Does magnetic resonance imaging morphology correlate with preoperative neurological status in operatively treated lumbar canal stenosis? A retrospective study
Article information
Abstract
Study Design
Retrospective observational study.
Purpose
To determine whether objective neurological deficits in lumbar canal stenosis (LCS) correlate with the commonly used radiological grading systems.
Overview of Literature
LCS is primarily a clinical diagnosis determined based on symptoms of neurological claudication and lower back pain, which is confirmed using radiological evidence. However, the current radiological measures of LCS do not often correlate with the clinical severity of the syndrome, particularly in terms of sensory, motor, and bladder/bowel dysfunction.
Methods
Preoperative clinical and imaging data of 100 consecutive cases of operated LCS were reviewed to determine symptom severity (using the Zurich Claudication Questionnaire [ZCQ]) and neurological dysfunction. Four radiological scores based on axial magnetic resonance imaging—Schizas’s, Lee’s, Miskin and Mandell’s (M&M), and Menon’s—were computed for each patient. The association between the severity of neurological dysfunction and radiological classification was assessed.
Results
All patients were graded as having moderate or severe grade LCS based on the ZCQ score. The Schizas score was significantly associated with motor (p=0.029) and sensory (p=0.034) deficits, and the M&M score was significantly associated with only motor deficits (p=0.012). The Lee and Menon’s classifications did not show significant associations with any neurological dysfunction (Lee: motor: p=0.258, sensory: p=0.615; Menon’s score: motor: p=0.1368, sensory: p=0.668).
Conclusions
Surgical decisions in LCS should not be based on imaging findings, but rather on the clinical severity of symptoms and presence of neurological dysfunction.
Introduction
Lumbar canal stenosis (LCS) is one of the most common clinical syndromes affecting the aging spine [1]. Described as a clinical syndrome caused by neural compression arising from developmental or acquired narrowing of the spinal canal, LCS manifests as neurogenic claudication with varying degrees of back pain [2]. While computed tomography (CT) scans proficiently demonstrate the configuration of the bony canal, magnetic resonance imaging (MRI) is considered the gold standard for diagnosing LCS owing to its superior anatomical resolution, visibility of the neural structures without the need for intrathecal contrast, delineation of soft tissue structures contributing to canal compromise, and the avoidance of radiation exposure [3]. However, it is often observed that the imaging-based diagnosis of spinal canal compression does not correlate linearly with the clinical severity of the syndrome; none of the quantitative parameters of spinal canal measurements have been found to have a bearing on the disease severity [4]. Notably, Boden et al. [5] reported that 21% of asymptomatic individuals over the age of 60 years have evidence of LCS on MRI scans.
Several classification/grading systems have been proposed to describe the severity (radiological and clinical) of LCS. Schizas et al. [6] devised a seven-stage grading system based on the ratio of the canal volume occupied by the rootlet to the cerebrospinal fluid (CSF) as a major determinant of severity. Lee et al. [7] proposed another four-grade schema based on the separation of the cauda equina and the CSF space in front of it. On the other hand, Miskin and Mandell’s classification claims to substantially reduce the inter- and intra-observer variability of LCS descriptions [8], while Menon et al. [9] described a grading score based on the MRI morphology to distinguish surgically treatable LCS from the minor variants. Apart from these radiological classifications, the clinical severity of LCS syndrome is evaluated using scores like the Swiss Spinal Stenosis (SSS) score or the Zurich Claudication Questionnaire (ZCQ). The symptom severity and physical function subscales of the SSS have been reported as valid and reliable, as well as reproducible for measuring clinical symptoms of LCS [10–12].
Despite the availability of multiple grading systems, the basic question regarding the radiological denominators of severe LCS remains unanswered. If the severity of clinical/neurological dysfunction is suggested as an indication of disease severity, i.e., the presence of sensory, motor, and bladder/bowel dysfunction are considered indicative of increasingly more severe disease, it can be reasonably assumed that the correlation of these clinical features with radiological appearance of LCS may be used to establish radiological criteria to define severe LCS, and eventually determine surgical indications. Accordingly, this study retrospectively reviewed the clinical and radiological features of 100 patients operated on for LCS and correlated their neurological status with the four established radiological schemas for LCS to develop diagnostic radiological criteria for severe LCS.
