5.5 Lumbar Spine Interpretation & Vertebral Exclusions
Key Takeaways
Adult lumbar BMD uses PA L1–L4. An anatomically abnormal level can be excluded if clearly non-assessable or if its T-score differs from adjacent vertebrae by more than 1.0.
Use all evaluable vertebrae. At least two levels are needed for diagnosis or monitoring, and they may be nonadjacent.
For a valid comparison using two or three levels, match the levels and use the facility LSC for that combination.
Review each level before accepting the composite
Use the PA L1–L4 region for adult lumbar bone mineral density (BMD). The International Society for Clinical Densitometry (ISCD) directs the interpreter to use all evaluable vertebrae and exclude those affected by local structural change or artifact. Inspect the image, level labels and region of interest (ROI) alongside the numerical results. A software-generated four-level score may be invalid if one of the included levels is distorted.
An anatomically abnormal vertebra may be excluded if it is clearly abnormal and non-assessable within the system's resolution, or if its T-score differs from adjacent vertebrae by more than 1.0. A discrepancy is a reason to check the anatomy, boundaries and image quality. It does not tell the reader which level is at fault by itself. A clearly non-assessable level can be excluded even if no valid adjacent score remains to calculate a difference.
| Finding | How it can affect the measurement | Interpretation action |
|---|---|---|
| Focal osteophytes or sclerosis | Additional projected mineral may increase the measured BMD | Inspect the severity and locality, and exclude affected levels when they are non-assessable |
| Compressed vertebral body | Changes in projected area and structure can affect BMD | Use the structural finding and technical validity to decide whether the level belongs in the composite |
| Fixation hardware or vertebral augmentation cement | Radiopaque material in the ROI can contaminate the measurement | Omit an affected result and state the reason |
| Incorrect level label or boundary | Samples different anatomy or assigns an incorrect reference | Correct the analysis before deciding which values are usable |
A height-loss percentage used when grading a vertebral deformity does not supply an independent DXA BMD exclusion rule. A compressed vertebra can be clinically important as a fracture finding even when it cannot contribute to the diagnostic BMD.
Use all remaining valid levels
With four evaluable vertebrae, report L1–L4. If one level is unusable, use the three remaining evaluable levels. If two are unusable, use the two remaining evaluable levels. The remaining vertebrae can be nonadjacent. The system derives a composite T-score from the BMD of the retained levels. Do not calculate it by averaging the printed individual T-scores. Name the reported levels and explain the exclusions. If only one level remains, use another valid site for diagnosis. ISCD also advises against using a single vertebra for monitoring.
Worked example with a focal artifact
A postmenopausal woman has T-scores of −2.6 at L1, −2.5 at L2, −0.8 at L3 and −2.4 at L4. L3 is clearly affected by focal osteophytes and sclerosis. Its T-score is 1.7 higher than L2 and 1.6 higher than L4. Exclude L3. The diagnostic spine BMD comes from L1, L2 and L4, including the valid nonadjacent level. A report should give the resulting composite BMD and applicable score without using L3's value as a diagnostic result.
Now consider a patient with hardware at L3 and L4 and a clearly non-assessable wedge deformity at L2. Only L1 is usable. The spine should not determine the densitometric classification. Use the valid femoral neck or total hip if available, and the one-third radius when indicated. Document the reason the spine is unusable. Lateral lumbar BMD is not a substitute for adult densitometric diagnosis.
Match the levels for follow-up
Use the same evaluable levels on both studies. If the prior exam reported all four levels and L3 is newly invalid, determine whether the prior images can be reanalyzed for L1, L2 and L4. Use the facility's least significant change for the resulting combination. Precision can worsen when fewer than four vertebrae are used, even if the retained levels are adjacent. Applying an L1–L4 threshold to a two- or three-level measurement can misstate statistical significance.
For example, the only valid matched levels are L1 and L4. If a facility has established a precision assessment for that pair, the LSC for the pair is the applicable threshold. If the old study cannot be reanalyzed on the same levels, or there is no applicable LSC, document the limitation and avoid a quantitative significance conclusion. This adult BMD method is separate from the rule for eligible trabecular bone score, which uses L1–L4 without excluding individual levels.
ISCD Adult Official Positions · ISCD U.S. Adult DXA Sample Report
A postmenopausal woman has L1, L2, L3 and L4 T-scores of −2.6, −2.5, −0.8 and −2.4. The image shows severe focal osteophytes and sclerosis affecting only L3. What is the correct analysis?
Retain all four levels to preserve the reference database.
Exclude L3 and use the composite BMD from the evaluable L1, L2 and L4 levels.
Exclude L1 as well and report only L2 and L4.
Exclude the entire spine whenever one level is invalid.
A 74-year-old man has hardware that makes L3 and L4 non-assessable and a wedge fracture that makes L2 non-assessable. Only L1 remains valid. What should the interpretation do?
Classify using the T-score from L1.
Combine L1 with one pedicle from the fractured vertebra.
Document the spine limitation and use another valid skeletal site for diagnosis.
Use the L1–L4 composite after estimating the contribution of the hardware.
Sections you finish are checked off in the contents.