Medical Image Analysis
As medical imaging techniques become more sophisticated it is becoming increasingly difficult to use traditional techniques to access and interpret the wealth of information that they contain. The goal of the Medical Image Analysis group is to develop new computational tools that assist in extracting clinically relevant information from multi-modality andmulti-dimensional images.
The main focus of our research is the analysis of dynamic contrast-enhanced images which can be used to assess a wide range of physiological parameters. For example in breast cancer, dynamic contrast-enhanced MRI images allow us to distinguish between benign and malignant lesions and in the liver ultrasound imaging using microbubbles as a contrast agent can be used to screen for tumours. The main challenge posed by dynamic imaging is in interpreting the large numbers of images that are typically acquired. Techniques for extracting relevant information in the form of parametric maps or classification images are needed to aid the clinician in making best use of the images. We also need to improve the quality of the image data by correcting for patient motion and reducing the effects of artifacts and noise. Finally we need to develop methods of validating new techniques using histopathology as the ?ground truth?.
Specific areas of research include:
- Deformable image registration algorithms to correct for patient motion and to assess longitudinal changes
- Multivariate techniques such as factor analysis and independent components analysis to extract features from dynamic contrast-enhanced images
- Computer Aided Diagnosis in breast cancer
- The use of contrast-enhanced imaging to assess therapeutic response in cancer
- Correlation of histopathology with in vivo imaging as a validation tool
- Cristina Gallego
- Hatef Mehrabian
- Anthony Kim (co-supervised)
- Anthony Lausch
- Jacob Levman
- Rushin Shojaii
Selected References:Link to Pubmed Publications
Lueck GJ, Kim TK, Burns PN, Martel AL. Hepatic perfusion imaging using factor analysis of contrast enhanced ultrasound. IEEE Trans Med Imaging. 2008 Oct;27(10):1449-57.
Levman J, Leung T, Causer P, Plewes D, Martel AL. Classification of dynamic contrast-enhanced magnetic resonance breast lesions by support vector machines. IEEE Trans Med Imaging. 2008 May;27(5):688-96.
- Martel AL, Chan RW, Ramsay E, Plewes DB. Removing undersampling artifacts in DCE-MRI studies using independent components analysis.Magn Reson Med. 2008 Apr;59(4):874-84.
- Martel AL, Froh MS, Brock KK, Plewes DB, Barber DC. Evaluating an optical-flow-based registration algorithm for contrast-enhanced magnetic resonance imaging of the breast. Phys Med Biol 2007;52(13):3803-3816.