Preoperative Predictors of Intramedullary Increased Signal Intensity and Risk Factors for Poor Recovery in Thoracic Ossification of the Ligamentum Flavum: A Single-Center Retrospective Study

Abstract

Study Design

Retrospective cohort study.

Objective

To comprehensively evaluate the clinical and radiological characteristics across intramedullary increased signal intensity (ISI) grades (Grade 0, 1, and 2) on T2-weighted magnetic resonance imaging (MRI) and identify preoperative predictors of ISI severity as well as risk factors for poor postoperative recovery in patients with thoracic ossification of the ligamentum flavum (T-OLF).

Methods

A total of 148 patients who underwent posterior laminectomy surgery for T-OLF between January 2017 and August 2023 was retrospectively analyzed. Demographic, imaging, surgical, and clinical outcome data were collected. ISI grades were independently assessed by three blinded spinal surgeons. Univariate and multivariate logistic regression analyses were performed to identify independent predictors and risk factors.

Results

Higher ISI grades were associated with worse preoperative neurological status, greater spinal canal compression, more complex OLF morphology, and increased intraoperative blood loss. Multivariate analysis identified a high canal occupying ratio (COR) (Odds ratio [OR] = 3.20, P = .012) and beak-type configuration on sagittal MRI (OR = 4.24, P = .003) as independent predictors of higher ISI grades. Moreover, multi-segment OLF (OR = 2.23, P = .028) and ISI Grade 2 (OR = 3.67, P = .044) independently predicted poor postoperative recovery.

Conclusion

Preoperative ISI grade serves as a critical prognostic indicator in T-OLF. High COR and beak-type configuration on sagittal MRI predict severe ISI grades, whereas multi-segment OLF and ISI Grade 2 independently predict poor postoperative outcomes. Incorporating these parameters into preoperative assessment may improve risk stratification, patient counseling, and surgical planning.

Keywords

thoracic spine ossification of the ligamentum flavum increased signal intensity neurological recovery

Introduction

Thoracic ossification of the ligamentum flavum (T-OLF) is a slowly progressive degenerative disorder that leads to spinal cord compression and subsequent motor weakness, sensory deficits, and sphincter dysfunction, substantially compromising patients’ quality of life.¹ Surgical decompression remains the only effective treatment to relieve compression and restore neurological function.^2,3 However, postoperative outcomes vary considerably, and a substantial proportion of patients fail to achieve satisfactory recovery.^2,4-6

Magnetic resonance imaging (MRI) provides valuable insights into intramedullary pathological changes associated with T-OLF.⁷ Increased signal intensity (ISI) on T2-weighted MRI has long been recognized as a radiological marker of spinal cord injury and has been correlated with postoperative neurological recovery.⁸ Nevertheless, most previous studies have simply regarded ISI as a binary variable (present or absent),^2,9-11 resulting in inconsistent conclusions, with some studies reporting ISI as a predictor of poor outcomes while others found no significant association.^1,5,9,12

To date, only one study by Zhang et al¹³ has specifically investigated the relationship between ISI grade (Grade 0, 1, and 2) and surgical outcomes in patients with T-OLF. Although their findings supported the prognostic significance of ISI grading, interpretation was limited by the small number of Grade 2 cases, short follow-up duration, and heterogeneity in MRI acquisition across 4 centers. Furthermore, the preoperative determinants of ISI severity remain unclear, and no large-scale study has systematically evaluated how ISI grades relate to postoperative outcomes.

To address these gaps, we conducted a large single-center retrospective cohort study with long-term follow-up. This study aimed to: (1) compare demographic, radiological, surgical, and clinical characteristics across different ISI grades; (2) identify independent preoperative predictors of ISI severity; and (3) determine independent risk factors for poor postoperative recovery. We hypothesized that higher ISI grades would correlate with more severe preoperative neurological deficits and predict worse postoperative outcomes.

Methods

Study Design and Population

This retrospective cohort study reviewed the medical records of 160 consecutive patients with T-OLF who underwent posterior laminectomy at our institution between January 2017 and August 2023. Ethical approval was obtained from our institutional review board, and the study adhered to the Declaration of Helsinki.¹⁴ Informed consent was waived due to the retrospective design. Inclusion criteria were as follows: (1) thoracic myelopathy caused solely by T-OLF confirmed by both computed tomography (CT) and MRI; (2) age ≥18 years; and (3) complete clinical data with a minimum follow-up of 2 years. Exclusion criteria included: (1) concurrent ventral compressive lesions such as thoracic disc herniation, kyphotic deformity, or ossification of the posterior longitudinal ligament; (2) incomplete data or loss to follow-up; (3) previous thoracic surgery; and (4) non-compressive intramedullary pathologies such as syringomyelia, inflammatory, or demyelinating myelopathies. All procedures were performed by the same senior spine surgery team, and posterior decompression was conducted using a standardized en-bloc laminectomy technique with instrumentation.

