Compensating for heterogeneities in proton radiation therapy.

Our method for predicting, and compensating for, the effects of surface irregularities and tissue heterogeneities in proton radiation therapy was evaluated by comparing the predicted and measured dose distributions. Two heterogeneity configurations in a D-shaped water-filled phantom were handled in exactly the same way as patients. Target volumes were designated on thin-section CT scans, a single en face portal was defined, compensating boli were designed and made, and the dose distribution behind the phantom measured and compared with that intended. The compensation was accurate to within 1 mm for the phantom with a single air heterogeneity and to within 2.5 mm for the phantom with multiple bone and air heterogeneities. The bolus and phantom were misaligned by 3 mm and the dramatic change in the dose distribution demonstrated the need to address the problems of patient motion and imperfect immobilisation through compensator design. A philosophy of 'expanding' the bolus is described, and dose distributions measured with the 'expanded' boli indicate that target volume treatment can be assured within prespecified repositioning and motion uncertainties. The uncertainty in the alignment of bolus and heterogeneities leads to corresponding uncertainty in the penetration of the protons. Ranges within which they will stop are calculated and shown to encompass adequately the measured distributions in both the aligned and misaligned cases.