Skip to content

Imaging and grading

TL;DR — MRI describes disc and endplate structure but does not diagnose a pain generator. The five-grade Pfirrmann system is reproducible for morphology (original interobserver κ 0.69–0.81) but is ordinal, age-sensitive and insensitive to early composition (Pfirrmann 2001, PMID 11568697). Degeneration, bulges and protrusions are common in asymptomatic people, so interpretation requires symptoms, examination and pre-test probability (Brinjikji 2015, PMID 25430861). Modic change, high-intensity zones and quantitative MRI can enrich phenotypes, but none is a stand-alone clinical biomarker (Hopayian 2023, PMID 36438174; Russo 2023, PMID 37247638).

When imaging answers a question

Imaging is most valuable when it can:

  • evaluate red flags or severe/progressive neurological deficits;
  • localize suspected nerve-root compression;
  • plan an intervention or surgery;
  • assess postoperative complications;
  • characterize deformity, instability or fracture;
  • establish research phenotype and longitudinal change.

Routine early imaging for uncomplicated nonspecific low-back pain exposes patients to incidental labels without improving the structural specificity of the diagnosis (Maus 2010, PMID 20977958; Qaseem 2017, PMID 28192789).

Modalities

Modality Best contribution Main limitation
Radiograph Alignment, height, osteophytes, dynamic views Poor soft tissue; radiation
MRI T1/T2 Disc, marrow, roots, canal, infection/tumor screen Incidental findings; sequence dependence
CT Bone, calcification, endplate, postoperative fusion Radiation; weak soft-tissue contrast
CT myelography Canal/root when MRI limited Invasive, radiation
DCE-MRI Experimental endplate enhancement Contrast/modeling; sparse validation
T2/T2* mapping Composition/collagen-water environment Protocol heterogeneity
T1ρ Proteoglycan-sensitive signal Limited availability/standardization
Sodium MRI Fixed-charge proxy Low signal, long acquisitions
Diffusion Water mobility Artifacts and uncertain thresholds

Conventional MRI features

Feature What is seen Possible substrate Not established by finding alone
Desiccation Low T2 signal Water/proteoglycan loss Pain
Height loss Narrow disc space Matrix loss/remodeling Instability
Bulge Broad circumferential extension Distributed deformation Root compression
Protrusion Focal base wider than dome Focal displacement Symptomatic herniation
Extrusion Dome wider than base/continuity loss Annular failure Neurological deficit
Sequestration Free fragment Migrated extrusion Need for surgery
Annular fissure/HIZ Focal high T2 signal Fissure/inflammation Discogenic pain
Endplate defect Focal discontinuity/irregularity Mechanical/transport lesion Pain
Modic 1 T1 low, T2 high marrow Edema-like/inflammatory change Infection exclusion or causality
Modic 2 T1 high, T2 iso/high Fatty marrow Active pain
Modic 3 T1/T2 low Sclerosis Symptom severity

Pfirrmann grading

The original system integrates nucleus structure, signal intensity, distinction between nucleus and annulus, and disc height on T2 MRI (Pfirrmann 2001, PMID 11568697).

Grade Structure/signal Nucleus–annulus distinction Height
I Homogeneous bright white Clear Normal
II Inhomogeneous with/without bands; bright Clear Normal
III Inhomogeneous gray Unclear Normal to slightly decreased
IV Inhomogeneous dark gray Lost Normal to moderately decreased
V Inhomogeneous black Lost Collapsed

In the original 300-disc study, intraobserver κ was 0.84–0.90, interobserver κ 0.69–0.81 and exact agreement 83.8%; only 1.3% differed by two or more grades (PMID 11568697).

Independent assessment has generally found acceptable agreement but highlights ambiguity between middle grades (Urrutia 2016, PMID 26879918). Grade is ordinal: the biological distance between II and III is not necessarily equal to IV and V.

Asymptomatic prevalence

Feature Age 20 Age 80 Source
Disc degeneration 37% 96% Systematic review, 3,110 asymptomatic people (PMID 25430861)
Disc bulge 30% 84% Same
Disc protrusion 29% 43% Same
Annular fissure 19% 29% Same

The review pooled scans performed for research or non-pain reasons, with heterogeneous definitions. It nonetheless establishes that degenerative morphology has limited specificity for pain.