Postoperative Management

Postoperative care was standardized across all patients according to our institutional protocol. Early mobilization was routinely encouraged, with ambulation initiated on the first postoperative day with thoracolumbar bracing under supervision of a rehabilitation therapist. Bedside physiotherapy focusing on gait training and lower-limb strengthening, which was progressively advanced. Wound drains were removed when the daily output fell below 50 mL, after which discharge planning was arranged.

For patients with cerebrospinal fluid (CSF) leakage, ambulation was temporarily restricted and conservative management including adequate hydration and hypertonic saline infusion was provided. Bedside functional rehabilitation was continued during bed rest. Drainage tubes were removed once CSF drainage became clear, followed by pressure dressing of the wound. All patients received structured rehabilitation instructions at discharge and were regularly followed postoperatively.

Data Collection

Demographic data included age, sex, body mass index (BMI), comorbidities (hypertension and diabetes mellitus), American society of Anesthesiologists (ASA) grade, preoperative modified Japanese Orthopaedic Association (mJOA) score, disease duration, nonambulatory status, sphincter dysfunction, and follow-up duration. Imaging data included the presence of multi-segment OLF, location of maximal stenosis, canal occupying ratio (COR), axial CT configuration, sagittal MRI configuration, and presence of diffuse idiopathic skeletal hyperostosis (DISH). Surgical variables included the number of decompressed segments, dural ossification, incidental durotomy, operative time per segment, and estimated blood loss (EBL) per segment. Postoperative complications included CSF leakage, transient neurological deterioration, epidural hematoma, wound infection, instrumented vertebrae fracture, and systemic complications. Clinical outcomes included neurological recovery and length of hospital stay.

Imaging Evaluation

Multi-segment OLF was defined as ossification extending across three or more vertebral levels. Axial CT configuration was classified as unilateral, bilateral, or bridged types¹⁵ (Figure 1), and sagittal MRI configuration as round or beak types^9,15 (Figure 2). DISH was diagnosed according to Resnick’s criteria.¹⁶ The cross-sectional area (CSA) of the spinal canal at the pedicle level of the most stenotic segment was measured on axial CT images and defined as the reference CSA¹⁷ (Figure 1). COR was calculated as: COR = (area of ossified mass at the narrowest level/reference CSA) × 100%. A COR >50% was defined as high. ISI grades were assessed on sagittal T2-weighted MRI according to Yukawa et al¹⁸: Grade 0 (none), Grade 1 (light or obscure), and Grade 2 (intense or bright) (Figure 2). Three experienced spine surgeons independently assessed all MRI scans in a blinded manner, and final grades were determined by consensus.

Figure 1.

OLF configuration type on axial CT and COR measurement.

Figure 2.

ISI Grades and OLF configuration type on sagittal T2-weighted MRI.

Neurological Assessment

Neurological function was assessed using the mJOA scores system both preoperatively and at final follow-up. The mJOA scores system excluded upper extremity function, with a maximum of 11 points.¹⁹ Recovery rate (RR) was calculated as: RR = (postoperative mJOA score - preoperative mJOA score)/(11 - preoperative mJOA score) × 100%.¹⁹ RR ≥50% was defined as good recovery, and RR <50% as poor recovery, consistent with previously thoracic myelopathy studies.^11,19,20

Statistical Analysis

All statistical analyses were conducted using R software (version 4.4.2; R Foundation for Statistical Computing, Vienna, Austria). Normality of continuous variables was tested using the Shapiro–Wilk test. Continuous data were expressed as mean ± standard deviation (SD) and compared using one-way analysis of variance, Kruskal–Wallis test, unpaired t test, or Wilcoxon rank-sum test as appropriate. Categorical variables were presented as counts (percentages) and compared using the Chi-square or Fisher’s exact test. Interobserver reliability for ISI grades was evaluated using the intraclass correlation coefficient (ICC) and 95% confidence intervals (CIs), with ICC >0.8 indicating excellent agreement. Variables with P < .05 in univariate analysis were included in multivariate logistic or ordinal regression analyses to identify independent predictors. Odds ratios (ORs) with 95% CIs were calculated. Multicollinearity was assessed using the generalized variance inflation factor (GVIF), with GVIF <5 considered acceptable. Sample size adequacy was evaluated according to the 10 events-per-variable principle, ensuring no overfitting in multivariable models. A two-tailed P < .05 was considered statistically significant.