Materials and Methods
In this retrospective cohort study, we reviewed 100 consecutive patients undergoing surgery for LCS at the Department of Orthopedics, Bharati Vidyapeeth (deemed to be University) Medical College, Dhankawadi, Pune, India between January 2020 and August 2022. Case-related data were retrieved from the hospital’s electronic records. This observational study was approved by the Bharati Hospital Research Ethics Committee (DHR Reg. No. EC/NEW/INST/2022/MH/0150; Ref. No. BVDUMC/IEC/13) on June 17, 2024. The requirement for informed consent from individual patients was omitted because of the retrospective design of this study.
All adult patients (>18 years of age) with clinical and radiological features of LCS due to degenerative spinal disorders undergoing index surgery were included in this study. Those presenting with tandem stenosis, or congenital or developmental stenosis, along with those undergoing revision surgeries, were excluded. Additionally, patients with a history of spinal infection, tumor pathologies, diabetic peripheral neuropathy, noncompressive neuropathy due to any cause, and peripheral vascular disease (including smoking) were excluded.
The operative procedures performed in the study patients included laminectomy, bilateral foraminotomy, instrumented posterolateral fusion, transforaminal lumbar interbody fusion (with cages or autograft), and posterior lumbar interbody fusion (with cages or autograft). This data was not considered relevant to the current study and hence not included in the analysis.
For all patients, sagittal, coronal, and axial plane MRI scans, including T1-weighted, T2-weighted, and short tau inversion recovery image sequences, were obtained using a 3-Tesla MRI machine (Siemens Healthineers, Erlangen, Germany; Philips Healthcare, Amsterdam, The Netherlands). The field of view was set as 31×31 cm for sagittal images and 18×15 cm for axial images. All images were stored directly as Digital Imaging and Communications in Medicine files on the hospital’s Picture Archiving and Communication System to be viewed on workstations. All levels from L1–S1 were studied. All available MRI studies were accepted, irrespective of the quality of the obtained images. The images were collected, tabulated, and reviewed together by a senior spinal surgery consultant and a spine fellow. In case of multilevel stenosis, the worst-affected level was considered for the scoring.
All patients were preoperatively evaluated using one clinical—the ZCQ, comprising a pain severity and a physical function score—and four radiological LCS assessment tools: Schizas, Lee, Miskin and Mandell, and Menon radiological grading systems. For the ease of statistical analysis, the total ZCQ score was divided into mild, moderate, and severe (pain severity: normal, 0–7; mild, 8–14; moderate, 15–25; and severe, 26–35; physical function: mild, 0–7; moderate, 8–14; and severe, 15–20).
The data were analyzed using IBM SSPS for Windows ver. 25.0 (IBM Corp., Armonk, NY, USA). Means and standard deviation (SD) were used to describe quantitative variables, and frequency and percentages (%) were used for categorical variables. The chi-square test was used to examine the associations between qualitative and quantitative variables. A p-value of <0.05 was considered to determine statistical significance.
Results
All 100 patients (53 males and 47 females; mean age, 61.11 years; range, 33–79 years) included in this study were diagnosed with degenerative LCS (Table 1). All patients presented with symptoms of neurogenic claudication and low back pain. Thirty-six patients had varying grades of motor weakness in their lower limbs, and 25 had sensory abnormalities; seven out of the 100 patients had bladder dysfunction. Regarding the spinal levels affected, 69 patients had multilevel LCS, while 31 patients had single-level LCS. The L4–L5 segment was most commonly affected, followed by L3–L4, L5–S1, L2–L3, and L1–L2. Out of the 53 male patients, 18 patients had single-level LCS and 35 had multilevel affections. Among the female patients, 15 were affected with single-level LCS, while 32 had multilevel LCS.
Baseline characteristics of patients with degenerative lumbar spinal stenosis (n=100)
Based on preoperative pain, 65 patients were graded as having moderate pain (ZCQ pain score: 15–25), and 35 patients had severe pain (ZCQ pain score: 26–35); none of the patients had mild pain (ZCQ pain score: 0–15). In terms of physical function, all patients had a score ≥8; 55 patients were graded as having severe functional disability due to LCS symptoms (ZCQ physical function score: 15–20), while 45 patients were in the moderate category (score: 8–14) (Table 1).
As mentioned above, a total of 68 patients had some sensory (n=25), motor (n=36), or bladder/bowel dysfunction (n=7) (Table 2). According to the Schizas classification, 13 cases were classified as moderate grade LCS, of which motor deficit was present only in one patient; 60 cases were graded as severe LCS, and 27 patients were in the extremely severe category (Table 3). Among the severe and extremely severe categories (n=87), 35 patients had some motor deficit in the lower limbs. Accordingly, there was a statistically significant association between Schizas’ grading and motor deficits (p=0.029) (Table 4). In patients with sensory deficits (n=25), all patients (100%) had severe stenosis, and none of them were graded as moderate LCS using the Schizas classification. The association between the Schizas classification and sensory symptoms was also found to be statistically significant (p=0.034) (Table 4).
