A Gaze Beyond the Surface: Acknowledging the Little We Know About Radiographic Parameters for Evaluation of Lumbar Spinal Stenosis

Mattei, Tobias A.

doi: 10.1227/NEU.0b013e3182752bb7

Peoria, Illinois

Article Outline

To the Editor:

We read with great interest the article by Wilkinson et al published at the July issue of Neurosurgery entitled “Failure of Percutaneous Remodeling of the Ligamentum Flavum and Lamina for Neurogenic Claudication”1 and the hot discussion that followed the presentation of the study's results.2,3

The initial reaction of the majority of the readers after analyzing such a study is to focus on the effectiveness (or lack of) and the possible role of the new proposed technique (minimally invasive lumbar decompression, so-called MILD technique) in the treatment of lumbar spinal stenosis, and after that, quickly position themselves in 1 of the 2 fronts and passionately defend their own opinion based on their previous clinical experience and personal beliefs. Nevertheless, we truly believe that the great contribution of such a small series about a still primitive and underdeveloped technique is not what this series shows “to” us, but what it shows “about” us.

Without taking any judgmental attitude regarding the possible (or more likely impossible) long-term benefits of the aforementioned technique, the clear discrepancy in the results found by such a study should flash a bright red light and alert us that something is wrong: not with the data obtained, not with the methods of the study, maybe neither with the new proposed technique, but probably with what we believe we know (and most likely do not) about the pathophysiology, imaging, and clinical outcomes of lumbar spinal canal stenosis.

The first remarkable finding of this prospective study, which aimed to provide an objective and impartial analysis of the results of this newly proposed minimally invasive technique, was that, although a satisfactory intraoperative decompression was achieved (as demonstrated by intraoperative fluoroscopy) in all 10 subjects, no significant difference in the cross-sectional area of the operated lumbar segments was observed in the 12-week postoperative magnetic resonance imaging (MRI). Apart from the discussion regarding whether a full-time continuous exposure to radiation during a blind lumbar decompression of the spinal canal is really less invasive than a small incision that proportions safe visualization of the important neural elements involved in such operation, the fact that the intraoperative fluoroscopy demonstrated such a poor correlation with postoperative MRI is at least interesting (not to say very uncomfortable to our so aimed straight-square, mathematically exact, fully controlled, and totally predictable neurosurgical world). In other words, if the fluoroscopy images, which have shown such clear improvement in the contrasted area of the epidural space during the surgical procedure, present no correlation with the expected improvement in the cross-sectional area of the lumbar canal in the postoperative MRI, how reliable (or how comparable to MRI) would be, for example, computed tomography (CT) myelography, which has been used as the decisive “gold-standard” diagnostic method for preoperative evaluation of patients with neurogenic claudication for decades? In other words, does CT myelography underestimate lumbar canal stenosis, does MRI overestimate it, or is this just a symptom that, even today, we still do not know the exact relation between the degree of lumbar canal stenosis, patients' symptoms, and postoperative outcomes?

A superficial appraisal of the issue might induce one to wonder that the true question is what is the exact relation between the degree of stenosis of the lumbar canal as evaluated by MRI and CT myelogram. Regarding such a comparison, it can be safely stated that:

- It has already been demonstrated that both MRI and CT myelogram of the lumbar spine present an acceptable intra- and interobserver reproducibility (>0.9 in the majority of the studies) in all measurement parameters studied up to now, with the exception of the estimation of ligamentum flavum thickness.4,5

- It has also been shown that the measurements of the spinal subarachnoid space (ie, dural area, dural anteroposterior diameter, dural right-left diameter, and cerebrospinal fluid space - anterior and posterior) are slightly, but significantly larger when obtained from CT myelogram than from MRI scans. In contrast, MRI measurements of neural elements (cord or cauda equina area, cord anteroposterior diameter, and cord right-left diameter) are slightly but significantly larger than those obtained by CT myelogram.5

Based on such information, one can easily perceive why there is a tendency to grade the degree of stenosis as significantly more severe in MRI than in CT myelogram. Furthermore, it has already been shown that, as lumbar spinal stenosis becomes more severe, the discrepancies between CT myelogram and MRI measurements become larger.4

Nevertheless, a direct comparison between MRI and CT myelogram, although providing meaningful information about the relation between both methods, does not provide a definitive answer to the question that really matters in the daily clinical practice. In fact, after analyzing such results, it can be easily perceived that more important than knowing whether MRI overestimates spinal canal stenosis (or CT myelogram underestimates it) is to know what is the exact relation between the degree of compression (as shown by MRI or CT myelogram) and claudication symptoms presented by the patients.