Results

Patient Demographics

A total of 148 patients (84 males and 64 females) were finally included, with 29 in Grade 0, 62 in Grade 1, and 57 in Grade 2. The interobserver reliability for ISI grades was excellent (ICC = 0.929, 95% CI: 0.901-0.948, P < .001). The mean follow-up duration was 54.7 ± 26.0 months. The mean age and BMI were 56.0 ± 12.1 years and 27.4 ± 4.0 kg/m², respectively. Hypertension and diabetes mellitus were present in 37.8% and 16.9% of patients, respectively. Neurological status differed significantly among ISI groups, with progressively lower preoperative mJOA scores (6.3 ± 2.0 vs 5.5 ± 2.2 vs 5.0 ± 1.8, P = .006) and higher proportions of nonambulatory patients (10.3% vs 22.6% vs 36.8%, P = .022) across the three groups. Other baseline demographic variables, including age, sex, BMI, comorbidities, ASA grade, disease duration, sphincter dysfunction, and follow-up duration, were comparable among the three groups (all P > .05) (Table 1).

Table 1.

Demographic Characteristics of Patients Among Different ISI Grades

	Total (N = 148)	Grade 0 (N = 29)	Grade 1 (N = 62)	Grade 2 (N = 57)	P
Age (years)	56.0 ± 12.1	59.1 ± 12.1	57.0 ± 11.7	53.3 ± 12.1	.075
Sex					.770
Male	84 (56.8%)	15 (51.7%)	37 (59.7%)	32 (56.1%)
Female	64 (43.2%)	14 (48.3%)	25 (40.3%)	25 (43.9%)
BMI (kg/m²)	27.4 ± 4.0	26.4 ± 3.7	27.5 ± 3.3	27.9 ± 4.7	.306
Hypertension	56 (37.8%)	10 (34.8%)	23 (37.1%)	23 (40.4%)	.858
Diabetic mellitus	25 (16.9%)	5 (17.2%)	13 (21.0%)	7 (12.3%)	.449
ASA grade					.900
I	21 (14.2%)	3 (10.3%)	9 (14.5%)	9 (15.8%)
II	121 (81.8%)	24 (82.8%)	51 (82.3%)	46 (80.7%)
III	6 (4.1%)	2 (6.9%)	2 (3.2%)	2 (3.5%)
Preoperative mJOA score	5.4 ± 2.1	6.3 ± 2.0	5.5 ± 2.2	5.0 ± 1.8	.006^**
Disease course (months)	25.0 ± 40.8	22.5 ± 25.3	23.4 ± 48.9	28.0 ± 37.7	.309
Nonambulatory status	38 (25.7%)	3 (10.3%)	14 (22.6%)	21 (36.8%)	.022^*
Sphincter dysfunction	56 (37.8%)	6 (20.7%)	25 (40.3%)	25 (43.9%)	.097
Follow-up duration (months)	54.7 ± 26.0	56.7 ± 27.6	51.7 ± 25.8	57.0 ± 25.6	.509

ISI, increased signal intensity; BMI, body mass index; ASA, American society of Anesthesiologists; mJOA, modified Japanese Orthopaedic Association. * represented P < .05; ** represented P < .01; *** represented P < .001.

Imaging Characteristics

Significant radiological differences were noted across ISI grades (Table 2). Patients with higher ISI grades demonstrated a significantly greater proportion of high COR (13.8% vs 51.6% vs 79.0%, P < .001), a higher prevalence of bridged-type OLF on axial CT (31.0% vs 48.4% vs 64.9%, P = .002), and a markedly increased rate of beak-type morphology on sagittal MRI (P < .001). In contrast, there were no significant differences among the three groups in the presence of DISH (P = .734) or the thoracic level of maximal compression (P = .526).

Table 2.

Comparisons of Imaging Features Among Different ISI Grades

	Total (N = 148)	Grade 0 (N = 29)	Grade 1 (N = 62)	Grade 2 (N = 57)	P
Multi-segment OLF	75 (50.7%)	14 (48.3%)	27 (43.5%)	34 (59.6%)	.206
Location of maximal stenosis					.526
Upper (T1-T4)	20 (13.5%)	6 (20.7%)	7 (11.3%)	7 (12.3%)
Middle (T5-T8)	11 (7.4%)	3 (10.3%)	3 (4.8%)	5 (8.8%)
Lower (T9-T12)	117 (79.1%)	20 (69.0%)	52 (83.9%)	45 (79.0%)
High COR	81 (54.7%)	4 (13.8%)	32 (51.6%)	45 (79.0%)	<.001^***
Axial CT configuration					.002^**
Unilateral	23 (15.5%)	10 (34.5%)	11 (17.7%)	2 (3.5%)
Bilateral	49 (33.1%)	10 (34.5%)	21 (33.9%)	18 (31.6%)
Bridged	76 (51.4%)	9 (31.0%)	30 (48.4%)	37 (64.9%)
Sagittal MRI configuration					<.001^***
Round	51 (34.5%)	21 (72.4%)	25 (40.3%)	5 (8.8%)
Beak	97 (65.5%)	8 (27.6%)	37 (59.7%)	52 (91.2%)
DISH	76 (51.4%)	16 (55.2%)	33 (53.2%)	27 (47.4%)	.734

