UNDISTORT Correction of Distortions in Diffusion MRI V1.0
- Conditions
- Prostate CancerMR Image Distortion Correction
- Registration Number
- NCT03151512
- Lead Sponsor
- University College, London
- Brief Summary
This is a three-year project funded by a Cancer Research UK Multidisciplinary Award and brings together a team from UCL Division of Medicine, Computer Science and University College London Hospital. The aim is to develop Magnetic Resonance (MR) sequences and mathematical algorithms to reduce the distortions in MR images, especially of the prostate.
- Detailed Description
This is a three-year project funded by a Cancer Research UK Multidisciplinary Award and brings together a team from UCL Division of Medicine, Computer Science and University College London Hospital. The aim is to develop Magnetic Resonance (MR) sequences and mathematical algorithms to reduce the distortions in MR images, especially of the prostate. Current NICE guidelines include a type of MR imaging called Diffusion Weighted MRI for the detection of tumour within the prostate, and for active surveillance of low risk confirmed disease. However, approximately 40% of prostate diffusion images suffer from severe localised distortions and this is most marked in the peripheral zone of the prostate where 75% of prostate cancers occur. The source of these distortions is magnetic field imperfections due to the presence of rectal gas or metallic hip implants.
The research study will ask both healthy volunteers and patients to undergo research MR scans and use the acquired data for analysis. For patients, the scans may be either additional sequences acquired during an extended clinical session, or a separate additional session entirely for research.
The output from the research will be modified ways to run an MR scanner and compute the final images.
The work should lead to improved diagnostic accuracy and a reduced number of non-diagnostic studies. It will have broader impact through application to diffusion imaging of other body sites, including whole-body diffusion MRI and non-cancer applications. If successful, the results would provide evidence for a larger trial with the eventual outcome being manufacturers incorporating modified MR sequences and data processing into clinical systems worldwide.
Recruitment & Eligibility
- Status
- COMPLETED
- Sex
- All
- Target Recruitment
- 37
- For the development and testing phases, there are no specific inclusion criteria. Assessment of the primary outcome measure (reduced distortions) also does not require specific inclusion criteria. The intended application of our methods is to prostate cancer and this is reflected in some of the secondary outcome measures. Recruitment will come from the UCLH imaging bookings system. This list will include many men having prostate scans and many of these will subsequently be found to have at least a suspicion of cancer. Note there is no requirement for a suspicion of cancer to be recruited for the study.
- Subjects unable to have an MRI scan due to contraindications for MRI, for example, pacemaker and certain other implants, severe claustrophobia.
- Subjects unable to give informed consent.
- Children and vulnerable populations.
Study & Design
- Study Type
- OBSERVATIONAL
- Study Design
- Not specified
- Primary Outcome Measures
Name Time Method Dice similarity score to assess distortions Three Years Dice score provides a measure of how similar is the distortion-corrected image to a reference.
- Secondary Outcome Measures
Name Time Method Diffusion coefficient consistency Three Years Diffusion coefficients (ADC) in relatively undistorted regions will be compared pre and post distortion correction to quantify any changes (if the algorithm is working correctly, none are expected in these regions).
Radiological scoring of image quality Three Years Diffusion images will be scored blinded to correction scheme on a scale: 1 - undiagnostic, 2- distorted but diagnostic, 3 - undistorted. The change in score following the proposed method will be reported.
Trial Locations
- Locations (1)
University College London Hospital
🇬🇧London, United Kingdom