Standing radiographs to screen severe lumbar disc herniation for timely referral

Abstract

Objectives:

The degree of spinal canal compromise in lumbar disc herniation (LDH) influences treatment decisions and outcomes. Although magnetic resonance imaging (MRI) is the gold standard for assessing LDH severity, limited access in primary and secondary care settings maÿ delay diagnosis and referral. This study aimed to examine whether radiological parameters measured on standing spine radiographs are associated with MRI-defined LDH severity and could support timely referral for advanced imaging.

Methods:

This retrospective cross-sectional study included patients diagnosed with LDH who underwent both standing whole-spine radiographs and lumbar MRI between June 2014 and January 2024. Radiographic parameters assessed were disc height index (DHI), pedicle width-to-sagittal vertebral body width ratio (PW:SBW), pelvic incidence (PI), lumbar lordosis (LL), and sagittal vertical axis (SVA). Canal compromise was quantified on MRI and classified as mild (<50%) or severe (≥50%). Binary logistic regression and receiver operating characteristic (ROC) curve analyses were used to evaluate association between radiographic parameters and severe canal compromise.

Results:

An SVA ≥50 mm was independently associated with severe canal compromise (odds ratio = 3.376; 95% confidence interval: 1.658–6.877). The optimal SVA cutoff was 52 mm, yielding a sensitivity of 41% and specificity of 83%. Other radiographic parameters, including DHI, PW:SBW, PI, and LL, were not significantly associated with LDH severity.

Conclusions:

Sagittal vertical axis measured on standing radiograph is associated with the severity of spinal canal compromise in LDH but demonstrates limited discriminatory performance. SVA should not be used as a screening or exclusion tool; however, it may provide adjunctive information to support clinical assessment and MRI referral decisions in selected patients, particularly in resource-limited healthcare settings. Prospective studies are required to validate its role within clinical care pathways.

Introduction

Lumbar disc herniation (LDH) is a common condition among middle-aged adults, characterized by displacement of disc material that may compress the spinal canal or nerve roots. Diagnosis typically relies on clinical assessment supported by imaging. While conservative management is effective for most patients, surgical intervention is indicated for those with severe symptoms such as motor weakness or cauda equina syndrome. The degree of spinal canal compromise is an important factor influencing treatment outcomes and the likelihood of requiring surgery.

Previous studies have shown that radiographic findings, particularly the extent of canal compromise, are strong predictors of conservative treatment failure (1–4). Reported degrees of compromise vary widely, ranging from approximately 20% (3, 4) to 50% in surgical cohorts (5). Factors such as reduced intervertebral disc space, smaller spinal canal dimensions, and the biomechanical loading environment of the lumbar spine may contribute to the severity of disc herniation. Adolescent patients with narrower spinal canals are more likely to undergo surgical treatment (6). In addition, spinopelvic alignment parameters—including pelvic incidence (PI) and sagittal vertical axis (SVA)—have been reported to differ between surgical and nonsurgical groups in LDH and other degenerative spine conditions (7–9).

Although magnetic resonance imaging (MRI) is the gold standard for evaluating disc displacement and canal compromise, it is costly and often unavailable in many primary and secondary care settings. In contrast, standing radiographs are universally accessible, even in resource-limited environments, and remain the first-line imaging modality for patients presenting with lumbar symptoms. Identifying radiolographic alignment parameters that correlate with MRI-determined LDH severity could help clinicians prioritize referrals, improve triage accuracy, and optimize MRI utilization where resources are constrained.

Given the limited evidence directly linking standing radiographic parameters with MRI-defined severity in LDH, we propose that altered spinopelvic alignment, decreased disc height, and reduced sagittal canal dimensions may be associated with the degree of herniation. To our knowledge, no prior study has evaluated whether standing spine radiographs can predict the severity of LDH. This study therefore aims to assess radiological parameters obtained from standing full-spine lateral radiographs as potential predictors of spinal canal compromise in healthcare settings with limited MRI access.

