Contours of the recti of 37 prostate cancer patients from Haukeland University Hospital years 2007-2009. The patients were part of a study on hypofractionated radiotherapy with integrated boost, as desribed in Ekanger et al. (2018). The data is now fully anonymous, as the identification keys have been deleted.

Each patient had a planning CT plus 6-10 CT repeat CT scans taken during the course of radiotherapy. The rectum was delineated on every scan by an expert physicist. The contours are represented as polyhedra, where each shape contour consist of a set of of 18063 points, with x, y and z-coordinates, on the organ surface. The points are tied together by 36168 triangular faces. The definition of of the triangular faces are given by an array of indices into the arrays of contour points. All contours have been deformably registered to a reference shape; therefore, there is point-to-point correspondence between all shapes in the dataset. This makes the data ideal for statistical modelling.

This data was used in the models of Rørtveit et al. (2021) and Rørtveit et al. (2023). A description of the patients, treatment and imaging is given in Hysing et al. (2018).

Two representations of the data are available in this dataset; firstly a Matlab .mat file which contains all the data. For users who do not use Matlab, the data is also made available as text files in the form of comma separated values.

Matlab, R2020a

Matterhorn, v.16