A brief review of the literature clearly shows some interesting (not to say surprising) facts to all those who suppose that the relation between symptomatology, imaging, and surgical outcomes in lumbar canal stenosis is clear and straightforward. First, it has already been shown that morphological changes on MRI in patients with lumbar canal stenosis correlate with patients' measurements of pain, function, and quality of life only to a very limited extent. In fact, it has been demonstrated that there is a poor correlation between walking distance, Oswestry Disability Index (ODI),6 the 36-Item Short Form Health Survey, EQ-5D, and the Visual Analog Scale for leg and back pain levels, on the one hand, and the dural sac cross-sectional area as measured by MRI on the other.7 Furthermore, it has also been demonstrated that none of the parameters of either MRI or myelo-CT (such as anteroposterior diameters or dural sac cross-sectional area) present any correlation to the “time to the first symptoms” and the “total ambulation time” as measured during the exercise treadmill test, a functional diagnostic tool that has been proposed as an attempt to quantify neurogenic claudication.8

Furthermore, the very validity of the results obtained with currently used survey instruments has been recently questioned, because it was demonstrated that “Walking Capacity” (what is actually evaluated by questionnaires such as the ODI, the Pain Disability Index, and the 36-Item Short Form Health Survey, and by laboratory walk testings such as the exercise treadmill test) and “Walking Performance” (as measured by an activity monitor during walking hours for 7 days in the patient's community environment) appear to be completely different constructs with no significant relation between themselves.9

Finally, it has already been demonstrated that the most important predictors of functional capacity (as evaluated by the exercise treadmill test) in patients with lumbar canal stenosis are not any MRI measurements of the degree of lumbar canal stenosis, but the Body Mass Index and the subjective functional back capacity, a self-rated score in which the patient evaluates how the presented symptoms actually influence in his perceived walking performance.10

All these results support the conclusion that, up to now, the correlation between lumbar canal dimensions and clinical symptoms in patients with neurogenic claudication is very poor, so that MRI by itself is not capable to differentiate between individuals with clinically meaningful spinal canal stenosis and asymptomatic control subjects with incidental finding of spinal canal measurements below the average expected levels for the normal population at that age.11 To avoid a complete nihilistic attitude regarding what we know about the issue and ending up denying that lumbar constriction of the lumbar spinal canal may lead to pain and functional deterioration in such patients (and that MRI may play an important role in the diagnostic and clinical decision-making scenario) it can be safely stated that there seems to be, at least, a critical cutoff value in the measurements of the cross-sectional area of the lumbar spinal canal (which has been estimated to be approximately 70 mm2) below which it has already been shown that patients present significantly greater functional disability.12

Nevertheless, the deeper underlying issue that may present important implications for the daily clinical practice is not what is the relation between MRI and CT-myelogram measurements in patients with spinal canal stenosis or whether such measurements do (or do not) correlate with pain, disability, or functional preoperative scores. The most important question that each neurosurgeon should ask before indicating a surgical procedure for decompression of lumbar canal stenosis is how does the preoperative radiographic evaluation of the severity of lumbar canal stenosis correlate with outcomes after surgical decompression.

Although, for the 2 other aforementioned issues, several studies have already been performed, a thorough search in the current literature as an attempt to answer this last question yielded only 1 result: a weak study with a very limited number of patients (27) in which the authors were only able to suggest in their conclusions that there appears to be a relationship between severity of stenosis and outcomes of decompressive surgery such that patients with a greater than 50% reduction in cross-sectional area are more likely to have a successful outcome.13

Such ignorance about the real preoperative predictive value of the degree of stenosis of the lumbar canal (as measured by either MRI or CT myelogram) on the postoperative outcomes of such patients might very well explain the discrepancy observed in the results presented by Wilkinson et al, in which, although it was possible to observe a significant improvement in the postoperative disability scores (as measured by the ODI) between 1 and 26 weeks, none of the MRI examinations performed at the 12th postoperative week revealed any evidence of significant decompression of the lumbar canal. Such results surely reveal that the proposed minimally invasive technique did not effectively decompress the lumbar spinal canal, at least from the radiological standpoint of MRI, the current gold-standard examination for evaluating such pathology. Moreover, regarding the clinical results, it seems that the improvement after such a procedure is temporary and not sustained in the long-term follow-up. Nevertheless, the most shameful fact for all neurosurgeons (both to those who believe and those who do not believe in such a minimally invasive approach) is that, from the radiological standpoint of the MRI, with a basis in our current knowledge, we are not even capable to know which of those patients really needed any decompressive procedure.

