| Office Location: | 2424 Erwin Road, Suite 302 |
| Email Address: |
|
| Web Page: | http://dailabs.duhs.duke.edu/person.php?id=470 |
Teaching (Fall 2009):
There are several research projects underway in my laboratory, focusing on the improvement of diagnostic accuracy in digital chest radiography and digital mammography. The first project uses a dual-energy digital radiography technique that forms images which have either the bones or soft-tissues canceled. By eliminating the ribs in the soft-tissue image, better accuracy in detecting nodules has been demonstrated in early clinical evaluations of the technique. The bone image is also useful in that it demonstrates whether a detected nodule is calcified, at a fraction of the cost and radiation exposure of sending the patient to CT. We have received six years of NIH grant funding to develop a practical screening method for nodular disease using the dual-energy technique. A second major area of research is digital tomosynthesis. This technique allows longitudinal slice images of a patient to be generated from a series of projection images taken at different angles. We are evaluating this technique for its potential to improve detection of pulmonary nodules. We have received seven years of NIH funding to develop and evaluate this technique. Preliminary evaluation in a pilot study using human subjects revealed that the detection rate for pulmonary nodules increased from about 22% to 70% in tomosynthesis slice images compared with conventional PA chest radiographs. Thus, the tomosynthesis technique would appear to hold promise as a method of improving both sensitivity and specificity of pulmonary nodule detection in vivo. We are also expanding this tomosynthesis work to evaluate its potential for improving detection of masses and calcifications in digital mammograms. Other projects include theoretical analysis of digital imaging performance and the measurement of image quality (DQE) in computed radiography and flat-panel systems.