A data and knowledge driven approach for SPECT using convolutional neural networks and iterative algorithms

We propose a data and knowledge driven approach for SPECT by combining a classical iterative algorithm of SPECT with a convolutional neural network. The classical iterative algorithm, such as ART and ML-EM, is employed to provide the model knowledge of SPECT. A modified U-net is then connected to exploit further features of reconstructed images and data sinograms of SPECT. We provide mathematical formulations for the architecture of the proposed networks. The networks are trained by supervised learning using the technique of mini-batch optimization. We apply the trained networks to the problems of simulated lung perfusion imaging and simulated myocardial perfusion imaging, and numerical results demonstrate their effectiveness of reconstructing source images from noisy data measurements.

Value of Dual-energy Lung Perfusion Imaging Using a Dual-source CT System for the Pulmonary Embolism

ObjectiveTo investigate the diagnostic value of dual-energy lung perfusion imaging (DEPI) using a dual-source CT system for the pulmonary embolism (PE).Methods50 patients in high acute PE prevalence were enrolled to accept the DEPI (lung perfusion image and CTA image of pulmonary artery acquired through the Dual Energy software) and emergent DSA angiography (golden diagnostic criterion).ResultsPatients using CT had significantly reduced examination duration and dosage of contrast agent than those using DSA examination, (P < 0.05). In total, 260 pulmonary arteries and 1020 pulmonary segments were examined through CTA, in which embolisms were identified in 50 lobes of lung, 108 pulmonary segments and 82 sub-segments. Reduction or lack of perfusion was identified through DEPI in 48 lobes of lung (concordance rate of 96.0%), 103 pulmonary segments (concordance rate of 95.4%) and 78 subsegments (concordance rate of 95.1%). The comparison of embolism quantity and morphological characteristics of pulmonary artery between CTA images and DEPI images showed no statistically significant difference.ConclusionBetter application value can be achieved in the diagnosis of pulmonary embolism by dual-energy lung perfusion imaging using a dual-source CT system.