Distribution of patients with sensory and motor deficit
Schizas grading system distribution of patients as per grading (two groups, n=100)
Association between motor deficit and sensory with Schizas grading system
According to Lee’s classification for LCS, 30 patients had moderate grade stenosis, of which eight patients (22.2%) had motor deficit and 22 (34.4%) were neurologically intact. The remaining 70 patients had severe stenosis, of which 28 (77.8%) had motor weakness, while 42 (65.6%) had intact neurology (Table 5). Regarding sensory deficit, six patients (24%) in the moderate LCS group and 19 patients (76%) in the severe LCS group had sensory deficits. The chi-square analysis revealed no statistically significant association of LCS severity graded using Lee’s classification system with motor symptoms (p=0.258) or sensory symptoms (p=0.615) (Table 6).
Lee grading system distribution of patient according to severity of lumbar canal stenosis (n=100)
Association between motor deficit and sensory deficit in Lee grading system
According to the Miskin and Mandell grading system, 10 out of 100 patients belonged to the mild category, 44 to the moderate category, and 46 to the severe category of LCS. Among patients with mild and moderate LCS (n=54), 13 (36.1%) demonstrated motor deficits, whereas 41 (64.1%) were neurologically intact. In the severe stenosis category, 23 patients (63.9%) had some motor deficit, and 23 (35.9%) had intact neurology (Table 7). Regarding sensory loss, 10 patients (40%) in the mild plus moderate LCS groups had sensory loss, whereas 44 patients (58.7%) had intact sensations. In the severe stenosis group, 15 patients (60%) had sensory deficits, whereas 31 cases (41.3 %) had normal sensory distribution. Based on these data, there was a statistically significant association between LCS severity grading using the Miskin and Mandell radiological scale and motor deficit (p=0.012); however, the association between the Miskin and Mandell radiological grading and sensory symptoms was not statistically significant (p=0.113) (Table 8).
Miskin and Mandell grading system (n=100)
Association between motor deficit and sensory deficit in Miskin and Mandell grading system
Further, the LCS grading using Menon’s MRI-based morphological classification did not show significant correlation between motor deficit and radiological grading (p=0.1368) (Table 9). According to this classification, the “pinhole” type morphology was seen on the MRI axial images of 19 patients; of these, only seven patients had motor deficit. Regarding the “large triangle,” “trefoil,” and “small triangle” morphologies, motor deficits were seen in three, two, and six cases, respectively. Four patients each in the “cat’s eye” and “complete block” (no hole) morphology categories had motor deficits. In case of sensory deficit, there was no correlation between the magnetic resonance (MR) morphology of stenosis and sensory deficit (p=0.668) (Table 9). Also, the bladder dysfunction did not correlate with the morphology of the stenosed canal (p=0.340) (Table 9).
Association between motor deficit and sensory deficit in Menon’s et all classification
Discussion
Summary of results
Several studies have attempted to establish a correlation between the clinical severity of LCS and radiological grading systems, but with little success [4,13–18]. To the best of our knowledge, no previous studies have compared canal morphology and MRI appearances with the severity of neurological deficit, both sensory and motor, in LCS. This study used three MRI-based grading systems—Schizas’s, Lee’s, and Miskin and Mandell’s—and one morphology-based scheme (Menon’s) to classify LCS and compared them with the preoperative neurological status of patients. This study clearly establishes that neurological dysfunction (motor and sensory) in LCS correlates linearly with the Schizas schema, and only motor deficits correlate with the Miskin and Mendall scheme. Other radiological classifications, Menon’s and Lee’s, do not correlate with neurological deficit.