Material and methods

Study design and setting

This retrospective cross-sectional study was conducted at Lerdsin Hospital, Department of Medical Services, Ministry of Public Health, Thailand. Data were collected from patients who visited the hospital between 1 June 2014, and 31 January 2024. Patients were identified using ICD-10 code M51 (intervertebral disc displacement) and M54.3 (sciatica).

Inclusion criteria

• Age more than 18 years

• Diagnosis of LDH confirmed by history, physical examination, and MRI

• Availability of both lumbar spine and standing whole-spine radiographs.

Exclusion criteria

• Previous lumbar spine surgery

• Foraminal or extraforaminal disc herniation

• Recurrent lumbar disc herniation

• Congenital lumbar disease, lumbar spondylolisthesis, spinal trauma, spinal tumor, or bilateral femoral head disease

• Lower limb joint disease affecting standing posture

Participants

A total of 259 participants were initially identified (Figure 1). Forty-three participants were excluded due to missing MRI (n = 6) or whole-spine radiographs (n = 37). An additional 49 patients were excluded because of multilevel disc herniation involving more than two levels (n = 20), foraminal or extraforaminal herniation (n = 17), prior spinal surgery (n = 3), or other spinal conditions (n = 9). After exclusions, 167 patients were included in the final analysis.

Figure 1

Variables

Demographic and clinical data included age, sex, body mass index (BMI), smoking history, nerve tension signs, and neurological deficit (motor weakness). Imaging measurements were performed using the Picture Archiving and Communication System (PACs) at Lerdsin Hospital.

Radiographic measurements

• Disc Height Index (DHI): Measured on lateral radiographs using the modified Farfan's method (10). DHI was calculated as the sum of anterior and posterior disc heights divided by the sum of superior and inferior disc depths at the affected level (Figure 2A).

• Pedicle Width-to-Sagittal Vertebral Body Width Ratio (PW:SBW): Used as a surrogate measure of spinal canal diameter. Pedicle width was defined as the distance from the posterior vertebral border to a line connecting the caudal and cephalad articular facets, and vertebral body width was measured at the mid-vertebral level (Figure 2B).

• Spinopelvic parameters: Pelvic Incidence (PI), Lumbar Lordosis (LL), and Sagittal Vertical Axis (SVA), were measured on standing whole-spine lateral radiographs (Figures 2C,D). SVA was defined as the horizontal distance from the C7 plumb line to the posterosuperior corner of the S1 endplate. LL was measured between the inferior endplate of T12 and the superior endplate of S1. PI was defined as the angle between a line perpendicular to the sacral endplate and a line connecting its midpoint to the bicoxofemoral axis (8).

MRI assessment and canal compromise measurement

Severe lumbar disc herniation was defined as ≥50% spinal canal compromise. Although the NASS-ASSR-ASNR Task Force classifies canal compromise as mild (<1/3), moderate (1/2–2/3), and severe (>2/3), previous studies have demonstrated inconsistent correlation between these categories and clinical outcomes. Several reports indicate that canal compromise of approximately 50% is commonly observed in surgical cohorts and is associated with higher rates of conservative treatment failure. Therefore, a ≥50% threshold was adopted, consistent with the mean value reported by Sutheerayongprasert et al. (5).

Figure 2

Canal compromise was quantified using the A-index, measured at the axial MRI slice demonstrating the greatest degree of canal compromise (Figure 3). The A-index represents the ratio of the area occupied by the herniated disc to the estimated original spinal canal area. Canal boundaries were defined posteriorly by the neural arch, laterally by the medial borders of the pedicles, and anteriorly by the posterior vertebral body or an extrapolated line based on Calisle's method when borders were indistinct (11). All MRI measurements were performed using EV Insite DICOM Viewer (version 3.11.1.500).

Figure 3

Study size

Sample size estimation followed the rule of 10 events per predictor variable. Five predictors were selected based on prior literature. Assuming a 39% prevalence of failed conservative treatment after 18.7 weeks (12), a minimum sample size 129 patients was determined to achieve adequate statistical power.

