Parallel MRI

          Magnetic Resonance Imaging (MRI) offers tremendous opportunities for advancing medical imaging research. MRI, unlike most other medical imaging modalities, can use a wide variety of contrast mechanisms, each of which provides unique information about not only anatomical structure, but also physiological function and in-vivo biochemistry. Unfortunately, the method MRI has used for encoding spatial information into images - the switching of magnetic field gradients - is very slow, making it difficult or impossible to access the full richness of potentially available information in reasonable times. The emerging technique of parallel MRI can alleviate this image encoding bottleneck and unlock the full potential of MRI. Parallel MRI uses arrays of radiofrequency (RF) receiver coils to replace a substantial fraction of the time-consuming image encoding done by switching magnetic field gradients. With many coils in an array, tenfold and higher efficiency gains in image encoding can be realized.

         This increased efficiency allows time formerly used for spatial encoding to be invested instead in achieving order-of-magnitude increases in imaging speed, spatial resolution, or anatomical coverage of standard MRI images. Depending on the application this could allow 3D imaging of the whole heart in a single heart beat, 3D breast MRI at resolutions nearly as high as X-ray mammography or sub-millimeter resolution 3D MRI of the entire abdomen in a single breath hold. These gains in spatial encoding efficiency can also enable otherwise prohibitively time consuming examinations/techniques for crucial measurements of disease biomarkers.

The goals of our research program in parallel MRI are as follows:

1. Construction of RF coil arrays with large numbers of coils explicitly designed for parallel MRI at high acceleration factors.
2. Adaptation and optimization of pulse sequences and reconstruction techniques for parallel imaging.
3. Translation of the developments of Specific Aims 1 and 2 into use for basic research and clinical studies of a wide variety of diseases, including respiratory disease, cancer, and fatty liver disease.

         Parallel MRI techniques are widely applicable to the complete array of MR methods that use gradients to encode spatial information. This includes standard proton (1H) imaging at any field strength, transmission of RF pulses, multinuclear (non 1H) imaging, spectroscopic imaging (chemical shift imaging), and imaging with hyperpolarized nuclei. Parallel MRI can be used to enhance virtually any existing study and can often make otherwise impractical studies feasible.

Click to view Lanette's abstract on Parallel MRI Acceleration of Dynamic and High Resolution Hyperpolarized 13C MRI

Click to view Linghe Yang's abstract on Quantification of Artefact in Parallel MRI Reconstruction by Point Spread Function Analysis

Fat Quantification

          The study of obesity and metabolic disease in medical research continues to be a hot topic. As the prevalence of obesity rises, the need for accurate tools to study obesity becomes more apparent. Studies have shown that adverse health effects are not just caused by the total amount of fat, but more importantly where the internal fat is located. It turns out that the fat around the organs (Visceral Fat) is strongly correlated with the development of illnesses such as cardiovascular disease, type II diabates, and non-alcoholic fatty liver disease (NAFLD). Internal fat is much more difficult to assess with traditional measurement techniques such as Body Mass Index, Waist-to-Hip ratio, and Bioelectric impedence analysis. As a result, internal imaging methods such as CT and MRI are required to differentiate between body fat depots. MRI provides a key advantage over X-ray, CT, and Dual-energy X-ray Absorptiometry (DXA) in that it allows for the scanning of patients without the use of harmful ionizing radiation. As a result, patients can been scanned multiple times without the worry of any adverse health effects making MRI a more favorable option.

          With the use of Parallel MRI and the IDEAL imaging technique we are able to simultaneously acquire Fat-only and Water-only images of the entire visceral cavity within a single breath-hold scan (~24 seconds). From these images we can automatically segment the tissue and calculate accurate volumes of subcutaneous fat, visceral fat, and water tissue (muscle and organs). See the 3D exploded view of the different tissues below. Click here for a quick video showing the 3D volumes

Prostate MRI

         Prostate cancer is the third most common cancer in the world and becomes more prevalent with age. Diagnosis of prostate cancer currently requires a biopsy, where transrectal ultrasound (TRUS) is often used to guide the biopsy needle. TRUS-guided biopsies often underestimate or fail to detect prostate cancers. However, Magnetic Resonance Imaging (MRI) has been shown to be extremely sensitive and specific for the detection of prostate cancer. Since ultrasound (US) is the most effective method to biopsy the large and increasing number of men, a technique is required to direct the biopsy to the suspicious lesion or to the portion of the detected tumor.

