Reading Spine MRI

Neuroradiology series · Part 2 of 4

Lumbar Central Stenosis

Grading the dural sac by what is inside it — and why that beats measuring it.

Companion to Part 1 and the timeline figures.

1Before you open the study

The cervical question is is the cord in danger. The lumbar question is different, and softer: does this picture explain this patient’s legs?

Nobody dies of lumbar stenosis. That changes what your read is for. You are not triaging a cord at risk — you are deciding whether the imaging supports the story the patient is telling, because the operation is offered for symptoms, not for pictures.

The one exception, and it is absolute. Saddle anaesthesia, new urinary retention or overflow, or new bilateral leg weakness is cauda equina syndrome until disproved — a same-day surgical problem. No grade on this page overrides those three symptoms, and no grade is needed to act on them.

Two habits will save you more than any grading system. Count the levels from the sacrum every time — transitional anatomy is common and a mislabelled level ends up in an operative note. And look at the patient’s age before the pictures: disc degeneration, bulges and facet arthropathy are near-universal past sixty, and finding them tells you almost nothing on its own.

2The seven-step read

Run it in this order every time. Step 1 first, because if the conus is low nothing else on the study means what you think it means.

The seven-step lumbar read Same discipline as the cervical read: front-loaded, so an interruption after step three still leaves you with what matters. 1 · Find the conus first Sagittal T2, mid-line. Count up from the sacrum. Below mid-L2 is abnormal at any age. 2 · Scroll the thecal sac Sagittal T2. Where does the CSF column narrow or vanish? Mark every level. 3 · Grade it, axial T2 Schizas at each narrowed level. The highest-yield single skill in lumbar imaging. 4 · Both foramina Parasagittal T2, left and right. The classic miss — invisible unless you go looking. 5 · Lateral recess Axial. Is the traversing root displaced, deviated or compressed? 6 · Muscle, side to side STIR or T2 fat-sat. Denervation oedema localises the level when the anatomy is ambiguous. 7 · Everything else Marrow, endplates, alignment, pars defects, and anything that is not degenerative. The safety rule — identical to the cervical read Your read exists to raise urgency, never to lower it. Saddle anaesthesia, new retention or bilateral leg weakness is a same-day problem no grade overrides.

3Schizas — the whole system on one page

Seven grades, built from a single observation: what is the ratio of rootlets to cerebrospinal fluid inside the dural sac? Nothing is measured. Nothing is calculated.

The rule that makes it learnable: look inside the sac, not at the bone around it. If you can see cerebrospinal fluid, it is A or B. If you cannot see any, it is C or D. That single split is the clinically important one, and it takes about two seconds.
The complete Schizas classification: schematic and matched MRI for grades A1 to D
Read it top to bottom as one continuous process: the rootlets start settled dorsally in a generous pool of fluid, gradually fill the sac, and finally the fluid disappears altogether.
Schizas et al., Spine 2010;35:1919–24, Figure 1 permission pending

4The four grades that change anything

A1 to A4 differ only in where the rootlets have settled under gravity — a distinction worth knowing exists, and not worth drilling. Learn the four letters.

Axial T2, Schizas grade A
Schizas et al., Spine 2010, Figure 1 permission pending

Grade A — fluid clearly visible

Axial T2 at the disc level

Cerebrospinal fluid is obvious inside the sac and individual rootlets can be picked out. The four subtypes (A1–A4) describe only where the rootlets have settled — dorsally, in a horseshoe, or centrally.

What it means: No or minor stenosis. If this is the narrowest level, the imaging does not support decompression, whatever the measured area says.
Axial T2, Schizas grade B
Schizas et al., Spine 2010, Figure 1 permission pending

Grade B — rootlets fill the sac, but still countable

Axial T2 at the disc level

The rootlets occupy the whole sac and are no longer separated by a clear pool of fluid, but you can still make out individual roots. Some fluid remains, giving the sac a grainy texture rather than a smooth one.

What it means: Moderate stenosis. The grainy look is the giveaway — graininess means fluid is still in there.
Axial T2, Schizas grade C
Schizas et al., Spine 2010, Figure 1 permission pending

Grade C — no fluid at all, fat still present

Axial T2 at the disc level

No rootlets can be identified. The sac is a homogeneous grey disc with no fluid signal anywhere inside. Posterior epidural fat is still visible behind it.

What it means: Severe stenosis. In the Schizas cohort, 20 of 37 surgical patients had this as their narrowest grade.
Axial T2, Schizas grade D
Schizas et al., Spine 2010, Figure 1 permission pending

Grade D — no fluid, no posterior fat either

Axial T2 at the disc level

As grade C — a homogeneous grey sac — and in addition the posterior epidural fat has been squeezed out. Nothing is left between the sac and the posterior arch.

What it means: Extreme stenosis. Not one patient in the low-back-pain-only group had a grade D level.

5Why the picture beats the number

Dural sac cross-sectional area is the conventional measurement: under 100 mm² is called relative stenosis, under 75 mm² absolute. Schizas tested that against morphology, and the number lost.

