Experimental investigation of neural network estimator and transfer learning techniques for K-edge spectral CT imaging.

PURPOSE: Spectral computed tomography (CT) material decomposition algorithms require accurate physics-based models or empirically derived models. This study investigates a machine learning algorithm and transfer learning techniques for Spectral CT imaging of K-edge contrast agents using simulated and experimental measurements. METHODS: A feed forward multilayer perceptron was implemented and trained on data acquired using a step wedge phantom containing acrylic, aluminum, and gadolinium materials. The neural network estimator was evaluated by scanning a rod phantom with varying dilutions of gadolinium oxide nanoparticles and by scanning a rat leg specimen with injected nanoparticles on a bench-top photon-counting computed tomography system. The algorithm decomposed each spectral projection measurement into path lengths of acrylic and aluminum and mass lengths of gadolinium. Each basis material sinogram was reconstructed into basis material images using filtered backprojection. Machine learning techniques of data standardization, transfer learning from aggregated pixel data, and transfer learning from simulations were investigated to improve image quality. The algorithm was compared to a previously published empirical method based on a linear approximation and calibration error look-up tables. RESULTS: The combined transfer learning techniques did not improve quantification in the rod phantom and provided only a small qualitative improvement in ring artifacts. Transfer learning from aggregated pixel data and from simulations improved the qualitative image quality of the rat specimen, for which the calibration data were limited. Transfer learning from aggregated pixel data and simulations estimated 3.26, 6.26, and 12.45 mg/mL Gd concentrations compared to true 2.72, 5.44, and 10.88 mg/mL concentrations in the rod phantom. Additionally, the neural networks were able to separate the soft tissue, bone, and gadolinium nanoparticles of the ex vivo rat leg specimen into the different basis images. CONCLUSIONS: The results demonstrate that empirical K-edge imaging from calibration measurements using machine learning and transfer learning is possible without explicit models of material attenuations, incident spectra, or the detector response.

Experimental comparison of empirical material decomposition methods for spectral CT.

Material composition can be estimated from spectral information acquired using photon counting x-ray detectors with pulse height analysis. Non-ideal effects in photon counting x-ray detectors such as charge-sharing, k-escape, and pulse-pileup distort the detected spectrum, which can cause material decomposition errors. This work compared the performance of two empirical decomposition methods: a neural network estimator and a linearized maximum likelihood estimator with correction (A-table method). The two investigated methods differ in how they model the nonlinear relationship between the spectral measurements and material decomposition estimates. The bias and standard deviation of material decomposition estimates were compared for the two methods, using both simulations and experiments with a photon-counting x-ray detector. Both the neural network and A-table methods demonstrated a similar performance for the simulated data. The neural network had lower standard deviation for nearly all thicknesses of the test materials in the collimated (low scatter) and uncollimated (higher scatter) experimental data. In the experimental study of Teflon thicknesses, non-ideal detector effects demonstrated a potential bias of 11-28%, which was reduced to 0.1-11% using the proposed empirical methods. Overall, the results demonstrated preliminary experimental feasibility of empirical material decomposition for spectral CT using photon-counting detectors.

The effects of extending the spectral information acquired by a photon-counting detector for spectral CT.

Photon-counting x-ray detectors with pulse-height analysis provide spectral information that may improve material decomposition and contrast-to-noise ratio (CNR) in CT images. The number of energy measurements that can be acquired simultaneously on a detector pixel is equal to the number of comparator channels. Some spectral CT designs have a limited number of comparator channels, due to the complexity of readout electronics. The spectral information could be extended by changing the comparator threshold levels over time, sub pixels, or view angle. However, acquiring more energy measurements than comparator channels increases the noise and/or dose, due to differences in noise correlations across energy measurements and decreased dose utilisation. This study experimentally quantified the effects of acquiring more energy measurements than comparator channels using a bench-top spectral CT system. An analytical and simulation study modeling an ideal detector investigated whether there was a net benefit for material decomposition or optimal energy weighting when acquiring more energy measurements than comparator channels. Experimental results demonstrated that in a two-threshold acquisition, acquiring the high-energy measurement independently from the low-energy measurement increased noise standard deviation in material-decomposition basis images by factors of 1.5-1.7 due to changes in covariance between energy measurements. CNR in energy-weighted images decreased by factors of 0.92-0.71. Noise standard deviation increased by an additional factor of [Formula: see text] due to reduced dose utilisation. The results demonstrated no benefit for two-material decomposition noise or energy-weighted CNR when acquiring more energy measurements than comparator channels. Understanding the noise penalty of acquiring more energy measurements than comparator channels is important for designing spectral detectors and for designing experiments and interpreting data from prototype systems with a limited number of comparator channels.