Statistical analysis

Categorical variables were summarized as frequencies and percentages, and continuous variables as means with standard deviations. Group comparisons were performed using the chi-square test for categorical variables and independent t-tests for continuous variables.

Univariate logistic regression was conducted to screen predictors of severe canal compromise (≥50%). Variables with p < 0.25 were entered into a multivariable logistic regression model, following the recommendation of Hosmer and Lemeshow. Forward stepwise selection was applied. Model fit was assessed using Nagelkerke R², and results were reported as odds ratios (ORs) with 95% confidence intervals (CIs).

Receiver operating characteristic (ROC) curves were generated for both the SVA-only model and the full multivariable model. The area under the curve (AUC) with 95% CI was reported, and the optimal SVA cutoff was determined using Youden's index.

All analyses were performed using SPSS version 28. A two-sided p-value < 0.05 was considered statistically significant. This study was approved by the Human Research and Ethics Committees of Lerdsin Hospital (LH671001) and the Institutional Review Board of the Faculty of Medicine, Chulalongkorn University (COA No. 1121/2024).

Results

Demographic and clinical characteristics

A total of 167 patients were analyzed, comprising 99 patients with mild canal compromise (<50%) and 68 patients with severe canal compromise (≥50%). Demographic and clinical characteristics are summarized in Table 1.

There were no significant differences between groups in sex (female 59.6% vs. 48.5%, p = 0.15), age (40.81 vs. 40.38 years, p = 0.82), BMI (25.71 vs. 25.07, p = 0.37), smoking status (18.3% vs. 12%, p = 0.98), nerve tension signs (88.6% vs. 82.3%, p = 0.26), or neurological deficits (29.8% vs. 35.3%, p = 0.46). Diabetes mellitus was more frequently observed in the mild group (7.3% vs. 0%, p = 0.04).

Radiographic predictors of LDH severity

The most frequently affected disc level was L4-L5, accounting for 57.6% of cases in the mild group and 62.7% in the severe group, followed by L5-S1 (39.4% vs. 35.8%). Central herniations were more common in the severe group (62.7%) then in the mild group (59.8%). With respect to herniation morphology, protrusion (49.5%) and extrusion (48.5%) occurred at similar frequencies in the mild group, whereas extrusion predominated in the severe group (73.5%). The mean degree of canal compromise was significantly greater in the severe group compared with the mild group (59.7% vs. 38.33%, p < 0.001).

No significant differences were observed between groups for DHI (0.22 vs. 0.21, p = 0.24), PW:SWB (0.32 vs. 0.33, p = 0.29), PI (49.27 vs. 50.56, p = 0.43), or LL (42.83 vs. 39.19, p = 0.12).

In contrast, SVA was significantly greater in the severe group (28.25 mm vs. 6.86 mm, p = 0.002). When categorized using a threshold of ≥50 mm, indicative of sagittal imbalance (13), a significantly higher proportion of patients in the severe group met this criterion compared with the mild group (41.2% vs. 17.2%, p < 0.001) (Table 2).

Multivariable analysis

Variables meeting the p < 0.25 threshold in univariable analysis (DHI, LL, and SVA category) were entered into the multivariable logistic regression model (Table 3). In the final model, SVA ≥50 mm was independently associated with severe canal compromise (OR = 3.376, 95% CI: 1.658–6.877). DHI and LL were not significant predictors.

ROC analysis

The ROC curve for SVA alone yielded an AUC of 0.60 (95%CI: 0.51–0.69). An optimal cutoff of ≥52 mm provided 41% sensitivity and 83% specificity for identifying severe canal compromise (Figure 4).

Figure 4

The full multivariable model, incorporating SVA category, DHI, LL, and PW:SBW, demonstrated a modestly higher AUC of 0.638 (95% CI: 0.55–0.73). Despite this improvement, SVA remained the primary contributor, and overall discriminatory performance was moderate.

