DUAL-ENERGY SCREENING FOR NODULAR DISEASE

点击翻译按钮获取中文摘要

The purpose of these studies is to develop and evaluate a practical screening method for nodular disease using dual-energy chest radiography. The dual-energy technique is used to subtract either bone or soft-tissue structures from a radiographic image, by utilizing the different energy dependence of x-ray attentuation in bone and soft-tissue. The technique uses a linear combination of two radiographic images acquired at different x-ray energies, with the weighting coefficient chosen to cancel either bone or soft-tissue. The ability to improve conspicuity of soft-tissue nodules and to determine calcium content of nodules are the diagnostic advantages of dual-energy. Four primary areas of research will be addressed. First, a new hybrid dual-energy cassette will be evaluated which permits simultaneous acquisition of a conventional film image and digital dual-energy images. Second, the signal-to-noise ratio (SNR) of dual-energy images will be improved by using x-ray equalization and a variety of imaging algorithms. Third, a scanning grid assembly will be added to the x-ray equalizer to provide improved scatter rejection and therefore improved SNR. Fourth, an evaluation of automated nodule detection algorithms will be performed on the dual-energy tissue images. In each of these areas of research, emphasis will be placed on detailed evaluation of technical improvement, practicality of implementation and clinically-related diagnostic improvemen The specific studies in this proposal include: implementation and evaluation of a hybrid film/digital dual-energy cassette, SNR improvement in dual-energy imaging with x-ray equalization, evaluation of inexpensive generic filters for low-cost equalization, construction of a scanning grid for the x-ray equalizer, multiple-plate integration for improved SNR and dose efficiency, matched filtering of imaging plates for further improved SNR, combination of the tissue and low-energy images for improved SNR, optimization of the correlated noise reduction technique, a contrast-detail study of the bone and soft-tissue images, a clinical nodule detection study, evaluation of the calcium determination in the bone image, and an evaluation of automated nodule detection algorithms (including artificial neural networks).