Where the number and the picture disagree Both directions of error happen, and they cancel out in opposite ways. This is the case for grading morphology rather than measuring area. What the measurement says What the morphology says How often Over-called "66 mm² — below the 75 mm² threshold for absolute stenosis." Grade A. Cerebrospinal fluid and individual rootlets plainly visible. No rootlet impingement at all. 40 levels in 35 patients Schizas 2010, Figure 3 Under-called "Over 100 mm² — above even the relative stenosis line." Grade C. A homogeneous grey sac with no cerebrospinal fluid signal anywhere inside it. 12 patients Schizas 2010, Results Why area measurement fails A cross-sectional area is one number describing a shape it cannot capture. Two sacs of identical area can hold rootlets floating in cerebrospinal fluid, or packed solid with none. The rootlet-to-fluid ratio is the thing that correlates with the patient — so grade it directly.
Axial T2 at L4-L5, dural sac area 66 square millimetres but grade A morphology
Schizas et al., Spine 2010, Figure 3 permission pending

The case that makes the argument

Axial T2 at L4–L5 — the paper’s own worked example

Measured dural sac area: 66 mm². That is below 75, so by the conventional numeric criterion this is absolute stenosis — the most severe label the measurement offers.

Now look at it. Cerebrospinal fluid is plainly visible. The rootlets are not impinged. It is a grade A sac, and grade A is where 23 of 27 low-back-pain-only patients sat.

The sac is small, but it is not full. A narrow sac with room inside it is not the same problem as a wide sac packed solid, and only one of those two facts is visible in a cross-sectional area.
Dural sac area ranges for each Schizas grade, showing wide overlap
The same data as a picture. Every grade spans a wide range of areas, and the ranges overlap heavily — grade C cases appear above 100 mm² while grade A cases appear below 50. There is no cut-off you could draw on this axis that would reproduce the grades.
Schizas et al., Spine 2010, Figure 2 permission pending

6Does the grade actually track the patient?

This is the question that decides whether a grading system is worth your time. Schizas applied it to three groups: people with back pain only, people treated conservatively, and people who went to surgery.

The grade tracks what happened to the patient Narrowest grade per patient, in the three groups Schizas studied. Each bar is one group; width is proportion of that group. Grade A Grade B Grade C Grade D Low back pain only no stenosis symptoms n = 27 A 23 C 3 11% Treated conservatively symptomatic, managed without surgery n = 31 A 16 B 6 C 6 D 3 29% Went to surgery symptomatic, decompressed n = 37 B 3 C 20 D 13 89% C or D Read the right-hand column downwards: 11%, 29%, 89%. The grade is not describing the picture — it is describing the patient.
What to take from this. Grades C and D account for 89% of the surgical group and 11% of the back-pain-only group. That is a real separation, achieved with no measurement and no calculation — just looking inside the sac. It is also why the A1–A4 subtypes are not worth drilling: they all sit in the same clinical bucket.

7Drill

Four axial T2s. Grade each one before you open the answer. Ask only: can I see cerebrospinal fluid inside the sac? — then, if not, is there still fat behind it?

Case 1
Axial T2 lumbar spine, unlabelled drill case 1
Show answer

Grade A

Fluid is clearly visible and rootlets are not impinged. Measured area 66 mm² — which is why this case is in the paper. The number says absolute stenosis; the morphology says A.

Case 2
Axial T2 lumbar spine, unlabelled drill case 2
Show answer

Grade C

A homogeneous grey sac with no fluid signal inside it, but posterior epidural fat is still present behind the sac. No fluid + fat still there = C.

Case 3
Axial T2 lumbar spine, unlabelled drill case 3
Show answer

Grade A3

Rootlets lie dorsally and occupy more than half the sac, but fluid is still clearly visible ventrally. Within grade A — the subtype is A3, which changes nothing clinically.

Case 4
Axial T2 lumbar spine, unlabelled drill case 4
Show answer

Grade D

No fluid inside the sac and no posterior epidural fat either. Nothing between the sac and the posterior arch. That second absence is the whole difference between C and D.

Coronal MR myelogram showing the cerebrospinal fluid column interrupted at two levels
The same patient, seen whole. This is the coronal myelographic sequence from cases 2–4. The bright fluid column is interrupted at exactly the levels graded C and D and flows past the level graded A3. The grading is not an abstraction — it is the block, seen on a single axial slice.
Schizas et al., Spine 2010, Figure 4 permission pending

8Why this system, and not one of the others

Schizas is not the newest — Lee is a year later and performs about as well. It is, by a wide margin, the most cited, and the reasons it won out are worth two minutes because they are reasons about evidence, not fashion.