ISI, increased signal intensity; OLF, ossification of the ligamentum flavum; COR, canal occupying ratio; CT, computed tomography; MRI, magnetic resonance imaging; DISH, diffuse idiopathic skeletal hyperostosis. * represented P < .05; ** represented P < .01; *** represented P < .001.

Surgical Findings and Postoperative Outcomes

The mean number of decompressed segments was 3.5 ± 1.2, with a mean operative time of 29.3 ± 9.7 minutes and EBL of 85.2 ± 48.4 mL per segment. Importantly, EBL per segment increased significantly with ISI severity (P = .016), whereas the incidence of dural ossification (P = .918) and incidental durotomy (P = .656) were comparable among the three groups. CSF leakage was the most common complication (32.4%), followed by transient neurological deterioration (7.4%), pulmonary complications (3.4%), gastrointestinal complications (3.4%), epidural hematoma (1.4%), wound infection (1.4%), and vertebral fracture (1.4%). Overall complication rates did not significantly differ among ISI groups (all P > .05) (Table 3).

Table 3.

Comparisons of Surgical Details, Complications and Clinical Outcomes Among Different ISI Grades

	Total (N = 148)	Grade 0 (N = 29)	Grade 1 (N = 62)	Grade 2 (N = 57)	P
Surgical details
No. of decompression segments	3.5 ± 1.2	3.6 ± 1.2	3.4 ± 1.3	3.6 ± 1.1	.272
Operative time per segment (min)	29.3 ± 9.7	27.6 ± 9.2	30.5 ± 10.0	29.0 ± 9.6	.210
EBL per segment (mL)	85.2 ± 48.4	63.3 ± 26.9	84.4 ± 47.1	97.1 ± 54.5	.016^*
Dural ossification	34 (23.0%)	6 (20.7%)	14 (22.6%)	14 (24.6%)	.918
Incidental durotomy	39 (26.4%)	6 (20.7%)	16 (25.8%)	17 (29.8%)	.656
Complications
CSF leakage	48 (32.4%)	7 (24.1%)	18 (29.0%)	23 (40.4%)	.238
Transient neurological deterioration	11 (7.4%)	1 (3.5%)	7 (11.3%)	3 (5.3%)	.419
Epidural hematoma	2 (1.4%)	1 (3.5%)	1 (1.6%)	0 (0.0%)	.501
Wound infection	2 (1.4%)	0 (0.0%)	1 (1.6%)	1 (1.8%)	1.000
Instrumented vertebrae fracture	2 (1.4%)	1 (3.5%)	0 (0.0%)	1 (1.8%)	.336
Pulmonary complication	5 (3.4%)	1 (3.5%)	3 (4.8%)	1 (1.8%)	.842
Gastrointestinal complication	5 (3.4%)	1 (3.5%)	3 (4.8%)	1 (1.8%)	.842
Clinical outcomes
Preoperative mJOA score	5.4 ± 2.1	6.3 ± 2.0	5.5 ± 2.2	5.0 ± 1.8	.006^**
Final mJOA score	8.1 ± 1.5	8.8 ± 1.8	8.2 ± 1.4	7.7 ± 1.4	<.001^***
RR (%)	47.9 ± 23.7	55.1 ± 31.3	46.3 ± 22.4	46.0 ± 20.2	.007^**
Poor recovery	60 (40.5%)	5 (17.2%)	26 (41.9%)	29 (50.9%)	.011^*
Length of hospital stay (days)	7.3 ± 2.0	7.7 ± 3.0	7.2 ± 1.9	7.3 ± 1.5	.678

ISI, increased signal intensity; EBL, estimated blood loss; CSF, cerebrospinal fluid; mJOA, modified Japanese Orthopaedic Association; RR, recovery rate. * represented P < .05; ** represented P < .01; *** represented P < .001.

Neurologically, the mJOA scores improved significantly from 5.4 ± 2.1 preoperatively to 8.1 ± 1.5 at final follow-up (P < .001) and improved significantly within each ISI subgroup (all P < .001). However, both the final mJOA score (P < .001) and RR (P = .011) differed significantly among the three groups, with Grade 2 ISI showing the poorest recovery. Overall, 40.5% of patients experienced poor recovery, which was most frequent in Grade 2 ISI (50.9%). Length of hospital stay was comparable across groups (P = .678) (Table 3).

