
TE D
M AN U
Assessment of differences based on region, geography and comfort with performing multilevel procedures Responses for Likert questions were compared between regions for categories with >50 respondents using Kruskal-Wallis tests. Additional post-hoc Mann-Whitney U tests with Bonferroni correction for each post-hoc test were performed to determine significant pairwise differences for statistically significant Kruskal-Wallis tests (p<0.05). Mann-Whitney U tests were also used to compare responses between all neurosurgeons and orthopedic surgeons. The frequency of choice for any anterior surgery, any posterior surgery or no influence for MRI factors was compared between all orthopedic surgeons and neurosurgeons via Mann-Whitney U tests. Univariate logistic regression was used to assess the effect of equal comfort in performing multilevel (≥3 levels) anterior and posterior surgery on a dichotomized choice of anterior or posterior for each Likert question.
AC C
EP
RESULTS A total of 513 responses were received, corresponding to an approximate response rate of 8.3%. Most respondents came from Europe (30.4%, n=156), followed by Asia (26.7%, n=137), South America (22.8%, n=117), North America (14.4%, n=74), Africa (3.9%, n=20) and Oceania (1.4%, n=7). Most respondents were spine fellowship trained orthopedic surgeons (41.5%, n=213) or neurosurgeons (17.7%, n=91), Table 1. The majority of the remaining respondents were non-fellowship trained neurosurgeons (19.1%) and orthopedic surgeons (10.1%). Fellows, residents and the “other” category comprised the remaining 11.6% of respondents. Most respondents (59.8%, 305/510) indicated that they were equally comfortable doing multilevel (≥3 levels) anterior and posterior cervical decompressive/reconstructive procedures, but did not feel comfortable in determining the surgical approach based only on MRI (61.5%, 312/507). There was no significant difference in the level of comfort for performing multilevel (≥3 levels) anterior or posterior approaches between neurosurgeons and orthopedic surgeons (p=0.48), nor did this affect the choice for anterior or posterior surgery for any of
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
SC
RI PT
the MRI features (p>0.05) with the exception of one case. Equal comfort for performing anterior and posterior multilevel procedures predicted a lower likelihood of performing a posterior approach for foraminal stenosis [OR=0.51 (0.33-0.78), p=0.002]. Respondents estimated that they perform an anterior-only approach, a posterior-only approach, and a combined approach at an average of 61%, 30% and 9%, respectively. All respondents indicated a comfort in treating more than one level from an anterior approach, but most surgeons (51.9%, 254/489) indicated that 3 levels was their maximum. There were 30.7% (150/489) who indicated that they would treat more than three levels by an anterior approach. Of the total group, 24.5% (120/489) were comfortable treating up to for 4 levels, 5.1% (25/489) for 5 levels and 1.0% (5/489) for 6 levels. Only 17.4% (85/489) indicated that 2 levels was the maximum that they would treat with an anterior approach.
M AN U
Likert Question Results Descriptive responses, the modes, mean, and overall perceived surgical influence for each MRI factor is reported in Table 2. Eight out of 18 MRI factors resulted in responses that favored an anterior approach, 7 favored a posterior approach and three were not significantly influential.
TE D
MRI Factors Favoring Anterior Surgical Influences A high degree of anterior compression was the strongest factor (Strong Influence) leading respondents to favor an anterior approach. Following this was strongmoderate influence for anterior surgery in cervical kyphosis, and moderate influence for anterior surgery for multilevel budging discs. Moderate-weak preference for anterior surgery was reported for focal OPLL, retro-vertebral disease and spondylolisthesis. MRI factors that had moderate-no influence or weak-no influence for anterior surgery were foraminal stenosis and Modic changes, respectively.
AC C
EP
MRI Factors Favoring Posterior Surgical Influences A high degree of posterior cord compression represented the strongest factor (strong-moderate influence) leading the respondents to favor a posterior approach. Moderate influence for posterior surgery was reported for multilevel pathology, multilevel OPLL, ligamentum flavum enlargement, and congenital stenosis. Moderate-weak and Weak-no influence for posterior surgery was reported for the presence of a short neck and Klippel-Feil Syndrome, respectively. MRI Factors That Did Not Clearly Favor An Anterior or Posterior Approach The presence of cervical scoliosis, spinal cord intramedullary signal changes, and previous anterior cervical surgery did not demonstrate a clear influence on the approach. However, there was substantial variability in the responses for cervical scoliosis and previous anterior surgery. Approximately two thirds of responses for these questions were something other than no influence.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
SC
RI PT
Variation by Training Surgical decision-making was significantly different between neurosurgeons and orthopedic surgeons training for 7 out of 18 Likert questions (greater number of vertebral levels with cord compression, multilevel OPLL, short neck, cervical scoliosis, cord signal intensity changes, previous anterior surgery, spondylolisthesis), and tended to be different in 6 other questions (p<0.10), (Table 3 and APPENDIX B). In all cases, neurosurgeons were more influenced toward anterior surgery or were less influenced toward posterior surgery than orthopedic surgeons. This was also the case for comparisons that tended to be different, with the exception of the presence of congenital cervical fusion (Klippel-Feil), which neurosurgeons favored posterior to a higher degree than orthopedic surgeons (p=0.054).
TE D
M AN U
Variation by Region Comparisons between Asia, Europe, North and South America showed significant differences between 15 of the 18 MRI features influencing surgical decision-making (p<0.05) (Table 3 and APPENDIX C). Differences in the degree of influence but similar approach (i.e anterior or posterior) represented 10 of the 15 factors. In 5 of the statistical differences, there was divergence between regional preference with 4 differing in terms of no influence and a surgical preference. In one case, when MRI is suggestive of focal OPLL, North America displayed a median preference for a posterior approach, whereas all other regions showed preference for an anterior approach (p=0.008). Differences existed most commonly between the regions Europe vs. Asia (11 MRI factors) and South America vs. Asia (7 MRI factors), whereas differences, less commonly between North America vs. Asia (5 factors) and North America vs. South America (5 factors), and least common between Europe vs. South America (2 factors) and Europe vs. North America (3 factors). Generally, Asia showed a greater degree of influence for posterior approaches and less influence toward anterior approaches when differences were observed.
AC C
EP
Additional Optional Comments Common themes were that multiple imaging modalities and physical examination play important roles in surgical decision-making. In response to the final question, “Do you have any suggestions for other ways in which you use MRI to decide on anterior versus posterior surgery for Degenerative Cervical Myelopathy?”, the most common point was that multiple imaging modalities should be considered (16.1%, 27/168) and that addition of dynamic MRI to the decision-making process may be useful (14.3%, 24/168). Other frequently mentioned comments and specific suggestions are listed in Table 4. For most questions, participants commonly also commented that sagittal alignment would impact their decision.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
AC C
EP
TE D
M AN U
SC
RI PT
DISCUSSION This is the first global survey to demonstrate that specific MRI pathology play a role in influencing surgeons’ choice for an anterior or posterior surgical approach, and that region and specialty training can affect these choices. Furthermore, this survey provides the first indication to what degree (i.e. strong, moderate, weak, no influence) specific factors influence a surgeon towards anterior or posterior surgery. These findings will play an important role for ongoing and future studies that are comparing the efficacy of anterior vs. posterior surgical procedures. There were 8 MRI factors favoring an anterior approach whereas 7 favored a posterior approach. The most impactful pathology to influence surgeons towards anterior was a high degree of anterior compression and presence of cervical kyphosis. This was expected as compressive disc pathology can be directly removed from the anterior and cord compression may not be sufficiently relieved by posterior approaches. In terms of cervical kyphosis, anterior surgery may be preferred because it allows for a certain degree of kyphosis correction by applying distraction while reconstructing the vertebral column, particularly in cases of fixed kyphosis.10 Further, there may be limited posterior migration of the cord if treated posteriorly without appropriate alignment correction. Also, there is a concern for progression of kyphosis after posterior approaches such as laminoplasty or laminectomy without instrumented fusion.11,12 Moderate-weak influences for anterior approaches were seen with focal OPLL, retro-vertebral disease, and spondylolisthesis in ascending order. It was notable that focal OPLL was the only MRI factor that showed marked divergence between regions, with the median North American response favoring a posterior approach while all other regions favored an anterior approach (p=0.008). Foraminal stenosis demonstrated a weak influence toward anterior; however, this was based on a mode of moderate influence anterior and a mean of no influence. This divergence may be partially due to the fact that foraminal stenosis can be easily treated from either anterior or posterior approaches. Additionally, the choice of approach for foraminal stenosis represented the only MRI factor that was influenced by a surgeons’ comfort in performing multilevel anterior procedures [OR=0.51, p=0.002], with surgeons that were not comfortable with a multilevel anterior approach more commonly indicating a preference for a posterior approach. The most impactful pathologies to influence a posterior approach were a high degree of posterior cord compression, multilevel compression or OPLL, ligamentum flavum enlargement, and congenital stenosis. Generally, it seems reasonable that a posterior approach was favored in these cases to allow for expansion of the spinal canal and posterior shift of the spinal cord. In addition to this, the comments provided by respondents suggest that multilevel pathology, frequently demarcated by respondents as >3 levels, is favored due to longer operative time, complexity, and concerns with fusion (based on clinical factors and bone quality) with multilevel anterior approaches. Although these concerns differ based on surgical experience, there are clearly differences in surgical preference as only 60% of surgeons felt equally comfortable performing ≥3 levels anteriorly or posteriorly, with most preferring posterior approaches with ≥3 levels. The influence of multilevel OPLL on choosing a posterior surgical approach is not surprising, but is
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
AC C
EP
TE D
M AN U
SC
RI PT
worthy of note, because treatment is aimed at cord decompression through canal expansion and posterior shift of the cord rather than the pathology itself. As a consequence, posterior approaches are dependent on the thickness of the OPLL with more thickness limiting how much further the spinal cord can shift. Indeed, it has been suggested that corpectomy may be preferable in patients with an occupying ratio ≥60%.13 As for ligamentum flavum enlargement, it is intuitive that a posterior approach is favored, as it can directly address the site of cord compression. The least influential factor on MRI was the presence of signal intensity changes in the spinal cord. This was not surprising given that the spinal cord is not involved from a surgical perspective. Interestingly though, a number of respondents commented that advanced MRI techniques such as functional, diffusion tensor and perfusion, which all relate to spinal cord pathology were potentially useful. Two other factors that were not considered to be influential were cervical scoliosis and a previous anterior surgery. However, the “no influence” response only comprised about 38% for each of these (compared to 61% for signal change), suggesting that there was great variability with about a third choosing anterior and posterior approaches for each question. Thus, respondents either answered the question at random or surgical decision-making in these cases may exist but may be surgeon dependent. Geographical region had a significant impact on the responses, with 15 out of 18 questions having statistically significant differences between regions. However, in most cases (14/15) the choice for surgical approach remained the same and the difference was based on the degree of influence, or there was a weak preference compared to no influence. However, focal OPLL influenced North American towards the posterior, while all other regions were influenced toward anterior (p=0.008). The reason for this is unclear, but this does suggest that North Americans prefer to treat all types of OPLL from the posterior, while other regions treat focal and multilevel OPLL differently. Significant variability between responses was also exhibited between neurosurgeons and orthopedic surgeons for 7 of the 18 MRI factors. Neurosurgeons were generally more influenced toward anterior and less influenced towards posterior approaches for when compared to orthopedic surgeons. While differences in training or surgical preference may be contributory. Our findings are in accordance with a previous survey-based study that found neurosurgeons tended to favor anterior approaches, 14 although statistical differences between surgical approaches were not shown due to the limited number of cases in that study. The most common survey respondent comments made were that multiple imaging modalities should be used in the assessment, and that dynamic (flexion/extension) MRI is useful in the diagnostic work-up. Indeed, in terms of dynamic MRI, it has been reported that dynamic stenosis can be uncovered in 8.3% of patients in extension and 1.6% in flexion. 15 Many respondents also suggested the use of a modified K-line to assess sagittal alignment on MRI. While alignment measures are not typically recommended for MRI, it has been reported that the modified K-line can predict surgical outcome.16 Another set of respondents also mentioned that they use MRI to assess the location and course of the vertebral artery as part of their surgical-decision making. Indeed, it has been reported that
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
M AN U
SC
RI PT
vertebral arteries can migrate medially into the vertebral body, increasing the risk for injury during anterior approaches.17 Conversely, variations in vertebral artery anatomy such as a ‘high riding vertebral’ artery or ‘ponticulus posticus’ may elevate the risk of posterior approaches when instrumented fusion is extended to C2.18 Currently, there is neither a standardized approach for surgical treatment nor are there guidelines to direct this decision-making process. Reasons for this include the need for flexibility in treatment given the heterogeneity of DCM presentation, familiarity with procedures, regional differences in surgical preference (i.e. laminoplasty in Asia), costs for instrumentation, and ambiguity regarding potential differences in effectiveness between procedures. In addition, clinical factors such as age as well as other physical characteristics may impact the decision-making process. While there is substantial consensus on how to treat certain types of presentations such as myelopathy due to single disc pathology, ambiguity remains with regard to how to manage conditions for which both approaches are perceived to be equally efficacious. Convincing evidence derived from prospective and multicenter research, has been shown that when surgical decision is left at the discretion of the attending surgeon, anterior and posterior surgery provide comparable efficacy.4 This is further corroborated by systematic reviews concluding that anterior and posterior procedures result in similar outcomes with regards to effectiveness and safety.3,6 However, it remains to be seen whether there are differences in surgical outcome between these surgical approaches in patients for whom equipoise exists for randomization toward an anterior or posterior procedure.