The best scientific evidence for the surgical treatment of lumbar spinal stenosis at the present moment is undoubtedly the Spine Patient Outcomes Research Trial (SPORT), a prospective randomized trial with a concurrent observational cohort arm. In relation to the MRI parameters for inclusion in such study, all patients with a cross-sectional dural area less than 75 mm2 (the current accepted cutoff for defining severe central lumbar spinal stenosis), and who also met the other clinical criteria, were included in the study.14 It is interesting that a recently published subgroup analysis of the SPORT trial demonstrated that the only radiographic factor that predicted a significant treatment effect (defined as the Change in ODIoperative arm minus the change in ODInonoperative arm) for patients with lumbar canal stenosis was the presence of neuroforaminal stenosis. Other factors found to have a significant association with greater treatment effect improvement in the multivariate analysis were baseline ODI (less than 56) and not smoking – the 2 with the greatest magnitude of impact in the outcomes – predominant leg pain, not lifting at work, and baseline neurological deficit.15 No other radiographic parameter (such as the presence or absence of central canal stenosis, lateral recess stenosis, or the number of levels presenting severe canal stenosis) had a significant correlation with treatment effect.

As demonstrated by a recent literature review on the issue, the radiological definition for central lumbar spinal stenosis in previous published studies in the literature may be considered anything (most likely a chaotic, random, and subjective study-based decision) but a consensus (Figure).16 It may be even the case that the current method of classifying lumbar spinal canal stenosis with a basis on absolute parameters (such as the cross-sectional area of the spinal canal) needs to be urgently revised. Such criteria are actually based on primitive cadaveric experiments from the 1980s (in which a clamp was placed around the roots of the cauda equina, and a pressure-recording catheter indicated at what degree of stenosis there was a pressure increase among the roots)17,18 and gained further acceptance in the clinical scenario after a weak retrospective study that used CT scan as the gold-standard preoperative imaging method (nothing wrong for more than 2 decades ago, but certainly very inappropriate for the current state-of-the-art of radiological evaluation methods).19

As an interesting alternative, a recent study, for example, proposed a classification of lumbar spinal stenosis with a basis on the separation degree of the cauda equina on T2-weighted axial images.20 Such an idea of grading lumbar canal stenosis with basis on the neural tissue impingement has already been applied by previous studies21 and assumes that the relative ratio of the cauda equina and the remaining CSF space of the dural sac would be the most important factor in determining patients’ symptoms and, therefore, surgical outcomes. Although such new classifications still require validation in clinical series (as well as further studies to demonstrate its supposed superiority to the currently used systems), for now they at least serve to alert us that maybe our approach of radiographically grading lumbar canal stenosis with a basis on absolute numbers may be inadequate, and, therefore, one of the main causes behind our failure in demonstrating a significant relation between symptomatology, imaging measurements, and surgical outcomes.

The great gaps (and the most pitiful ones) in our basic knowledge in neurosurgery are not those related to new technological discoveries or to advanced frontline research techniques (such as some revolutionary discovery in Optogenetics with further neurosurgical implications),22 but those related to our most basic knowledge about prevalent diseases in the daily neurosurgical practice such as lumbar canal stenosis. And worst, although there is a great chance that new technological advances may receive significant attention owing to their novelty and promising future applications, the founding basic questions of our daily neurosurgical practice tend to be ignored and approached with a naïve, unscientific, and ultimately proud attitude by most neurosurgeons. After all, it is so logical (but yet unproven) and so clear (but certainly unknown, at least to which degree) that lumbar canal stenosis as demonstrated by MRI leads to pain and disability and deserves surgical decompression. In other words, we do not really know what we believe we already know (a classical epistemological mistake commonly referred to as “trust in unfounded beliefs”), most likely because we ask the wrong questions while gazing at the surface of the problem, as demonstrated by the main focus of previously published studies.

As the English iconoclastic author of the Victorian period Samuel Butler (1835-1902) once said: “A blind man knows he cannot see, and is glad to be led, though it be by a dog; but he that is blind in his understanding, which is the worst blindness of all, believes he sees as the best, and scorns a guide.”

In summary, we congratulate the authors for their serious, objective, and unbiased evaluation of such a new minimally invasive technique for the treatment of lumbar canal stenosis. It may be true that, for the majority of the neurosurgeons, the presented data will probably not change what they already believe (or most likely disbelieve) about such a new minimally invasive approach, nor result in any meaningful change in their current practice. Nevertheless, the interesting (not to say odd and incoherent) results of such a study serve at least to alert us about what we should know (but do not) about the patients we operate on and what we do not known (but could) if our clinical studies were more properly designed to answer the questions that really matter.

The great king Solomon (a man who, according to the Holy Scriptures surpassed in wisdom and knowledge all those who preceded him in the ancient Israel) once stated: “The glory of God is in concealing: but the honor of the kings in searching them out - Proverbs 25:2.” God has done his job very well! Unfortunately, we have not…

Back to Top | Article Outline


The author has no personal financial or institutional interest in any of the drugs, materials, or devices described in this article.