A complementary meta-analysis found degeneration and several endplate/disc features more common in symptomatic adults than controls (Brinjikji 2015, PMID 26359154). Imaging therefore modifies probability but cannot close the causal inference.

Modic changes

Modic change describes vertebral marrow adjacent to an endplate, not the disc itself. Types can convert and coexist.

Type MRI signal Common interpretation Natural-history note
1 T1 low/T2 high Edema-like, vascular/inflammatory May convert to type 2
2 T1 high/T2 iso/high Fatty replacement Most common in community cohort
3 T1 low/T2 low Sclerosis Rare

In a community cohort, type 2 affected 27.8% at baseline; 18 incident lesions developed in seven of 56 follow-up participants, and resolution was uncommon (Teichtahl 2017, PMID 27324605).

Systematic review concluded that study heterogeneity prevents treating Modic change as a universal causal diagnosis (Hopayian 2023, PMID 36438174). A contrary phenotype-focused review argues heterogeneous sampling obscures a real association, particularly for type 1 (Czaplewski 2023, PMID 37170132). This disagreement should remain explicit.

High-intensity zone and annular fissure

A posterior annular high-intensity zone may reflect fluid/granulation tissue in a fissure. It is biologically compatible with annular pain but also occurs without symptoms. Reader definition, slice thickness and field strength alter detection (Brayda-Bruno 2014, PMID 23978994).

Longitudinal twin data found some HIZs appear and disappear: new sagittal HIZs in 0.5% of discs and disappearance in 1.6% over five years (Videman 2006, PMID 16540872). A static HIZ is therefore not a permanent molecular label.

Endplate and marrow imaging

Endplate defects may alter load distribution and nutrient transport. Conventional MRI is limited for thin cartilage; CT better depicts sclerosis and osseous defects but adds radiation (Urban 2007, PMID 17260404).

DCE-MRI in nine patients found enhancement differences associated with degeneration grade and spinal level, but the 45-disc study was exploratory and did not validate perfusion thresholds (Muftuler 2015, PMID 25421547).

Quantitative MRI

Technique Candidate biological sensitivity Barriers
T2 mapping Water/collagen organization Magic-angle and sequence effects
T2* mapping Collagen and hydration Field/sequence dependence
T1ρ Proteoglycan-related macromolecules Hardware and standardization
ADC/diffusion Water mobility Motion/distortion
Sodium MRI Glycosaminoglycan fixed charge Low spatial resolution
MR spectroscopy Metabolites/lipids Small voxels and reproducibility
Ultrashort echo Short-T2 endplate/annulus tissues Emerging availability

A systematic review found growing evidence that quantitative techniques detect early compositional changes, but acquisition, segmentation, reference values and longitudinal clinical validity were heterogeneous (Russo 2023, PMID 37247638).

Clinical sodium-MRI investigations continue to test feasibility, but acquisition burden and absence of outcome-linked reference ranges keep the method investigational (Nakahashi 2025, PMID 40522507).

T2 mapping correlates with Pfirrmann grade in small studies but overlapping distributions limit patient-level thresholds (Stefanou 2023, PMID 39119364). Multi-parameter studies have not established a clinically approved pain classifier (Xiong 2018, PMID 29511438).

Automated grading and radiomics

Deep-learning models can automate disc localization and Pfirrmann grading, often using internal retrospective datasets. Performance against reader labels does not prove biological or prognostic validity (Gao 2021, PMID 33094867; Liawrungrueang 2023, PMID 36832151).

Systematic reviews of artificial-intelligence grading likewise emphasize dataset and reference-label limitations (Liawrungrueang 2024, PMID 39659374).

Kinematic MRI can display position-dependent morphology and motion, but trends observed in selected cohorts have not established a routine instability or pain threshold (Roberts 2021, PMID 33940491).

Evaluation should report:

  • external site/scanner validation;
  • patient-level data separation;
  • class imbalance by grade;
  • calibration, not only accuracy;
  • disagreement adjudication;
  • performance by age, sex and ancestry;
  • incremental value over clinical variables;
  • effect on decisions and outcomes.