AC C
EP
TE D
Limitations The cross-sectional survey of the international AOSpine community is the largest assessment of the factors influencing surgeon preference for a particular approach to surgically manage degenerative cervical myelopathy. While the response rate for our international survey is somewhat lower than the average response rates for other web based surveys,19 it is comparable to other surveys of the international AOSpine community that have been recently published.20-22 Our response rate is likely the result of a number of factors including: (1) not all AOSpine members may have received the email (filtered out as spam or wrong email address), (2) lack of English language proficiency, and (3) non-surgeon members. The most important limitation to this survey was that each of the factors questioned was asked to be considered in isolation. The impact of concomitant pathologies may alter one’s perspective on the most effective approach. Moreover, nearly two-thirds of individuals indicated that they were not comfortable using MRI alone to determine surgical approach and many commented that their decision would be swayed by sagittal alignment. Upright lateral radiographs would therefore be a valuable addition to MRI and can provide Cobb angle, T1 slope and sagittal vertical axis (SVA) measurements. Furthermore, all anterior and posterior approaches are not the same and the choice for a specific procedure was not considered. It should also be noted that results might have been impacted by the inclusion of residents and fellows, who are likely still in the process of learning to optimize their decisionmaking. Lastly, reliance on MRI to make surgical decisions assumes that this
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
RI PT
modality is readily available, however, the global nature of this survey and recognition that some regions may not have this accessibility, may impact responses from these regions.
CONCLUSION
M AN U
SC
A large degree of anterior cord compression and cervical kyphosis strongly influence surgeons toward an anterior approach. Conversely, a high degree of posterior cord compression, congenital stenosis, and multilevel compression or OPLL were the strongest factors influencing surgeons toward a posterior surgery. It is also evident that the surgeons’ preference is influenced by region and type of training. Additional research through investigation of case-based surgical decisionmaking is recommended to study this topic further.
FIGURE LEGEND
2.
3.
4.
Nouri A, Tetreault L, Singh A, Karadimas S, Fehlings M. Degenerative Cervical Myelopathy: Epidemiology, Genetics and Pathogenesis. Spine (Phila Pa 1976). 2015;40(12):E675-693. Lawrence BD, Shamji MF, Traynelis VC, et al. Surgical management of degenerative cervical myelopathy: a consensus statement. Spine (Phila Pa 1976). 2013;38(22 Suppl 1):S171-172. Luo J, Cao K, Huang S, et al. Comparison of anterior approach versus posterior approach for the treatment of multilevel cervical spondylotic myelopathy. Eur Spine J. 2015;24(8):1621-1630. Fehlings MG, Barry S, Kopjar B, et al. Anterior versus posterior surgical approaches to treat cervical spondylotic myelopathy: outcomes of the prospective multicenter AOSpine North America CSM study in 264 patients. Spine (Phila Pa 1976). 2013;38(26):2247-2252.
AC C
1.
EP
REFERENCES
TE D
Figure 1. Two patients with preoperative and postoperative sagittal T2 weighted MRIs and clinically confirmed DCM are presented. A) Patient 1 presents with spinal cord compression by the C5-6 disc. B) A 6-months post-operative MRI of patient 1 showing that the patient was treated with an anterior discectomy and plate at C5-6. C) Patient 2 presents with multilevel spinal cord compression with the origin of stenosis coming from both the anterior and posterior in the upper and middle cervical spine. D) A 24-months post-operative MRI of patient 2 showing that the patient was treated with multilevel posterior laminectomy and fusion from C2-T3.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making 5.
RI PT
6.
Ren H, Liu F, Yu D, et al. Patterns of Neurological Recovery After Anterior Decompression With Fusion and Posterior Decompression With Laminoplasty for the Treatment of Multilevel Cervical Spondylotic Myelopathy. Clin Spine Surg. 2016. Lawrence BD, Jacobs WB, Norvell DC, Hermsmeyer JT, Chapman JR, Brodke DS. Anterior versus posterior approach for treatment of cervical spondylotic myelopathy: a systematic review. Spine (Phila Pa 1976). 2013;38(22 Suppl 1):S173-182. Ghogawala Z, Martin B, Benzel EC, et al. Comparative effectiveness of ventral vs dorsal surgery for cervical spondylotic myelopathy. Neurosurgery. 2011;68(3):622-630; discussion 630-621. Seng C, Tow BP, Siddiqui MA, et al. Surgically treated cervical myelopathy: a functional outcome comparison study between multilevel anterior cervical decompression fusion with instrumentation and posterior laminoplasty. Spine J. 2013;13(7):723-731. Ghogawala Z, Benzel EC, Heary RF, et al. Cervical spondylotic myelopathy surgical trial: randomized, controlled trial design and rationale. Neurosurgery. 2014;75(4):334-346. Emery SE. Anterior approaches for cervical spondylotic myelopathy: which? When? How? Eur Spine J. 2015;24 Suppl 2:150-159. Kaptain GJ, Simmons NE, Replogle RE, Pobereskin L. Incidence and outcome of kyphotic deformity following laminectomy for cervical spondylotic myelopathy. J Neurosurg. 2000;93(2 Suppl):199-204. Suk KS, Kim KT, Lee JH, Lee SH, Lim YJ, Kim JS. Sagittal alignment of the cervical spine after the laminoplasty. Spine (Phila Pa 1976). 2007;32(23):E656-660. Chen Z, Liu B, Dong J, et al. Comparison of anterior corpectomy and fusion versus laminoplasty for the treatment of cervical ossification of posterior longitudinal ligament: a meta-analysis. Neurosurg Focus. 2016;40(6):E8. Irwin ZN, Hilibrand A, Gustavel M, et al. Variation in surgical decision making for degenerative spinal disorders. Part II: cervical spine. Spine (Phila Pa 1976). 2005;30(19):2214-2219. Hayashi T, Wang JC, Suzuki A, et al. Risk factors for missed dynamic canal stenosis in the cervical spine. Spine (Phila Pa 1976). 2014;39(10):812-819. Taniyama T, Hirai T, Yoshii T, et al. Modified K-line in magnetic resonance imaging predicts clinical outcome in patients with nonlordotic alignment after laminoplasty for cervical spondylotic myelopathy. Spine (Phila Pa 1976). 2014;39(21):E1261-1268. Eskander MS, Drew JM, Aubin ME, et al. Vertebral artery anatomy: a review of two hundred fifty magnetic resonance imaging scans. Spine (Phila Pa 1976). 2010;35(23):2035-2040. Elgafy H, Pompo F, Vela R, Elsamaloty HM. Ipsilateral arcuate foramen and high-riding vertebral artery: implication on C1-C2 instrumentation. Spine J. 2014;14(7):1351-1355. Shih T-H, Fan X. Comparing response rates from web and mail surveys: A meta-analysis. Field methods. 2008;20(3):249-271.
7.
11.
12.
13.
14.
AC C
15.
TE D
10.
EP
9.
M AN U
SC
8.
16.
17.
18.
19.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
21.
AC C
EP
TE D
M AN U
SC
22.
Tetreault L, Nouri A, Singh A, Fawcett M, Nater A, Fehlings MG. An Assessment of the Key Predictors of Perioperative Complications in Patients with Cervical Spondylotic Myelopathy Undergoing Surgical Treatment: Results from a Survey of 916 AOSpine International Members. World Neurosurg. 2015;83(5):679-690. Schroeder GD, Kurd MF, Kepler CK, et al. The Development of a Universally Accepted Sacral Fracture Classification: A Survey of AOSpine and AOTrauma Members. Global Spine J. 2016;6(7):686-694. Nater A, Tetreault LL, Davis AM, Sahgal AA, Kulkarni AV, Fehlings MG. Key Preoperative Clinical Factors Predicting Outcome in Surgically Treated Patients with Metastatic Epidural Spinal Cord Compression: Results from a Survey of 438 AOSpine International Members. World Neurosurg. 2016;93:436-448 e415.
