The DR THERAPAT Demonstrator Software System (DSS)

The DR THERAPAT project is a multi-site project with geographically distributed contributors with the main goal to bring together advances in MR imaging and radiation therapy in order to advance treatment plan personalization for prostate and cervical cancer patients.

The key idea to personalization addressed in DR THERAPAT has been coined “Dose painting”, which describes the process of identifying those regions in the image dataset that benefit from a modification of the regionally homogenous radiation dose distribution that is normally applied to the whole organ and precisely sculpting their dose prescription while sparing the surrounding organs.

In order to support testing, adaptation and demonstration of the numerous algorithms that come together to extract relevant information from the patient’s MR images. The image processing algorithms developed for and used in DR THERAPAT is depicted here

DR THERAPAT Demonstrator 1

They specifically cover
• Spatial co-registration of the individual scans to account for patient and organ motion
• MR image quality improvements
• Target Organ and Organs-at-risk segmentation
• Tumor Tissue Classification
• Tumor Control Probability (TCP) and Normal Complication Tissue (NTCP) Modeling
• Dose Boosting

TCP modeling is implemented in two ways, the first being the standard Linear-Quadratic Tufts model, the other a computational cancer response simulation approach called biological multi-scale modeling.

The data flow through these algorithms is mostly linear, with certain steps in co-registration and image quality improvement optional and others such as the TCP modeling alternatives logically parallel. Not all the steps can be fully automated, meaning they need human input. The chaining of the algorithms thus requires additional architecture to pass status signals between algorithms.

The software system targeted here is a new development specific for the DR THERAPAT collaborative project. There are several general-purpose frameworks that provide a subset of functions needed by the [DSS]. We built the [DSS] on top of [XNAT], an open source imaging informatics platform produced by the NRG Lab at the Washington University School of Medicine. The components of the overall system, subsystems, and external interfaces for XNAT are shown here

DR THERAPAT Demonstrator 2

Pipeline engine

The XNAT Pipeline Engine is a core piece to the DR THERAPAT demonstrator platform. Technically, it is a standalone Java-based framework that links sequential activities, whether triggered and executed by human or programmatically, into a defined process flow and manages how data moves from step to step in that flow, based on the results of each step. The Pipeline Engine facilitates both fully automated and semi-automated workflows. The atoms of the workflow are the pipelines. In DR THERAPAT, each image processing algorithm is mapped to a separate pipeline.

Public report downloads

DR THERAPAT Deliverable 6.31: Platform for dose painting by numbers

DR THERAPAT Deliverable 6.32: Platform for dose painting by contours