The reliability of optimization under dose-volume limits☆

摘要: Abstract Purpose: An optimization algorithm improves the distribution of dose among discrete points in tissues, but tolerance depends on across a continuous volume. This report asks whether an exact can be completed when enough are taken to accurately model dose-volume constraint. Methods and Materials: Trials were performed using 3-dimensional conformal therapy lung cancer Trial repeated with different limits placed fraction which could receive > 20 Gy. Bounds cord target inhomogeneity. A mixed integer was used find feasible set beam weights would maximize tumor dose. Tests feasibility optimality introduced check solution accuracy. Results: Solutions optimal for tissues. accuracy 3–4% volume condition obtained models 450–600 points. The error improved 2% 800 lung. Solution times increased six-fold at this level Conclusion: method optimum respect conditions usually acceptable times. If constraints violated by excessive amount, should rerun more