Evaluation of CT Image Synthesis Methods:From Atlas-based Registration to Deep Learning

Lauritzen, Andreas D.; Papademetris, Xenophon; Turovets, Sergei; Onofrey, John A.

doi:10.48550/arxiv.1906.04467

Cited by 3 publications

(5 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For global image similarity between sCT and ground truth CT, mean absolute error (MAE) and peak-signal-to-noise-ratio (PSNR) metrics were utilized to evaluate the HU intensity space, which is important to preserve clinically relevant electron density information. The contrast and the anatomy were evaluated using the mean structural similarity (MSSIM) (Dong et al 2017), Pearson cross-correlation (PCC) coefficient (Lauritzen et al 2019), Frechet inception distance (FID) (Heusel et al 2017), and sliced Wasserstein distance (SWD) (Deshpande et al 2018) measures. We also evaluated the shape quality of the different HU thresholded areas of the sCT image using the Dice score (Crum et al 2006, Milletari et al 2016, where the sCT image is segmented into to three different regions corresponding to bone (HU > 300), air (HU < −100) and soft tissues (−100 < HU < 300).…”

Section: Training and Evaluationmentioning

confidence: 99%

Multi-planar dual adversarial network based on dynamic 3D features for MRI-CT head and neck image synthesis

Touati,

Trung Le,

Kadoury

2024

Phys. Med. Biol.

View full text Add to dashboard Cite

Objective. Head and neck radiotherapy planning requires electron densities from different tissues for dose calculation.Dose calculation from imaging modalities such as MRI remains an unsolved problem since this imaging modality does not provide information about the density of electrons. Approach. We propose a generative adversarial network (GAN) approach that synthesizes CT (sCT) images from T1-weighted MRI acquisitions in head and neck cancer patients. Our contribution is to exploit new features that are relevant for improving multimodal image synthesis, and thus improving the quality of the generated CT images. More precisely, we propose a Dual branch generator based on the U-Net architecture and on an augmented multi-planar branch. The augmented branch learns specific 3D dynamic features, which describe the dynamic image shape variations and are extracted from different view-points of the volumetric input MRI. The architecture of the proposed model relies on an end-to-end convolutional U-Net embedding network.Results. The proposed model achieves a mean absolute error (MAE) of $18.76 (5.167)$ in the target Hounsfield unit (HU) space on sagittal head and neck patients, with a mean structural similarity (MSSIM) of $0.95 (0.09)$ and a Frechet inception distance (FID) of $145.60 (8.38)$. The model yields a MAE of $26.83 (8.27)$ to generate specific primary tumor regions on axial patient acquisitions, with a Dice score of $0.73 (0.06)$ and a FID distance equal to $122.58 (7.55)$. The improvement of our model over other state-of-the-art GAN approaches is of $3.8\%$, on a tumor test set. On both sagittal and axial acquisitions, the model yields the best peak signal-to-noise ratio (PSNRs) of $27.89 (2.22)$ and $26.08 (2.95)$ to synthesize MRI from CT input.Significance. The proposed model synthesizes both sagittal and axial CT tumor images, used for radiotherapy treatment planning in head and neck cancer cases. The performance analysis across different imaging metrics and under different evaluation strategies demonstrates the effectiveness of our dual CT synthesis model to produce high quality sCT images compared to other state-of-the-art approaches. Our model could improve clinical tumor analysis, in which a further clinical validation remains to be explored.

show abstract

Section: Training and Evaluationmentioning

confidence: 99%

Multi-planar dual adversarial network based on dynamic 3D features for MRI-CT head and neck image synthesis

Touati,

Trung Le,

Kadoury

2024

Phys. Med. Biol.

View full text Add to dashboard Cite

show abstract

“…Specifically, while MAE and PSNR metrics are direct error and image quality measures computed at a pixel-wise level (Huynh et al 2016, Dong et al 2017, Oulbacha and Kadoury 2020, MSSIM, FID and SWD are used to evaluate the contrast and the anatomy of the sCT image, which are important to assess tumor burden in medical imaging. PCC allows us to evaluate how similar the generated anatomies are at a global image-level (Lauritzen et al 2019). BD is an interesting metric for its ability to compare histograms, which in the case of CT images have meaningful HU values, important for downstream dose calculation tasks.…”