Discussion

In this study, we examined whether radiographic alignment parameters measurable on standing full-spine lateral radiographs are associated with the severity of LDH. Among all evaluated parameters, only SVA demonstrated a statistically significant association with severe canal compromise (≥50%). In contrast, DHI, pelvic parameters, and PW:SBW showed no significant association with the degree of canal compromise.

We selected a ≥50% threshold for canal compromise based on the representative mean value reported in surgical cohorts by Sutheerayongprasert et al. (5). Demographic characteristics—including sex, age, BMI, smoking status, and neurological deficits—did not differ significantly between patients with mild (<50%) and severe (≥50%) canal compromise. The higher prevalence of diabetes mellitus observed in the mild group (7.3% vs. 0%, p = 0.04) is likely attributable to the small sample size. This finding is consistent with Kaiser et al. (14), who reported no association between demographic factors and severe high canal compromise presentations such as cauda equina syndrome. Although smoking, diabetes, and obesity have been associated with recurrent LDH (15), our results suggest these factors may not substantially influence the initial severity of canal compromise.

Herniation characteristics differed between groups. Central and extrusion-type herniations were more common in the severe group, whereas paracentral and protrusion-type herniations predominated in the mild group. The mean canal compromise was significantly greater in the severe group (59.7%) than in the mild group (38.33%), consistent with previous reports (5). DHI did not differ significantly between groups, suggesting that although reduced disc height has been linked to recurrence risk (16), it may not correlate with the initial severity of canal compression.

Regarding spinopelvic parameters, PI and LL did not differ significantly between groups; however, SVA was markedly higher in patients with severe canal compromise (28.25 mm vs. 6.86 mm, p = 0.002). This finding is consistent with prior studies by Endo et al. (7) and Fei et al. (17), which reported increased SVA in LDH cohorts. LL tended to be higher in the mild group, in line with earlier observations (17). The finding that PI values were within normal ranges aligns with reports by Barrey et al. (18) and normative data from the Thai population reported by Pluemvitayaporn et al. (19). Collectively, these findings suggest that while PI and LL may not be directly associated with canal compromise severity, increased SVA may reflect compensatory sagittal alignment changes in patients with more substantial canal compression.

Importantly, SVA should not be interpreted as a risk factor for canal compromise. Rather, it is a global alignment parameter that appears to be associated with MRI-defined severity. This association must be interpreted cautiously, as SVA can be influenced by pain-related postural adaptation, compensatory mechanisms, and individual differences in pain tolerance. These factors may limit the generalizability of SVA as a predictive measure. Moreover, the discriminatory performance of SVA was modest (AUC = 0.60). Although multivariable analysis showed that SVA ≥50 mm was associated with severe canal compromise (OR = 3.376, 95% CI: 1.658–6.877), its sensitivity was low (41%) despite relatively high specificity (83%) at a cutoff of 52 mm.

From clinical perspective, these findings do not support the use of SVA as a stand-alone screening or diagnostic tool. Instead, SVA may serve as an adjunctive parameter in settings where access to MRI is limited. In patients with clinically suspected LDH, an elevated SVA (≥52 mm) may prompt consideration of earlier MRI to clarify the extent of canal compromise. Early identification of patients with severe canal compromise is relevant because such patients are more likely to require surgical intervention (3, 4, 11) and delays in surgery have been associated with slower recovery and increased healthcare costs (12, 20, 21). Nevertheless, these implications must be interpreted caution given the limited sensitivity of SVA, its susceptibility to postural variation, and the absence of clinical severity measures in this study (3, 4, 11, 12, 20, 21).

From a health system perspective, SVA assessment may assist triage decision in primary and secondary care environments with limited MRI availability, where referral timing is often uncertain. While MRI is clearly indicated for patients with progressive neurological deficits or cauda equina syndrome, guidance for other cases is less clearly defined and often recommends MRI only after four weeks of unsuccessful conservative treatment (22, 23). In this context, SVA may help identify patients who could benefit from earlier evaluation, but it should not replace clinical judgment or MRI when available.

