Our lab is interested in discovering and applying novel image analysis and image-based computational techniques for improved detection, follow-up and treatment of cancer. Much of our work is highly translational, with direct application to the treatment of patients within the Department of Radiation Oncology. We also do perform animal and in vitro studies at times to validate and improve upon our discoveries.
Novel 3D tumor detection approaches are being applied for lung, brain and breast screening, and in the future for virtual colonoscopy. Finite element modeling and deformable image registration have been used to quantify the deformation during needle biopsy. Quantitative analysis of CT image changes following radiation treatment is being used to assess radiation damage to healthy tissue surrounding lung tumors with eventual use in assaying new drugs to protect normal tissue from radiation. MR diffusion-weighting imaging and MRI spectroscopy is being used in both patient and animal studies to model microscopic tumor spread for improved treatment of aggressive brain cancers.
Monitoring and Modeling Cancer Cell Migration in the Brain
White matter architecture can be measured non-invasively using MR Diffusion Tensor Imaging (DTI) and we have recently shown a significant correlation between the paths of fiber tracts leading from the primary tumor and the sites of tumor recurrence in humans. This result suggests a direct and unique utility for MR-DTI tractography in probing the mechanisms of cancer cell migration. Our objective in this research on human subjects on our 3T system is to utilize DTI to constrain a computational pseudo-random walk model of cancer cell migration to predict the location of future tumor recurrence
We are currently recruiting patients for a prospective trail using our computational model to predict the location of tumor recurrence after surgery and radiation treatment. The initial results look very encouraging. However, patient data provides only organ-scale observations of cancer recurrence patterns whereas to advance scientific understanding of the effects of tissue architecture on cell migratory behavior more quantitative and direct measures of cell spread in animal models are needed.
In vivo anatomical MR images of rat brains are being obtained at 3T just prior to engraftment with rat-native and GFP-labeled infiltrative glioma cells. A high-resolution ex-vivo rat brain DTI dataset has been acquired using our 9.4T scanner. The in vivo images from the engrafted rats will be registered to the ex vivo DTI data and the computational cell migration model will be run on these datasets. Histological analysis at 10-days post-trans-plantation will be used to directly monitor cell dispersion at the tissue level to compare with model predictions. Moreover, we propose to extend this animal model by labeling individual glioma cells with super-paramagnetic iron-oxide (SPIO) particles.
The increased SPIO sensitivity at high field strength will enable us to visualize and track individual cells twice daily for 10 days post-implant. The knowledge gained from the SPIO-labeling experiments will enable us to provide our migration model with more physiologically-appropriate parameters for step size, persistence of step direction, fiber affinity, and probability of a cell leaving a fiber bundle once it has become associated with it.
Monitoring and Modeling Cancer Cell Migration in the Rat Brain
Investigators: W. O'Dell, Anitha Krishnan, Divya Raman, Sharmistha Chaudhuri, Scott Kennedy, John Olschowka, Bruce Fenton, Sammy N'dive.
While patient data provides valuable observations of cancer recurrence patterns (see our patient MR-DTI/MRS page for further details), to advance our scientific understanding of the effects of tissue architecture on cell migratory behavior and thereby improve the accuracy of our model, more quantitative and direct measures of cell spread in animal models are needed. For an animal model of high-grade human glioma, we are using a unique, genetically-enhanced infiltrating native rat brain cancer cell line (HEBAB/brevican CNS-1 Rat Glioma cells) prepared by our colleagues Drs. Hong Zhang and Russell Matthews.
In vivo anatomical (T2-weighted) MR images in a rat brain were acquired at 3T with 0.27x0.27x1.0 mm resolution and the same brain was then extracted, fixed and scanned ex vivo with a DTI sequence on the 9.4T system with a resolution of 117x163x117 um. The MR-DTI from the above ex vivo study was processed to generate 3D fiber maps of the major fiber bundles of the rat. The computational model was then run on this data set, starting from a point adjacent to a major fiber bundle.
A second rat brain was MR-DTI imaged ex vivo as above and histologically sectioned. A representative set of six sequential histological sections, without staining, were photographed, each slice 40?m thick and separated by 920?m. The approximate MR section was selected by inspection for comparison. A close match is observed before application of any high level registration.
In vivo MR-DTI at 3T will be performed as demonstrated by us in Figure 1A prior to the engraftment of approximately 100,000 cells. The brains will be carefully harvested and imaged ex vivo at 9.4T to aid in registration, followed by histological analysis of sectioned brain slices to identify the spread of GFP-labeled cells. Custom image processing software for automatic detection and statistical assessment of labeled cells will be created. Registration of the histological sections with the ex vivo MR image will be performed using a b-spline grid-based deformable registration method developed previously by the PI and colleagues.
Metastatic Cancer Early Detection on Lung and Brain
Investigators: W. O'Dell, Robert Ambrosini (BME PhD '09), Peng Wang (BME PhD '08).
Novel applications of conformal stereotactic radiation therapy championed by our clinical colleagues have renewed the clinical motivation for early detection of metastatic cancer to brain and lung. The major challenge in the lung is that the small nodules of interest resemble in a 2D slice the numerous blood vessels coursing through the lung tissue. In the brain one has to deal with differential uptake of contrast in T1-weighted post-contrast images. We have developed and patented a novel 3D template matching approach that seem to far outperform any existing commercial CAD system or any other method presented in the literature.
Detection of nodules in the body is not equivalent to a diagnosis of metastatic disease. For that, accurate quantification of tumor growth over time from sequential images is needed. Expert radiologists are notoriously poor at estimating tumor volumes, as evidenced by the results of the recent Lung Image Database Consortium (LIDC) multi-institutional study. We have been working on an automated method for computing tumor volumes, and thereby growth, using the matching features of our 3D template algorithm. This project is described in detail on our Tumor Volume Estimation page.
We have applied this approach successfully to images of a breast phantom scanned using a new breast-dedicated cone-beam CT scanner, and less successfully to detection of lymph nodes in the upper chest and neck. We hope to expand this method to the detection of tumors of the liver and colon polyps a part of a virtual colonoscopy scheme.
Make an Appointment
If you would like to make an appointment or consult our physicians for a second opinion, please contact us at one of the following locations:
- Wilmot Cancer Center
- Highland Hospital
- Cancer Center at Park Ridge
- Sands Cancer Center