ARRT Bone Densitometry cheat sheet

Dual-energy separation: a K-edge filter (typically cerium or samarium) splits the spectrum into low-energy (~70 kVp peak) and high-energy (~140 kVp peak) components; differential attenuation permits soft-tissue subtraction and isolation of bone mineral.

Pencil-beam geometry: single collimated detector in rectilinear raster scanning - slower, high precision, negligible magnification. Fan-beam geometry: linear detector array with a wide beam - faster, higher spatial resolution, slight magnification that varies with object-to-detector distance.

Areal BMD (g/cm²) = bone mineral content (g) ÷ projected area (cm²); because the measurement is projected, a large bone reads higher and a small bone lower at identical true volumetric density.

Typical effective dose is roughly 1–10 µSv for a lumbar spine or hip scan (fan beam at the upper end); scatter is low enough that the operator may remain in the scan room without a lead barrier.

The QC phantom is an anthropomorphic spine phantom with hydroxyapatite-equivalent inserts; its measured BMD is plotted on a control chart on the manufacturer-recommended schedule (usually daily).

A common patient table weight limit is about 300 lb (136 kg); exceeding the stated limit risks table damage and invalid data, so the manufacturer's specification is checked before scanning.

DXA is the reference standard for BMD; QCT reports volumetric density in mg/cm³, and calcaneal QUS reports speed of sound and broadband attenuation - QUS values cannot be converted into DXA T-scores.

Collimation, detector, or alignment faults present as nonuniform phantom BMD or an unstable control chart; phantom BMD trending beyond tolerance indicates service is required before patient scanning.

T-score = (patient BMD − young-adult mean) ÷ young-adult SD. Z-score = (patient BMD − age-matched mean) ÷ age-matched SD.

WHO categories by T-score: −1.0 or above = normal; between −1.0 and −2.5 = low bone mass (osteopenia); −2.5 or lower = osteoporosis; −2.5 or lower with a fragility fracture = severe (established) osteoporosis.

Diagnosis uses the lowest T-score among lumbar spine, femoral neck, total hip, and 33% radius (the radius only when spine and hip are unavailable or invalid).

Z-scores, not T-scores, are reported for premenopausal women, men younger than 50, and children; Z ≤ −2.0 is termed below the expected range for age and prompts secondary workup.

Metal hardware (zippers, buttons, coins, keys, snaps) is removed before acquisition; recent barium or iodinated contrast and recent radionuclide administration can invalidate lumbar spine analysis and follow the facility's deferral interval.

Pregnancy is a relative contraindication: DXA is deferred unless the clinical indication clearly outweighs the small fetal dose.

History needed for interpretation: prior fragility fracture, glucocorticoid use of 3 months or more, parental hip fracture, smoking, alcohol intake, height loss, and previous DXA results.

Standing height and weight are measured at each visit; they are required for reference database matching, Z-score derivation, FRAX calculation, and detection of height loss.

Default lumbar spine region of interest is L1–L4; exclude vertebrae that are fractured, contain focal sclerosis or artifact, have vertebroplasty or laminectomy, or show severe degenerative change.

A minimum of two evaluable lumbar vertebrae is required; if only one remains, the spine result is non-diagnostic and the hip is used instead.

Patient positioning for the spine: supine, centered near L3, with the legs resting on a foam block to reduce lumbar lordosis and open the disc spaces.

Hip acquisition: internal rotation of the femur about 15–25° with the foot secured in the positioning device so the femoral neck lies parallel to the table and is measured at full length.

Reported hip regions are femoral neck, total hip, and trochanter; Ward's area is displayed but is excluded from diagnosis because of poor reproducibility.

When the two hips differ, the lower BMD and lower T-score at the femoral neck or total hip govern diagnosis and monitoring; a single available hip may be used.

Forearm site: 33% radius of the non-dominant arm, a cortical site used when the spine and hip are unavailable, invalid, or uninterpretable.

Vertebral numbering landmarks: lowest ribs align with T12 and the iliac crest with the L4–L5 interspace; consistent labeling is mandatory for serial comparison.

A single vertebra whose BMD differs from its neighbors by more than about 1.0 T-score (roughly 0.1 g/cm²) is inspected for artifact and excluded if the cause is confirmed.

Precision CV (%) = (standard deviation ÷ mean) × 100, derived from repeat in vivo measurements or repeated phantom scans.

Least significant change at 95% confidence = 2.77 × precision CV; a follow-up difference smaller than the LSC lies within measurement error and is not reported as change.

ISCD precision assessment: duplicate scans in 30 patients, or 15 patients repositioned between scans, per site and per technologist; recalculate after a change in technologist, software, or equipment.

Daily phantom values are plotted on a control chart; investigate drift exceeding about 1.5% from baseline or any point outside ±3 SD before scanning patients.

Reference databases: femoral neck T-scores in the U.S. use NHANES III data for non-Hispanic white women aged 20–29; spine reference data come from the manufacturer - the same database version must be used for serial scans.

Cross-calibration after equipment replacement or major software change: scan a phantom and/or 10–30 patients on both systems and derive a conversion before comparing serial results.

Repeat DXA is generally not indicated sooner than 2 years unless therapy changes or clinical findings warrant it, since typical annual change often falls below the LSC.

FRAX estimates 10-year major osteoporotic and hip fracture probability from femoral neck BMD, age, sex, BMI, prior fracture, parental hip fracture, smoking, glucocorticoids, rheumatoid arthritis, alcohol, and secondary osteoporosis.

