RATIONALE AND OBJECTIVES: Urinary calculi are now commonly detected with helical computed tomography (CT), and it has been proposed that stone composition can be determined from CT attenuation values. However, typical scans are made with a beam collimation of 5 mm or more, resulting in volume averaging and reduction in accuracy of attenuation measurement. The authors tested a model for correction of errors in attenuation values, even at section widths larger than the width of the object. MATERIALS AND METHODS: Human urinary stones were scanned with helical CT at different beam collimation widths. A computer model was used to predict the effect of beam width and stone size on accuracy of measured attenuation. RESULTS: At 3-mm collimation, the model corrected the attenuation readings with an underestimation of 12% +/- 1 (compared with values at 1-mm collimation; 127 stones; diameters of 1.7-11.3 mm). With attenuation measured at 10-mm collimation, the model underestimated the true value by 34% +/- 3 (103 stones), with a significant negative correlation with stone diameter on magnitude of error (diameters of 3.0-11.3 mm). Correlation of data from patient scans with subsequent in vitro scanning of the same stones confirmed the validity of the model, but corrected in vivo scans consistently yielded lower values for the stones than in vitro. CONCLUSION: Volume averaging effects on attenuation in helical CT are predictable in vitro for urinary calculi--and presumably for other roughly spherical structures--as long as section width does not excessively exceed the diameter of the structure.
RATIONALE AND OBJECTIVES: Urinary calculi are now commonly detected with helical computed tomography (CT), and it has been proposed that stone composition can be determined from CT attenuation values. However, typical scans are made with a beam collimation of 5 mm or more, resulting in volume averaging and reduction in accuracy of attenuation measurement. The authors tested a model for correction of errors in attenuation values, even at section widths larger than the width of the object. MATERIALS AND METHODS:Human urinary stones were scanned with helical CT at different beam collimation widths. A computer model was used to predict the effect of beam width and stone size on accuracy of measured attenuation. RESULTS: At 3-mm collimation, the model corrected the attenuation readings with an underestimation of 12% +/- 1 (compared with values at 1-mm collimation; 127 stones; diameters of 1.7-11.3 mm). With attenuation measured at 10-mm collimation, the model underestimated the true value by 34% +/- 3 (103 stones), with a significant negative correlation with stone diameter on magnitude of error (diameters of 3.0-11.3 mm). Correlation of data from patient scans with subsequent in vitro scanning of the same stones confirmed the validity of the model, but corrected in vivo scans consistently yielded lower values for the stones than in vitro. CONCLUSION: Volume averaging effects on attenuation in helical CT are predictable in vitro for urinary calculi--and presumably for other roughly spherical structures--as long as section width does not excessively exceed the diameter of the structure.
Authors: Andrew N Primak; Joel G Fletcher; Terri J Vrtiska; Oleksandr P Dzyubak; John C Lieske; Molly E Jackson; James C Williams; Cynthia H McCollough Journal: Acad Radiol Date: 2007-12 Impact factor: 3.173