RI PT
20.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Survey on MRI features and DCM Surgical-Decision Making
ACCEPTED MANUSCRIPT
Table 1. Training Level of Respondents
AC C
EP
TE D
RI PT
M AN U
Resident Fellow, Spine Surgery Fellow, Other discipline Neurosurgeon Neurosurgeon, fellowship trained in spine Orthopedic Surgeon Orthopedic Surgeon, fellowship trained in spine Other Total
Frequency (number of respondents) 2.9% (n=15) 6.2% (n=32) 0.2% (n=1) 19.1% (n=98) 17.7% (n=91) 10.1% (n=52) 41.5% (n=213) 2.1% (n=11) 513
SC
Type of Training
ACCEPTED MANUSCRIPT
Table 2. Results from the Likert Scale questions. Modes are bolded. Overall influence was classified based on both the mode and mean.
Anterior Approach Strong (+3) Moderate Weak (+2) (+1)
Neither (0)
19.7% (n=101)
3.9% (n=20)
40.6% (n=207)
33.7% (n=172)
7.1% (n=36)
5.1% (n=26)
30.3% (n=154)
31.6% (n=161)
9.0% (n=46)
Overall Influence
0% (n=0)
2.65 (n=512)
Strong Anterior
2.9% (n=15)
7.5% (n=38)
3.1% (n=16)
1.69 (n=510)
StrongModerate Anterior
17.7% (n=90)
3.3% (n=17)
5.9% (n=30)
2.2% (n=11)
1.41 (n=509)
Moderate Anterior
11.6% (n=59)
15.3% (n=78)
5.1% (n=26)
10.2% (n=52)
4.1% (n=21)
1.03 (n=509)
Moderateweak Anterior
TE D
32.0% (n=163)
Mean (m)
0.4% (n=2)
EP
21.6% (n=110)
1.8% (n=9)
0% (n=0)
M AN U
74.1% (n=380)
AC C
1. High degree of anterior compression (e.g. large bulging disc). 9. Presence of cervical kyphosis on MRI (Interpreted as extent of kyphosis in a neutral to extended neck position) 4. Greater number of bulging discs that are not causing cord compression (i.e. cord is only compressed at 1 level but discs show degeneration at other levels). 18. Presence of Spondylolisthesis (Defined as ≥3mm displacement) 5. MRI suggestive of focal ossification of the posterior longitudinal ligament (affecting one segment) 8. Presence of cord compression due to retrovertrebral disease (midvertebral level). 14. Presence of foraminal stenosis 13. Presence of Modic changes in the vertebral bodies
Posterior Approach Moderate Strong (-2) (-3)
SC
Anterior approach favored
Weak (-1)
RI PT
Question - Please respond as to how the following factors (in isolation), as seen on MRI, influence your surgical decision toward anterior or posterior surgery?
21.9% (n=111)
27.0% (n=137)
7.3% (n=37)
7.3% (n=37)
8.3% (n=42)
18.5% (n=94)
9.7% (n=49)
0.53 (n=507)
Moderateweak Anterior
23.2% (n=117)
25.8% (n=130)
8.5% (n=43)
11.5% (n=58)
5.6% (n=28)
17.5% (n=88)
7.9% (n=40)
0.65 (n=504)
Moderateweak Anterior
11.6% (n=59)
24.9% (n=126) 15.7% (n=80)
15.4% (n=78) 14.6% (n=74)
22.3% (n=113) 52.4% (n=266)
6.3% (n=32) 1.6% (n=8)
13.2% (n=67) 3.0% (n=15)
6.3% (n=32) 0.6% (n=3)
0.48 (n=507) 0.73 (n=508)
Moderate-No Influence Weak Anterior-No Influence
12.2% (n=62)
ACCEPTED MANUSCRIPT
Posterior approach favored 5.7% (n=29)
1.2% (n=6)
1.4% (n=7)
7.7% (n=39)
29.1% (n=148)
50.7% (n=258)
-1.92 (n=509)
Strong to Moderate Posterior
2.0% (n=10)
3.7% (n=19)
0.6% (n=3)
11.8% (n=60)
9.8% (n=50)
38.6% (n=196)
33.5% (n=170)
-1.73 (n=508)
Moderate Posterior
5.7% (n=29)
7.3% (n=37)
4.3% (n=22)
3.1% (n=16)
28.9% (n=147)
44.7% (n=227)
-1.62 (n=508)
Moderate Posterior
2.0% (n=10)
2.6% (n=13) 8.1% (n=41) 8.7% (n=44)
1.8% (n=9) 1.8% (n=9) 7.1% (n=36)
9.6% (n=49) 4.9% (n=25) 19.8% (n=100)
25.1% (n=128) 8.4% (n=43) 19.8% (n=100)
39.3% (n=200) 41.8% (n=213) 30.1% (n=152)
19.6% (n=100) 29.1% (n=148) 11.3% (n=57)
-1.50 (n=509) -1.44 (n=509) -0.80 (n=505)
Moderate Posterior Moderate Posterior Moderateweak Posterior
8.3% (n=42)
5.6% (n=28)
32.1% (n=162)
12.9% (n=65)
25.0% (n=126)
12.9% (n=65)
-0.70 (n=504)
Weak Posterior-No Influence
3.2% (n=16)
SC 5.9% (n=30)
M AN U
TE D
5.9% (n=30) 3.2% (n=16)
RI PT
4.3% (n=22)
EP
2. High degree of posterior cord compression (e.g. ligamentum flavum inbuckling). 17. Presence of congenital canal stenosis (e.g. cord occupies more than 70% of the canal at C2 or C7 in the absence of disc/ligament pathology at these levels) 6. MRI suggestive of multilevel ossification of the posterior longitudinal ligament (>1 segment) 12. Presence of ligamentum flavum enlargement 3. Greater number of vertebral levels with cord compression. 7. Presence of a short neck length and/or chin obstructing the most rostral (upper) disc level requiring surgery. 15. Presence of autofusion (or Klippel-Feil)
10. Presence of cervical scoliosis on MRI 11. Presence of cord signal intensity changes (T2weighted hyper-intensity, T1weighted hypo-intensity) 16. Previous anterior surgery (e.g. single-level ACDF)
AC C
No Influence
4.9% (n=25) 9.1% (n=46)
5.7% (n=29)
15.0% (n=76) 11.5% (n=58)
4.9% (n=25) 2.8% (n=14)
38.1% (n=193) 61.1% (n=309)
10.3% (n=52) 4.0% (n=20)
18.3% (n=93) 6.7% (n=34)
8.5% (n=43) 4.9% (n=25)
-0.23 (n=507) 0.21 (n=506)
No Influence
16.2% (n=83)
6.1% (n=31)
38.7% (n=198)
13.9% (n=71)
13.5% (n=69)
5.9% (n=30)
-0.03 (n=511)
No Influence
No Influence
ACCEPTED MANUSCRIPT
Table 3. Assessment of Variation by Region and Training. Detailed cross-comparisons are available in APPENDIX B and C.
6. MRI suggestive of multilevel ossification of the posterior longitudinal ligament (>1 segment)
7. Presence of a short neck length and/or chin obstructing the most rostral (upper) disc level requiring surgery. 8. Presence of cord compression due to retrovertrebral disease (mid-vertebral level).
Training, Neurosurgeon (NS) vs. Orthopedic Surgeon (OS) P=0.076 NS – Mean Rank = 217.23 (n=189); OS – Mean Rank = 233.99 (n=264) P=0.162 NS – Mean Rank = 216.63 (n=187); OS – Mean Rank = 232.64 (n=264)
RI PT
Group – P=0.008 NA vs. SA p=0.018; NA vs. EU p=0.01; NA vs. Asia p=0.027 North America was influenced toward a posterior approach whereas all other groups were influenced toward an anterior approach Group – P<0.001 NA vs. SA p=0.015; NA vs. EU p<0.001; NA vs. Asia p=0.035 North America was most strongly influenced toward a posterior approach Group - P=0.265
AC C
5. MRI suggestive of focal ossification of the posterior longitudinal ligament (affecting one segment)
TE D
4. Greater number of bulging discs that are not causing cord compression (i.e. cord is only compressed at 1 level but discs show degeneration at other levels).
EP
3. Greater number of vertebral levels with cord compression.
Group – P=0.001 EU vs. Asia p=0.001; SA vs. Asia p=0.016 Asia was less influenced toward an anterior approach. Group – P<0.001 NA vs. Asia p=0.005; SA vs. Asia p=0.021; EU vs. Asia p=0.001 Asia was more strongly influenced toward a posterior approach Group – P<0.001 EU vs. SA p=0.001; EU vs. Asia p=0.001 Europe was less strongly influenced toward a posterior approach Group – P=0.008 SA vs. NA p=0.017; SA vs. Asia p=0.035 South America was more strongly influenced toward an anterior approach
M AN U
2. High degree of posterior cord compression (e.g. ligamentum flavum inbuckling).
Region
SC
Please respond as to how the following factors (in isolation), as seen on MRI, influence your surgical decision toward anterior or posterior surgery? 1. High degree of anterior compression (e.g. large bulging disc).
