Adaptive Planning and Delivery to Account for Anatomical Changes Induced by Radiation Therapy of Head and Neck Cancer

Reviewer: Charles Wood, MD
Abramson Cancer Center of the University of Pennsylvania
Last Modified: October 18, 2005

Presenter: M.B. Sharpe
Presenter's Affiliation: Princess Margaret Hospital, Toronto, ON
Type of Session: Plenary


Radiation-induced changes manifested during treatment can theoretically increase the risk of local failure and alter predicted normal tissue toxicity. For this reason, a framework is necessary to assess and potentially intervene in instances of organ movement and deformation. This study assessed the potential benefits of an online and offline adaptive planning system for treatment of patients with head and neck cancer.

Materials and Methods

  • An adaptive framework was developed that utilized the components of a commercially-available treatment planning system (image registration, dose computation, IMRT optimization) and added software elements to support deformable organ modeling, dose evaluation in a changing patient geometry, and replanning techniques taking into consideration dose previously delivered; these tools were employed to allow simulation of different adaptation scenarios
  • Pretreatment CT and MRI were acquired for 16 head and neck cancer patients and, weekly MRI for each patient was performed during radiotherapy
  • Pretreatment CT and MRI images were fused and manually contoured according to RTOG H-0022
  • Weekly MRI images were likewise manually contoured, and the pretreatment MRI was propagated and conformed to the corresponding structures from the weekly MRI images to represent a 4D model of targets and clinical structures during the radiation course
  • IMRT was planned to meet the H-0022 dose-volume criteria, and the dose distribution corresponding to the geometry of each MRI was determined
  • 3 clinical scenarios were considered:
    • IMRT delivery with systematic and random variations over 30 treatment fractions
    • IMRT delivery with daily online correction of isocenter position
    • IMRT delivery with daily online correction of isocenter position and weekly offline adaptive replanning based on weekly MRI scanning
  • The latter 2 scenarios were repeated without a PTV margin to assess for limiting sensitivity of online/offline adaptation


  • Set-up uncertainty and anatomy deformation produced significant dose variations over the radiation treatment course
  • In cases in which a PTV margin was included, CTV coverage was generally assured with online correction; however, online correction resulted in increases (up to 30%) in mean salivary gland dose and increases (up to 10%) in maximum cord dose versus the initial plan
  • The mixed online/offline strategy produced equivalent CTV coverage without the necessity of a PTV margin and reduced the maximum cord dose and mean salivary gland dose by 5% and 12%, respectively

Author's Conclusions

  • Accurate dose accumulation data is required for proper treatment assessment
  • PTV margins achieve target coverage but at a cost of increased normal tissue dose
  • Implementation of an online/offline adaptive framework can preclude the necessity of a PTV margin and reduce dose to adjacent normal tissue

Clinical/Scientific Implications

This novel online/offline adaptive framework illustrates the dosimetric importance of changes in patient anatomy during treatment. However, although dosimetric gains were achieved, the clinical relevance of these endpoints is unclear. Further study is warranted to evaluate the significance of such dosimetric improvements to justify an approach that is considerably more labor-intensive.