No abstract

The aim is to develop and evaluate machine learning methods for generating quantitative parametric maps of effective atomic number (Z), relative electron density (ρ ), mean excitation energy (I ), and relative stopping power (RSP) from clinical dual-energy CT data. The maps could be used for material identification and radiation dose calculation. Machine learning methods of historical centroid (HC), random forest (RF), and artificial neural networks (ANN) were used to learn the relationship between dual-energy CT input data and ideal output parametric maps calculated for phantoms from the known compositions of 13 tissue substitutes. After training and model selection steps, the machine learning predictors were used to generate parametric maps from independent phantom and patient input data. Precision and accuracy were evaluated using the ideal maps. This process was repeated for a range of exposure doses, and performance was compared to that of the clinically-used dual-energy, physics-based method which served as the reference. The machine learning methods generated more accurate and precise parametric maps than those obtained using the reference method. Their performance advantage was particularly evident when using data from the lowest exposure, one-fifth of a typical clinical abdomen CT acquisition. The RF method achieved the greatest accuracy. In comparison, the ANN method was only 1% less accurate but had much better computational efficiency than RF, being able to produce parametric maps in 15 s. Machine learning methods outperformed the reference method in terms of accuracy and noise tolerance when generating parametric maps, encouraging further exploration of the techniques. Among the methods we evaluated, ANN is the most suitable for clinical use due to its combination of accuracy, excellent low-noise performance, and computational efficiency.

Purpose Accurate photon attenuation assessment from MR data remains an unmet challenge in the thorax due to tissue heterogeneity and the difficulty of MR lung imaging. As thoracic tissues encompass the whole physiologic range of photon absorption, large errors can occur when using, for example, a uniform, water‐equivalent or a soft‐tissue‐only approximation. The purpose of this study was to introduce a method for voxel‐wise thoracic synthetic CT (sCT) generation from MR data attenuation correction (AC) for PET/MR or for MR‐only radiation treatment planning (RTP). Methods Acquisition: A radial stack‐of‐stars combining ultra‐short‐echo time (UTE) and modified Dixon (mDixon) sequence was optimized for thoracic imaging. The UTE‐mDixon pulse sequence collects MR signals at three TE times denoted as UTE, Echo1, and Echo2. Three‐point mDixon processing was used to reconstruct water and fat images. Bias field correction was applied in order to avoid artifacts caused by inhomogeneity of the MR magnetic field. Analysis: Water fraction and R2* maps were estimated using the UTE‐mDixon data to produce a total of seven MR features, that is UTE, Echo1, Echo2, Dixon water, Dixon fat, Water fraction, and R2*. A feature selection process was performed to determine the optimal feature combination for the proposed automatic, 6‐tissue classification for sCT generation. Fuzzy c‐means was used for the automatic classification which was followed by voxel‐wise attenuation coefficient assignment as a weighted sum of those of the component tissues. Performance evaluation: MR data collected using the proposed pulse sequence were compared to those using a traditional two‐point Dixon approach. Image quality measures, including image resolution and uniformity, were evaluated using an MR ACR phantom. Data collected from 25 normal volunteers were used to evaluate the accuracy of the proposed method compared to the template‐based approach. Notably, the template approach is applicable here, that is normal volunteers, but may not be robust enough for patients with pathologies. Results The free breathing UTE‐mDixon pulse sequence yielded images with quality comparable to those using the traditional breath holding mDixon sequence. Furthermore, by capturing the signal before T2* decay, the UTE‐mDixon image provided lung and bone information which the mDixon image did not. The combination of Dixon water, Dixon fat, and the Water fraction was the most robust for tissue clustering and supported the classification of six tissues, that is, air, lung, fat, soft tissue, low‐density bone, and dense bone, used to generate the sCT. The thoracic sCT had a mean absolute difference from the template‐based (reference) CT of less than 50 HU and which was better agreement with the reference CT than the results produced using the traditional Dixon‐based data. Conclusion MR thoracic acquisition and analyses have been established to automatically provide six distinguishable tissue types to generate sCT for MR‐based AC of PET/MR and for MR‐only RTP.