Convolutional neural network enhancement of fast-scan low-dose cone-beam CT images for head and neck radiotherapy

Abstract

To improve image quality and CT number accuracy of fast-scan low-dose cone-beam computed tomography (CBCT) through a deep-learning convolutional neural network (CNN) methodology for head-and-neck (HN) radiotherapy. Fifty-five paired CBCT and CT images from HN patients were retrospectively analysed. Among them, 15 patients underwent adaptive replanning during treatment, thus had same-day CT/CBCT pairs. The remaining 40 patients (post-operative) had paired planning CT and 1st fraction CBCT images with minimal anatomic changes. A 2D U-Net architecture with 27-layers in 5 depths was built for the CNN. CNN training was performed using data from 40 post-operative HN patients with 2080 paired CT/CBCT slices. Validation and test datasets include 5 same-day datasets with 260 slice pairs and 10 same-day datasets with 520 slice pairs, respectively. To examine the impact of differences in training dataset selection and network performance as a function of training data size, additional networks were trained using 30, 40 and 50 datasets. Image quality of enhanced CBCT images were quantitatively compared against the CT image using mean absolute error (MAE) of Hounsfield units (HU), signal-to-noise ratio (SNR) and structural similarity (SSIM). Enhanced CBCT images reduced artifact distortion and improved soft tissue contrast. Networks trained with 40 datasets had imaging performance comparable to those trained with 50 datasets and outperformed those trained with 30 datasets. Comparison of CBCT and enhanced CBCT images demonstrated improvement in average MAE from 172.73 to 49.28 HU, SNR from 8.27 to 14.25 dB, and SSIM from 0.42 to 0.85. The image processing time is 2 s per patient using a NVIDIA GeForce GTX 1080 Ti GPU. The proposed deep-leaning methodology was fast and effective for image quality enhancement of fast-scan low-dose CBCT. This method has potential to support fast online-adaptive re-planning for HN cancer patients.

Improving CBCT quality to CT level using deep learning with generative adversarial network

“…The style loss was used to control the similarity of image styles and was defined as the Euclidian distance between the stylistic feature maps from original and synthesized images of each layer:where Gram matrix was defined as:where m and n represent different output channels from the same layer. So the loss function becomesβ 1 and β 2 are the weights.In addition, we also compared our methods with previously published models as U‐net 17,18 and cycleGAN 19 . U‐net is a popular algorithm in image processing field and some investigators have explored its use in this context 17,18,20 .…”

Clinical Enhancement in AI-Based Post-processed Fast-Scan Low-Dose CBCT for Head and Neck Adaptive Radiotherapy

“…A 2D U-Net shape architecture with 19-layers in 5 depths was specially optimized and trained using a total of 2080 CT and CBCT slice. The network design and architecture were described in the previous study ( Yuan et al, 2019 ). Additional 15 patients with pCT, and replanning CT (rCT) 3–4 weeks into treatment with the same-day CBCT in relation to rCT were selected for DCNN validation.…”

“…Deep convolutional neural networks (DCNN) can denoise images, reduce blurring, and improve soft tissue contrast resolution ( Jain, 2008 ; Dong et al, 2016 ). Specifically for those fast-scan-low-dose CBCT scans, a U-NET based DCNN was developed for enhancing image quality for HNC patients, with improved HU accuracy, signal-to-noise ratio, and small anatomical structure preservation ( Yuan et al, 2019 ). Such image quality enhancement should bring clinical benefits specifically for ART, including improved CT-CBCT image registration accuracy, thus improved contour propagation accuracy and better visualization for identifying organs at risk on CBCT images.…”

Medical Physics and Imaging–A Timely Perspective

“…For tumor outlining and organs-at-risk segmentation, AI technology based on deeplearning is at the same level as human expertise [71,72]. Further, image reconstruction approaches and radiation treatment plan optimization algorithms also benefit from AI-driven solutions to allow for fast and accurate results [73][74][75]. Finally, the complexities of integrating the various hardware and software components of standard and advanced RT workflows can be addressed only through a multi-disciplinary team effort that supports active interactions between medical physicists, medical doctors and technologists, along with engineers, IT and datascience engineers [76][77][78][79].…”