The effect of setup uncertainty on normal tissue sparing with IMRT for head-and-neck cancer

Intensity-modulated radiotherapy (IMRT) is being evaluated in the management of head-and-neck cancers at several institutions, and a Radiation Therapy Oncology Group study of its utility in parotid sparing is under development. There is an inherent risk that the sharper dose gradients generated by IMRT amplify the potentially detrimental impact of setup uncertainty. The International Commission on Radiation Units and Measurements Report 62 (ICRU-62) defined planning organ-at-risk volume (PRV) to account for positional uncertainties for normal tissues. The purpose of this study is to quantify the dosimetric effect of employing PRV for the parotid gland and to evaluate the use of PRV on normal-tissue sparing in the setting of small clinical setup errors.

The optimized nine-beam IMRT plans for three head-and-neck cancer patients participating in an institutional review board approved parotid-sparing protocol were used as reference plans. A second optimized plan was generated for each patient by adding a PRV of 5 mm for the contralateral parotid gland. The effect of these additions on the quality of the plans was quantified, in terms of both target coverage and normal-tissue sparing. To test the value of PRV in a worst-case scenario, systematic translational setup uncertainties were simulated by shifting the treatment isocenter 5 mm superiorly, inferiorly, left, right, anteriorly, and posteriorly, without altering optimized beam profiles. At each shifted isocenter, dose distributions were recalculated, producing a total of six shifted plans without PRV and six shifted plans with PRV for each patient. The effect of setup uncertainty on parotid sparing and the value of PRV in compensating for the uncertainty were evaluated.

The addition of the PRV and reoptimization did not significantly affect the dose to gross tumor volume, spinal cord, or brainstem. In contrast, without any shift, the PRV did increase parotid sparing and reduce coverage of the nodal region adjacent to the parotid gland. As expected, when the plans were shifted, the greatest increase in contralateral parotid irradiation was noted with shifts toward the contralateral parotid gland. With these shifts, the average volume of contralateral parotid receiving greater than 30 Gy was reduced from 22% to 4% when a PRV was used. This correlated with a reduction in the average normal-tissue complication probability (NTCP) from 22% to 7%.

The use of PRV may limit the volume of normal tissue structures, such as the parotid gland, exceeding tolerance dose as a result of setup errors. Consequently, it will be important to incorporate the nomenclature of ICRU-62 into the design of future IMRT studies, if the clinical gains of increased normal-tissue sparing are to be realized.