Report documentation: scanner model, software version, technologist, sites analyzed, excluded vertebrae or regions, artifacts, and comparison with prior studies.

Dual-energy separation uses a K-edge filter to produce distinct low- and high-energy spectra.

Pencil beam: single detector, slow, high precision, negligible magnification.

Fan beam: array detector, fast, higher resolution, slight magnification and dose increase.

Areal BMD (g/cm²) = bone mineral content ÷ projected bone area, so it is size-dependent.

Effective dose for spine or hip DXA is roughly 1–10 µSv.

QC phantom is an anthropomorphic spine phantom with hydroxyapatite-equivalent inserts.

Typical table weight limit is about 300 lb (136 kg); verify the manufacturer's specification.

QCT gives volumetric mg/cm³; calcaneal QUS gives SOS and BUA, not interchangeable with DXA.

Phantom BMD drift beyond tolerance indicates a collimation, detector, or alignment problem.

BMD is expressed in g/cm² because DXA measures a projected, areal density.

T-score compares patient BMD to the young-adult mean; Z-score compares to the age-matched mean.

WHO thresholds: normal ≥ −1.0; low bone mass −1.0 to −2.5; osteoporosis ≤ −2.5; severe osteoporosis if a fragility fracture is present.

Diagnosis uses the lowest T-score of spine, femoral neck, total hip, or 33% radius.

Premenopausal women, men under 50, and children are reported with Z-scores, not T-scores.

Metal hardware is removed; recent barium, contrast, or radionuclides can invalidate spine analysis.

Pregnancy is a relative contraindication to DXA.

History items relevant to interpretation include glucocorticoid use, prior fragility fracture, and family history.

Height and weight are recorded at each visit for reference data and FRAX.

DXA involves no injection and typical total visit time is 15–30 minutes.

Standard spine ROI is L1–L4; fractured, sclerotic, or artifact-containing vertebrae are excluded.

At least two evaluable vertebrae are required for a valid spine result.

Legs elevated on a block reduce lumbar lordosis during spine scanning.

Femur is internally rotated 15–25° so the femoral neck is measured at full length.

Reported hip regions: femoral neck, total hip, trochanter; Ward's area is not diagnostic.

Use the lower of the two hips when they disagree.

33% radius of the non-dominant forearm is the alternative site.

Lowest ribs = T12; iliac crest = L4–L5 interspace for vertebral numbering.

TBS is a texture index from the spine image that adds fracture risk information beyond BMD.

CV (%) = (SD ÷ mean) × 100 quantifies precision.

LSC at 95% confidence = 2.77 × precision CV.

ISCD precision protocol: 30 patients with duplicate scans, or 15 patients repositioned.

Daily phantom BMD is plotted on a control chart; ±1.5% or ±3 SD triggers investigation.

Femoral neck T-scores use NHANES III non-Hispanic white women aged 20–29.

Cross-calibration is required after scanner or software change.

Repeat scans generally are not done before 2 years unless clinically indicated.

FRAX uses femoral neck BMD plus clinical risk factors to give 10-year fracture probability.

Reports must document scanner, software version, sites, exclusions, and artifacts.

K-edge filter materials commonly used in DXA include cerium and samarium.

Pencil-beam scanning produces essentially no magnification of the measured region.

Fan-beam magnification increases as the object moves farther from the detector.

Bone mineral content is reported in grams; BMD in g/cm²; area in cm².

DXA dose is far lower than a chest radiograph (about 100 µSv).

The spine phantom is scanned on a fixed schedule to verify scanner stability.

Calcaneal QUS predicts fracture but is not used for diagnosis or monitoring by DXA criteria.

Table weight and height limits are published by the manufacturer and must be observed.

T-score = (patient BMD − young-adult mean BMD) ÷ young-adult SD.

Z-score = (patient BMD − age-matched mean BMD) ÷ age-matched SD.

Z ≤ −2.0 is below the expected range for age.

Exclude metal objects before scanning to avoid focal BMD artifact.

Defer spine scanning after recent barium or radionuclide studies per protocol.

Document height, weight, and relevant medications such as glucocorticoids.

Explain that DXA measures bone density and estimates fracture risk, not current fracture.

A 33% radius T-score may be used for diagnosis only when spine and hip are unusable.

Premenopausal women should not be diagnosed with osteoporosis from a T-score alone.

Vertebrae with vertebroplasty cement or surgical hardware are excluded from the spine ROI.

Degenerative sclerosis and aortic calcification artificially raise spine BMD.

The femoral neck should appear long and narrow on a well-positioned hip image.

Total hip ROI includes the femoral neck, trochanter, and intertrochanteric region.

The 33% radius site is predominantly cortical bone.

Inconsistent vertebral labeling causes comparison of different vertebrae over time.

A vertebra differing by more than about 1.0 T-score from neighbors is checked for artifact.

Low TBS indicates degraded trabecular microarchitecture at a given BMD.

Single-energy or mis-centered scans produce non-diagnostic images and require repeat acquisition.

Precision is expressed as CV in percent or as absolute SD in g/cm².

LSC must be exceeded before a BMD change is called significant.

Precision values are site-specific and technologist-specific.

Control charts use Shewhart-type rules to detect drift and shifts.

The same reference database version is maintained for all serial scans.

Conversion formulas from cross-calibration are stored with the patient record.

Clinical risk factors beyond BMD are captured for FRAX reporting.

Software upgrades require repeat phantom scanning and review of prior comparisons.

Archived scan files should be retained per facility policy for future re-analysis.