SU-E-T-461: Fractionation Schedule Optimization for Lung Cancer Treatments Using Radiobiological and Dose Distribution Characteristics.

PURPOSE Lung cancer radiotherapy treatments employ a wide variety of fractionation protocols. The choice among protocols mostly depends on the size of the target volume (GTV or ITV) and the volume of normal tissue receiving a critical dose. Rigorous mathematical criteria for normal tissue (NT) dose distributions were derived to determine the type of dose per fraction schedule that maximizes linear-quadratic tumor effect. METHODS Selecting the individual doses per fraction that maximize a linear-quadratic effect in the tumor while constraining the normal tissue complication probability according to the Lyman-Kutcher-Burman model leads to an optimization problem. For time-independent parameters, the solution is always an equal dose per fraction schedule; depending on parameter values, two different class solutions are suggested: minimal number of fractions clinically realized with hypo-fractionation, or minimizing dose per fraction clinically realized with standard- or hyper-fractionation. The value of a single scale-free "bifurcation" number, derived from the DVH of the NT dose distribution suggests which solution is preferred for a given plan with respect to a given normal tissue. The clinical relevance of the bifurcation number in selecting fractionation schemes was tested for 30 patients previously treated for non-small-cell lung cancer according to various fractionation protocols. RESULTS The bifurcation numbers for both lung and esophagus were a good classifier for the hypofractionated and the conventional fractionation groups. The variability of the numbers within patients of the conventional fractionation group was much smaller than the variability of the treated ITV volumes or the ITV to lung volume ratios. The prescribed fractionations were also consisted with the currently accepted alpha-beta values for tumor (10) and radiation-induced pneumonities in the lung (4). CONCLUSIONS Model-based criteria such as the bifurcation number may replace the more empirical volume criteria to decide the optimal fractionation protocol once the dose distribution has been optimized.