This study has several limitations. The retrospective design introduces potential selection bias and treatment details were not systematically collected, limiting the ability to correlate canal compromise with clinical outcomes. Although intra- and inter-observer reliability were not recalculated, the use of established measurement methods supports reproducibility. Chen et al. (24) demonstrated high reliability for the modified Farfan's method used to measure DHI, and Yamada et al. (25) reported high reliability for spinopelvic parameters, including PI, SS, and PT. Prospective studies incorporating clinical severity measures, symptom-based stratification, and longitudinal outcomes are needed to confirm these findings and define the clinical role of SVA more precisely.

Conclusion

This study demonstrates that increased sagittal vertical axis is associated with more severe spinal canal compromise in patients with lumbar disc herniation. Although SVA alone cannot reliably determine the severity of canal compromise and should not be used as a screening or exclusion tool, it may provide adjunctive information in clinical assessment, particularly in settings where access to MRI is limited. When interpreted alongside clinical findings, SVA help identify patients who could benefit from earlier MRI evaluation. Further prospective studies are required to validate these findings and to define the appropriate role of SVA within clinical decision-making pathways.

References

• 1.

  KrishnanVRajasekaranSAiyerSNKannaRShettyAP. Clinical and radiological factors related to the presence of motor deficit in lumbar disc prolapse: a prospective analysis of 70 consecutive cases with neurological deficit. Eur Spine J. (2017) 26(10):2642–9. 10.1007/s00586-017-5019-5

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 2.

  KalidindiKKVSathSVishwakarmaGChhabraHS. Magnetic resonance imaging findings in intervertebral disc herniation: comparison of canal compromise and canal size in patients with and without cauda equina syndrome. Surg Neurol Int. (2020) 11:171. 10.25259/SNI_242_2020

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 3.

  KoniecznyMRReinhardtJSchleichCProstMKrauspeR. MRI based analysis of grade of spinal canal stenosis and grade of compression of nerve root by lumbar disc herniation as tools to predict probability to need surgical treatment. J Spine Surg. (2020) 6(2):356–62. 10.21037/jss-19-424

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 4.

  VarlottaCGGeDHStekasNFrangellaNJManningJHSteinmetzLet alMRI radiological predictors of requiring microscopic lumbar discectomy after lumbar disc herniation. Global Spine J. (2020) 10(1):63–8. 10.1177/2192568219856345

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 5.

  SutheerayongprasertCPaiboonsirijitSKuansongthamVAnuraklekhaSHiranyasthitiNNetiS. Factors predicting failure of conservative treatment in lumbar-disc herniation. J Med Assoc Thai. (2012) 95(5):674–80.

  • Pubmed Abstract
  • Google Scholar

• 6.

  WileyMRHee JoCKhaleelMAMcIntoshAL. Size matters: which adolescent patients are most likely to require surgical decompression for lumbar disk herniations?J Pediatr Orthop. (2019) 39(10):e791–e5. 10.1097/BPO.0000000000001371

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 7.

  EndoKSuzukiHTanakaHKangYYamamotoK. Sagittal spinal alignment in patients with lumbar disc herniation. Eur Spine J. (2010) 19(3):435–8. 10.1007/s00586-009-1240-1

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 8.

  BaeJLeeSHShinSHSeoJSKimKHJangJS. Radiological analysis of upper lumbar disc herniation and spinopelvic sagittal alignment. Eur Spine J. (2016) 25(5):1382–8. 10.1007/s00586-016-4382-y

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 9.

  BeckJBrisbyHBarantoAWestinO. Low lordosis is a common finding in young lumbar disc herniation patients. J Exp Orthop. (2020) 7(1):38. 10.1186/s40634-020-00253-7

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 10.

  AkedaKYamadaTInoueNNishimuraASudoA. Risk factors for lumbar intervertebral disc height narrowing: a population-based longitudinal study in the elderly. BMC Musculoskelet Disord. (2015) 16:344. 10.1186/s12891-015-0798-5

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 11.