Predictors of ISI Severity

Variables with P < .05 in the univariate analysis were included in multivariate ordinal logistic regression. The parallel regression assumption was satisfied (P = .528). High COR (OR = 3.20, 95% CI = 1.29 - 7.94, P = .012) and beak-type morphology on sagittal MRI (OR = 4.24, 95% CI = 1.66 - 10.85, P = .003) were identified as independent predictors of higher ISI grade (Table 4).

Table 4.

Multivariate Ordinal Logistic Regression Analysis for ISI Grade

	OR	95% CI	P
Preoperative mJOA score	0.90	0.73-1.11	.325
High COR	3.20	1.29-7.94	.012^*
Axial CT configuration
Unilateral	Reference
Bilateral	1.40	0.49-3.99	.530
Bridged	1.04	0.33-3.26	.953
Sagittal MRI configuration
Round	Reference
Beak	4.24	1.66-10.85	.003^**
Nonambulatory status	1.61	0.59-4.39	.351

ISI, increased signal intensity; OR, odds ratio; 95% CI, 95% confidence interval; mJOA, modified Japanese Orthopaedic Association; COR, canal occupying ratio; CT, computed tomography; MRI, magnetic resonance imaging. * represented P < .05; ** represented P < .01; *** represented P < .001.

Risk Factors for Poor Recovery

Comparative analysis revealed that poor recovery was significantly associated with multi-segment OLF (P = .011), high COR (P = .002), beak-type morphology (P = .046), and higher ISI grade (P = .011) (Table 5). Multivariate logistic regression further demonstrated that multi-segment OLF (OR = 2.23, 95% CI: 1.09 - 4.55, P = .028) and ISI Grade 2 (OR = 3.67, 95% CI: 1.04 - 12.89, P = .044) were independent risk factors for poor postoperative neurological recovery (Table 6).

Table 5.

Univariate Analysis Between Good and Poor Recovery Groups

	Good recovery (N = 88)	Poor recovery (N = 60)	P
Age (years)	56.2 ± 12.2	55.6 ± 11.9	.763
Sex			.488
Male	52 (59.1%)	32 (53.3%)
Female	36 (40.9%)	28 (46.7%)
BMI (kg/m²)	27.6 ± 4.6	27.2 ± 3.0	.566
Hypertension	30 (34.1%)	26 (43.3%)	.255
Diabetic mellitus	17 (19.3%)	8 (13.3%)	.340
ASA grade			.113
I	11 (12.5%)	10 (16.7%)
II	71 (80.7%)	50 (83.3%)
III	6 (6.8%)	0 (0.0%)
Preoperative mJOA score	5.6 ± 2.0	5.2 ± 2.2	.496
Disease course (months)	26.3 ± 47.6	23.0 ± 28.0	.537
Nonambulatory status	18 (20.5%)	20 (33.3%)	.078
Sphincter dysfunction	34 (38.6%)	22 (36.7%)	.808
Multi-segment OLF	37 (42.1%)	38 (63.3%)	.011^*
High COR	39 (44.3%)	42 (70.0%)	.002^**
Axial CT configuration			.398
Unilateral	16 (18.2%)	7 (11.7%)
Bilateral	26 (29.6%)	23 (38.3%)
Bridged	46 (52.3%)	30 (50.0%)
Sagittal MRI configuration			.046^*
Round	36 (40.9%)	15 (25.0%)
Beak	52 (59.1%)	45 (75.0%)
ISI grade			.011^*
Grade 0	24 (27.3%)	5 (8.3%)
Grade 1	36 (40.9%)	26 (43.3%)
Grade 2	28 (31.8%)	29 (48.3%)
DISH	46 (52.3%)	30 (50.0%)	.786
Dura ossification	16 (18.2%)	18 (30.0%)	.093
Location of maximal stenosis			.108
Upper (T1-T4)	15 (17.1%)	5 (8.3%)
Middle (T5-T8)	4 (4.6%)	7 (11.7%)
Lower (T9-T12)	69 (78.4%)	48 (80.0%)
Follow-up duration (months)	54.7 ± 26.9	54.8 ± 24.9	.911
Length of hospital stay (days)	7.4 ± 2.2	7.3 ± 1.8	.504

BMI, body mass index; ASA, American society of Anesthesiologists; mJOA, modified Japanese Orthopaedic Association; OLF, ossification of the ligamentum flavum; COR, canal occupying ratio; CT, computed tomography; MRI, magnetic resonance imaging; ISI, increased signal intensity; DISH, diffuse idiopathic skeletal hyperostosis. * represented P < .05; ** represented P < .01; *** represented P < .001.

