Appositional long bone growth: Implications for measuring cross‐sectional geometry

Objectives As growth at the periosteal and endosteal surfaces varies with age, cross-sectional geometric (CSG) properties derived from periosteal ("solid") contours may not produce comparable results to those from endosteal and periosteal contours ("true"), contrary to findings from adults. Error in CSG properties derived from the "solid" sections is compared with "true" sections in a sample of archeologically derived skeletons with estimated dental ages ranging from 1.5 months to 23.5 years. Materials and Methods Cross sections were extracted from 3D surface models, and endosteal contours were located from biplanar radiographs for 56 femora and 59 humeri. Polar second moment of area (J), cross-sectional shape (I-max/I-min), and polar section modulus (Z(p)) were calculated from solid and true sections. Relationships between solid and true properties were examined with least squares regression. Multiple regression examined the effect of age and % cortical area on solid section CSG error. Results While correlations were high (R-2 = 0.72-0.99, all p < 0.001), the results indicate that solid CSG properties are not within an acceptable error range (%SEE of <= 8.0, and %PE of <= 5.0) of true CSG. Error was most affected by %CA, while estimated age was not correlated with %CA, %PE, or percent difference of true-solid CSG. Discussion Periosteal contours alone should not be used to calculate CSG properties from individuals during the period of growth and development. Variation in bone growth and/or adaptive responses independent of age may account for the inconsistent age effects.