Category: Bone Densitometry or Body Composition

24 - Mirror scans are not an accurate reflection of a total body composition study.

Background: Mirror scans have been proposed to estimate body composition when subjects do not fit within the scan window. This study examines the effectiveness of a mirror scan analysis attempting to estimate full body bone, lean and fat in individual subjects.

A total body scan was obtained on 15 subjects weighing between 79kg and 235kg on the Norland Elite scanner. Data on the left and right side of the raw scan files was shaved off to produce a raw mirror study lacking material on one side or the other. This process provided undisturbed and identical raw data to the three copies that could undergo analysis to reveal the true affect of missing left or right side data. The three copies of the scan--Full, Missing Left and Missing Right--were analyzed to compute bone mineral content, lean mass and fat mass. Agreement between the total body assessment by the Full Study and Missing Left or Missing Right Studies was assessed by Linear Regression and Equality Plots.

Regressions showed that the Missing Left or Missing Right Studies showed significant relationships to the Full Studies for Bone (Left: y = 0.858X + 761.12, r = 0.9758; Right: y = 0.935X + 84.854, r = 0.9803), Lean (Left: y = 0.8795X + 6705, r = 0.9726; Right: y = 0.851X + 1832.8, r = 0.9841) and Fat (Left: y = 1.0238X + 3835.4, r = 0.9981; Right: y = 0.8998X + 4520, r = 0.9962). Nonetheless, Missing Left and Missing Right Studies did not show equality to Full Study results for Bone (Left: Difference between +6.7% to -17.1%; Right: Difference +9.8% to -5.8%), Lean (Left: Difference +13.5% to -2.7%; Right: Difference +17.7% to +6.3%) and Fat (Left: Difference -2.6% to -49.5%; Right: Difference +3.6% to -34.9%). Most of the error introduced by the Missing Left or Missing Right Studies seemed to originate in assessing composition as seen in the Full Body Weight regressions (y = 0.9842X - 416.26, r = .9997) where the Mirror Studies showed a tighter equality between +3.9% and 0.0%.

While Mirror Studies show reasonably good relationships to results obtained from a Total Body Study, composition results obtained on Mirror Studies can prove to be faulty. Given the substantial composition differences seen in the Left and Right Side Mirror Studies we suggest that the DXA study--a powerful tool in evaluating body composition--apply only the true Total Body Study when assessing DXA-based body composition studies for individual subjects.

Patrick Cunniff

Applications Specialist
Norland at Swissray
Fort Atkinson, Wisconsin

Chad A Dudzek

Software Engineer
Norland at Swissray
Fort Atkinson, Wisconsin

Joe Joyce

Director of Research Markets
Norland at Swissray
Fort Atkinson, Wisconsin

Jingmei Wang

Director of Densitometry Research
Norland at Swissray
Beijing, Beijing, China (People's Republic)

Tom V. Sanchez

Chief Science Officer
Norland at Swissray
Socorro, New Mexico