Reliability of measuring volume by different methods for tumors of the musculoskeletal system

In the present study different methods for determining the volume of a tumor were evaluated. For 12 models, the real volume, the volume according to measurement of the surface area on MRI, an ellipsoid and a cylindrical approximation of the volume, as well as the maximum diameter, were determined. There appeared to be a good correlation between all calculated volumes and the real volume. The error (mean: 17%) and the standard deviation (SD: 14%) on this error were smallest if the volume was determined by means of determination of the surface area. The ellipsoid approximation resulted in a smaller error (mean: 0%) but a higher standard deviation (SD: 27%). The cylindrical approximation resulted in unacceptable deviations (mean: 51%; SD: 40%). Volume was significantly related to the maximal diameter to the power of 2.3. Volume calculated according to this power resulted in an error of 18%. Standard deviation in this case however was unacceptable (SD: 89%). Volume calculation based upon the determination of the surface area has given the best and most reliable results. Ellipsoid approximation was less reliable, but faster and cheaper. Cylindrical approximation was unacceptable. Size, expressed as maximal diameter of the tumor, was also unacceptable as a parameter for volume.