Table 6.

Multivariate Logistic Regression Analysis of Risk Factors for Poor Recovery

	OR	95% CI	P
Multi-segment OLF	2.23	1.09-4.55	.028^*
High COR	2.12	0.84-5.35	.110
Sagittal MRI configuration
Round	Reference
Beak	0.79	0.29-2.17	.652
ISI grades
Grade 0	Reference
Grade 1	3.10	0.97-9.91	.055
Grade 2	3.67	1.04-12.89	.044^*

OR, odds ratio; 95% CI, 95% confidence interval; OLF, ossification of the ligamentum flavum; COR, canal occupying ratio; MRI, magnetic resonance imaging; ISI, increased signal intensity. * represented P < .05; ** represented P < .01; *** represented P < .001.

Discussion

To our knowledge, this retrospective cohort represents the largest study and longest follow-up study to comprehensively evaluate the clinical and radiological implications of preoperative ISI grades in patients with T-OLF undergoing posterior laminectomy. Our findings demonstrated that preoperative ISI grades served as a significant prognostic indicator, correlating with disease severity, radiological morphology, intraoperative complexity, and postoperative neurological recovery. In particular, a high COR and beak-type configuration on sagittal MRI independently predicted higher ISI grades, whereas multi-segment OLF and ISI Grade 2 were identified as independent risk factors for poor neurological recovery.

Consistent with previous evidence in cervical myelopathy, intramedullary hyperintensity on T2-weighted MRI is generally regarded as a surrogate of underlying spinal cord pathology, such as edema, ischemia, myelomalacia, demyelination, or necrosis.^21,22 In this cohort, 80.4% of patients exhibited intramedullary T2 hyperintensity on T2-weighted MRI, consistent with prior studies reporting incidence of 41% - 79%.²³ The strong association between higher ISI grades and high COR aligns with the hypothesis that mechanical compression plays a major role in the development of intramedullary signal alterations.²³ Moreover, progressive worsening of preoperative neurological status and higher rates of nonambulatory status were observed with increasing ISI grades. Patients with Grade 2 ISI also experienced significantly greater intraoperative blood loss, which may be mechanistically linked to the higher prevalence of complex OLF configurations in this subgroup, potentially contributing to venous plexus congestion. These findings are in line with Zhang et al¹³ who also demonstrated that preoperative ISI classification was associated with clinical severity, disease duration, degree of spinal cord compression, and surgical outcomes.

Multivariate ordinal logistic regression analysis further confirmed that a high COR and beak-type configuration on sagittal MRI were independent predictors of higher ISI grades. Previous literature by Kuh et al¹⁵ has likewise suggested that beak-type OLF is associated with greater surgical difficulty and worse prognosis compared with round-type morphology. We speculate that the focal, sharp compression produced by beak-type OLF, combined with severe stenosis, may result in microvascular compromise and irreversible neural injury, ultimately manifesting as intense intramedullary hyperintensity (Grade 2) on MRI.

With regard to surgical outcomes, 40.5% of patients experienced poor recovery at a mean follow-up of 54.7 months. Multivariate analysis identified multi-segment OLF and preoperative ISI Grade 2 as independent predictors of unfavorable postoperative neurological recovery. The adverse impact of extensive OLF involvement on recovery is consistent with the findings of Nagesh et al¹⁹ Moreover, the independent predictive value of ISI Grade 2 is supported by neuropathological evidence indicating that Grade 1 ISI generally reflects reversible intramedullary edema,⁴ whereas Grade 2 ISI reflect corresponds to more advanced and irreversible histopathological changes, including necrosis and cavitation.^24,25 Collectively, these findings highlight the clinical importance of ISI grades as a noninvasive marker of irreversible spinal cord injury in T-OLF.

Our findings provide clinically meaningful guidance for patient management. Preoperatively, identifying a high COR, beak-type morphology, multi-segment OLF, and particularly ISI Grade 2 enables more accurate risk stratification and prognostic counseling. Intraoperatively, recognizing these high-risk features allows surgeons to better anticipate technical challenges, optimize surgical planning, and prepare adequate blood management strategies, including autologous blood salvage systems and appropriate blood product availability. Moreover, the strong association between ISI Grade 2 and irreversible cord damage underscores the potential importance of timely surgical intervention before progression to advanced intramedullary pathology.

