Deep convolutional neural network for reduction of contrast-enhanced region on CT images.

Iori Sumida,Kyohei Yamashita,Indra J Das,Yuji Yamada,Hisao Kizaki,Keiko Aboshi,Taiki Magome,Yuji Seo,Hajime Yamaguchi,Hideki Kitamori,Fumiaki Isohashi,Kazuhiko Ogawa

doi:10.1093/jrr/rrz030

Abstract

This study aims to produce non-contrast computed tomography (CT) images using a deep convolutional neural network (CNN) for imaging. Twenty-nine patients were selected. CT images were acquired without and with a contrast enhancement medium. The transverse images were divided into 64 × 64 pixels. This resulted in 14 723 patches in total for both non-contrast and contrast-enhanced CT image pairs. The proposed CNN model comprises five two-dimensional (2D) convolution layers with one shortcut path. For comparison, the U-net model, which comprises five 2D convolution layers interleaved with pooling and unpooling layers, was used. Training was performed in 24 patients and, for testing of trained models, another 5 patients were used. For quantitative evaluation, 50 regions of interest (ROIs) were selected on the reference contrast-enhanced image of the test data, and the mean pixel value of the ROIs was calculated. The mean pixel values of the ROIs at the same location on the reference non-contrast image and the predicted non-contrast image were calculated and those values were compared. Regarding the quantitative analysis, the difference in mean pixel value between the reference contrast-enhanced image and the predicted non-contrast image was significant (P < 0.0001) for both models. Significant differences in pixels (P < 0.0001) were found using the U-net model; in contrast, there was no significant difference using the proposed CNN model when comparing the reference non-contrast images and the predicted non-contrast images. Using the proposed CNN model, the contrast-enhanced region was satisfactorily reduced.

Highlights

Computed tomography (CT) imaging has been used for diagnostic imaging and planning of radiation therapy
We proposed here a convolutional neural network (CNN) structure for reducing the contrastenhanced region from contrast-enhanced CT images, and compared it with another CNN model named the U-net
The image pairs for training and testing were made from image patches based on the same method as shown in the studies conducted by Nishio et al [15] and Chen et al

Summary

Introduction

Computed tomography (CT) imaging has been used for diagnostic imaging and planning of radiation therapy. CT examination can be conducted with images obtained both with and without a contrast medium to identify contrast-enhanced regions, to delineate the contours of structures and to calculate patient doses. Using contrast images changes the reading on the Hounsfield unit (HU) scale and produces streaking artifacts. This creates problems in treatment planning, leading to inaccurate dosimetry for both the target volume and the organs at risk. To overcome this problem, two sequential images are acquired without and with contrast medium. Taking two images is more expensive and increases the radiation dose to the patient [1, 2]

Objectives

Methods

Results

Conclusion