Abstract Book

ESTRO 37

S550

CT, then all the images were exported to 3D Slicer v4.5.0 where rigid registration was performed and the ROIs were translated from the reference CT to the second. Next the images and ROIs of both CT studies were imported into IBEX v1.0 where the RFs were calculated. After selecting the most reproducible RFs based on the concordance correlation coefficient (CCC) and the coeffcient of variation (CV), an hierarchical cluster analysis was carried out to discard redundant features. A total of 177 RFs pertaining to three main categories (Intensity, Shape and Texture) were evaluated. In the Texture group in turn we calculated 3 different subgroups related to 3 calculated matrices: Neighborhood Intensity Difference Matrix, Gray Level CoOcurrence Matrix and Gray Level Run Length Matrix.

PO-0990 A Clinical Decision Support Tool Based on Active Appearance Modelling for Prostate Segmentation K. Cheng 1 , H. Li 2 , D. McLaren 3 , S. McLaughlin 4 , W. Nailon 5 1 Beijing University of Posts & Telecommunications, School of Electronic Engineering, Beijing- 100876, China 2 The University of Edinburgh, Institute for Digital Communications, Edinburgh, United Kingdom 3 Edinburgh Cancer Centre, Department of Clinical Oncology, Edinburgh, United Kingdom 4 Heriot Watt University, School of Engineering and Physical Sciences, Edinburgh, United Kingdom 5 Edinburgh Cancer Centre, Department of Oncology Physics, Edinburgh, United Kingdom Purpose or Objective In prostate cancer there is an ever-increasing amount of research into the use automatic segmentation methods for 1) identifying the gross tumour volume (GTV) for radiotherapy and 2) tracking disease pro gression on magnetic resonance imaging (MRI). The Active Appearance Model (AAM) has the potential to accurately segment the GTV and organs at risk (OARs) on MRI. However, when the conventional AAM model is used to assess post-brachytherapy MRI the model analyses the whole prostate GTV, which suffers from appearance distortions caused by the inhomogeneous distribution of disease and the pre-implanted brachytherapy seeds. The aim of this study was to improve the robustness and accuracy of the AAM model for prostate GTV definition on MRI scans of patients receiving brachytherapy. Material and Methods The training set used contained 50 pre-contoured transversal T2-weighted MR images of the prostate GTV. The image re-sampling process was guided by the image resolution and in this data set all training images were re-sampled into a 1x1x1 mm/pixel isotropic frame. To remove the appearance distortion caused by brachytherapy seeds and inhomogeneous foci distribution, a three-dimension (3D) volume was extracted at a distance of 10 pixels perpendicular to the GTV surface in both the interior and exterior directions. This 3D appearance patch inherently discards the distortion within the prostate gland and includes potential lesions in surrounding structures. The GTV appearance model interprets the GTV surface as the central surface of the appearance patch extracted, which deforms to minimise the appearance difference with the training set. Training was performed on 44 randomly selected cases and testing on the remaining 6 cases. The Dice similarity coefficient was used as a metric to evaluate the performance of the model. Results Figure 1 shows the proposed AAM search at different iterations against the ground truth GTV where the mesh color was calculated using a pairwise Euclidean distance (mm) based on the point correspondence. Table 1 shows a summary of the results obtained on the 6 test cases where the lowest Dice similarity was observed in the case with smallest GTV volume. In the cases with larger GTV volume the model performance significantly improved.

Results The result of the TEST-RETEST showed that 91% of the RFs were reproducible according to CCC. With respect to the influence of the modification of the CT acquistion parameters on the reproducibility of the RFs (intra-CT) the results ranged from 89.3% to 43.1% where the pitch factor and the reconstruction algorithm were modified respectively. Inter-CT RFs reproducibility also showed large material differences: from 85.3% (wood) to only 15.8% (polyurethane). Reproducible RFs were obtained only under changes of FOV, mAs, kVp and for the PMMA as the most water-like material. After the hierarchical cluster analysis(Image 2) 10 reproducible and non- redundant RFs were left(Kurtosis or Global Median among others)

Conclusion After evaluating the reproducibility of 177 RFs under the TEST-RETEST, intra-CT and inter-CT comparison we conclude that care must be taken since a lot of RFs must be discarded. CT image acquisition parameters influence Radiomics results, so tight acquisition protocols must be established. If all the CT parameters are fixed except FOV, mAs and kVp (which could be varied from patient to patient in clinical routine), the 10 RFs obtained should be reproducible and informative for clinical studies. Results suggest caution in interpreting multicenter Radiomics studies or even studies carried out in a sole institution but involving several scanners.