Graduate Research in Nuclear Engineering

In order to grow, tumors need a blood supply that is provided by angiogenesis (Judah Folkman). A growth factor for angiogenesis is vacular endothelial growth factor (VEGF) which both stimulates capillary production and causes "leakiness" of the vessels. These leaks cause an increase in the permeability of the capilaries and allow molecules in the blood, such as a contrast agent to cross from the plasma to the extravascular extracellular space (EES). The EES is the area outside the vasculature between the cells.

Figure 1

Tumor contrast agent extraction factors. Specifically these are permeabilty (P), surface area (S), flow (F), interstial tumor transit time, distribution half-life (T_{1/2 DIST}), and excretion half-life (T_{1/2 EXCRE})

Weidner showed that capillary density in a tumor slice measured at specific magnefications was diagnostic of tumor type and grade. The argument in DCE MRI is that tumor extravasation of a contrast agent, specifically the PS, is also diagnostic of tumor type and grade since permeability is related to angiogensis which is seen as increased capillary density. What I have been studying then is the effect of the spatial resolution or the "sharpness" of DCE MRI on the diagnostic accuracy and some solutions to the spatial and temporal resolution limitations of clinical DCE MRI.

Basically, the goal was to quantitate the transfer rate, on a voxel-by-voxel basis (the elements that compose an MR image), of a contrast agent into and out of a tumor and determining its diagnostic importance. These transfer rates should translated into an diagnosis of the tumor's type (i.e. is it benign or malignant) and grade (i.e. the degree of malignancy). Currently, the only clinically approved method of this is to remove the tumor and perform a biopsy.The overall goal of my graduate research under Dr. Erik Wiener was to improve the sensitivity and specificity of breast cancer screening methods using DCE MRI. That is, trying to improve MRI detection of tumors and develop diagnostic techniques with MRI for determining if a tumor is benign or malignant. DCE MRI has the potential for better sensitivity and specificity than the current gold standard, X-ray mammography. With any DCE MRI, a series of images can be made of the inside of a tumor following injection of a contrast agent into the blood; the agent enters the tissue and is then gradually removed. Over a period of time, each image will have a different amount of contrast agent in it. The physiological effects of the tumor on the agent can be modeled, and from this model, potentially, the tumor type can be established. The most important parameter of the model for this is the contrast agent transfer rate between the vasculature (i.e. the plasma) and the interstitial space of the tumor (i.e. the area outside the blood vessels between the cells) or K^TRANS. This K^TRANS is composed of the permeability-surface area product, PS, and the blood perfusion or flow, F. The PS reflects the size and number of holes for the contrast agent to slip from the capillaries into the tumor.

For my Master's I sought to quantify K^TRANS of mammary tumors as a function of contrast agent size, charge, and tumor type, as well as identifying any intratumor heterogeneity for use in characterizing tumor specificity. My Ph.D. took a more basic MRI science approach by testing the hypothesis that poor spatial resolution used in clinical DCE MRI results in partial volume effects that yield inaccurate K^TRANS which results in erroneous diagnostic information. This involved wet lab work, shop work, animal handling/perfusions, MR imaging, and computer modeling/programming.

Figure 2

Looped animation of coronal (from head to tail through the shoulders) MRI slices of rat following a bolus injection of contrast agent. Notice the images go from dark to bright as the heart (center toward the top) lights up followed by the rest of the body. The lighter area to left of the three images to the left is a mammary tumor. The bright spot that appears near the bottom is the bladder filling. The kidneys can be seen on the image to the farthest right.

During my graduate career I worked in other areas as well as some small collaborations. One project was imaging mice with orally dosed Gd microspheres. Other research projects have included imaging of folate targeted dendrimer agents in nude mice with cell-induced tumors. Some tumors express folate receptors that this agent could bind to, increasing the contrast of the tumor and showing potential as a targeted therapeutic agent.

I've also worked on a modeling program for looking at the Solomon-Bloembergen-Morgan (SBM) equations (never quite had time to finish it). These equations govern the relaxation behavior of the proton spins in the presence of a paramagnetic ion (e.g. gadolinium). The relaxation behavior is important, because this behavior governs the change in signal intensity due to the presence of a paramagnetic species. This was related to another experimental project I worked on, trying to determine the binding time of a paramagnetic ion to the oxygen in water using ¹⁷O NMR.

I've done a lot of MR imaging: mice with brain tumors, rabbits, perfusion studies looking for angiogenesis in the brain due to exercise, and all sorts of rat models of cancer. Finally, I've also done some work with a field cycling relaxometer, a machine that measure the longitudinal relaxation rates at different field strengths.