  CarlisleELunaMTsouPMWangJC. Percent spinal canal compromise on MRI utilized for predicting the need for surgical treatment in single-level lumbar intervertebral disc herniation. Spine J. (2005) 5(6):608–14. 10.1016/j.spinee.2005.05.384

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 12.

  PeulWCvan HouwelingenHCvan den HoutWBBrandREekhofJATansJTet alSurgery versus prolonged conservative treatment for sciatica. N Engl J Med. (2007) 356(22):2245–56. 10.1056/NEJMoa064039

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 13.

  SchwabFJBlondelBBessSHostinRShaffreyCISmithJSet alRadiographical spinopelvic parameters and disability in the setting of adult spinal deformity: a prospective multicenter analysis. Spine (Phila Pa 1976). (2013) 38(13):E803–12. 10.1097/BRS.0b013e318292b7b9

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 14.

  KaiserRKrajcovaAWaldaufPSrikandarajahNMakelMBenesV. Are there any risk factors associated with the presence of cauda equina syndrome in symptomatic lumbar disk herniation?World Neurosurg. (2020) 141:e600–e5. 10.1016/j.wneu.2020.05.260

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 15.

  ShepardNChoW. Recurrent lumbar disc herniation: a review. Global Spine J. (2019) 9(2):202–9. 10.1177/2192568217745063

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 16.

  KimKTParkSWKimYB. Disc height and segmental motion as risk factors for recurrent lumbar disc herniation. Spine (Phila Pa 1976). (2009) 34(24):2674–8. 10.1097/BRS.0b013e3181b4aaac

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 17.

  FeiHLiWSSunZRMaQWChenZQ. Analysis of spino-pelvic sagittal alignment in young Chinese patients with lumbar disc herniation. Orthop Surg. (2017) 9(3):271–6. 10.1111/os.12340

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 18.

  BarreyCJundJNosedaORoussoulyP. Sagittal balance of the pelvis-spine complex and lumbar degenerative diseases. A comparative study about 85 cases. Eur Spine J. (2007) 16(9):1459–67. 10.1007/s00586-006-0294-6

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 19.

  PluemvitayapornTPenratWKunakornsawatSSurapuchongSRatanakoosakulWTiracharnvutKet alCharacteristics of sagittal spinopelvic alignment in Thai population: a cross-sectional study. Spine Deform. (2024) 13:617–23. 10.1007/s43390-024-00997-7

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 20.

  SiccoliAWispelaereMPSchroderMLStaartjesVE. Timing of surgery in tubular microdiscectomy for lumbar disc herniation and its effect on functional impairment outcomes. Neurospine. (2020) 17(1):204–12. 10.14245/ns.1938448.224

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 21.

  DandurandCMashayekhiMSMcIntoshGSinghSPaquetJChaudhryHet alCost consequence analysis of waiting for lumbar disc herniation surgery. Sci Rep. (2023) 13(1):4519. 10.1038/s41598-023-31029-5

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 22.

  ChouRQaseemASnowVCaseyDCrossJTJrShekellePet alDiagnosis and treatment of low back pain: a joint clinical practice guideline from the American college of physicians and the American pain society. Ann Intern Med. (2007) 147(7):478–91. 10.7326/0003-4819-147-7-200710020-00006

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 23.

  BardinLDKingPMaherCG. Diagnostic triage for low back pain: a practical approach for primary care. Med J Aust. (2017) 206(6):268–73. 10.5694/mja16.00828

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 24.

  ChenXSimaSSandhuHSKuanJDiwanAD. Radiographic evaluation of lumbar intervertebral disc height index: an intra and inter-rater agreement and reliability study. J Clin Neurosci. (2022) 103:153–62. 10.1016/j.jocn.2022.07.018

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 25.

  YamadaKAotaYHigashiTIshidaKNimuraTSaitoT. Accuracies in measuring spinopelvic parameters in full-spine lateral standing radiograph. Spine (Phila Pa 1976). (2015) 40(11):E640–6. 10.1097/BRS.0000000000000904

  • Pubmed Abstract
  • CrossRef
  • Google Scholar