The evaluation of the reduction of radiation dose via deep learning-based reconstruction for cadaveric human lung CT images

Abstract

To compare the quality of CT images of the lung reconstructed using deep learning-based reconstruction (True Fidelity Image: TFI ™; GE Healthcare) to filtered back projection (FBP), and to determine the minimum tube current–time product in TFI without compromising image quality. Four cadaveric human lungs were scanned on CT at 120 kVp and different tube current–time products (10, 25, 50, 75, 100, and 175 mAs) and reconstructed with TFI and FBP. Two image evaluations were performed by three independent radiologists. In the first experiment, using the same tube current–time product, a side-by-side TFI and FBP comparison was performed. Images were evaluated with regard to noise, streak artifacts, and overall image quality. Overall image quality was evaluated in view of whole image quality. In the second experiment, CT images reconstructed using TFI and FBP with five different tube current–time products were displayed in random order, which were evaluated with reference to the 175 mAs-FBP image. Images were scored with regard to normal structure, abnormal findings, noise, streak artifacts, and overall image quality. Median scores from three radiologists were statistically analyzed. Quantitative evaluation of noise was performed by setting regions of interest (ROIs) in air. In first experiment, overall image quality was improved, and noise was decreased in images of TFI compared to that of FBP for all tube current–time products. In second experiment, scores of all evaluation items except for small vessels in images of 25 mAs-TFI were almost the same as that of 175 mAs-FBP (all p > 0.31). Using TFI instead of FBP, at least 85% radiation dose reduction could be possible without any degradation in the image quality.

Patient-derived PixelPrint phantoms for evaluating clinical imaging performance of a deep learning CT reconstruction algorithm

“…Studies evaluating other DLR algorithms have also demonstrated reduced noise and improved lesion detectability in DLR compared to IR or FBP (Son et al 2022 , Miyata et al 2022 , Park et al 2022 , Greffier et al 2022b , Koetzier et al 2023 ). The exact percentage of dose reduction reported was heavily dependent upon many factors including the clinical scenario, the reference dose, reference reconstruction algorithm, DLR algorithm and denoising strength, and specific metrics evaluated.…”

“…With the rise of commercially available DLR algorithms, there has been an increase in studies evaluating DLR. Multiple patient and phantom studies have demonstrated that DLR can improve image quality at low doses through enhanced lesion detectability and reduced noise (Akagi et al 2019 , Nakamura et al 2019 , Nagayama et al 2021 , Sun et al 2021 , Greffier et al 2022a , Miyata et al 2022 , Park et al 2022 , Greffier et al 2022b , Mikayama et al 2022 , Son et al 2022 , Greffier et al 2023a ). These studies utilize quantitative metrics such as signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), noise, detectability index (d’), and noise power spectrum (NPS).…”

Improved image quality in CT pulmonary angiography using deep learning-based image reconstruction

“…A phantom study demonstrated higher accuracy of ultra-low-dose chest with DLIR for volumetric assessment of ground glass nodules compared to model-based iterative reconstruction and hybrid iterative reconstruction 31 . A study on cadaveric human lungs reported that with DLIR dose could be reduced by up to 85% with preserved image quality compared to FBP 32 . Others have used combinations of deep learning-based image denoising and iterative reconstruction and reported improved assessment of pulmonary nodules when both were combined 33 .…”

Patient-derived PixelPrint phantoms for evaluating clinical imaging performance of a deep learning CT reconstruction algorithm

“…A patient study by Greffier et. al 24 18 leading to a range of dose reduction estimates between 30-80% 1,20,21 . Overall, our results align closely with the results presented in previous studies evaluating DLR, which supports the validity of using PixelPrint phantoms for evaluation of DLR algorithms.…”