1. Wilkinson JS, Fourney DR. Failure of percutaneous remodeling of the ligamentum flavum and lamina for neurogenic claudication. Neurosurgery. 2012;71(1):86–92.
2. Chopko BW. Regarding failure of percutaneous remodeling of the ligamentum flavum and lamina for neurogenic claudication. Neurosurgery. 2012;71(2):E525–E526.
3. Fourney DR. In Reply. Neurosurgery. 2012;71(2):E525–E528.
4. Ogura H, Miyamoto K, Fukuta S, Naganawa T, Shimizu K. Comparison of magnetic resonance imaging and computed tomography-myelography for quantitative evaluation of lumbar intracanalar cross-section. Yonsei Med J. 2011;52(1):137–144.
5. Naganawa T, Miyamoto K, Ogura H, Suzuki N, Shimizu K. Comparison of magnetic resonance imaging and computed tomogram-myelography for evaluation of cross sections of cervical spinal morphology. Spine (Phila Pa 1976). 2011;36(1):50–56.
6. Sirvanci M, Bhatia M, Ganiyusufoglu KA, et al.. Degenerative lumbar spinal stenosis: correlation with Oswestry Disability Index and MR imaging. Eur Spine J. 2008;17(5):679–685.
7. Sigmundsson FG, Kang XP, Jönsson B, Strömqvist B. Correlation between disability and MRI findings in lumbar spinal stenosis: a prospective study of 109 patients operated on by decompression. Acta Orthop. 2011;82(2):204–210.
8. Moon ES, Kim HS, Park JO, et al.. Comparison of the predictive value of myelography, computed tomography and MRI on the treadmill test in lumbar spinal stenosis. Yonsei Med J. 2005;46(6):806–811.
9. Conway J, Tomkins CC, Haig AJ. Walking assessment in people with lumbar spinal stenosis: capacity, performance, and self-report measures. Spine J. 2011;11(9):816–823.
10. Zeifang F, Schiltenwolf M, Abel R, Moradi B. Gait analysis does not correlate with clinical and MR imaging parameters in patients with symptomatic lumbar spinal stenosis. BMC Musculoskelet Disord. 2008;9:89.
11. Geisser ME, Haig AJ, Tong HC, et al.. Spinal canal size and clinical symptoms among persons diagnosed with lumbar spinal stenosis. Clin J Pain. 2007;23(9):780–785.
12. Athiviraham A, Yen D, Scott C, Soboleski D. Clinical correlation of radiological spinal stenosis after standardization for vertebral body size. Clin Radiol. 2007;62(8):776–780.
13. Weiner BK, Patel NM, Walker MA. Outcomes of decompression for lumbar spinal canal stenosis based upon preoperative radiographic severity. J Orthop Surg Res. 2007;2:3.
14. Weinstein JN, Tosteson TD, Lurie JD, et al.. Surgical versus nonoperative treatment for lumbar spinal stenosis four-year results of the Spine Patient Outcomes Research Trial. Spine (Phila Pa 1976). 2010;35(14):1329–1338.
15. Pearson A, Lurie J, Tosteson T, Zhao W, Abdu W, Weinstein J. Who should have surgery for spinal stenosis? [published online ahead of print June 11, 2012] Spine (Phila Pa 1976). doi: 10.1097/BRS.0b013e3182634b04.
16. Steurer J, Roner S, Gnannt R, Hodler J. Quantitative radiologic criteria for the diagnosis of lumbar spinal stenosis: a systematic literature review. BMC Musculoskelet Disord. 2011;12:175.
17. Schönström N, Bolender NF, Spengler DM, Hansson TH. Pressure changes within the cauda equina following constriction of the dural sac. An in vitro experimental study. Spine (Phila Pa 1976). 1984;9(6):604–607.
18. Schönström N, Hansson T. Pressure changes following constriction of the cauda equina. An experimental study in situ. Spine (Phila Pa 1976). 1988;13(4):385–388.
19. Schumacher M. Die Belastungsmyelographie. Fortschr Rontgensrt. 1986;145:642–648.
20. Lee GY, Lee JW, Choi HS, Oh KJ, Kang HS. A new grading system of lumbar central canal stenosis on MRI: an easy and reliable method. Skeletal Radiol. 2011;40(8):1033–1039.
21. Schizas C, Theumann N, Burn A, et al.. Qualitative grading of severity of lumbar spinal stenosis based on the morphology of the dural sac on magnetic resonance images. Spine (Phila Pa 1976). 2010;35(21):1919–1924.
22. Mattei TA. “…MAB®”: can we dream with a future monoclonal antibody therapy for lumbar canal stenosis? World Neurosurg. 2012;78(3-4):197–199.
Copyright © by the Congress of Neurological Surgeons