United multi-task learning for abdominal contrast-enhanced CT synthesis through joint deformable registration. (April 2023)

• Record Type: Journal Article
• Authors: Zhong, Liming; Huang, Pinyu; Shu, Hai; Li, Yin; Zhang, Yiwen; Feng, Qianjin; Wu, Yuankui; Yang, Wei
• Is Part Of: Computer methods and programs in biomedicine. Volume 231(2023)
• Publication Date: 2023-04
• Subjects: Deep learning -- Multi-task learning -- Abdominal contrast-enhanced CT synthesis -- Generative adversarial network; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2023.107391
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: We simultaneously achieve the registration and synthesis of abdominal CECT. The joint-training can reduce misalignment errors in the synthesis model training. We adopt a multi-loss function to improve the accuracy of synthetic CECT images. The learning of enhanced components facilitates the synthesis model training. Superior performance is achieved on multi-resolution and multi-center CECT datasets. Abstract: Synthesizing abdominal contrast-enhanced computed tomography (CECT) images from non-enhanced CT (NECT) images is of great importance, in the delineation of radiotherapy target volumes, to reduce the risk of iodinated contrast agent and the registration error between NECT and CECT for transferring the delineations. NECT images contain structural information that can reflect the contrast difference between lesions and surrounding tissues. However, existing methods treat synthesis and registration as two separate tasks, which neglects the task collaborative and fails to address misalignment between images after the standard image pre-processing in training a CECT synthesis model. Thus, we propose an united multi-task learning (UMTL) for joint synthesis and deformable registration of abdominal CECT. Specifically, our UMTL is an end-to-end multi-task framework, which integrates a deformation field learning network for reducing the misalignment errors and a 3D generator for synthesizing CECT images. Furthermore, the learning of enhanced component images and the multi-loss function are adopted for enhancing the performance of synthetic CECT images. The proposed method is evaluated on two different resolution datasets and a separate test dataset from another center. The synthetic venous phase CECT images of the separate test dataset yield mean absolute error (MAE) of 32.78 ± 7.27 HU, mean MAE of 24.15 ± 5.12 HU on liver region, mean peak signal-to-noise rate (PSNR) of 27.59 ± 2.45 dB, and mean structural similarity (SSIM) of 0.96 ± 0.01 . The Dice similarity coefficients of liver region between the true and synthetic venous phase CECT images are 0.96 ± 0.05 (high-resolution) and 0.95 ± 0.07 (low-resolution), respectively. The proposed method has great potential in aiding the delineation of radiotherapy target volumes.