Optimization of Biopsy Strategy by a Statistical Atlas of Prostate Cancer Distribution

The statistical atlas of prostate cancer distribution, also called asummary statistical map (SSM) is designed to facilitate and optimize cancer detection during prostate biopsy. As a probability distribution for prostate cancer distribution, targets can be chosen with the greatest likelihood to contain malignant tissues, more efficiently detecting disease.

MR-Compatible Robotics

Robotics is an emerging field in MR-therapy and will become increasingly important as the field evolves. Intraprocedural robotics may allow formore accurate and effective clinical treatment. For example, robotic needle guidance makes it possible to achieve very precise and flexible alignment with needle trajectories, which is more accurate than human hands can be.

Needle Placement Accuracy

Needle placement accuracy during prostate biopsy procedures is difficult to achieve because it is limited by how precise a human hand can be, the type and size of the needle, as well as the physiological characteristics of the tissues through which the needle must penetrate. Through robotics and needle tracking, higher accuracy can be achieved.

Image Registration

Registrationis a technique that allows medical images, from modalities such as MRI, taken at different times and from different positions to still bevisualized as the same person/prostate. For example, when an organchanges shape and treatment targeting must match biopsy data, imageregistration can be very useful. We are developing a non-rigidregistration scheme that can account for glandular position changes aswell as differences over time.

Signal Inhomogeneity Correction

The signal inhomogeneity correction project is concerned with creating more accurate and consistent MRI images. For MRI an endorectal coil is effectively used to boost signal, but the farther away from the coil, the less sharp the image tends to be. Instead of brighter patchescorresponding solely to the type of tissue, it also corresponds to the distance from the coil. To correct this, an algorithm can be introduced to reduce this proximity bias.

Image Segmentation

Image segmentation can be accomplished by manually outlining targeted tissues such as the prostate or by automated anotomical segmentation. Segmentation is useful for many reasons including ascertaining information such as the volume of the prostate or allowing for color-coded 3D rendering of targeted tissues. This can help clinicians more accurately evaluate where the position of tumors in relation to vital organs and tissues and more effecient treatment.

Statistical Validation and Support, Kelly Zou, Ph.D.

MR-Spectroscopy, Robert Mulkern, Ph.D.