Simulations to Evaluate Accuracy and Patient Dose in Neutron-Stimulated, Emission-Computed Tomography (NSECT) for Diagnosis of Breast Cancer

摘要

We have developed a new tomographic technique called Neutron Stimulated Emission Computed Tomography (NSECT) for early detection of breast cancer. NSECT is sensitive to metabolic changes in trace element concentrations that are seen in tumors at very early stages of development. Detecting and measuring these element concentrations has the potential to detect breast cancer very early. While the use of neutrons as the imaging radiation leads to concerns about radiation dose, preliminary experiments suggest that it may be possible to perform NSECT scans with dose levels comparable to mammography. A key aspect of successful clinical translation is to deliver the minimum dose possible. This project aims at evaluating the effects of NSECT dose-reduction techniques on the accuracy of detecting breast cancer. Four dose-reduction techniques are under evaluation: reduction of - neutron flux; spatial projections; angular positions; and the use of multiple detectors. As separate evaluation of each factor using experimental studies is prohibitively time consuming, we have investigated each effect using Monte-Carlo simulations as an alternative. We developed a Monte-Carlo simulation of the NSECT tomographic scanning system in GEANT4 and designed phantoms of benign and malignant breast tissue. The simulation was tested and validated against experimentally acquired data from several different types of phantoms. After successful validation, the simulation was used to generate spectral data from NSECT scans of the benign and malignant breast and obtain an estimate of the radiation dose. Tomographic images were then generated from a single-element disease model using a maximum- likelihood algorithm. Finally, we analyzed the effects of the four dose- reduction techniques on system accuracy for both, nontomographic acquisitions and tomographic images.