Literature
Currently, surgical decisions regarding LCS are made based on subjective and objective clinical features; radiological evidence is only used to confirm the clinical diagnosis, and its severity grading has no bearing on therapeutic decisions. Schizas et al. [6] described a LCS classification system based on MRI morphology; according to this system, only grades C and D LCS warrant surgery, whereas patients with grade A or B stenosis presenting with claudication do not warrant surgery. However, the authors did not comment on the correlation between neurological deficit and radiological severity of stenosis. Likewise, Lee et al. [7] classified LCS into mild, moderate, and severe grades, based on the extent of obliteration of the anterior CSF space and aggregation of the cauda equina, and showed excellent inter- and intra-observer agreement. However, the authors did not demonstrate any correlation with the clinical presentation or neurological deficit. Similarly, the other two classification systems proposed by Miskin et al. [8] and Menon et al. [9] also do not comment on any correlation with clinical presentation and neurological deficit. Kuittinen et al. [18] reported a significant correlation between MRI-assessed severity of lateral recess stenosis and abnormal electromyography findings, but failed to correlate the severity of stenosis with the patient’s symptoms.
Clinically, the severity of neurogenic claudication, the hallmark of LCS, is measured by the SSS or the ZCQ scores; these scales have been proven to be reliable assessment tools and have prognostic value as well [10–13,19]. However, like our study, other authors have also reported findings establishing the non-correlation of clinical and radiological features of LCS. Jonsson et al. [20] reported that they did not find any correlation between MRI morphology severity and clinical feature severity.
Importance of the study
Currently, LCS is a clinical plus radiological diagnosis, primarily based on canal appearance on axial T2-weighted MR images; however, the decision to confer this diagnosis based on imaging alone has been previously challenged. Likewise, the decision to offer incremental treatment protocols based on imaging alone has also been presumed dubious and never established unequivocally. Clinically, motor weakness and autonomic dysfunction, less often sensory diminution, are primarily considered evidence of significant neural compression and therefore constitute surgical indications; it is not mandatory to obtain matching radiological grading on axial MR imaging. The present study corroborates these findings that defining imaging thresholds for surgical intervention may not be appropriate in the context of LCS.
Limitations
The study has certain limitations. First, it is a single-center retrospective analysis including only surgical patients, with all the inherent limitations in patient selection and performance of the surgical procedures; patients with severe radiological stenosis, but asymptomatic and managed conservatively, were not considered. Second, we did not assess for inter-rater reliability of stenosis grading using different classifications. Third, the fact that patients with significant neurological dysfunction were automatically considered for surgery may have been a source of potential selection bias. Fourth, several confounding variables, such as diabetes, smoking, and body mass index, were not considered. Finally, all possible radiological scoring systems for LCS were not considered in this study; we only used the more frequently employed classification systems. Also, we were unable to confirm whether the current surgical threshold based on the clinical scoring system was appropriate in terms of subjective and objective neurological recovery of these patients.
Future scope of research
What then is the role of MRI in LCS? Our study verifies that the current role of MRI is only to establish the diagnosis of neural compression in any patient complaining of neurogenic claudication. Despite the many scoring systems established to putatively grade the severity of stenosis, there was little correlation between the imaging grades and symptom severity [13,14]. The current study also establishes that the preoperative neurological status has little bearing on the radiological severity of the disease. It is therefore recommended that surgical decisions in LCS should not be based on imaging findings, but rather on clinical severity of symptoms and presence of neurological dysfunction.
Conclusions
In LCS, the clinical severity of symptoms, as measured by the ZCQ score, does not correlate with radiological severity. Assuming that neurological deficit is the most severe form of clinical symptoms, we sought to determine whether it correlates with radiological severity in operated LCS cases. Of the various classifications described in literature, Schizas’ classification was significantly associated with neurological deficit, both motor and sensory, whereas Miskin and Mandell’s classification was only associated with motor deficit. Menon’s morphology-based score and Lee’s radiological classification failed to show any statistically significant correlation with neurological deficit.
It is noteworthy that this study does not seek to determine the outcomes of surgery in relation to the imaging findings or the preoperative clinical neurological status.
Key Points
Schizas classification correlates best with neurological deficit both motor and sensory.
Miskin and Mandell’s classification statistically correlates with motor deficit only.
Whereas Menon’s and Lee’s classifications do not correlates with neurological deficit.
Notes
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Author Contributions
Conceptualization: VM, AVM, MB. Methodology: VM, AVM, MB, SR, UK, AK. Data curation: VM, MB, SR, UK, AK. Investigation: AVM. Resources: AVM, MB, SR, UK. Validation: VM, AVM. Visualization: AVM, UK. Software: VM, AVM, MB, SR, UK. Funding acquisition: AVM. Administration: AVM, SR, UK, AK. Writing–original draft: all authors participated in writing. Writing–review and editing: VM, AVM, TD, AK. Supervision: all authors participated. Final approval of the manuscript: all authors.