Adaptive intensity modulated radiation therapy for pharyngo-laryngeal squamous cell carcinoma pre-clinical validation

Over the years, it has been recognized that using pre-treatment imaging as a “snapshot” of the tumors and/or normal tissues is an oversimplification of reality, as location, morphology and physiology can change during treatment. Indeed, the tumor ecosystem is not static; continuous changes occur in the tumor, particularly in response to radiotherapy (RT), e.g. well-known radiobiological phenomena of re-oxygenation, repopulation, re-assortment of cells within the cell cycle and repair of sub-lethal damage. The dynamic character of the tumor ecosystem and consequential changes in its biological properties cast doubt upon uniform treatment. Performing anatomic and functional imaging prior and during the RT course followed by treatment plan adaption may also tailor temporal and spatial changes in the tumor and become a step forward in treatment individualization. In head and neck (HN) tumors, it has been shown that the tumor volume can dramatically decrease during a 7-week treatment. Similarly, substantial variation in normal-tissue volume has also been reported during RT of other site. Adaptive RT, where patients are re-imaged and re-planned several times during treatment, is a possible strategy to improve treatment delivery by taking into account anatomic modifications over time. First, we make a review of the literature on adaptive RT by reviewing the techniques of adaptive RT and discussing their potential for the maximization of the therapeutic index. Modifications of patient’s anatomy may be evaluated by performing several imaging during treatment. When the shapes or relative positions of target volumes (TV) and organs at risk (OAR) change, simple rigid-body correction techniques may not be sufficient for accurate registration. The solution is the use of deformable registration algorithms (DRA) for the alignment of data sets that are mismatched in a nonlinear or non-uniform way. Applications of DRAs in adaptive RT include 1) automatic time propagation of regions of interest from one image to another also called “Automatic Recontouring” and 2) deformation of the re-planned dose maps to integrate the accumulated dose given to the patient. We describe a strict and objective methodology to compare different algorithms and strategies for deformable registration to be used on images from HN cancer patients. Thereafter, we describe an automatic recontouring method by using the DRA previously validated. Using such a precise and reproductible method, we quantified the volumetric and positional changes in HN gross tumor volumes (GTV), clinical target volumes (CTV) and OARs. By applying the planned treatment on images acquired during the treatment, per-treatment dose maps are obtain. In order to be able to accumulate the dose received by each patient’s volume elements, the data sets need to be non-rigidly registered. In the case of inter-fraction accumulation, the effects of fractionated irradiation on cells killing are not linear and it is generally accepted that it follows the linear-quadratic model. This means that in terms of cell survival within each patient’s volume element, fractionated doses cannot theoretically be accumulated by a simple addition. In other words, even if we have an exact knowledge of the dose given to each voxel at every fraction and if the dose is accumulated simply by summing all fractions registered to the same reference, the corresponding total dose map could not be representative of the radiobiological effect of the dose as it will lose the information of dose per fraction variation. We evaluate the error made by performing simple summation of dose vs. a method that accumulate the dose while taking biological effects in consideration. At least, we develop an adaptive methodology to try to adapt the dose distribution to the changing anatomy. The objectives of this study were: 1) to assess the consequences of anatomic modification in dose distribution for both TVs and OARs; 2) to assess the potential benefit of adaptive strategies using a DRA applied to both volume and dose variations. As our group has already demonstrated the value of fluoro-deoxy-glucose (FDG) positron emission tomography (PET) for GTV delineation of the primary HN tumors and its potential consequences on dose distribution, we also include a comparison between adaptive RT based on anatomic (computed tomography - CT) and functional (FDG PET) imaging.(MED 3) -- UCL, 201