Group - P=0.106
P=0.007 NS – Mean Rank = 207.45 (n=188); OS – Mean Rank = 239.26 (n=263) P=0.099 NS – Mean Rank = 213.93 (n=187); OS – Mean Rank = 233.73 (n=263) P=0.052 NS – Mean Rank = 211.17 (n=186); OS – Mean Rank = 234.78 (n=263)
P=0.003 NS – Mean Rank = 205.29 (n=186) OS – Mean Rank = 239.87 (n=263) P=0.001 NS – Mean Rank = 200.94 (n=187) OS – Mean Rank = 239.79 (n=259) P=0.843 NS – Mean Rank = 224.91 (n=185)
ACCEPTED MANUSCRIPT
SC
AC C
13. Presence of Modic changes in the vertebral bodies
Group – P<0.001 Asia vs. NA p<0.001; Asia vs. SA p=0.02; Asia vs. EU p<0.001 Asia was most strongly influenced toward posterior approach, while the other regions displayed less influence toward posterior. Group – P<0.001 NA vs. SA p=0.043; NA vs. EU p=0.013; Asia vs. SA p=0.021; Asia vs. EU p=0.004 North America and Asia were generally not influenced, but South America and Europe were slight influenced toward anterior. Group – P=0.001 EU vs. Asia p<0.001 Europe had a slight influence toward anterior, while Asia was most commonly not influenced (but had high variability) Group - P=0.539
EP
12. Presence of ligamentum flavum enlargement
TE D
11. Presence of cord signal intensity changes (T2weighted hyper-intensity, T1-weighted hypointensity)
M AN U
10. Presence of cervical scoliosis on MRI
14. Presence of foraminal stenosis
Group – P<0.001 SA vs. EU p=0.024; SA vs. Asia p<0.001; NA vs. Asia p=0.007 South America was most strongly influenced toward anterior. Asia was less strongly influenced toward anterior. Group – P=0.013 EU vs. Asia p=0.012 All regions had a median of no influence, but EU had more respondents favouring anterior and Asia had more favouring posterior. Group – P=0.017 EU vs. Asia p=0.03 Asia showed no influence, but EU was slightly influenced toward anterior
RI PT
9. Presence of cervical kyphosis on MRI (Interpreted as extent of kyphosis in a neutral to extended neck position)
OS – Mean Rank = 222.50 (n=261) P=0.603 NS – Mean Rank = 211.17 (n=186) OS – Mean Rank = 234.78 (n=263)
15. Presence of autofusion (or Klippel-Feil)
P=0.036 NS – Mean Rank = 210.28 (n=186) OS – Mean Rank = 235.41 (n=263) P=0.023 NS – Mean Rank = 210.25 (n=188) OS – Mean Rank = 234.80 (n=260) P=0.074 NS – Mean Rank = 213.07 (n=187) OS – Mean Rank = 234.33 (n=263)
P=0.372 NS – Mean Rank = 219.54 (n=187) OS – Mean Rank = 229.74 (n=263)
P=0.612 NS – Mean Rank = 221.39 (n=187) OS – Mean Rank = 227.57 (n=262) P=0.054 NS – Mean Rank = 237.09 (n=185) OS – Mean Rank = 213.87 (n=261)
ACCEPTED MANUSCRIPT
16. Previous anterior surgery (e.g. single-level ACDF)
AC C
EP
TE D
M AN U
SC
18. Presence of Spondylolisthesis (Defined as ≥3mm displacement)
P=0.007 NS – Mean Rank = 207.61 (n=189) OS – Mean Rank = 240.07 (n=263) P=0.090 NS – Mean Rank = 213.83 (n=187) OS – Mean Rank = 233.80 (n=263) P<0.001 NS – Mean Rank = 196.88 (n=189) OS – Mean Rank = 247.01 (n=262)
RI PT
17. Presence of congenital canal stenosis (e.g. cord occupies more than 70% of the canal at C2 or C7)
Group – P=0.014 EU vs. Asia p=0.028 Europe had a slight influence toward anterior, while Asia was most commonly not influenced (but had high variability) Group – P=0.014 EU vs. Asia p=0.020 Europe was less strongly influenced toward posterior than Asia which was most influenced toward posterior Group – P<0.001 SA vs. NA p<0.001; SA vs. Asia p<0.001; EU vs. Asia p<0.001 South America was most influenced toward anterior, while Asia was least influenced toward anterior
ACCEPTED MANUSCRIPT
Table 4. Selected additional comments by respondents in response to “Do you have any suggestions for other ways in which you use MRI to decide on anterior versus posterior surgery for Degenerative Cervical Myelopathy?”. Selected comments provided for specific questions are also included. Comments have been modified for grammar and context.
RI PT
Q - Do you have any suggestions for other ways in which you use MRI to decide on anterior versus posterior surgery for Degenerative Cervical Myelopathy? Dynamic (flexion/extension) MRI should be considered in the work up of patients to assess movement dependent cord compression. Standing and/or flexion radiographs, and/or CT with or without contrast should complement MRI in decision-making.
Collated from multiple responses
SC
Collated from multiple responses
MRI is useful for assessing the course and location of the vertebral artery. MRI measurement of the modified K-line is useful for assessment of sagittal alignment. Advanced MRI techniques (functional, diffusion tensor, perfusion) can be used in surgical decision-making. Clinical risk factors for non-union such as smoking and diabetes may sway me to consider a posterior approach. For patients likely needing a combined approach, I favour an anterior approach initially with followed up. If there is progression or lack of improvement I then add a posterior approach.
M AN U
Collated from multiple responses Collated from multiple responses Collated from multiple responses Specific Comment
Specific Comment
TE D
Specific Comment
Alignment assessment for MRI should be standardized as some are taken with pillows, making alignment assessment difficult.
Circumferential compression on MRI should be approached posteriorly. Specific Comment Measurement of disc height at the site of pathology in comparison to disc height at other sites. Specific Comment Level of and area of T2 signal change could be assessed. Q. High degree of anterior compression (e.g. large bulging disc) Collated from multiple If the compression is very large, a posterior approach done first responses and followed by an anterior approach is sometimes done. Q. Greater number of vertebral levels with cord compression Collated from multiple Most respondents commented that >3 levels should be responses approached posteriorly. Q. Greater number of bulging discs that are not causing cord compression (i.e. cord is only compressed at 1 level but discs show degeneration at other levels) Collated from multiple Treat only the level of pathology; consider arthroplasty to responses preserve motion. Q. Presence of cervical kyphosis on MRI (Interpreted as extent of kyphosis in a neutral to extended neck position) Collated from multiple A combined approach is favoured.
AC C
EP
Specific Comment
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
responses Q. Presence of cervical scoliosis on MRI Collated from multiple A combined approach is favoured. responses Q. Previous anterior surgery (e.g. single-level ACDF) Specific Comment 1) Depends on the “age” of the previous surgery 2) Depends on if compression is at previous surgical site or other segment. Q. Presence of Spondylolisthesis (Defined as ≥3mm displacement) Collated from multiple A combined approach is favoured. responses Q. Do you feel equally comfortable doing multilevel (3 or more levels) anterior and posterior cervical decompressive/reconstructive procedures? Collated from multiple 1) Posterior multilevel procedures are less difficult and favoured responses for >3 levels. 2) Concern for prolonged operating time for multilevel anterior procedures and potential swelling
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
(3 - 5 bullet points, 85 characters incl spaces or less)
Survey Bullet Points
• •
AC C
EP
TE D
M AN U
•
RI PT
•
Specific pathology influence the choice for an anterior/posterior surgical approach Neurosurgeons chose anterior more and posterior less vs. orthopedic surgeons 59.8% were equally comfortable doing multilevel (≥3) anterior & posterior surgery Degree of influence between differed between region & training for 15 & 6 factors Further investigation through case-based surgical decision-making is recommended
SC
•
ACCEPTED MANUSCRIPT
ABBREVIATIONS
Degenerative Cervical Myelopathy
OPLL =
Ossification of the Posterior Longitudinal Ligament
MRI
=
Magnetic Resonance Imaging
CT
=
Computed Tomography
AC C
EP
TE D
M AN U
SC
RI PT
DCM =
ACCEPTED MANUSCRIPT
Conflict of Interest and Disclosures
AC C
EP
TE D
M AN U
SC
RI PT
The authors have no relevant conflicts of interest. Dr. Fehlings wishes to disclose consulting agreements with Pfizer, Zimmer Biomet and InVivo Therapeutics
S1878-8750(17)30917-8
DOI:
10.1016/j.wneu.2017.06.025
Reference:
WNEU 5896
To appear in:
World Neurosurgery
Received Date: 4 May 2017 Revised Date:
2 June 2017
Accepted Date: 5 June 2017
Please cite this article as: Nouri A, Martin AR, Nater A, Witiw CD, Kato S, Tetreault L, ReihaniKermani H, Santaguida C, Fehlings MG, The Influence of MRI Features on Surgical-Decision Making in Degenerative Cervical Myelopathy: Results From a Global Survey of AOSpine International Members, World Neurosurgery (2017), doi: 10.1016/j.wneu.2017.06.025. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
SC
RI PT
The Influence of MRI Features on SurgicalDecision Making in Degenerative Cervical Myelopathy: Results From a Global Survey of AOSpine International Members
M AN U
Aria Nouri MD MSc1, Allan R. Martin MD1, Anick Nater MD1, Christopher D. Witiw MD MS1, So Kato MD1, Lindsay Tetreault PhD1, Hamed Reihani-Kermani MD2, Carlo Santaguida MD3, Michael G. Fehlings MD PhD1,4
TE D
Affiliations 1Division of Neurosurgery, Toronto Western Hospital, University Health Network, Toronto, Ontario, Canada 2Department of Neurosurgery, Kerman University of Medical Sciences, Kerman, Iran 3McGill University Health Centre, Montreal, Quebec, Canada 4Department of Surgery, University of Toronto, Toronto, Ontario, Canada
EP
Corresponding Author:
AC C
Michael G. Fehlings MD PhD FRCSC FACS Professor of Neurosurgery Halbert Chair in Neural Repair and Regeneration University of Toronto 399 Bathurst St. Toronto Western Hospital, Toronto, Ontario Canada Email: [email protected] Phone: 416-603-5229 Fax: (416) 603-5627
ACCEPTED MANUSCRIPT
RI PT
Survey on MRI features and DCM Surgical-Decision Making
ABSTRACT
SC
Objective: We conducted a survey to understand how specific pathologic features on MRI influence surgeons toward an anterior or posterior surgical approach in Degenerative Cervical Myelopathy (DCM) .
M AN U
Methods: A questionnaire was sent out to 6,179 AOSpine International members via email. This included 18 questions on a 7-point Likert scale regarding how MRI features influence the respondent’s decision to perform an anterior or posterior surgical approach. Influence was classified based on the mean and mode. Variations in responses were assessed by region and training.
