2.1

Data Description

Thirteen high quality cardiac patient cases acquired from a UHR-CT system were used as the learning target. The acquisition energy and dose varied from 120 ~ 140 kVp and 150 ~ 190 mAs. Different reconstruction field of views and doses were applied and simulated for enriching data diversity. All target image dimensions reconstructed with model-based iterative reconstruction (MBIR) method were 1024x1024 pixels by 0.25 mm thickness. The acquired high-resolution data were first converted to the pre-log count domain, and then 2x2 data binning was applied to mimic the NR data. For more realistic data, additional white noise can be added at this step to compensate for the electronic noise. All simulated NR data were reconstructed with filtered-back projection (FBP) and standard ramp filter in the dimensions of 1024x1024 pixels and 0.25mm reconstruction pitch. The reconstructed SR-DLR images are compared with FBP and MBIR images. For quantitative comparison, the Catphan-600 phantom with different modules were used to evaluate the resolution improvement in terms of MTF (modulation transfer function) and line pairs (lp/cm) analysis. One true UHR-CT patient data (140kVp, 164mAs, 150mm FOV, 0.25mm reconstruction pitch) also used for comparison as shown in Figure 4. NR simulation process as stated in previous publication [7]. Another three real normal resolution (NR-CT) patient scans (see Table 1) were reconstructed to test the image quality through visual inspection. All images in result section reconstructed in default 512x512 pixels (SR-DLR in 1024x1024 pixels) with 240-mm field of view and 0.25 mm reconstruction pitch.

Table 1

Dose information for three representative patient cases

2.2

Network Architecture, Training and Implementation

The SR-DLR network adopts the U-NET structure with residual 3D convolutional blocks as basic building units, which enables deep structure without gradient vanishing and maintains proper receptive field for capturing large spatial features in the image (see Figure 2) [8, 9]. In network training we used small patches of size 32x128x128. Data argumentation such as patch flipping were also performed during the training. The training was mainly conducted using the TensorFlow-Keras (ver. 2.3) framework with Nvidia Titan RTX GPUs for acceleration. We chose the ADAM optimizer and the mean absolute error loss function. A total 200 epochs were run for effective convergence.

Figure 2

The proposed SR-DLR network which includes a 3D-UNET structure with residual blocks. The number beside the convolutional layer represents the number of filters.

3.1

Quantitative Evaluation

SR-DLR imaging shows superior 200mm FOV MTF performance as compared to FBP and MBIR reconstruction among all three materials. For low-contrast polystyrene (~-30HU), SR-DLR improves 10%-MTF to 1.04 lp/mm compares to 0.77 in FBP and 0.80 in MBIR. For mid-contrast Delrin (~330HU), 10%-MTF in SR-DLR is 1.15 lp/mm compares to 0.86 in FBP and 0.91 in MBIR. For high-contrast Teflon (~900HU), 10%-MTF in SR-DLR is 1.24 lp/mm compares to 0.86 in FBP and 1.11 in MBIR. (see Figure 3).

Figure 3

Modulation transfer function measurement as a function of line pair per millimeter with three different representative materials (polystyrene, Delrin and Teflon) in standard CATPHAN phantom (CTP404 module). Three different reconstruction methods (FBP in blue, MBIR in orange and proposed SR-DLR in green) listed from 0 to 1.25 lp/mm for comparison.

3.2

Comparison with True UHR-CT Images

Figure 4 compares the SR-DLR image to the closest-ground-truth image (e.g. true UHR-CT MBIR image) so that the resolution performance can be examined. The simulated-NR SR-DLR phantom image has better bar resolution compared to the FBP image (~9 lp/cm vs. ~8 lp/cm). The image contrast as well as the bar intensity of SR-DLR is also closer to the ground truth image compared to the FBP reconstruction. The simulated NR SR-DLR patient image shows better resolution in stent and higher contrast in soft tissue compares to the FBP image. No artificial feature created when we compare NR SR-DLR to the true UHR-CT patient image.

Figure 4

Resolution phantom module (CTP528 module) and patient examples. Upper row: two reconstruction methods with normal resolution (NR) data to the true UHR-CT MBIR image. The 9 lp/cm contrast bars are zoomed in the bottom-right small image. Window level and width: 1000 and 3000 HU. Lower row: two simulated NR data for FBP and SR-DLR reconstructions to the true UHR-CT MBIR patient image. Window level and width: 350 and 1500 HU

3.3

Compare SR-DLR Image to Other Reconstructions

In Figure 5, we are able to see the noise and resolution differences when comparing SR-DLR images to FBP and MBIR reconstructions. Patient (A) images shows that the SR-DLR image has better resolution for the aortic valve and better contrast of small vessels. Patient (B) images further show cleaner microstructure of the implanted stent. Patient case (C) shows the SR-DLR image has sharper calcification contrast and boundaries compared to other two reconstruction methods. In addition, SR-DLR has lowest noise and sharpest soft tissue boundaries among these three methods.

Figure 5

Three representative patient cases (case A: aortic valve; case B: stent; case C: calcium) with three different reconstruction methods. Window level and width for case A and C: 300 and 800 HU; for case B: 350 and 1500 HU.