3.6 Areal BMD & Projection Geometry
Key Takeaways
Areal BMD divides estimated BMC in grams by projected area in cm².
Fan-beam magnification depends on the source-to-object geometry and device. The BMD ratio may change less than its individual components.
An adult height or areal DXA value does not directly supply volumetric density.
For pediatric short stature or growth delay, adjust spine using BMAD or height Z-score and total body less head using height Z-score.
Follow the path from the bone image to areal BMD
On a dual-energy X-ray absorptiometry (DXA) scan, software identifies the bone within the region of interest (ROI). It estimates the bone mineral content (BMC) in grams and the area of the bone projected onto the image in square centimeters. The report's bone mineral density is areal BMD, calculated as
| Example | BMC | Projected area | Calculated areal BMD |
|---|---|---|---|
| First region of interest | 36.0 g | 48.0 cm² | 0.750 g/cm² |
| Second region of interest | 36.0 g | 45.0 cm² | 0.800 g/cm² |
Both regions have the same estimated mineral mass. The second gives a larger areal density because its projected area is smaller. The difference by itself says nothing about why the areas differ. The images might cover different anatomy or the software might draw a different boundary. Review the actual images before using the difference as evidence of mineral change.
The projection does not contain an independent measurement of depth. A small bone can have less projected mineral per area even with the same volumetric density as a larger bone. A calibrated quantitative CT estimate in mg/cm³ answers a different question and cannot be derived by rescaling the DXA mass units. Body size can influence interpretation, but there is no fixed change to a DXA result based on an adult's height alone.
Check the acquisition geometry
A pencil-beam system samples a narrow beam as it moves across the measurement area. This reduces divergent-beam magnification compared with a broad fan. A fan-beam system can sample multiple detector elements at once, which can shorten acquisition and improve image detail. The scan path, timing and detector design depend on the instrument and mode. Narrow and wider fans differ in how much anatomy is covered in a pass and in geometric and scatter effects.
For a divergent projection, an object closer to the X-ray source projects larger on the detector when source-to-detector distance is fixed. The relationship is to source-to-object distance. Moving farther from the table can put the object closer to the source on one design and farther away on another. Follow the scanner's positioning protocol and reproduce the prior position when comparing studies. Changing a support or the position of the hip can change both geometry and the ROI captured.
Worked comparability decision
In a teaching example, a projected area increases from 48.0 to 50.4 cm² and estimated BMC increases from 36.0 to 37.8 g. Both rise by 5%. The resulting aBMD is still 0.750 g/cm². Equal proportional changes cancel in the ratio. This illustrates why a nearly unchanged aBMD does not guarantee identical projected geometry or measurements. A study of fan-beam magnification likewise found that BMC and area can change together. A real comparison still requires the same valid region and a technically comparable acquisition.
If the area changes while BMC does not, inspect anatomical coverage, rotation, motion and edge detection. Correct the analysis when the acquisition supports it. Repeat a technically inadequate acquisition only if the question warrants it. Document the ROI and measurement units. A lower T-score in a different region does not show that BMD fell in the original ROI. Use the site's least significant change after establishing that the examinations can be compared. A precision threshold cannot correct a mismatch in anatomy or technique.
Interpret growth and short stature appropriately
A 10-year-old with growth delay has an unadjusted lumbar spine aBMD Z-score of −2.9. The measured value is areal, and a smaller vertebral projection can contribute to a low score. The 2019 International Society for Clinical Densitometry (ISCD) pediatric positions call for adjustment in children with short stature or growth delay. Use either bone mineral apparent density (BMAD) or a height Z-score adjustment at the spine, and use a height Z-score adjustment for total body less head. Record the adjustment method in the report. BMAD estimates a size-adjusted quantity from a model. It does not reconstruct a directly measured vertebral volume.
Interpret the adjusted pediatric result with fracture history, scan quality and clinical context. A low pediatric Z-score alone cannot establish osteoporosis. A verified vertebral compression fracture without local disease or high-energy trauma can establish the pediatric diagnosis independently of BMD.
ISCD Pediatric Official Positions · International DXA technical guideline · Fan-beam geometry study
Two ROIs have BMC of 36.0 g and projected areas of 48.0 and 45.0 cm². What is the areal BMD of the second ROI, and what does the calculation establish?
0.750 g/cm² and identical geometry.
0.800 mg/cm³ and a higher volumetric density.
0.800 g/cm². Review the ROI and images before inferring mineral change.
1.250 g/cm² and a 5% increase in mineral mass.
In a fan-beam projection with the detector and X-ray source fixed, what happens when an object moves toward the source?
Its projected image gets larger. The effect of changing table height depends on the source location.
Its image always gets smaller because the object is farther from the detector.
There is no magnification because every ray is parallel.
Projected size changes only if the patient has gained mineral.
A 10-year-old with growth delay has an unadjusted lumbar spine areal BMD Z-score of −2.9. What is the appropriate next step for interpreting the spine result?
Assign WHO osteoporosis from the Z-score.
Apply the appropriate BMAD or height Z-score adjustment and interpret with fracture history and quality.
Convert the reported g/cm² directly to mg/cm³.
Use the lowest hip T-score instead.
Sections you finish are checked off in the contents.