AC C
EP
TE D
Results: Of 513 respondents, 51.7% were orthopedic surgeons, 36.8% neurosurgeons and the remainder were fellows, residents or “other”. In ascending order, multilevel bulging discs, cervical kyphosis and a high degree of anterior cord compression had a moderate to strong influence toward an anterior approach. A high degree of posterior cord compression had a strong-moderate influence, while multilevel compression, OPLL, ligamentum flavum enlargement, and congenital stenosis had a moderate influence toward a posterior approach. Neurosurgeons chose anterior approaches more and posterior approaches less, in comparison to orthopedic surgeons (p<0.01). Of note, 59.8% of respondents were equally comfortable performing multilevel (≥3 levels) anterior and posterior procedures, while 61.5% did not feel comfortable in determining the surgical approach based MRI alone. Conclusion: Specific DCM pathology influences the choice for anterior or posterior surgical approach. These data highlight factors based on surgeon experience, training and region of practice. They will be helpful in defining future areas of investigation in an effort to provide individualized surgical strategies and optimize patient outcomes.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
AC C
EP
TE D
M AN U
SC
RI PT
KEYWORDS: Cervical Spondylotic Myelopathy (CSM), Ossification of the posterior longitudinal ligament (OPLL), Degenerative Cervical Myelopathy (DCM), surgical approach, anterior, posterior, imaging
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making INTRODUCTION
METHODS
TE D
M AN U
SC
RI PT
Degenerative Cervical Myelopathy (DCM) encompasses a set of age-related or genetically determined degenerative changes of the cervical spine involving multiple spinal structures that result in progressive spinal cord dysfunction due to cord compression and dynamic injury.1 Patients with DCM typically present with a wide range of neurological deficits including numbness, weakness and gait impairment. While the diagnosis of DCM is based on clinical presentation, MRI is necessary for both confirmation of a clinical diagnosis and surgical planning. Although radiographs and CT may be acquired as well, the clear depiction of the cord and spinal structures makes MRI a powerful tool. Once the decision is made to treat DCM with surgery, surgeons may opt for an anterior and/or posterior approach to achieve spinal cord decompression and cervical spine stabilization depending on the underlying pathology, their personal preference and various other factors2. As a consequence, there has been considerable interest to determine which approach may be more effective when both can be used.3-8 Indeed, this topic is currently being investigated in a multicenter study (CSM-S),1 which is randomizing patients into anterior or posterior surgical treatment when there is equipoise for undertaking either procedure.9 Given that anatomical factors have an important role in the surgical-decision making process, we conducted an international survey of AOSpine International members to evaluate which specific MRI features influenced surgeons’ choice between an anterior or posterior approach and whether type of specialty and region of practice influenced surgical decision making.
AC C
EP
A group of nine spine researchers, neurosurgery residents, orthopedic surgeons and neurosurgeons discussed the various MRI features that could potentially influence the choice for an anterior or posterior surgical approach for treatment of DCM (Figure 1). The group composed 18 questions based on a seven-point Likert scale to indicate no preference or either a strong, moderate or weak preferences for an anterior or posterior surgical approach (See Appendix A). Respondents were also asked to provide additional answers about their surgical preferences, as well as their specialty, region of practice and their highest level of training. An option was provided to allow for additional comment or suggestions. The English language questionnaire was constructed using Survey Monkey (www.surveymonkey.net, Palo Alto, USA) and a link asking to participate in the survey was sent out to the AOSpine International community (6,179 members) via email in August 2016. Two additional reminders were sent, and the survey was closed in October 2016.
Statistical Analysis Likert scale responses were approximated by a linear scale and attributed a value of 3, 2 and 1 corresponding to strong, moderate and weak preference for anterior 1
https://clinicaltrials.gov/ct2/show/NCT02076113
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making surgery, respectively. Conversely, responses were attributed a value of -3, -2 and -1, corresponding to strong, moderate and weak preference for posterior surgery, respectively. A response of no preference for either anterior or posterior was valued as 0.
SC
RI PT
Degree of influence Overall influence of a MRI factor was estimated based on the mode and mean. The mode was identified by the most frequent response. The mean (m) was computed in the following manner: No influence was determined if the mean approximated 0 (0.50≤m≤0.50); Increasing degree of influence were apportioned equal ranges of 0.83 away from the no influence range in the following manner: (1) Strong, moderate and weak influence for anterior surgery was considered for values of m>2.16, 1.33
TE D
M AN U
Assessment of differences based on region, geography and comfort with performing multilevel procedures Responses for Likert questions were compared between regions for categories with >50 respondents using Kruskal-Wallis tests. Additional post-hoc Mann-Whitney U tests with Bonferroni correction for each post-hoc test were performed to determine significant pairwise differences for statistically significant Kruskal-Wallis tests (p<0.05). Mann-Whitney U tests were also used to compare responses between all neurosurgeons and orthopedic surgeons. The frequency of choice for any anterior surgery, any posterior surgery or no influence for MRI factors was compared between all orthopedic surgeons and neurosurgeons via Mann-Whitney U tests. Univariate logistic regression was used to assess the effect of equal comfort in performing multilevel (≥3 levels) anterior and posterior surgery on a dichotomized choice of anterior or posterior for each Likert question.
AC C
EP
RESULTS A total of 513 responses were received, corresponding to an approximate response rate of 8.3%. Most respondents came from Europe (30.4%, n=156), followed by Asia (26.7%, n=137), South America (22.8%, n=117), North America (14.4%, n=74), Africa (3.9%, n=20) and Oceania (1.4%, n=7). Most respondents were spine fellowship trained orthopedic surgeons (41.5%, n=213) or neurosurgeons (17.7%, n=91), Table 1. The majority of the remaining respondents were non-fellowship trained neurosurgeons (19.1%) and orthopedic surgeons (10.1%). Fellows, residents and the “other” category comprised the remaining 11.6% of respondents. Most respondents (59.8%, 305/510) indicated that they were equally comfortable doing multilevel (≥3 levels) anterior and posterior cervical decompressive/reconstructive procedures, but did not feel comfortable in determining the surgical approach based only on MRI (61.5%, 312/507). There was no significant difference in the level of comfort for performing multilevel (≥3 levels) anterior or posterior approaches between neurosurgeons and orthopedic surgeons (p=0.48), nor did this affect the choice for anterior or posterior surgery for any of
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
SC
RI PT
the MRI features (p>0.05) with the exception of one case. Equal comfort for performing anterior and posterior multilevel procedures predicted a lower likelihood of performing a posterior approach for foraminal stenosis [OR=0.51 (0.33-0.78), p=0.002]. Respondents estimated that they perform an anterior-only approach, a posterior-only approach, and a combined approach at an average of 61%, 30% and 9%, respectively. All respondents indicated a comfort in treating more than one level from an anterior approach, but most surgeons (51.9%, 254/489) indicated that 3 levels was their maximum. There were 30.7% (150/489) who indicated that they would treat more than three levels by an anterior approach. Of the total group, 24.5% (120/489) were comfortable treating up to for 4 levels, 5.1% (25/489) for 5 levels and 1.0% (5/489) for 6 levels. Only 17.4% (85/489) indicated that 2 levels was the maximum that they would treat with an anterior approach.
M AN U
Likert Question Results Descriptive responses, the modes, mean, and overall perceived surgical influence for each MRI factor is reported in Table 2. Eight out of 18 MRI factors resulted in responses that favored an anterior approach, 7 favored a posterior approach and three were not significantly influential.
TE D
MRI Factors Favoring Anterior Surgical Influences A high degree of anterior compression was the strongest factor (Strong Influence) leading respondents to favor an anterior approach. Following this was strongmoderate influence for anterior surgery in cervical kyphosis, and moderate influence for anterior surgery for multilevel budging discs. Moderate-weak preference for anterior surgery was reported for focal OPLL, retro-vertebral disease and spondylolisthesis. MRI factors that had moderate-no influence or weak-no influence for anterior surgery were foraminal stenosis and Modic changes, respectively.
AC C
EP
MRI Factors Favoring Posterior Surgical Influences A high degree of posterior cord compression represented the strongest factor (strong-moderate influence) leading the respondents to favor a posterior approach. Moderate influence for posterior surgery was reported for multilevel pathology, multilevel OPLL, ligamentum flavum enlargement, and congenital stenosis. Moderate-weak and Weak-no influence for posterior surgery was reported for the presence of a short neck and Klippel-Feil Syndrome, respectively. MRI Factors That Did Not Clearly Favor An Anterior or Posterior Approach The presence of cervical scoliosis, spinal cord intramedullary signal changes, and previous anterior cervical surgery did not demonstrate a clear influence on the approach. However, there was substantial variability in the responses for cervical scoliosis and previous anterior surgery. Approximately two thirds of responses for these questions were something other than no influence.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
SC
RI PT
Variation by Training Surgical decision-making was significantly different between neurosurgeons and orthopedic surgeons training for 7 out of 18 Likert questions (greater number of vertebral levels with cord compression, multilevel OPLL, short neck, cervical scoliosis, cord signal intensity changes, previous anterior surgery, spondylolisthesis), and tended to be different in 6 other questions (p<0.10), (Table 3 and APPENDIX B). In all cases, neurosurgeons were more influenced toward anterior surgery or were less influenced toward posterior surgery than orthopedic surgeons. This was also the case for comparisons that tended to be different, with the exception of the presence of congenital cervical fusion (Klippel-Feil), which neurosurgeons favored posterior to a higher degree than orthopedic surgeons (p=0.054).
TE D
M AN U
Variation by Region Comparisons between Asia, Europe, North and South America showed significant differences between 15 of the 18 MRI features influencing surgical decision-making (p<0.05) (Table 3 and APPENDIX C). Differences in the degree of influence but similar approach (i.e anterior or posterior) represented 10 of the 15 factors. In 5 of the statistical differences, there was divergence between regional preference with 4 differing in terms of no influence and a surgical preference. In one case, when MRI is suggestive of focal OPLL, North America displayed a median preference for a posterior approach, whereas all other regions showed preference for an anterior approach (p=0.008). Differences existed most commonly between the regions Europe vs. Asia (11 MRI factors) and South America vs. Asia (7 MRI factors), whereas differences, less commonly between North America vs. Asia (5 factors) and North America vs. South America (5 factors), and least common between Europe vs. South America (2 factors) and Europe vs. North America (3 factors). Generally, Asia showed a greater degree of influence for posterior approaches and less influence toward anterior approaches when differences were observed.