Several limitations should be acknowledged. First, the retrospective design and single-center nature of this study may have introduced selection and information bias. Second, the study included only patients with T-OLF, so the conclusions may not be generalizable to other etiologies of thoracic myelopathy. Third, although dichotomizing RR at a 50% threshold is clinically intuitive and widely adopted in previous literature, this approach may reduce outcome granularity and potentially attenuate statistical power. Finally, although the ISI grading system demonstrates excellent interobserver reliability, it remains a qualitative assessment and is unable to directly quantify the severity of intramedullary pathology. The absence of quantitative MRI metrics, such as signal intensity ratios or radiomics-based texture parameters, limits the biological interpretation of ISI alterations. Future studies integrating advanced quantitative MRI techniques, together with prospective multicenter investigations, are warranted to validate these findings and further clarify their clinical implications.

Conclusions

In conclusion, preoperative ISI grade on T2-weighted MRI represents a critical prognostic biomarker in patients with T-OLF. Higher ISI grades are associated with greater canal compression, more complex radiological morphology, and poorer postoperative neurological recovery. High COR and beak-type sagittal MRI configuration independently predict severe ISI changes, whereas multi-segment OLF and ISI Grade 2 independently are independent risk factors for unfavorable postoperative outcomes. Early surgical intervention before progression to high-grade ISI may help prevent irreversible spinal cord injury and optimize neurological recovery.

Footnotes

ORCID iDs

Juncai Lei

Panpan Hu

Feng Wei

Xiaoguang Liu

Ethical Considerations

This study was approved by the Ethics Committee of Peking University Third Hospital (No. M20241067) and conducted in accordance with the principles of the Declaration of Helsinki.

Consent to Participate

All identifiable individual data presented in this article were obtained with informed consent from the participants.

Consent for Publication

Informed consent was waived due to the retrospective nature of the study.

Authors Contribution

Juncai Lei contributed to the study design, data analysis, and manuscript drafting. Panpan Hu contributed to data analysis and materials support. Yanchao Tang and Xiao Liu assisted in data collection and material supports. Zhongjun Liu contributed to study design and material supports. Hua Zhou, Feng Wei, and Xiaoguang Liu conceived the study, supervised the manuscript. All authors read and approved the final manuscript. Juncai Lei and Panpan Hu contributed equally. Hua Zhou, Feng Wei, and Xiaoguang Liu shared corresponding authorship.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the National Natural Science Foundation of China (No. 82372451) and Beijing Natural Science Foundation (No. 4252025).

Correction (June 2026):

This article has been updated to correct the values in the “Disease course (months)” row of Table 1.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The datasets generated and analyzed during the current study are available from Peking University Third Hospital. However, restrictions apply to the availability of these data, which were used under institutional license and are not publicly accessible.*

Appendix

References

Hussain

Hou

. Clinical and prognostic analysis of ossified ligamentum flavum in a Chinese population. J Neurosurg Spine. 2005;3(5):348-354. doi:10.3171/spi.2005.3.5.0348.

Ren

Zhao

, et al. Clinical characteristics and surgical outcome of thoracic myelopathy caused by ossification of the ligamentum flavum: a retrospective analysis of 85 cases. Spinal Cord. 2016;54(3):188-196. doi:10.1038/sc.2015.139.

Osman

Cheung

Hussain

, et al. Outcomes and complications following laminectomy alone for thoracic myelopathy due to ossified ligamentum flavum: a systematic review and meta-analysis. Spine (Phila Pa 1976). 2018;43(14):E842-E848. doi:10.1097/BRS.0000000000002563.

Zhong

Meng

, et al. Clinical outcomes after decompressive laminectomy for symptomatic ossification of ligamentum flavum at the thoracic spine. J Clin Neurosci. 2016;28:77-81. doi:10.1016/j.jocn.2015.09.023.

Kang

Lee

Shin

Park

Chung

. Ossification of the ligamentum flavum of the thoracic spine in the Korean population. J Neurosurg Spine. 2011;14(4):513-519. doi:10.3171/2010.11.SPINE10405.

Onishi

Yasuda

Yamamoto

Iwaki

Ota

. Outcomes of surgical treatment for thoracic myelopathy: a single-institutional study of 73 patients. Spine (Phila Pa 1976). 2016;41(22):E1356-E1363. doi:10.1097/BRS.0000000000001622.

Lee

Park

Jeon

Rhim

Roh

. Clinically significant radiographic parameter for thoracic myelopathy caused by ossification of the ligamentum flavum. Eur Spine J. 2019;28(8):1846-1854. doi:10.1007/s00586-018-5750-6.

Park

Min

, et al. Surgical outcome of thoracic myelopathy secondary to ossification of ligamentum flavum. Jt Bone Spine. 2007;74(6):600-605. doi:10.1016/j.jbspin.2007.01.033.

