Dual-energy CT (DECT) was introduced to address the inability of conventional single-energy computed tomography (SECT) to distinguish materials with similar absorbances but different elemental compositions. However, material decomposition algorithms based purely on the physics of the underlying attenuation process have several limitations, leading to low signal-to-noise ratio (SNR) in the derived material-specific images. To overcome these, we trained a convolutional neural network (CNN) to develop a framework to reconstruct non-contrast SECT images from DECT scans. We show that the traditional physics-based decomposition algorithms do not bring to bear the full information content of the image data. A CNN that leverages the underlying physics of the DECT image generation process as well as the anatomic information gleaned via training with actual images can generate higher fidelity processed DECT images.

A fter an intracerebral hemorrhage (ICH), hematoma expansion by 33% or more has been reported in at least 38% of patients within 24 hours after symptom onset (1), and in up to 50% of patients who are on anticoagulation therapy (2). Early hematoma expansion is strongly associated with neurologic deterioration, worse functional outcome, and mortality after ICH (3). Thus, identification of patients at risk for hematoma expansion may help to direct management, in particular with respect to selecting candidates for early targeted medical or surgical intervention (4). Several imaging markers have been proposed to assess greater risk of hematoma expansion, including spot sign at CT angiography (5), defined as foci of enhancement within a hematoma due to active contrast media extravasation (6). However, the test characteristics of spot sign are not optimal for predicting hematoma expansion (7,8). For example, in the large Prediction of Haemotoma Growth and Outcome in Patients with Intracerebral Haemorrhage using the CT-angiography Spot Sign (PREDICT) trial, when the spot sign was assessed at arterial phase imaging, it demonstrated sensitivity only slightly greater than 50% (9,10). In other words, although presence of the spot sign is highly predictive of hematoma expansion, there is room for improvement in its sensitivity (even including assessment on delayed images and at dynamic CT angiography) so that it may be used as a robust predictor of both stability and expansion of ICH. One of the difficulties in spot sign reading is differentiating hyperdense hemorrhage from contrast media staining of the brain parenchyma due to spotty or diffuse contrast media extravasation from leaky blood vessels (11,12). Dual-energy CT can be helpful in these situations through

No abstract