AC C
EP
Additional Optional Comments Common themes were that multiple imaging modalities and physical examination play important roles in surgical decision-making. In response to the final question, “Do you have any suggestions for other ways in which you use MRI to decide on anterior versus posterior surgery for Degenerative Cervical Myelopathy?”, the most common point was that multiple imaging modalities should be considered (16.1%, 27/168) and that addition of dynamic MRI to the decision-making process may be useful (14.3%, 24/168). Other frequently mentioned comments and specific suggestions are listed in Table 4. For most questions, participants commonly also commented that sagittal alignment would impact their decision.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
AC C
EP
TE D
M AN U
SC
RI PT
DISCUSSION This is the first global survey to demonstrate that specific MRI pathology play a role in influencing surgeons’ choice for an anterior or posterior surgical approach, and that region and specialty training can affect these choices. Furthermore, this survey provides the first indication to what degree (i.e. strong, moderate, weak, no influence) specific factors influence a surgeon towards anterior or posterior surgery. These findings will play an important role for ongoing and future studies that are comparing the efficacy of anterior vs. posterior surgical procedures. There were 8 MRI factors favoring an anterior approach whereas 7 favored a posterior approach. The most impactful pathology to influence surgeons towards anterior was a high degree of anterior compression and presence of cervical kyphosis. This was expected as compressive disc pathology can be directly removed from the anterior and cord compression may not be sufficiently relieved by posterior approaches. In terms of cervical kyphosis, anterior surgery may be preferred because it allows for a certain degree of kyphosis correction by applying distraction while reconstructing the vertebral column, particularly in cases of fixed kyphosis.10 Further, there may be limited posterior migration of the cord if treated posteriorly without appropriate alignment correction. Also, there is a concern for progression of kyphosis after posterior approaches such as laminoplasty or laminectomy without instrumented fusion.11,12 Moderate-weak influences for anterior approaches were seen with focal OPLL, retro-vertebral disease, and spondylolisthesis in ascending order. It was notable that focal OPLL was the only MRI factor that showed marked divergence between regions, with the median North American response favoring a posterior approach while all other regions favored an anterior approach (p=0.008). Foraminal stenosis demonstrated a weak influence toward anterior; however, this was based on a mode of moderate influence anterior and a mean of no influence. This divergence may be partially due to the fact that foraminal stenosis can be easily treated from either anterior or posterior approaches. Additionally, the choice of approach for foraminal stenosis represented the only MRI factor that was influenced by a surgeons’ comfort in performing multilevel anterior procedures [OR=0.51, p=0.002], with surgeons that were not comfortable with a multilevel anterior approach more commonly indicating a preference for a posterior approach. The most impactful pathologies to influence a posterior approach were a high degree of posterior cord compression, multilevel compression or OPLL, ligamentum flavum enlargement, and congenital stenosis. Generally, it seems reasonable that a posterior approach was favored in these cases to allow for expansion of the spinal canal and posterior shift of the spinal cord. In addition to this, the comments provided by respondents suggest that multilevel pathology, frequently demarcated by respondents as >3 levels, is favored due to longer operative time, complexity, and concerns with fusion (based on clinical factors and bone quality) with multilevel anterior approaches. Although these concerns differ based on surgical experience, there are clearly differences in surgical preference as only 60% of surgeons felt equally comfortable performing ≥3 levels anteriorly or posteriorly, with most preferring posterior approaches with ≥3 levels. The influence of multilevel OPLL on choosing a posterior surgical approach is not surprising, but is
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
AC C
EP
TE D
M AN U
SC
RI PT
worthy of note, because treatment is aimed at cord decompression through canal expansion and posterior shift of the cord rather than the pathology itself. As a consequence, posterior approaches are dependent on the thickness of the OPLL with more thickness limiting how much further the spinal cord can shift. Indeed, it has been suggested that corpectomy may be preferable in patients with an occupying ratio ≥60%.13 As for ligamentum flavum enlargement, it is intuitive that a posterior approach is favored, as it can directly address the site of cord compression. The least influential factor on MRI was the presence of signal intensity changes in the spinal cord. This was not surprising given that the spinal cord is not involved from a surgical perspective. Interestingly though, a number of respondents commented that advanced MRI techniques such as functional, diffusion tensor and perfusion, which all relate to spinal cord pathology were potentially useful. Two other factors that were not considered to be influential were cervical scoliosis and a previous anterior surgery. However, the “no influence” response only comprised about 38% for each of these (compared to 61% for signal change), suggesting that there was great variability with about a third choosing anterior and posterior approaches for each question. Thus, respondents either answered the question at random or surgical decision-making in these cases may exist but may be surgeon dependent. Geographical region had a significant impact on the responses, with 15 out of 18 questions having statistically significant differences between regions. However, in most cases (14/15) the choice for surgical approach remained the same and the difference was based on the degree of influence, or there was a weak preference compared to no influence. However, focal OPLL influenced North American towards the posterior, while all other regions were influenced toward anterior (p=0.008). The reason for this is unclear, but this does suggest that North Americans prefer to treat all types of OPLL from the posterior, while other regions treat focal and multilevel OPLL differently. Significant variability between responses was also exhibited between neurosurgeons and orthopedic surgeons for 7 of the 18 MRI factors. Neurosurgeons were generally more influenced toward anterior and less influenced towards posterior approaches for when compared to orthopedic surgeons. While differences in training or surgical preference may be contributory. Our findings are in accordance with a previous survey-based study that found neurosurgeons tended to favor anterior approaches, 14 although statistical differences between surgical approaches were not shown due to the limited number of cases in that study. The most common survey respondent comments made were that multiple imaging modalities should be used in the assessment, and that dynamic (flexion/extension) MRI is useful in the diagnostic work-up. Indeed, in terms of dynamic MRI, it has been reported that dynamic stenosis can be uncovered in 8.3% of patients in extension and 1.6% in flexion. 15 Many respondents also suggested the use of a modified K-line to assess sagittal alignment on MRI. While alignment measures are not typically recommended for MRI, it has been reported that the modified K-line can predict surgical outcome.16 Another set of respondents also mentioned that they use MRI to assess the location and course of the vertebral artery as part of their surgical-decision making. Indeed, it has been reported that
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
M AN U
SC
RI PT
vertebral arteries can migrate medially into the vertebral body, increasing the risk for injury during anterior approaches.17 Conversely, variations in vertebral artery anatomy such as a ‘high riding vertebral’ artery or ‘ponticulus posticus’ may elevate the risk of posterior approaches when instrumented fusion is extended to C2.18 Currently, there is neither a standardized approach for surgical treatment nor are there guidelines to direct this decision-making process. Reasons for this include the need for flexibility in treatment given the heterogeneity of DCM presentation, familiarity with procedures, regional differences in surgical preference (i.e. laminoplasty in Asia), costs for instrumentation, and ambiguity regarding potential differences in effectiveness between procedures. In addition, clinical factors such as age as well as other physical characteristics may impact the decision-making process. While there is substantial consensus on how to treat certain types of presentations such as myelopathy due to single disc pathology, ambiguity remains with regard to how to manage conditions for which both approaches are perceived to be equally efficacious. Convincing evidence derived from prospective and multicenter research, has been shown that when surgical decision is left at the discretion of the attending surgeon, anterior and posterior surgery provide comparable efficacy.4 This is further corroborated by systematic reviews concluding that anterior and posterior procedures result in similar outcomes with regards to effectiveness and safety.3,6 However, it remains to be seen whether there are differences in surgical outcome between these surgical approaches in patients for whom equipoise exists for randomization toward an anterior or posterior procedure.
AC C
EP
TE D
Limitations The cross-sectional survey of the international AOSpine community is the largest assessment of the factors influencing surgeon preference for a particular approach to surgically manage degenerative cervical myelopathy. While the response rate for our international survey is somewhat lower than the average response rates for other web based surveys,19 it is comparable to other surveys of the international AOSpine community that have been recently published.20-22 Our response rate is likely the result of a number of factors including: (1) not all AOSpine members may have received the email (filtered out as spam or wrong email address), (2) lack of English language proficiency, and (3) non-surgeon members. The most important limitation to this survey was that each of the factors questioned was asked to be considered in isolation. The impact of concomitant pathologies may alter one’s perspective on the most effective approach. Moreover, nearly two-thirds of individuals indicated that they were not comfortable using MRI alone to determine surgical approach and many commented that their decision would be swayed by sagittal alignment. Upright lateral radiographs would therefore be a valuable addition to MRI and can provide Cobb angle, T1 slope and sagittal vertical axis (SVA) measurements. Furthermore, all anterior and posterior approaches are not the same and the choice for a specific procedure was not considered. It should also be noted that results might have been impacted by the inclusion of residents and fellows, who are likely still in the process of learning to optimize their decisionmaking. Lastly, reliance on MRI to make surgical decisions assumes that this
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
RI PT
modality is readily available, however, the global nature of this survey and recognition that some regions may not have this accessibility, may impact responses from these regions.
CONCLUSION
M AN U
SC
A large degree of anterior cord compression and cervical kyphosis strongly influence surgeons toward an anterior approach. Conversely, a high degree of posterior cord compression, congenital stenosis, and multilevel compression or OPLL were the strongest factors influencing surgeons toward a posterior surgery. It is also evident that the surgeons’ preference is influenced by region and type of training. Additional research through investigation of case-based surgical decisionmaking is recommended to study this topic further.
FIGURE LEGEND
2.
3.
4.
Nouri A, Tetreault L, Singh A, Karadimas S, Fehlings M. Degenerative Cervical Myelopathy: Epidemiology, Genetics and Pathogenesis. Spine (Phila Pa 1976). 2015;40(12):E675-693. Lawrence BD, Shamji MF, Traynelis VC, et al. Surgical management of degenerative cervical myelopathy: a consensus statement. Spine (Phila Pa 1976). 2013;38(22 Suppl 1):S171-172. Luo J, Cao K, Huang S, et al. Comparison of anterior approach versus posterior approach for the treatment of multilevel cervical spondylotic myelopathy. Eur Spine J. 2015;24(8):1621-1630. Fehlings MG, Barry S, Kopjar B, et al. Anterior versus posterior surgical approaches to treat cervical spondylotic myelopathy: outcomes of the prospective multicenter AOSpine North America CSM study in 264 patients. Spine (Phila Pa 1976). 2013;38(26):2247-2252.
AC C
1.
EP
REFERENCES
TE D
Figure 1. Two patients with preoperative and postoperative sagittal T2 weighted MRIs and clinically confirmed DCM are presented. A) Patient 1 presents with spinal cord compression by the C5-6 disc. B) A 6-months post-operative MRI of patient 1 showing that the patient was treated with an anterior discectomy and plate at C5-6. C) Patient 2 presents with multilevel spinal cord compression with the origin of stenosis coming from both the anterior and posterior in the upper and middle cervical spine. D) A 24-months post-operative MRI of patient 2 showing that the patient was treated with multilevel posterior laminectomy and fusion from C2-T3.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making 5.
RI PT
6.
