In radiotherapy for prostate cancer the dose at the treatment planning for the bladder may be a bad surrogate of the actual delivered dose as the bladder presents the largest inter-fraction shape variations during treatment. This paper presents PCA models as a virtual tool to estimate dosimetric uncertainties for the bladder produced by motion and deformation between fractions. Our goal is to propose a methodology to determine the minimum number of modes required to quantify dose uncertainties of the bladder for motion/deformation models based on PCA. We trained individual PCA models using the bladder contours available from three patients with a planning computed tomography (CT) and on-treatment cone-beam CTs (CBCTs). Based on the above models and via deformable image registration (DIR), we estimated two accumulated doses: firstly, an accumulated dose obtained by integrating the planning dose over the Gaussian probability distribution of the PCA model; and secondly, an accumulated dose obtained by simulating treatment courses via a Monte Carlo approach. We also computed a reference accumulated dose for each patient using his available images via DIR. Finally, we compared the planning dose with the three accumulated doses, and we calculated local dose variability and dose-volume histogram uncertainties.
In radiotherapy for prostate cancer the bladder presents the largest inter-fraction shape variations during treatment resulting in random geometric uncertainties that may increase the risk of developing side-effects. In this setting, our interest is thus to propose a hierarchical population model, based on longitudinal data, to characterize bladder motion and deformation between fractions. This method is based on a principal component analysis (PCA) of bladder shapes to obtain the dominant eigenmodes that describe bladder geometric variations between fractions. However, PCA may not properly capture the latent structure of complex data like longitudinal data of organs with large inter and intra-patient shape variations. With this, we propose hierarchical modes to separate intra- and inter-patient bladder variability of the longitudinal data following a dimensionality reduction by means of spherical harmonics (SPHARM). The training data base was used to derive a top-level PCA model that describes the entire structure of the bladder surface space. This space was subsequently divided into subspaces by lower-level PCA models that describe their internal structures. The model was evaluated using a reconstruction error and compared with a conventional PCA model following leave-one-out cross validation.
Conference Committee Involvement (1)
Tenth International Symposium on Medical Information Processing and Analysis