Research Of Key Technologies For Multi-modality X-ray Fluorescence Computed Tomography Based On An X-ray Tube Source

Cancer is one of the main diseases that threatens the human health significantly.Chemotherapy and radiation treatment are two common ways to deal with cancer.These two methods are useful for early-stage cancers in which the cancer cells are small.However,chemotherapy and radiation treatment is harmful to normal cells,so that developing methods with better tumor targeting and fewer side effects has been a research hotspot.In recent years,the precision medicine,personalized medicine and smart medicine are attracting more and more attention.Expecially,the development of nanotechnology has provided new opportunities for early diagnosis and targeted therapy of tumors.Gold NanoParticles(GNPs)are easier to aggregate inside the tumor due to their minimal biotoxicity and target specificity.The radiation dose or drug concentration inside the tumor tissue is increased but the damage to the surrounding normal tissue is reduced during treatment.Knowledge of the distribution and concentration of GNPs within the tumor is the key to the success of the above methods.X-ray Fluorescence Computed Tomography(XFCT),which combines X-ray fluorescence(XRF)analysis with Computed Tomography(CT),can identify and quantify the concentration and distribution of metal elements by detecting the emission of XRF and using specific reconstruction algorithms.Compared to synchrotron radiation used in regular XFCT,an X-ray tube has a lower cost and a smaller size.Therefore,XFCT imaging based on an X-ray tube has become a research hotspot in precision medicine and personalized medicine.For this reason,the author completed thesis study work under the supervision of Prof.Sang Hyun Cho in the Department of Radiation Physics,University of Texas MD Anderson Cancer Center,USA.This was after receiving the scholarship granted by China Scholarship Council(Admission No.201606050045,File No.[2016] 3100).The study work was partially supported by U.S.NIH Grant R01EB020658,National Natural Science Youth Foundation of China(No.61401049),Graduate Innovation Project of Chongqing(No.CYB16044),etc.Research of key technologies for multi-modality XFCT based on an x-ray tube source was completed under the supervision of a domestic and an overseas professor for two years each.The research contents of this thesis primarly include:(1)Studied parallel-collimated XFCT imaging theory and reconstruction algorithms.Focusing on the problem of preventing parallel-hole collimator from detecting the XRF signal produced outside of each detector elements' detection line,we proposed a Detector's Eye View(DEV)-based Order-subset Expectation Maximization(OSEM)method that considers the rates of the XRF signal originating from the same location within the imaging object that are seen by neighboring detector elements.Using experimental data,this algorithm was validated to improve image spatial resolution through comparisan with filter back projection and traditional OSEM.(2)Studied Locally Linear Embedding(LLE)-based geometry calibration.Using the measured projection data and the re-projected/calculated projection from the reconstructed images,this method completed calibration through updating the geometric parameters with LLE.Because any misalignment in the detection geometry would make the system matrix inaccurate and consequently affect the quality of reconstructed images,LLE-based geometry calibration was developed and tested to calibrate the offset between detector and rotational stage in the parallel-collimated XFCT system.Additionally,this method was used to calibrate the rigid patient motion and produce high quality XFCT images.(3)Fully and systematically studied the theory and performance of multi-pinhole XFCT systems based on an x-ray tube source.Two multi-pinhole XFCT systems,one parallel and one rotational,were built with Geant4 simulations.The parallel multi-pinhole XFCT system used a fan beam to scan a slice of an object and could obtain the 2D distribution with one scan.The rotational multi-pinhole XFCT system used a cone beam to scan am entire object and was able to do 3D imaging by rotating the object.The multi-pinhole diameter was changed to see the spatial resolution in the parallel multi-pinhole XFCT system.The detection limits and image qualities of these two systems were investigated and compared by scanning a polymethyl methacrylate phantom with different concentrations of GNPs and changing the pinhole-detector-distance.The performance of single pinhole and multi-pinhole were also compared,and it was shown that multi-pinhole systems performed better than single pinhole systems in terms of detection limits and GNPs insert shapes.(4)Fully and systematically studied the performance of a CT component in an experimental dual-mode benchtop XFCT and transmission CT system.The goal was to ensure that an experimental cone-beam CT/XFCT system utilizing a high power industrial x-ray source with considerably large focal spot sizes could produce CT images showing reasonable anatomical details for multimodal(CT/XFCT)preclinical imaging,while meeting the dose constraint for a separate CT scan.We used different tube voltages,currents and scanning methods(short and sparse scans)to do the imaging and calculated the contra-to-noise ratio,noise,detection limit,resolution and dose.The study showed that the transmission CT component within the experimental benchtop XFCT system produced images deemed acceptable for multimodal(CT + XFCT)imaging purposes,with low dose.(5)Studied CT and XFCT fusion.In order to combine the anatomical information from CT and the metal information from XFCT,we used the average,the Pulse Coupled Neural Network(PCNN)-based algorithm and the method based on Non-aliasing Contourlet(NACT)and PCNN to do a CT + XFCT fusion.Average,entropy,spatial frequency and edge preservation were calculated and analyzed to show that the fusion method based on NACT and PCNN is the best.This study worked on XFCT imaging based on an X-ray tube source,which has very important significance no matter in theory or in practical application.