Room: Exhibit Hall
Purpose: Tumor-size is an important factor for predicting outcomes with lung SBRT. The aim of this report is to validate a size-adjusted biologically effective dose (s-BED) model. This model incorporates tumor-size into the standard linear-quadratic (LQ) model and Acuros-based calculations to predict local-control rates for variable fractionation schemes.
Methods: A total of 51 early-stage NSCLC patients who underwent Acuros-based lung SBRT with highly-conformal non-coplanar VMAT at Truebeam using 6MV-FFF beam in our clinic were included. Average tumor-sizes (PTV) generated form 4D-CT scans were 2.7Â±0.8cm(2.0â€“4.2cm), 3.6Â±0.7cm(2.5â€“4.7cm), 3.4Â±0.9cm(2.4â€“5.2cm)and 3.7Â±1.0cm(2.4â€“5.3cm) for corresponding fractionation schemes of 30Gyx1(n=10); 18Gyx3(n=17); 12Gyx4(n=10) and 10Gyx5(n=14), respectively; prescribed to PTV(D95)=100%. Assuming, BED decreases nearly linearly with increasing tumor-diameter; s-BED=BED10-10Ã—L was defined for tumor-diameter, L (centimeter) and Î±/Î²=10Gy. For each dose-fractionation scheme, more realistic s-BED was calculated using PTV-D99 from each DVH and compared to nominal prescription. Crude local-control rates â‰¥ 2-year, as a function of s-BED was computed with: TCP=(EXP[s-BED-TCD50]/k)Ã·(1.0 + EXP[s-BED-TCD50]/k), where parameters TCD50 and k defined the shape of the sigmoid curve.
Results: For s-BED(PTV-D99), our predicted TCP indicates that s-BED values of 88Â±9Gy (1fx), 108Â±8Gy (3fxs), 64Â±9Gy (4fxs) and 58Â±10Gy (5fxs) schemes provided 94Â±2%, 97Â±1%, 88Â±3% and 86Â±4%, on average, chances of tumor-control rates at 2-year, respectively. For s-BED (nominal prescription), our predicted TCP indicates that s-BED values of 93Â±8Gy(1fx), 116Â±7Gy(3fxs), 71Â±9Gy(4fxs) and 63Â±10Gy(5fxs) provided 95Â±1%, 98Â±1%, 91Â±3% and 88Â±3%, on average, chances of tumor-control rates at 2-year, respectivelyâ€“suggesting that nominal prescription overestimated TCP by about 1-3% compared to s-BED(PTV-D99). However, regardless of tumor-size, BED10 overestimated TCP by more than 4%(1fx), 2%(3fxs), 9%(4fxs) and 10%(5fxs), on average, compared to s-BED(PTV-D99).
Conclusion: More accurate prediction of local-control rates is highly desirable in the clinic-potentially allowing for target dose escalation/de-escalation. Larger cohorts of lung SBRT patients will be followed up to determine predicted local-control rates correlate clinical outcomes.