TU-FF-A2-03: Effects of CT Calibration On Tissue Inhomogeneity Correction in Radiotherapy

Abstract

Purpose: To quantify the effect of CTcalibration on tissue inhomogeneity correction in radiotherapy.Method and Materials: Materials with known electron densities relative to water ranging from 0 to 2.34 were scanned with two GELight Speed CTs at the four available KVs of 80, 100, 120 and 140 KV and a GE 9800 CT. The effect of using mismatched calibration curves on calculation of tissue inhomogeneity corrections were evaluated using the effective TMR method. Results: All CTcalibration curves agree with each other between air and water region. Above the density of water, the calibration curves for lower kV were significantly higher than those for higher kV, by up to 560 HU for cortical bone (ρe= 1.69). Thus, using a generic calibration curves would lead to an error in the determination of electron density and hence the tissue inhomogeneity correction. The effect of CT number error (ΔHU) on dose calculation is approximated as ΔD/t [%/cm] = TMR′* ΔHU/2000, for HU > 1000, where t is the physical thickness of the dense material layer and TMR′ is the TMR gradient. For example if a cortical bone is scanned with GELight Speed at 80 KV but the CTcalibration curve of 140 KV is used in dose calculation and the treatmentradiation beam is the 6 MV at field size of 1×1 cm2, the dose calculation error down stream of the bone layer will be about 1.2% per cm of bone. The effect is smaller for the 18 MV beam due to the smaller TMR gradient. Conclusion: For best accuracy of tissue inhomogeneity correction, the KV specific, instead of a generic, CTcalibration curve should be used.