Section: Implementation Details and Experimental Settingmentioning

confidence: 99%

“…The batch size is fixed to 1.2.6. Evaluation metricsFor the quantitative evaluation, the quality of the synthesized CT images compared to the reference CT is evaluated using different imaging metrics(Bae and Kim 2015, Huynh et al 2016, Dong et al 2017, Lauritzen et al 2019, Oulbacha and Kadoury 2020. We calculate the mean absolute error (MAE) and peak-signal-to-noiseratio (PSNR), between intensity values of the ground truth CT image and the sCT image in the HU intensity space.…”

mentioning

confidence: 99%

“…We also calculate the mean structural similarity (MSSIM) index, which does not consider the intensity difference, but allows us to evaluate the contrast level and the anatomical structure of the sCT image compared with the corresponding ground truth(Dong et al 2017). The last metric measures the Pearson cross-correlation (PCC) coefficient(Lauritzen et al 2019). Besides image quality metrics, we also compute the Frechet inception distance (FID)…”

mentioning

confidence: 99%

See 1 more Smart Citation

A feature invariant generative adversarial network for head and neck MRI/CT image synthesis

Touati

Kadoury

2021

Phys. Med. Biol.

View full text Add to dashboard Cite

With the emergence of online MRI radiotherapy treatments, MR-based workflows have increased in importance in the clinical workflow. However proper dose planning still requires CT images to calculate dose attenuation due to bony structures. In this paper, we present a novel deep image synthesis model that generates in an unsupervised manner CT images from diagnostic MRI for radiotherapy planning. The proposed model based on a generative adversarial network (GAN) consists of learning a new invariant representation to generate synthetic CT (sCT) images based on high frequency and appearance patterns. This new representation encodes each convolutional feature map of the convolutional GAN discriminator, leading the training of the proposed model to be particularly robust in terms of image synthesis quality. Our model includes an analysis of common histogram features in the training process, thus reinforcing the generator such that the output sCT image exhibits a histogram matching that of the ground-truth CT. This CT-matched histogram is embedded then in a multi-resolution framework by assessing the evaluation over all layers of the discriminator network, which then allows the model to robustly classify the output synthetic image. Experiments were conducted on head and neck images of 56 cancer patients with a wide range of shape sizes and spatial image resolutions. The obtained results confirm the efficiency of the proposed model compared to other generative models, where the mean absolute error yielded by our model was 26.44(0.62), with a Hounsfield unit error of 45.3(1.87), and an overall Dice coefficient of 0.74(0.05), demonstrating the potential of the synthesis model for radiotherapy planning applications.

show abstract

“…As an alternative to working at the pixel-level, highlevel classification or detection approaches have been proposed [42], [43], [44], which can allow medical imaging experts to rapidly locate areas of infection, thus speeding up the diagnosis process. Though two CT image synthesis methods have been previously reported [45], [46], they did not focus on COVID-19 or lung CT imaging; to the best of our knowledge, our proposed method is the first designed specifically for COVID-19 CT image synthesis.…”

Section: Related Workmentioning

confidence: 99%

COVID-19 CT Image Synthesis with a Conditional Generative Adversarial Network

Jiang

Chen

Loew

et al. 2020

Preprint

View full text Add to dashboard Cite

Coronavirus disease 2019 (COVID-19) is an ongoing global pandemic that has spread rapidly since December 2019. Real-time reverse transcription polymerase chain reaction (rRT-PCR) and chest computed tomography (CT) imaging both play an important role in COVID-19 diagnosis. Chest CT imaging offers the benefits of quick reporting, a low cost, and high sensitivity for the detection of pulmonary infection. Recently, deep-learning-based computer vision methods have demonstrated great promise for use in medical imaging applications, including X-rays, magnetic resonance imaging, and CT imaging. However, training a deep-learning model requires large volumes of data, and medical staff faces a high risk when collecting COVID-19 CT data due to the high infectivity of the disease. Another issue is the lack of experts available for data labeling. In order to meet the data requirements for COVID-19 CT imaging, we propose a CT image synthesis approach based on a conditional generative adversarial network that can effectively generate high-quality and realistic COVID-19 CT images for use in deep-learningbased medical imaging tasks. Experimental results show that the proposed method outperforms other state-of-the-art image synthesis methods with the generated COVID-19 CT images and indicates promising for various machine learning applications including semantic segmentation and classification.

show abstract

Evaluation of CT Image Synthesis Methods:From Atlas-based Registration to Deep Learning

Cited by 3 publications

References 17 publications

Multi-planar dual adversarial network based on dynamic 3D features for MRI-CT head and neck image synthesis

Multi-planar dual adversarial network based on dynamic 3D features for MRI-CT head and neck image synthesis

A feature invariant generative adversarial network for head and neck MRI/CT image synthesis

COVID-19 CT Image Synthesis with a Conditional Generative Adversarial Network

Contact Info

Product

Resources

About