Room: AAPM ePoster Library
Purpose: determine the effects of tissue inhomogeneity and physical properties of tandem and ovoid (T&O) applicators on dose distributions in 3D high dose rate (HDR) planning using TG43 and Acuros algorithms.
Methods: Planning was performed using Eclipse BrachyVision 13.6 and five CT-simulation scans. CT-based volumes were contoured using criteria defined by GEC-ESTRO working groups and Viswanathan et al. The High Risk CTV (HR-CTV) and Intermediate Risk CTV (IR-CTV) were optimized to D90 and D100. Plans with manually digitized applicators were created, then calculated using TG43 and Acuros with identical dwell times. Plans using model-based applicator structures were also created, then calculated using TG43 and Acuros with the same dwell times as manually digitized plans. The target D90 and D100, and D2cc of sigmoid colon, rectum, and bladder, were collected for comparisons.
Results: For plans with digitized applicators, average Acuros D90 was 3.4% and 3.3% lower than TG43 for HR-CTV and IR-CTV, respectively, and 1.4 % lower for HR-CTV D100. Average Acuros D2cc of rectum, sigmoid colon, and bladder decreased by 4.0%, 3.4%, and 1.6%, respectively. For plans with model-based applicators, average Acuros D90 was 4.1% and 3.8% lower than TG43 for HR-CTV and IR-CTV, respectively, and 2.5% lower for HR-CTV D100. Average Acuros D2cc of rectum, sigmoid colon, and bladder decreased by 4.8%, 3.4%, and 2.1%, respectively. Use of model-based applicators resulted in statistically significant decreases in predicted dose compared to digitized applicators for all structures excluding IR-CTV D100 and sigmoid colon.
Conclusion: Acuros incorporates tissue inhomogeneity in calculations and its overall dose was lower than that of TG43 algorithm, which is likely due to increased attenuation. Use of model-based applicators incorporates applicator-specific material properties and showed a statistically significant difference compared to manually digitized applicators. Use of Acuros and library applicators may likely be more accurate for predicted dose.