Sanghvi

Chhabra

Mascarenhas

Mittal

Sangondimath

. Thoracic myelopathy due to ossification of ligamentum flavum: a retrospective analysis of predictors of surgical outcome and factors affecting preoperative neurological status. Eur Spine J. 2011;20(2):205-215. doi:10.1007/s00586-010-1423-9.

10.

Ando

Imagama

Kobayashi

, et al. Comparative study of surgical treatment and nonsurgical follow up for thoracic ossification of the posterior longitudinal ligament: radiological and clinical evaluation. Spine (Phila Pa 1976). 2017;42(6):407-410. doi:10.1097/BRS.0000000000001769.

11.

Bagga

Shetty

Viswanathan

Reddy

Kanna

Rajasekaran

. Thoracic myelopathy in ossified ligamentum flavum: surgical management and long-term outcome following 2 different techniques of surgical decompression. Glob Spine J. 2023;13(3):659-667. doi:10.1177/21925682211003061.

12.

Miyakoshi

Shimada

Suzuki

, et al. Factors related to long-term outcome after decompressive surgery for ossification of the ligamentum flavum of the thoracic spine. J Neurosurg. 2003;99(3 Suppl):251-256. doi:10.3171/spi.2003.99.3.0251.

13.

Zhang

Yang

Wang

, et al. Preoperative magnetic resonance imaging signal intensity classification is associated with clinical presentation and surgical outcomes in myelopathy caused by thoracic ossification of ligamentum flavum: a multicenter study. Glob Spine J. 2023;13(7):1971-1980. doi:10.1177/21925682211067752.

14.

World Medical Association . World Medical Association declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. 2013;310(20):2191-2194. doi:10.1001/jama.2013.281053.

15.

Kuh

Kim

Cho

, et al. Contributing factors affecting the prognosis surgical outcome for thoracic OLF. Eur Spine J. 2006;15(4):485-491. doi:10.1007/s00586-005-0903-9.

16.

Kim

Moon

Yoon

, et al. Prevalence of diffuse idiopathic skeletal hyperostosis diagnosed by whole spine computed tomography: a preliminary study. Clin Orthop Surg. 2018;10(1):41-46. doi:10.4055/cios.2018.10.1.41.

17.

Qiu

Zhao

. The relationship between dural ossification and spinal stenosis in thoracic ossification of the ligamentum flavum. J Bone Joint Surg Am. 2019;101(7):606-612. doi:10.2106/JBJS.17.01484.

18.

Yukawa

Kato

Yoshihara

Yanase

Ito

. MR T2 image classification in cervical compression myelopathy: predictor of surgical outcomes. Spine (Phila Pa 1976). 2007;32(15):1675-1678. doi:10.1097/BRS.0b013e318074d62e. discussion 1679.

19.

Nagesh

Mishra

Prabhuraj

, et al. Predictive factors of poor surgical outcome in thoracic ossified ligamentum flavum: a multivariate analysis. Clin Neurol Neurosurg. 2025;250:108815. doi:10.1016/j.clineuro.2025.108815.

20.

Ando

Imagama

Ito

, et al. Predictive factors for a poor surgical outcome with thoracic ossification of the ligamentum flavum by multivariate analysis: a multicenter study. Spine (Phila Pa 1976). 2013;38(12):E748-E754. doi:10.1097/BRS.0b013e31828ff736.

21.

Zhong

Wang

Huang

Luo

. Risk factors for rapid progressive neurological deterioration in patients with cervical spondylotic myelopathy. J Orthop Surg Res. 2021;16(1):75. doi:10.1186/s13018-021-02227-6.

22.

Arvin

Kalsi-Ryan

Mercier

Furlan

Massicotte

Fehlings

. Preoperative magnetic resonance imaging is associated with baseline neurological status and can predict postoperative recovery in patients with cervical spondylotic myelopathy. Spine (Phila Pa 1976). 2013;38(14):1170-1176. doi:10.1097/BRS.0b013e31828e23a8.

23.

Inamasu

Guiot

. A review of factors predictive of surgical outcome for ossification of the ligamentum flavum of the thoracic spine. J Neurosurg Spine. 2006;5(2):133-139. doi:10.3171/spi.2006.5.2.133.

24.

Mizuno

Nakagawa

Inoue

Hashizume

. Clinicopathological study of “snake-eye appearance” in compressive myelopathy of the cervical spinal cord. J Neurosurg. 2003;99(2 Suppl):162-168. doi:10.3171/spi.2003.99.2.0162.

25.

Ellingson

Salamon

Woodworth

Holly

. Correlation between degree of subvoxel spinal cord compression measured with super-resolution tract density imaging and neurological impairment in cervical spondylotic myelopathy. J Neurosurg Spine. 2015;22(6):631-638. doi:10.3171/2014.10.SPINE14222.