Generation of prostate IMAT plans adaptable to the inter-fractional changes of patient geometry

Properties of the anisotropy of dose contributions: A planning study on prostate cases

Purpose

To characterize the static properties of the anisotropy of dose contributions for different treatment techniques on real patient data (prostate cases). From this, we aim to define a class of treatment techniques with invariant anisotropy distribution carrying information of target coverage and organ‐at‐risk (OAR) sparing. The anisotropy presumably is a helpful quantity for plan adaptation problems.

Methods

The anisotropy field is analyzed for different intensity modulated radiotherapy (IMRT) and volumetric modulated arc therapy (VMAT) techniques for a total of ten planning CTs of prostate cases. Primary irradiation directions ranged from 5 to 15. The uniqueness of anisotropy was explored: In particular, the anisotropy distribution inside the planning treatment volume (PTV) and in its vicinity was investigated. Furthermore, deviations of the anisotropy under beam rotations were explored by direct plan comparison as an indicating the susceptibility of each planned technique to changes in the geometric plan configuration. In addition, plan comparisons enabled the categorization of treatment techniques in terms of their anisotropy distribution.

Results

The anisotropy profile inside the PTV and in the transition between OAR and PTV is independent of the treatment technique as long as a sufficient number of beams contribute to the dose distribution. Techniques with multiple beams constitute a class of almost identical and technique‐independent anisotropy distribution. For this class of techniques, substructures of the anisotropy are particularly pronounced in the PTV, thus offering good options for applying adaptation rules. Additionally, the techniques forming the mentioned class fortunately allow a better OAR sparing at constant PTV coverage. Besides the characterization of the distribution, a pairwise plan comparison reveals each technique's susceptibility to deviations which decreases for an increasing number of primary irradiation directions.

Conclusions

Techniques using many irradiation directions form a class of almost identical anisotropy distributions which are assumed to provide a basis for improved adaptation procedures. Encouragingly, these techniques deliver quite invariant anisotropy distributions with respect to rotations correlated with good plan qualities than techniques using few gantry angles. The following will be the next steps toward anisotropy‐based adaptation: first, the quantification of anisotropy regarding organ deformations; and second, establishing the interrelation between the anisotropy and beam shaping.

Anisotropy of dose contributions-An instrument to upgrade real time IMRT and VMAT adaptation?

PURPOSE

To suggest a definition of dose deposition anisotropy for the purpose of ad hoc adaptation of intensity modulated arc therapy (IMRT) and volumetric arc therapy (VMAT), particularly in the vicinity of important organs at risk (OAR), also for large deformations.

METHODS

Beam's-eye-view (BEV) based fluence warping is a standard adaptation method with disadvantages for strongly varying OAR shapes. 2-Step-adaptation overcomes these difficulties by a deeper analysis of the 3D properties of adaptation processes, but requires separate arcs for every OAR to spare, which makes it impractical for cases with multiple OARs. The authors aim to extend the 2-Step method to arbitrary intensity modulated plan by analyzing the anisotropy of dose contributions. Anisotropy was defined as a second term of Fourier transformation of gantry angle dependent dose contributions. For a cylindrical planning target volume (PTV) surrounding an OAR of varying diameter, the anisotropy and the dose-normalized anisotropy were analyzed for several scenarios of optimized fluence distributions. 2-Step adaptation to decreasing and increasing OAR diameter was performed, and compared to a usual fluence based adaptation method. For two clinical cases, prostate and neck, the VMAT was generated and the behavior of anisotropy was qualitatively explored for deformed organs at risk.

RESULTS

Dose contribution anisotropy in the PTV peaks around nearby OARs. The thickness of the "anisotropy wall" around OAR increases for increasing OAR radius, as also does the width of 2-Step dose saturating fluence peak adjacent to the OAR [K. Bratengeier et al., "A comparison between 2-Step IMRT and conventional IMRT planning," Radiother. Oncol. 84, 298-306 (2007)]. Different optimized beam fluence profiles resulted in comparable radial dependence of normalized anisotropy. As predicted, even for patient cases, anisotropy was inflated even more than increasing diameters of OAR.

CONCLUSIONS

For cylindrically symmetric cases, the dose distribution anisotropy defined in the present work implicitly contains adaptation-relevant information about 3D relationships between PTV and OAR and degree of OAR sparing. For more complex realistic cases, it shows the predicted behavior qualitatively. The authors claim to have found a first component for advancing a 2-Step adaptation to a universal adaptation algorithm based on the BEV projection of the dose anisotropy. Further planning studies to explore the potential of anisotropy for adaptation algorithms using phantoms and clinical cases of differing complexity will follow.