Clinical Developments

–Ongoing Projects–

Development of A Software Based Automatic Exposure Control For Varian On-Board Imaging

Modern image guided external beam radiation therapy involves the use of a mounted on-board imager (OBI) on external beam units to take images of a patient’s position before each treatment. These images are used to determine the treatment couch shifts required for ideal alignment based on digitally reconstructed radiographs created in treatment planning. The lack of an automatic exposure control (AEC) on Varian OBI systems requires x-ray exposure factors to be selected manually from technique factors tables which are based on general population data. This may result in over or under exposed images for certain patients and compromise the accuracy of the image matching. A software based AEC system has been developed to predict the optimal, patient specific exposure factors based on patient planning data. This system allows us to accurately choose the techniques that will give the highest quality image to use in image matching, while eliminating the need for multiple unnecessary exposures. 

DICOM De-Identifier Software

To observe and diagnose cancer in modern times, medical professionals rely heavily on imaging instruments, such as CT and MR. The images generated by these instruments have been standardized into a digital file format to contain all the relevant information. The Digital Imaging and Communications in Medicine (DICOM) Standard was the result, which dictates proper handling, storing, printing and transmitting of medical imaging information. It contains a file format definition and a network communications protocol. The DICOM file (.dcm) carries with it both pixel data to generate an image, as well as important information regarding the patient and the imaging procedure. This information allows an individual’s electronic health record to easily be maintained and updated. However, some of this information, referred to as protected health information (PHI), is legally protected from all but the patient or attending healthcare professional. This means that researchers must find a way to detach the image from the individual while still retaining important information for research.

The DICOM De-Identifier aims to accomplish just that; to remove all recognized PHIs from a DICOM file while still allowing the researcher to access the relevant data. The process is currently being done by hand; therefore, this program will offer an affordable solution for faster and more streamlined de-identification, ultimately allowing larger data sets to be used. It will also be designed such that any health researcher will be able to use it, adjusting the level of de-identification as they require, and due to the generality of DICOM, it can be used for any imaging modality.

Mammography Quality Control

Daily and weekly quality assurance tests are a critical component of maintaining a high standard of care within the Screening Mammography program of BC. Many of these procedures tend to be simple tests that involve a significant manual work, such as selecting and measuring regions of interest on phantom exposures and recording results in Microsoft Excel. Additionally, coordinating reports across multiple centers in a large organization such as the BC Cancer Agency can be slow and error-prone. Such work can be readily automated by software, and the results of the tests stored in a centralized database in order to facilitate region-wide tracking of mammography unit performance.

We have developed a software platform, mammoQC, which takes advantage of British Columbia’s province-wide transfer grid network to provide automated quality control measurement and recording to digital mammography centers. This system minimizes the time spend by technologists, efficiently consolidates test data and provides a visualization platform for quality assurance purposes for both individual units and an entire program.

Patient-Linac Collision Detection Tool for Radiation Therapy

During external beam radiation therapy the linear accelerator rotates around the patient, who is lying on the treatment table, and administers concentrated radiation to the cancer. One problem that may occur during treatment is a collision with the linear accelerator and the patient; this results in the treatment being stopped and re-planned, which is a time and resource consuming process that affects both the radiation therapists and the patient. This study involves developing a tool that simulates the radiation treatment on a computer, before the patient is brought in for the actual treatment. This will allow radiation therapists to check for collisions and adjust the treatment plan prior to treatment delivery. This tool hopes to save time and avoid repetition of treatment planning, and its clinical use is currently being evaluated in order to determine its value and effectiveness. Since the start of its regular use in early 2015 it has proved to be very useful tool.

Xbox Kinect Integration

Over the course of the 2014-2015 school year, a team of Computer Science students have been developing a program that will allow CT staff to capture 3D models of patients in real-time. These models will be used with the collision detection simulator and this program looks to replace the current method of taking physical measurements of the patients pose and extremities in order to estimate their 3D model.  

This software is demonstrated in the following video:

Kinect 3D Patient Modeling Demo from Duncan Szarmes on Vimeo.

Capstone project demo. This software uses an Xbox Kinect sensor to capture 3D models of patients, which can be used in external beam radiation therapy simulations to predict collision