Reliability and measurement error

Source of variation Control
Time of day/loading Standardize scan timing/activity
Supine unloading Record rest before scan
Field strength/sequence Harmonize acquisition
Slice thickness/angle Protocol and multiplanar confirmation
Reader thresholds Training atlas and blinded repeats
Patient positioning Standard positioning
Ordinal grade ceiling/floor Add continuous measures
Multi-level clustering Use appropriate statistical model

Small apparent longitudinal changes can fall within reader or acquisition error. Studies should predefine a smallest detectable change rather than labeling every grade difference “progression.”

Imaging–pain inference

Evidence combination Causal confidence
MRI finding alone Low
Finding plus compatible distribution Low to moderate
Finding plus exam and exclusion of alternatives Moderate
Finding plus convergent independent marker Potentially higher; marker not validated
Treatment response Supportive but vulnerable to circularity/placebo

Imaging of discogenic and vertebrogenic pain increasingly emphasizes phenotype triangulation rather than one sign (Abel 2024, PMID 38272616).

Reporting template

  1. State level and morphology.
  2. Separate degeneration from displacement.
  3. Describe neural compromise explicitly.
  4. Describe endplate/Modic findings by type.
  5. Compare with prior study using the same feature.
  6. Avoid causal language unless supported clinically.
  7. Note common age-associated findings neutrally.
  8. Identify urgent/nondegenerative abnormalities.

Qualitative synthesis shows patients and clinicians can interpret imaging as definitive damage; report language is therefore an intervention with potential benefit and harm (Sharma 2020, PMID 32830105; Alhowimel 2022, PMID 32746671).

Incidental degeneration in young pregnant women provides another reminder that reproductive-age imaging findings require the same clinical-context rule (Schwarz-Nemec 2020, PMID 31934808).

Evidence deepening: discriminating findings (2026-08-30)

The added evidence below was selected to change interpretation, not merely increase citation count. Each result is kept within its studied phenotype and design.

Evidence Quantified or mechanistic finding Consequence for interpretation
MRI-Based Grading Systems for Assessing Lumbar Disc Degeneration: A Scoping Review (Esposito 2025, PMID 40959741) A scoping review found 569 lumbar-MRI degeneration-grading studies and catalogued which systems had reliability, validity and sensitivity-to-change evidence. Frequency of use is not equivalent to validated longitudinal responsiveness.
The correlation between the high-intensity zone on a T2-weighted MRI and positive outcomes of discography: a meta-analysis (Fang 2017, PMID 28178999) A meta-analysis associated high-intensity zones with abnormal disc morphology (OR 47.79, 95% CI 17.07–133.77) and pain reproduction (OR 8.65, 95% CI 6.01–15.23). Large enrichment against discography does not overcome incorporation and selection bias.
Ultrashort time-to-echo MR morphology of cartilaginous endplate correlates with disc degeneration in the lumbar spine (Finkenstaedt 2023, PMID 37195362) Cadaveric 3-T ultrashort-echo imaging evaluated 547 cartilage endplates adjacent to 284 discs. Endplate morphology can be measured beyond conventional spin-echo signal, but clinical pain validity remains separate.
Robust Radiomic Signatures of Intervertebral Disc Degeneration From MRI (McSweeney 2025, PMID 40539398) A population-cohort radiomics study analyzed 1,397 adults and extracted 737 features after automated disc segmentation. High-dimensional signatures need external validation and prespecified clinical utility.
Results of an International Survey on Spinal Imaging by the ASNR/ASSR/ESNR/ESSR "Nomenclature 3.0" Working Group (D'Anna 2023, PMID 37816364) An international survey obtained 600 responses across 63 countries about whether lumbar-disc nomenclature 2.0 requires revision. Even consensus terminology needs adoption and update surveillance.
Revalidating Pfirrmann's Magnetic Resonance Image-Based Grading of Lumbar Nerve Root Compromise by Calculating Reliability among Orthopaedic Residents (Kaliya-Perumal 2018, PMID 29854345) Five orthopaedic residents graded 50 nerve-root images with mean agreement 80%±15.1%. Reader agreement varies by task and experience; nerve-root compromise grades should not be conflated with disc-degeneration grades.

Controversy carried forward

These additions narrow several claims but do not create a diagnostic gold standard. Where an imaging, molecular or treatment-response signal conflicts with sham-controlled, longitudinal or population evidence, the conflict is retained as a selection and transportability problem rather than resolved by vote.