Ren H, Liu F, Yu D, et al. Patterns of Neurological Recovery After Anterior Decompression With Fusion and Posterior Decompression With Laminoplasty for the Treatment of Multilevel Cervical Spondylotic Myelopathy. Clin Spine Surg. 2016. Lawrence BD, Jacobs WB, Norvell DC, Hermsmeyer JT, Chapman JR, Brodke DS. Anterior versus posterior approach for treatment of cervical spondylotic myelopathy: a systematic review. Spine (Phila Pa 1976). 2013;38(22 Suppl 1):S173-182. Ghogawala Z, Martin B, Benzel EC, et al. Comparative effectiveness of ventral vs dorsal surgery for cervical spondylotic myelopathy. Neurosurgery. 2011;68(3):622-630; discussion 630-621. Seng C, Tow BP, Siddiqui MA, et al. Surgically treated cervical myelopathy: a functional outcome comparison study between multilevel anterior cervical decompression fusion with instrumentation and posterior laminoplasty. Spine J. 2013;13(7):723-731. Ghogawala Z, Benzel EC, Heary RF, et al. Cervical spondylotic myelopathy surgical trial: randomized, controlled trial design and rationale. Neurosurgery. 2014;75(4):334-346. Emery SE. Anterior approaches for cervical spondylotic myelopathy: which? When? How? Eur Spine J. 2015;24 Suppl 2:150-159. Kaptain GJ, Simmons NE, Replogle RE, Pobereskin L. Incidence and outcome of kyphotic deformity following laminectomy for cervical spondylotic myelopathy. J Neurosurg. 2000;93(2 Suppl):199-204. Suk KS, Kim KT, Lee JH, Lee SH, Lim YJ, Kim JS. Sagittal alignment of the cervical spine after the laminoplasty. Spine (Phila Pa 1976). 2007;32(23):E656-660. Chen Z, Liu B, Dong J, et al. Comparison of anterior corpectomy and fusion versus laminoplasty for the treatment of cervical ossification of posterior longitudinal ligament: a meta-analysis. Neurosurg Focus. 2016;40(6):E8. Irwin ZN, Hilibrand A, Gustavel M, et al. Variation in surgical decision making for degenerative spinal disorders. Part II: cervical spine. Spine (Phila Pa 1976). 2005;30(19):2214-2219. Hayashi T, Wang JC, Suzuki A, et al. Risk factors for missed dynamic canal stenosis in the cervical spine. Spine (Phila Pa 1976). 2014;39(10):812-819. Taniyama T, Hirai T, Yoshii T, et al. Modified K-line in magnetic resonance imaging predicts clinical outcome in patients with nonlordotic alignment after laminoplasty for cervical spondylotic myelopathy. Spine (Phila Pa 1976). 2014;39(21):E1261-1268. Eskander MS, Drew JM, Aubin ME, et al. Vertebral artery anatomy: a review of two hundred fifty magnetic resonance imaging scans. Spine (Phila Pa 1976). 2010;35(23):2035-2040. Elgafy H, Pompo F, Vela R, Elsamaloty HM. Ipsilateral arcuate foramen and high-riding vertebral artery: implication on C1-C2 instrumentation. Spine J. 2014;14(7):1351-1355. Shih T-H, Fan X. Comparing response rates from web and mail surveys: A meta-analysis. Field methods. 2008;20(3):249-271.
7.
11.
12.
13.
14.
AC C
15.
TE D
10.
EP
9.
M AN U
SC
8.
16.
17.
18.
19.
ACCEPTED MANUSCRIPT Survey on MRI features and DCM Surgical-Decision Making
21.
AC C
EP
TE D
M AN U
SC
22.
Tetreault L, Nouri A, Singh A, Fawcett M, Nater A, Fehlings MG. An Assessment of the Key Predictors of Perioperative Complications in Patients with Cervical Spondylotic Myelopathy Undergoing Surgical Treatment: Results from a Survey of 916 AOSpine International Members. World Neurosurg. 2015;83(5):679-690. Schroeder GD, Kurd MF, Kepler CK, et al. The Development of a Universally Accepted Sacral Fracture Classification: A Survey of AOSpine and AOTrauma Members. Global Spine J. 2016;6(7):686-694. Nater A, Tetreault LL, Davis AM, Sahgal AA, Kulkarni AV, Fehlings MG. Key Preoperative Clinical Factors Predicting Outcome in Surgically Treated Patients with Metastatic Epidural Spinal Cord Compression: Results from a Survey of 438 AOSpine International Members. World Neurosurg. 2016;93:436-448 e415.
RI PT
20.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Survey on MRI features and DCM Surgical-Decision Making
ACCEPTED MANUSCRIPT
Table 1. Training Level of Respondents
AC C
EP
TE D
RI PT
M AN U
Resident Fellow, Spine Surgery Fellow, Other discipline Neurosurgeon Neurosurgeon, fellowship trained in spine Orthopedic Surgeon Orthopedic Surgeon, fellowship trained in spine Other Total
Frequency (number of respondents) 2.9% (n=15) 6.2% (n=32) 0.2% (n=1) 19.1% (n=98) 17.7% (n=91) 10.1% (n=52) 41.5% (n=213) 2.1% (n=11) 513
SC
Type of Training
ACCEPTED MANUSCRIPT
Table 2. Results from the Likert Scale questions. Modes are bolded. Overall influence was classified based on both the mode and mean.
Anterior Approach Strong (+3) Moderate Weak (+2) (+1)
Neither (0)
19.7% (n=101)
3.9% (n=20)
40.6% (n=207)
33.7% (n=172)
7.1% (n=36)
5.1% (n=26)
30.3% (n=154)
31.6% (n=161)
9.0% (n=46)
Overall Influence
0% (n=0)
2.65 (n=512)
Strong Anterior
2.9% (n=15)
7.5% (n=38)
3.1% (n=16)
1.69 (n=510)
StrongModerate Anterior
17.7% (n=90)
3.3% (n=17)
5.9% (n=30)
2.2% (n=11)
1.41 (n=509)
Moderate Anterior
11.6% (n=59)
15.3% (n=78)
5.1% (n=26)
10.2% (n=52)
4.1% (n=21)
1.03 (n=509)
Moderateweak Anterior
TE D
32.0% (n=163)
Mean (m)
0.4% (n=2)
EP
21.6% (n=110)
1.8% (n=9)
0% (n=0)
M AN U
74.1% (n=380)
AC C
1. High degree of anterior compression (e.g. large bulging disc). 9. Presence of cervical kyphosis on MRI (Interpreted as extent of kyphosis in a neutral to extended neck position) 4. Greater number of bulging discs that are not causing cord compression (i.e. cord is only compressed at 1 level but discs show degeneration at other levels). 18. Presence of Spondylolisthesis (Defined as ≥3mm displacement) 5. MRI suggestive of focal ossification of the posterior longitudinal ligament (affecting one segment) 8. Presence of cord compression due to retrovertrebral disease (midvertebral level). 14. Presence of foraminal stenosis 13. Presence of Modic changes in the vertebral bodies
Posterior Approach Moderate Strong (-2) (-3)
SC
Anterior approach favored
Weak (-1)
RI PT
Question - Please respond as to how the following factors (in isolation), as seen on MRI, influence your surgical decision toward anterior or posterior surgery?
21.9% (n=111)
27.0% (n=137)
7.3% (n=37)
7.3% (n=37)
8.3% (n=42)
18.5% (n=94)
9.7% (n=49)
0.53 (n=507)
Moderateweak Anterior
23.2% (n=117)
25.8% (n=130)
8.5% (n=43)
11.5% (n=58)
5.6% (n=28)
17.5% (n=88)
7.9% (n=40)
0.65 (n=504)
Moderateweak Anterior
11.6% (n=59)
24.9% (n=126) 15.7% (n=80)
15.4% (n=78) 14.6% (n=74)
22.3% (n=113) 52.4% (n=266)
6.3% (n=32) 1.6% (n=8)
13.2% (n=67) 3.0% (n=15)
6.3% (n=32) 0.6% (n=3)
0.48 (n=507) 0.73 (n=508)
Moderate-No Influence Weak Anterior-No Influence
12.2% (n=62)
ACCEPTED MANUSCRIPT
Posterior approach favored 5.7% (n=29)
1.2% (n=6)
1.4% (n=7)
7.7% (n=39)
29.1% (n=148)
50.7% (n=258)
-1.92 (n=509)
Strong to Moderate Posterior
2.0% (n=10)
3.7% (n=19)
0.6% (n=3)
11.8% (n=60)
9.8% (n=50)
38.6% (n=196)
33.5% (n=170)
-1.73 (n=508)
Moderate Posterior
5.7% (n=29)
7.3% (n=37)
4.3% (n=22)
3.1% (n=16)
28.9% (n=147)
44.7% (n=227)
-1.62 (n=508)
Moderate Posterior
2.0% (n=10)
2.6% (n=13) 8.1% (n=41) 8.7% (n=44)
1.8% (n=9) 1.8% (n=9) 7.1% (n=36)
9.6% (n=49) 4.9% (n=25) 19.8% (n=100)
25.1% (n=128) 8.4% (n=43) 19.8% (n=100)
39.3% (n=200) 41.8% (n=213) 30.1% (n=152)
19.6% (n=100) 29.1% (n=148) 11.3% (n=57)
-1.50 (n=509) -1.44 (n=509) -0.80 (n=505)
Moderate Posterior Moderate Posterior Moderateweak Posterior
8.3% (n=42)
5.6% (n=28)
32.1% (n=162)
12.9% (n=65)
25.0% (n=126)
12.9% (n=65)
-0.70 (n=504)
Weak Posterior-No Influence
3.2% (n=16)
SC 5.9% (n=30)
M AN U
TE D
5.9% (n=30) 3.2% (n=16)
RI PT
4.3% (n=22)
EP
2. High degree of posterior cord compression (e.g. ligamentum flavum inbuckling). 17. Presence of congenital canal stenosis (e.g. cord occupies more than 70% of the canal at C2 or C7 in the absence of disc/ligament pathology at these levels) 6. MRI suggestive of multilevel ossification of the posterior longitudinal ligament (>1 segment) 12. Presence of ligamentum flavum enlargement 3. Greater number of vertebral levels with cord compression. 7. Presence of a short neck length and/or chin obstructing the most rostral (upper) disc level requiring surgery. 15. Presence of autofusion (or Klippel-Feil)
10. Presence of cervical scoliosis on MRI 11. Presence of cord signal intensity changes (T2weighted hyper-intensity, T1weighted hypo-intensity) 16. Previous anterior surgery (e.g. single-level ACDF)
AC C
No Influence
4.9% (n=25) 9.1% (n=46)
5.7% (n=29)
15.0% (n=76) 11.5% (n=58)
4.9% (n=25) 2.8% (n=14)
38.1% (n=193) 61.1% (n=309)
10.3% (n=52) 4.0% (n=20)
18.3% (n=93) 6.7% (n=34)
8.5% (n=43) 4.9% (n=25)
-0.23 (n=507) 0.21 (n=506)
No Influence
16.2% (n=83)
6.1% (n=31)
38.7% (n=198)
13.9% (n=71)
13.5% (n=69)
5.9% (n=30)
-0.03 (n=511)
No Influence
No Influence
ACCEPTED MANUSCRIPT
Table 3. Assessment of Variation by Region and Training. Detailed cross-comparisons are available in APPENDIX B and C.
