Investigation of low-contrast tumor detection in algorithm-enabled low-dose CBCT

Flat-panel-based X-ray cone-beam computed tomography (CBCT) can provide images of high utility in medical applications such as image-guided surgery and radiotherapy by offering location and contrast information concerning the target and surrounding region. A potential issue in these applications of CBCT is the imaging dose involved because repeated scans are necessary in the image-guided procedures. We investigate algorithm-enabled low-dose CBCT imaging for potential applications to image-guided surgery and radiotherapy procedures, involving detection of low-contrast tumor structures from sparse-view CBCT. Results of our study suggest that images of quality comparable to that of the FDK-reference image can be reconstructed from data much less than full data currently used.