Open questions

  • Can harmonized quantitative MRI predict pain persistence beyond age and conventional morphology? (Russo 2023, PMID 37247638)
  • Which Modic transitions predict symptoms prospectively? (Teichtahl 2017, PMID 27324605)
  • Does time-standardized loading-sensitive imaging improve reproducibility? (Urban 2007, PMID 17260404)
  • Can external validation distinguish automated label replication from clinically useful prediction? (Gao 2021, PMID 33094867)
  • Which reporting language reduces fear without reducing recognition of concordant pathology? (Sharma 2020, PMID 32830105)

References

  1. Pfirrmann CW, Metzdorf A, Zanetti M, et al. Magnetic resonance classification of lumbar intervertebral disc degeneration. Spine. 2001;26(17):1873-8. PMID 11568697
  2. Brinjikji W, Luetmer PH, Comstock B, et al. Systematic literature review of imaging features of spinal degeneration in asymptomatic populations. AJNR. American journal of neuroradiology. 2015;36(4):811-6. PMID 25430861
  3. Hopayian K, Raslan E, Soliman S. The association of modic changes and chronic low back pain: A systematic review. Journal of orthopaedics. 2023;35:99-106. PMID 36438174
  4. Russo F, Ambrosio L, Giannarelli E, et al. Innovative quantitative magnetic resonance tools to detect early intervertebral disc degeneration changes: a systematic review. The spine journal : official journal of the North American Spine Society. 2023;23(10):1435-1450. PMID 37247638
  5. Maus T. Imaging the back pain patient. Physical medicine and rehabilitation clinics of North America. 2010;21(4):725-66. PMID 20977958
  6. Qaseem A, Wilt TJ, McLean RM, et al. Noninvasive Treatments for Acute, Subacute, and Chronic Low Back Pain: A Clinical Practice Guideline From the American College of Physicians. Annals of internal medicine. 2017;166(7):514-530. PMID 28192789
  7. Urrutia J, Besa P, Campos M, et al. The Pfirrmann classification of lumbar intervertebral disc degeneration: an independent inter- and intra-observer agreement assessment. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 2016;25(9):2728-33. PMID 26879918
  8. Brinjikji W, Diehn FE, Jarvik JG, et al. MRI Findings of Disc Degeneration are More Prevalent in Adults with Low Back Pain than in Asymptomatic Controls: A Systematic Review and Meta-Analysis. AJNR. American journal of neuroradiology. 2015;36(12):2394-9. PMID 26359154
  9. Teichtahl AJ, Finnin MA, Wang Y, et al. The natural history of Modic changes in a community-based cohort. Joint bone spine. 2017;84(2):197-202. PMID 27324605
  10. Czaplewski LG, Rimmer O, McHale D, et al. Modic changes as seen on MRI are associated with nonspecific chronic lower back pain and disability. Journal of orthopaedic surgery and research. 2023;18(1):351. PMID 37170132
  11. Brayda-Bruno M, Tibiletti M, Ito K, et al. Advances in the diagnosis of degenerated lumbar discs and their possible clinical application. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 2014;23 Suppl 3:S315-23. PMID 23978994
  12. Videman T, Battié MC, Ripatti S, et al. Determinants of the progression in lumbar degeneration: a 5-year follow-up study of adult male monozygotic twins. Spine. 2006;31(6):671-8. PMID 16540872
  13. Urban JP, Winlove CP. Pathophysiology of the intervertebral disc and the challenges for MRI. Journal of magnetic resonance imaging : JMRI. 2007;25(2):419-32. PMID 17260404
  14. Muftuler LT, Jarman JP, Yu HJ, et al. Association between intervertebral disc degeneration and endplate perfusion studied by DCE-MRI. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 2015;24(4):679-85. PMID 25421547
  15. Nakahashi Y, Saida T, Yoshida M, et al. Evaluation of lumbar intervertebral disc degeneration using 23Na-MRI in clinical settings. Skeletal radiology. 2025;54(12):2725-2734. PMID 40522507