6. MRI suggestive of multilevel ossification of the posterior longitudinal ligament (>1 segment)
7. Presence of a short neck length and/or chin obstructing the most rostral (upper) disc level requiring surgery. 8. Presence of cord compression due to retrovertrebral disease (mid-vertebral level).
Training, Neurosurgeon (NS) vs. Orthopedic Surgeon (OS) P=0.076 NS – Mean Rank = 217.23 (n=189); OS – Mean Rank = 233.99 (n=264) P=0.162 NS – Mean Rank = 216.63 (n=187); OS – Mean Rank = 232.64 (n=264)
RI PT
Group – P=0.008 NA vs. SA p=0.018; NA vs. EU p=0.01; NA vs. Asia p=0.027 North America was influenced toward a posterior approach whereas all other groups were influenced toward an anterior approach Group – P<0.001 NA vs. SA p=0.015; NA vs. EU p<0.001; NA vs. Asia p=0.035 North America was most strongly influenced toward a posterior approach Group - P=0.265
AC C
5. MRI suggestive of focal ossification of the posterior longitudinal ligament (affecting one segment)
TE D
4. Greater number of bulging discs that are not causing cord compression (i.e. cord is only compressed at 1 level but discs show degeneration at other levels).
EP
3. Greater number of vertebral levels with cord compression.
Group – P=0.001 EU vs. Asia p=0.001; SA vs. Asia p=0.016 Asia was less influenced toward an anterior approach. Group – P<0.001 NA vs. Asia p=0.005; SA vs. Asia p=0.021; EU vs. Asia p=0.001 Asia was more strongly influenced toward a posterior approach Group – P<0.001 EU vs. SA p=0.001; EU vs. Asia p=0.001 Europe was less strongly influenced toward a posterior approach Group – P=0.008 SA vs. NA p=0.017; SA vs. Asia p=0.035 South America was more strongly influenced toward an anterior approach
M AN U
2. High degree of posterior cord compression (e.g. ligamentum flavum inbuckling).
Region
SC
Please respond as to how the following factors (in isolation), as seen on MRI, influence your surgical decision toward anterior or posterior surgery? 1. High degree of anterior compression (e.g. large bulging disc).
Group - P=0.106
P=0.007 NS – Mean Rank = 207.45 (n=188); OS – Mean Rank = 239.26 (n=263) P=0.099 NS – Mean Rank = 213.93 (n=187); OS – Mean Rank = 233.73 (n=263) P=0.052 NS – Mean Rank = 211.17 (n=186); OS – Mean Rank = 234.78 (n=263)
P=0.003 NS – Mean Rank = 205.29 (n=186) OS – Mean Rank = 239.87 (n=263) P=0.001 NS – Mean Rank = 200.94 (n=187) OS – Mean Rank = 239.79 (n=259) P=0.843 NS – Mean Rank = 224.91 (n=185)
ACCEPTED MANUSCRIPT
SC
AC C
13. Presence of Modic changes in the vertebral bodies
Group – P<0.001 Asia vs. NA p<0.001; Asia vs. SA p=0.02; Asia vs. EU p<0.001 Asia was most strongly influenced toward posterior approach, while the other regions displayed less influence toward posterior. Group – P<0.001 NA vs. SA p=0.043; NA vs. EU p=0.013; Asia vs. SA p=0.021; Asia vs. EU p=0.004 North America and Asia were generally not influenced, but South America and Europe were slight influenced toward anterior. Group – P=0.001 EU vs. Asia p<0.001 Europe had a slight influence toward anterior, while Asia was most commonly not influenced (but had high variability) Group - P=0.539
EP
12. Presence of ligamentum flavum enlargement
TE D
11. Presence of cord signal intensity changes (T2weighted hyper-intensity, T1-weighted hypointensity)
M AN U
10. Presence of cervical scoliosis on MRI
14. Presence of foraminal stenosis
Group – P<0.001 SA vs. EU p=0.024; SA vs. Asia p<0.001; NA vs. Asia p=0.007 South America was most strongly influenced toward anterior. Asia was less strongly influenced toward anterior. Group – P=0.013 EU vs. Asia p=0.012 All regions had a median of no influence, but EU had more respondents favouring anterior and Asia had more favouring posterior. Group – P=0.017 EU vs. Asia p=0.03 Asia showed no influence, but EU was slightly influenced toward anterior
RI PT
9. Presence of cervical kyphosis on MRI (Interpreted as extent of kyphosis in a neutral to extended neck position)
OS – Mean Rank = 222.50 (n=261) P=0.603 NS – Mean Rank = 211.17 (n=186) OS – Mean Rank = 234.78 (n=263)
15. Presence of autofusion (or Klippel-Feil)
P=0.036 NS – Mean Rank = 210.28 (n=186) OS – Mean Rank = 235.41 (n=263) P=0.023 NS – Mean Rank = 210.25 (n=188) OS – Mean Rank = 234.80 (n=260) P=0.074 NS – Mean Rank = 213.07 (n=187) OS – Mean Rank = 234.33 (n=263)
P=0.372 NS – Mean Rank = 219.54 (n=187) OS – Mean Rank = 229.74 (n=263)
P=0.612 NS – Mean Rank = 221.39 (n=187) OS – Mean Rank = 227.57 (n=262) P=0.054 NS – Mean Rank = 237.09 (n=185) OS – Mean Rank = 213.87 (n=261)
ACCEPTED MANUSCRIPT
16. Previous anterior surgery (e.g. single-level ACDF)
AC C
EP
TE D
M AN U
SC
18. Presence of Spondylolisthesis (Defined as ≥3mm displacement)
P=0.007 NS – Mean Rank = 207.61 (n=189) OS – Mean Rank = 240.07 (n=263) P=0.090 NS – Mean Rank = 213.83 (n=187) OS – Mean Rank = 233.80 (n=263) P<0.001 NS – Mean Rank = 196.88 (n=189) OS – Mean Rank = 247.01 (n=262)
RI PT
17. Presence of congenital canal stenosis (e.g. cord occupies more than 70% of the canal at C2 or C7)
Group – P=0.014 EU vs. Asia p=0.028 Europe had a slight influence toward anterior, while Asia was most commonly not influenced (but had high variability) Group – P=0.014 EU vs. Asia p=0.020 Europe was less strongly influenced toward posterior than Asia which was most influenced toward posterior Group – P<0.001 SA vs. NA p<0.001; SA vs. Asia p<0.001; EU vs. Asia p<0.001 South America was most influenced toward anterior, while Asia was least influenced toward anterior
ACCEPTED MANUSCRIPT
Table 4. Selected additional comments by respondents in response to “Do you have any suggestions for other ways in which you use MRI to decide on anterior versus posterior surgery for Degenerative Cervical Myelopathy?”. Selected comments provided for specific questions are also included. Comments have been modified for grammar and context.
RI PT
Q - Do you have any suggestions for other ways in which you use MRI to decide on anterior versus posterior surgery for Degenerative Cervical Myelopathy? Dynamic (flexion/extension) MRI should be considered in the work up of patients to assess movement dependent cord compression. Standing and/or flexion radiographs, and/or CT with or without contrast should complement MRI in decision-making.
Collated from multiple responses
SC
Collated from multiple responses
MRI is useful for assessing the course and location of the vertebral artery. MRI measurement of the modified K-line is useful for assessment of sagittal alignment. Advanced MRI techniques (functional, diffusion tensor, perfusion) can be used in surgical decision-making. Clinical risk factors for non-union such as smoking and diabetes may sway me to consider a posterior approach. For patients likely needing a combined approach, I favour an anterior approach initially with followed up. If there is progression or lack of improvement I then add a posterior approach.
M AN U
Collated from multiple responses Collated from multiple responses Collated from multiple responses Specific Comment
Specific Comment
TE D
Specific Comment
Alignment assessment for MRI should be standardized as some are taken with pillows, making alignment assessment difficult.
Circumferential compression on MRI should be approached posteriorly. Specific Comment Measurement of disc height at the site of pathology in comparison to disc height at other sites. Specific Comment Level of and area of T2 signal change could be assessed. Q. High degree of anterior compression (e.g. large bulging disc) Collated from multiple If the compression is very large, a posterior approach done first responses and followed by an anterior approach is sometimes done. Q. Greater number of vertebral levels with cord compression Collated from multiple Most respondents commented that >3 levels should be responses approached posteriorly. Q. Greater number of bulging discs that are not causing cord compression (i.e. cord is only compressed at 1 level but discs show degeneration at other levels) Collated from multiple Treat only the level of pathology; consider arthroplasty to responses preserve motion. Q. Presence of cervical kyphosis on MRI (Interpreted as extent of kyphosis in a neutral to extended neck position) Collated from multiple A combined approach is favoured.
AC C
EP
Specific Comment
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
responses Q. Presence of cervical scoliosis on MRI Collated from multiple A combined approach is favoured. responses Q. Previous anterior surgery (e.g. single-level ACDF) Specific Comment 1) Depends on the “age” of the previous surgery 2) Depends on if compression is at previous surgical site or other segment. Q. Presence of Spondylolisthesis (Defined as ≥3mm displacement) Collated from multiple A combined approach is favoured. responses Q. Do you feel equally comfortable doing multilevel (3 or more levels) anterior and posterior cervical decompressive/reconstructive procedures? Collated from multiple 1) Posterior multilevel procedures are less difficult and favoured responses for >3 levels. 2) Concern for prolonged operating time for multilevel anterior procedures and potential swelling
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
(3 - 5 bullet points, 85 characters incl spaces or less)
Survey Bullet Points
• •
AC C
EP
TE D
M AN U
•
RI PT
•
Specific pathology influence the choice for an anterior/posterior surgical approach Neurosurgeons chose anterior more and posterior less vs. orthopedic surgeons 59.8% were equally comfortable doing multilevel (≥3) anterior & posterior surgery Degree of influence between differed between region & training for 15 & 6 factors Further investigation through case-based surgical decision-making is recommended
SC
•
ACCEPTED MANUSCRIPT
ABBREVIATIONS
Degenerative Cervical Myelopathy
OPLL =
Ossification of the Posterior Longitudinal Ligament
MRI
=
Magnetic Resonance Imaging
CT
=
Computed Tomography
AC C
EP
TE D
M AN U
SC
RI PT
DCM =
ACCEPTED MANUSCRIPT
Conflict of Interest and Disclosures
AC C
EP
TE D
M AN U
SC
RI PT
The authors have no relevant conflicts of interest. Dr. Fehlings wishes to disclose consulting agreements with Pfizer, Zimmer Biomet and InVivo Therapeutics