one-beam computed tomography noise reduction method based on U-Net with convolutional block attention module in proton therapy

one-beam computed tomography (CBCT) is mostly used for position verification during the treatment pro<br />cess. However, severe image artifacts in CBCT hinder its direct use in dose calculation and adaptive radiation<br />therapy re-planning for proton therapy. In this study, an improved U-Net neural network named CBAM-U-Net<br />was proposed for CBCT noise reduction in proton therapy, which is a CBCT denoised U-Net network with con<br />volutional block attention modules. The datasets contained 20 groups of head and neck images. The CT images<br />were registered to CBCT images as ground truth. The original CBCT denoised U-Net network, sCTU-Net, was<br />trained for model performance comparison. The synthetic CT(SCT) images generated by CBAM-U-Net and the<br />original sCTU-Net are called CBAM-SCT and U-Net-SCT images, respectively. The HU accuracies of the CT,<br />CBCT, and SCT images were compared using four metrics: mean absolute error (MAE), root mean square error<br />(RMSE), peak signal-to-noise ratio (PSNR), and structure similarity index measure (SSIM). The mean values of<br />the MAE, RMSE, PSNR, and SSIM of CBAM-SCT images were 23.80 HU, 64.63 HU, 52.27 dB, and 0.9919,<br />respectively, which were superior to those of the U-Net-SCT images. To evaluate dosimetric accuracy, the range<br />accuracy was compared for a single-energy proton beam. The γ-index pass rates of a 4 cm × 4 cm scanned<br />field and simple plan were calculated to compare the effects of the noise reduction capabilities of the original<br />U-Net and CBAM-U-Net on the dose calculation results. CBAM-U-Net reduced noise more effectively than<br />sCTU-Net, particularly in high-density tissues. We proposed a CBAM-U-Net model for CBCT noise reduction<br />in proton therapy. Owing to the excellent noise reduction capabilities of CBAM-U-Net, the proposed model<br />provided relatively explicit information regarding patient tissues. Moreover, it can be used in dose calculation<br />and adaptive treatment planning in the future.