Author: Rex A. Cardan, Carlos E. Cardenas, John B Fiveash, Joel A. Pogue, Richard A. Popple, Farnaz Rahim Li, Rodney J. Sullivan, Natalie N. Viscariello, Christopher Willey 👨🔬
Affiliation: The University of Alabama at Birmingham, University of Alabama at Birmingham 🌍
Purpose: vBrain radiosurgery toxicity-rate is estimated via clinically-relevant isodose volume (IDV) thresholds according to HyTEC (e.g., V12Gy≤10cc). HyperArcTM (HA) automated planning allows accurate IDV prediction from individual target volume(TV) via simple power laws (IDV=A·TVB), due to high planning consistency. Here, we test the hypothesis that this methodology enables a priori isotoxic prescription dose selection.
Methods: This analysis included 1241 single-isocenter HA plans from our institution (2700 targets without 50% IDV bridging), making it one of the largest single-institutional SRS dosimetry cohorts. Eight IDVs for relative isodose levels (RIL) between 50%-100% were calculated using ratios of HyTEC brain toxicity doses and common prescriptions (e.g., 50% = 12Gy/24Gy). Five RIL were modeled to generate exponential fits of parameters A/B (slope/power) versus RIL, allowing the prediction of three testing RIL power laws to be evaluated. Lastly, HyTEC estimates that 12Gy/14Gy IDV of 5 cc, 10cc and 20cc result in Grade 1-3 toxicity rates of 3.6%/4.1%, 4.8%/6.0%, and 8.6%/12.1% respectively, allowing comparison of clinical data and an analytic solution of prescription dose resulting in toxicity as a function of TV.
Results: IDV for five modelling RIL (50%/66.67%/75%/80%/100%) were well described by power laws (R2≥0.979), which classified targets as below HyTEC IDV thresholds from TV with accuracy/precision≥98.3%/99.1%. This allowed estimation of parameters A(slope) and B(power) for three testing RIL (60%/72.59%/81.33%) via exponential relationships (A=21.26·e(-0.0307·RIL) (R^2=0.983); B=0.721·e(0.0033·RIL) (R^2=0.981)). Finally, differences between predicted and actual prescription doses resulting in HyTEC toxicity rates were ≤0.5Gy/≤0.8Gy/≤1.0Gy for 1fx/3fx/5fx.
Conclusion: By analyzing a very large data set, we have demonstrated that power laws allow for the estimation of IDV from TV, enabling the prescription dose leading to each HyTEC toxicity level to be accurately solved as a function of TV. The results presented enable comprehensive isotoxic planning, and thus fractionation scheme selection and/or prescription dose escalation prior to treatment planning.