Timothy E. Doyle, Ph.D. - US grants

Two all-sky imagers sampling the OH nightglow are initially positioned in the Northern Utah, and Wyoming region. These imagers generate all-sky images distinct by virtue of observation geometry, which are then tomographically inverted to produce two-dimensional intensity maps of short period (less one hour) gravity waves in the upper mesosphere. Classical algebraic tomographic inversions and stochastic inversion algebraic inversions are compared with system (integral) and geometric tomographic models using the actual airglow data to isolate the best technique for wave structure extraction using two OH imagers. Extension to 3-dimension descriptions of wave motion and vertical structure in the upper mesosphere and lower thermosphere is finally accomplished using coincident measurements of the OH airglow, the O2 airglow, and the OI airglow - which arise from different layers between 80-155 km.

DESCRIPTION (provided by applicant): The early and reliable detection of cancer in vivo is critical to improving its treatment and prognosis. A safe, portable, inexpensive, and user-friendly diagnostic system such as ultrasound would have considerable value for real-time cancer detection during many routine procedures such as oral exams, endoscopies, biopsies, and surgeries. Tissue margin assessments during breast conservation surgery (BCS) are one example. Up to 50% of BCS patients endure additional surgery due to cancer-containing margins that go undetected during surgery. Ultrasound's specificity and sensitivity for cancer, however, require significant improvement. To this end, this project will develop physics-based, three-dimensional (3D) simulation tools to quantitatively determine tissue histology using ultrasound. These tools will be considerably advanced over current empirical and analytical models since they will directly simulate wave scattering and propagation in heterogeneous tissue with greater histological accuracy. This approach includes hierarchical tissue structures, multiple scattering, and wave-mode conversion, which have been largely ignored by previous methods. A simulation approach is attractive for modeling malignant cell infiltration, nuclear pleomorphism, and other microstructural changes associated with cancer since it has the ability to include heterogeneous distributions of cells with specific histological features and internal structures. By directly simulating neoplastic changes in the structure and material properties of cells and tissues, the proposed models will provide a radically new, mechanistic-based paradigm linking ultrasonic measurements to the underlying histology of tumors. The computational tools that will be developed in this project are based on recent models that simulate ultrasonic interactions at the cellular and subcellular levels. The feasibility of the approach has been demonstrated with computer simulations of complex tissue structures containing up to 2137 nucleated spherical cells. The results indicate that neoplastic changes at both the tissue and cell levels will be detectable with ultrasound at frequencies that can penetrate tissues up to 1 cm in depth. The studies also indicate that "ultrasonic signals" from malignancies involving as few as 300 cells will be resolvable from the background of normal tissue. The specific aims are to increase the capabilities of the current models to simulate nonspherical cells, compound tissue structures, and tissue volumes of 1 mm3. The models will be refined by ultrasonic testing of 3D tissue cultures of invasive ductal carcinoma of the breast. Finally, the models will be validated with ex vivo ultrasonic testing of BCS tissue specimens. The proposed models will provide new, more quantitative tools for detecting malignant tissues in vivo. Using intra-operative surgical margin evaluations as an example, they could reduce the number of BCS re-excisions by up to 60,000 per year and improve the quality of life and prognosis for many patients. PUBLIC HEALTH RELEVANCE: This project will develop unprecedented physics-based computer programs that will model how ultrasound interacts with cells and tissues at the microscopic level. Such programs can be used to predict how changes in cell and tissue structure associated with cancer will affect clinical ultrasound measurements, thereby providing new in vivo capabilities for detecting microscopic cancer. These capabilities promise improved cancer screening, diagnosis, and treatment during routine procedures such as surgeries, biopsies, endoscopies, and oral examinations.