  16. Stefanou DE, Velonakis G, Karavasilis E, et al. T2 mapping of lumbar intervertebral disc: quantitative evaluation of degeneration in relation to Pfirrmann grading system and a template for intervertebral disc segmentation. Hippokratia. 2023;27(3):75-81. PMID 39119364
  17. Xiong X, Zhou Z, Figini M, et al. Multi-parameter evaluation of lumbar intervertebral disc degeneration using quantitative magnetic resonance imaging techniques. American journal of translational research. 2018;10(2):444-454. PMID 29511438
  18. Gao F, Liu S, Zhang X, et al. Automated Grading of Lumbar Disc Degeneration Using a Push-Pull Regularization Network Based on MRI. Journal of magnetic resonance imaging : JMRI. 2021;53(3):799-806. PMID 33094867
  19. Liawrungrueang W, Kim P, Kotheeranurak V, et al. Automatic Detection, Classification, and Grading of Lumbar Intervertebral Disc Degeneration Using an Artificial Neural Network Model. Diagnostics (Basel, Switzerland). 2023;13(4). PMID 36832151
  20. Liawrungrueang W, Cholamjiak W, Sarasombath P, et al. Artificial Intelligence Classification for Detecting and Grading Lumbar Intervertebral Disc Degeneration. Spine surgery and related research. 2024;8(6):552-559. PMID 39659374
  21. Roberts S, Gardner C, Jiang Z, et al. Analysis of trends in lumbar disc degeneration using kinematic MRI. Clinical imaging. 2021;79:136-141. PMID 33940491
  22. Abel F, Altorfer FCS, Rohatgi V, et al. Imaging of Discogenic and Vertebrogenic Pain. Radiologic clinics of North America. 2024;62(2):217-228. PMID 38272616
  23. Sharma S, Traeger AC, Reed B, et al. Clinician and patient beliefs about diagnostic imaging for low back pain: a systematic qualitative evidence synthesis. BMJ open. 2020;10(8):e037820. PMID 32830105
  24. Alhowimel A, Alotaibi M, Coulson N, et al. Psychosocial consequences of diagnosing nonspecific low-back pain radiologically: a qualitative study. Physiotherapy theory and practice. 2022;38(7):890-896. PMID 32746671
  25. Schwarz-Nemec U, Friedrich KM, Prayer D, et al. Lumbar Intervertebral Disc Degeneration as a Common Incidental Finding in Young Pregnant Women as Observed on Prenatal Magnetic Resonance Imaging. Journal of women's health (2002). 2020;29(5):713-720. PMID 31934808
  26. Esposito D, Brown B, Hancock MJ, King SSG, Searant IGT, Jenkins H, et al. MRI-Based Grading Systems for Assessing Lumbar Disc Degeneration: A Scoping Review. JOR spine. 2025;8(3):e70113. PMID 40959741
  27. Fang C, Zhang W, Chen L, Li H. The correlation between the high-intensity zone on a T2-weighted MRI and positive outcomes of discography: a meta-analysis. Journal of orthopaedic surgery and research. 2017;12(1):26. PMID 28178999
  28. Finkenstaedt T, Siriwananrangsun P, Masuda K, Bydder GM, Chen KC, Bae WC, et al. Ultrashort time-to-echo MR morphology of cartilaginous endplate correlates with disc degeneration in the lumbar spine. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 2023;32(7):2358-2367. PMID 37195362
  29. McSweeney T, Tiulpin A, Kowlagi N, Määttä J, Karppinen J, Saarakkala S, et al. Robust Radiomic Signatures of Intervertebral Disc Degeneration From MRI. Spine. 2025;50(24):1737-1746. PMID 40539398
  30. D'Anna G, Shah L, Kranz PG, Hirsch JA, Khan M, Johnson M, et al. Results of an International Survey on Spinal Imaging by the ASNR/ASSR/ESNR/ESSR "Nomenclature 3.0" Working Group. Seminars in musculoskeletal radiology. 2023;27(5):561-565. PMID 37816364
  31. Kaliya-Perumal AK, Ariputhiran-Tamilselvam SK, Luo CA, Thiagarajan S, Selvam U, Sumathi-Edirolimanian RP, et al. Revalidating Pfirrmann's Magnetic Resonance Image-Based Grading of Lumbar Nerve Root Compromise by Calculating Reliability among Orthopaedic Residents. Clinics in orthopedic surgery. 2018;10(2):210-215. PMID 29854345