          Various magnetic resonance imaging (MRI) techniques have been evaluated using body and endorectal coils, contrast enhancement, and different pulse sequences resulting in disease detection rate sensitivity and specificity of 80-88% and 75-95%. Although progress has been made with the improvement of these techniques the accuracy of the results is not ideal; thus an invasive method is required to give a precise diagnosis and stage of the diseases. In clinical practice, the combination of imaging techniques is an option that can be explored and improved on. Fused MRI and 3D TRUS images have the potential to dramatically reduce the false negative rate of biopsies guided by TRUS alone and can improve the evaluation of cancer location, size, and extent, while providing an indication of tumor aggressiveness.

Hyperpolarized MRI

         Magnetic Resonance Imaging (MRI) has been used for many years to acquire images of the soft tissue inside the body. Typically, MRI measures hydrogen (1H) nuclei, but it is possible to observe other nuclei, such as carbon-13 (13C) and nitrogen-15 (15N). These other nuclei are interesting, because they are involved in metabolic processes in the body. It is known that changes in metabolic processes often cause and/or precede diseases, so being able to observe these changes in the body is very important for early diagnosis and understanding diseases. Acquiring spectra or images inside the body using nuclei other than 1H is difficult. The low natural abundance of 13C and 15N and the poor polarisation of these nuclei in a magnetic field results in very little signal. In order to acquire the simplest spectra or images of these nuclei would require many, many hours.

         A new development in MRI, called Dynamic Nuclear Polarisation (DNP), has enabled acquisition of spectra and images from nuclei other than 1H in more reasonable times. The problem of low natural abundance can be overcome by replacing the 12C with 13C in a molecule of interest. The problem of poor polarisation is overcome by the DNP process. This involves placing the 13C-enriched molecule in a magnet and supplying the electrons surrounding the 13C with energy. The electrons transfer this energy to the 13C nuclei resulting in increased polarisation, or hyperpolarisation, of the 13C. When the 13C-enriched, hyperpolarised molecule is injected into the body, a 10,000-fold increase in signal is obtained. With this increase, spectra or images can be acquired in a matter of minutes enabling the ability to monitor metabolic processes in the body.

         Our research involves the development of the techniques involved in Dynamic Nuclear Polarisation and MRI of hyperpolarised substances. We are applying these techniques to observing the metabolic processes in diseases such as multiple sclerosis, spinal cord injuries, and cancer.

Click to view Lanette's abstract on The Effects of Contrast Agents on Hyperpolarised [1-13C]-Pyruvic Acid

Liver MRI

         Magnetic Resonance Imaging (MRI) is one of the most diverse imaging modalities in medicine and offers the ability to take images of specific biological tissues. The two most common tissues within the human body are water and fat. Recently it has been discovered that when some organs (specifically the liver) are unhealthy, they will contain high levels of fat. In order to diagnose diseases, it is critical to determine an accurate estimate of the amount of fat within the liver (outlined in green).

         The current gold standard is to use a liver-biopsy, however this is an invasive method that only samples a small portion of the liver and can lead to misdiagnosis. The objective of our research is to use MRI to quantify the amount of fat within the entire liver. MRI techniques have been developed in the past that are able to form “purely-fat” and “purely-water”(normal liver tissue) images, however this makes the MRI scan longer. The increase in scan time is a problem for any type of abdominal imaging, as a patient can only hold their breath for a relatively short time. Therefore in order to produce the clearest image possible, an MRI abdominal scan is limited to the length of a patient “breath-hold”, which is on the order of 20 seconds.

         New MRI techniques such as Parallel MRI, allow us to generate a complete image using processing techniques and only a fraction of the data usually required. Using Parallel MRI can speed up an MRI acquisition to less than a “breath-hold”, however it can cause lower quality images if the fraction of data we acquire is not selected carefully. Ultimately, we will compare our MRI fat maps with liver biopsies to determine the accuracy of our non-invasive technique.

          Breakthroughs in MRI are now making it possible to not only image human anatomy, but to also look at human physiology by following metabolic processes. We hope to develop MRI based tools that can track metabolic processes and lead to earlier and more accurate diagnosis of disease.

Click to view Shawn Kisch's abstract on Non-Alcoholic Fatty Liver Disease