SystemYearWhat it looks at GradesWhere it stands now
Canal diameter
Verbiest lineage		 1954– Front-to-back diameter of the bony canal 2 bands Superseded. Steurer’s review found cut-offs of <10, <12 and <15 mm all in active use, none of them validated.
Dural sac cross-sectional area 1980s– Measured area of the sac on one axial slice continuous Still quoted in reports. Disagrees with morphology in both directions (§5), and repeat measurements of the same slice vary by 6–9 mm².
Schizas 2010 Rootlet-to-fluid ratio inside the sac, plus dorsal epidural fat 7
A1–A4, B, C, D		 488 citations. κ 0.44 between observers in the original paper; ICC 0.83–0.98 in later work with trained readers.
Lee 2011 Obliteration of the anterior fluid space, and whether individual roots can still be separated 4
0–3		 147 citations. ICC 0.84–0.98 — statistically indistinguishable from Schizas when the two were run head to head.
Nerve root sedimentation sign
Barz		 2010 Do the roots settle dorsally under gravity, or stay suspended? binary An adjunct rather than a grade. Head to head against Schizas: AUC 0.853 vs 0.943, losing mostly on specificity (76% vs 88%).

Why it won

1 · It arrived exactly as measurement was collapsing. One year after Schizas, Steurer published a systematic review of quantitative criteria for lumbar stenosis. It found ten different parameters in use, with incompatible cut-offs, and concluded there were no validated criteria at all — only 4 of 63 primary studies even reported the radiological thresholds they had used to enrol patients. Schizas offered something that needed no threshold.
2 · It was validated against what happened to the patient. Most grading papers report only whether readers agree with each other. Schizas tested the grade against three real groups — back pain only, treated conservatively, and operated (§6). Agreement between readers tells you a system is consistent; only that three-group design tells you it is about anything.
3 · Learnability was tested deliberately, not assumed. The original paper had it applied by junior orthopaedic trainees, and by a surgeon and radiologist in a different country who received only the written description with no briefing. That is an unusual thing to put in your own paper, and it is the reason this is a reasonable skill to teach to an audience of APPs in twenty minutes.
4 · It was published where surgeons read. Spine, not a radiology journal. The people who decide whether to operate encountered it directly.
The honest caveat, because you may hear otherwise. A 2023 review describes the Lee system as “now widely accepted in clinical practice, included in web-based radiology resources, and cited in many studies.” That review is co-authored by Joon Woo Lee, a developer of the Lee system — so treat it as a claim by an interested party rather than a neutral finding. On citation count Schizas leads 488 to 147. Both systems are defensible and they ask nearly the same question.

Practical answer: if your radiologists report in Lee grades, learn Lee. Grade 0–1 maps to Schizas A–B, and grade 2–3 to C–D. The underlying skill — can I see cerebrospinal fluid inside the sac? — is identical, and it is the skill that transfers.

9Honest limits

Interobserver agreement is only fair. Across all readers in the original study, κ was 0.44 between observers and 0.65 within the same observer. Among readers at the unit that devised it, those rose to 0.67 and 0.77 — and the authors saw a learning effect between readings. Two honest conclusions: the system is teachable and improves with practice, and a single grade from a single reader is not a hard number. Where a grade drives a decision, say who graded it.
Call it and act on itDefer it, always
The Schizas grade at each narrowed levelAny mass, marrow replacement or enhancing lesion
Which level is narrowestAnything that is not degenerative in origin
Whether fluid is present inside the sacInfection, and any destructive endplate change
Conus position, and whether it is normalThe final operative level — that is the surgeon’s call
That the imaging does or does not match the legsWhether an area measurement means anything here
The sentence to write. "Grade C at L4–5, grade A3 at L3–4; narrowest level L4–5. Symptoms are bilateral and positional and fit that level." That is a defensible, useful read from an APP. What is not defensible is "severe stenosis" with no level and no grade attached.
Source. Schizas C, Theumann N, Burn A, Tansey R, Wardlaw D, Smith FW, Kulik G. Qualitative grading of severity of lumbar spinal stenosis based on the morphology of the dural sac on magnetic resonance images. Spine 2010;35(21):1919–1924. All grading definitions, reliability figures and group distributions on this page are taken from that paper and were checked against its full text.

Comparison sources (§8). Lee GY, Lee JW, Choi HS, Oh KJ, Kang HS. A new grading system of lumbar central canal stenosis on MRI: an easy and reliable measurement. Skeletal Radiol 2011;40(8):1033–9. · Ko Y-j, Lee E, Lee JW, et al. Clinical validity of two different grading systems for lumbar central canal stenosis: Schizas and Lee classification systems. PLoS ONE 2020;15(5):e0233633. · Steurer J, Roner S, Gnannt R, Hodler J. Quantitative radiologic criteria for the diagnosis of lumbar spinal stenosis: a systematic literature review. BMC Musculoskelet Disord 2011;12:175. · Seo J, Lee JW. Magnetic resonance imaging grading systems for central canal and neural foraminal stenoses. Korean J Radiol 2023;24(3):224–234. · Qian G, et al. Value of nerve root sedimentation sign in diagnosis and surgical indication of lumbar spinal stenosis. BMC Musculoskelet Disord 2023;24:336. · Citation counts from Europe PMC, retrieved for this handout.

Figures reproduced from the same paper pending permission. Every image on this page is clickable.