Optimization of high precision stereotactic body radiotherapy with photons and ions for non-small-cell-lung cancer

This work presents a contribution in two different aspects required for the implementation of scanned-beam particle therapy for lung tumors. The first part of this work investigates the reproducibility of the calculated particle therapy dose distribution for early stage non-small cell lung cancer...

Full description

Saved in:
Bibliographic Details
Main Author: Santiago García, Alina
Contributors: Wittig, Andrea (Prof.) (Thesis advisor)
Format: Dissertation
Language:English
Published: Philipps-Universität Marburg 2016
Klinik für Strahlentherapie
Subjects:
Online Access:PDF Full Text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This work presents a contribution in two different aspects required for the implementation of scanned-beam particle therapy for lung tumors. The first part of this work investigates the reproducibility of the calculated particle therapy dose distribution for early stage non-small cell lung cancer (NSCLC) tumors in a clinical scenario. These calculations were carried out based on data sets of patients treated with single dose photon stereotactic body radiotherapy (SBRT) under high frequency jet ventilation (HFJV) in order to achieve near-total tumor fixation. A dosimetric evaluation of calculated proton and carbon ion plans was performed, to fulfill clinical plan acceptance criteria with emphasis on target coverage. By simulating the inter-fractional anatomical changes in a short time scale between planning and delivery-time anatomies as imaged by the planning and localization computed tomography (CT) data sets, we carried out an investigation of the deterioration in target coverage. The anatomical changes (e.g. tumor position, patient setup) were quantified through water equivalent path length (WEPL) calculations within the beam entrance channels and correlated with the loss in dosimetric coverage. In addition, we identified beam and planning settings, which also help to reduce dosimetric deterioration, such as best choice of beam angle, higher number of beams, larger spot sizes and larger allowances for beam spots outside the target. We demonstrated reproducible tumor fixations through HFJV. Such technique warranted excellent target coverage in proton SBRT in the majority of the investigated patients. However, for a minor number of cases, unacceptable dosimetric deviations were observed, illustrating the need for imaging prior to each dose delivery with dedicated protocols, together with the development of intervention thresholds in case of anatomical discrepancies based on their potential impact on the dose distribution. HFJV seems a suitable technique to reduce interplay effects. Newer assisted ventilation techniques which do not require use of anesthesia might be more suitable for fractionated radiotherapy. Biological treatment planning for carbon ion therapy requires a model of the radiobiological effects of high linear energy transfer (LET) radiation. One approach in the context of scanned beam ion therapy is built upon the local effect model (LEM). Within this approach, the description of the radiosensitivity and the behavior versus fractionated photon radiotherapy of both tumor and normal tissue requires input of α/β ratios, usually obtained from in vitro studies. Obtaining tumor-specific, realistic, clinical α/β values is urgently required. This topic is also relevant in hypofractionated photon radiotherapy, where there is an ongoing discussion, if the linear-quadratic (LQ) model represents adequately dose responses at high doses per fraction or if the linear-quadratic-linear (LQ-L) correction is necessary, and which α/β ratio describes better the fractionation effect for NSCLC tumors. The second part of this work presents a review of local control data of early stage NSCLC and models of these dose response data using the LQ and LQ-L approaches. Both, the LQ and LQ-L models can be fitted to clinical normo- and hypofractionated NSCLC outcome data. The LQ-L model yielded a significant value for the Dt of 11.0 Gy for the model based on biologically effective dose (BED) at the isocenter with α/β equal to 10 Gy for the full hypofractionation range; it produced a comparable tumor control probability (TCP) fit to the LQ model. We found a clear dose-effect relationship, which in the high BED region was weaker due to considerable dispersion in the data. For the application of BED (α/β=10 Gy) in the range of 100–150 Gy in three fractions or more, the differences in isoeffects predicted by both models can be neglected. Our findings therefore do not allow us to suggest use of the LQ-L model for an improved fitting compared to the LQ model of local control data in case of hypofractionation. A tentative analysis to establish the optimal α/β ratio in the frame of the LQ model for the full fractionation range did not produce significant estimates, although it showed a trend for α/β values lower than 10 Gy.
DOI:https://doi.org/10.17192/z2017.0393