ACCURACY OF VOLUME MEASUREMENTS OF COMPUTED-TOMOGRAPHY AND MAGNETIC-RESONANCE-IMAGING PHANTOMS BY 3-DIMENSIONAL RECONSTRUCTION AND PRELIMINARY CLINICAL-APPLICATION

RATIONALE AND OBJECTIVES. Using an independent three-dimensional workstation, the accuracy of volume measurements of phantoms was assessed using three-dimensional reconstruction of two-dimensional computed tomography (CT) and magnetic resonance (MR) images. METHODS. Round, cylindrical, and irregularly shaped high-contrast phantoms of known volume were imaged in a water bath. The effect of object contrast on volume estimation was tested using phantoms of known volume diluted serially with a contrast agent. The effect of changing field of view and slice thickness was assessed. A clinical application was performed, in which nine shoulders were injected with a known quantity of contrast material, to test the accuracy of the technique in vivo. RESULTS. A strong paired correlation (r = .99) between estimated and true volumes was obtained for high-contrast phantoms ranging from 17 to 128 mL. The weighted average absolute error was 1.42 mL (MR) and 3.50 mL (CT). Accuracy of the serially diluted 27-mL phantoms was essentially unaffected by contrast differences greater than 133 units (MR) and 102 units (CT). The weighted average absolute error was 1.33 mL (MR) and 1.56 mL (CT). Changing field of view had no effect on accuracy, but increasing the slice thickness resulted in overestimation of volume. The mean error for the clinical application was 4.4% (range: 1.7%-8.3%). CONCLUSION. Under certain circumstances, three-dimensional reconstructive volume estimation can be a convenient and accurate method for volume determination.