Generation of short-term follow-up chest CT images using a latent diffusion model in COVID-19

被引：0

作者：

Kawata, Naoko ^{[1
,2
]}

Iwao, Yuma ^{[3
,4
]}

Matsuura, Yukiko ^{[5
]}

Higashide, Takashi ^{[6
,7
]}

Okamoto, Takayuki ^{[3
]}

Sekiguchi, Yuki ^{[2
]}

Nagayoshi, Masaru ^{[5
]}

Takiguchi, Yasuo ^{[5
]}

Suzuki, Takuji ^{[1
]}

Haneishi, Hideaki ^{[3
]}

机构：

[1] Chiba Univ, Grad Sch Med, Dept Respirol, 1-8-1 Inohana,Chuo Ku, Chiba, Chiba 2608677, Japan

[2] Chiba Univ, Grad Sch Sci & Engn, Chiba 2638522, Japan

[3] Chiba Univ, Ctr Frontier Med Engn, 1-33 Yayoi cho,Inage ku, Chiba, Chiba 2638522, Japan

[4] Inst Quantum Med Sci, Natl Inst Quantum Sci & Technol, 4-9-1 Anagawa,Inage Ku, Chiba, Chiba 2638555, Japan

[5] Chiba Aoba Municipal Hosp, Dept Resp Med, 1273-2 Aoba Cho,Chuo Ku, Chiba, Chiba 2600852, Japan

[6] Chiba Univ Hosp, Dept Radiol, 1-8-1 Inohana,Chuo Ku, Chiba, Chiba 2608677, Japan

[7] Japanese Red Cross Narita Hosp, Dept Radiol, 90-1 Iida Cho, Narita, Chiba 2868523, Japan

来源：

JAPANESE JOURNAL OF RADIOLOGY | 2025年 / 43卷 / 04期

基金：

日本学术振兴会;

关键词：

COVID-19; Latent diffusion model; Chest CT images; Prognostic image generation; Deep learning;

D O I：

10.1007/s11604-024-01699-w

中图分类号：

R8 [特种医学]; R445 [影像诊断学];

学科分类号：

1002 ; 100207 ; 1009 ;

摘要：

PurposeDespite a global decrease in the number of COVID-19 patients, early prediction of the clinical course for optimal patient care remains challenging. Recently, the usefulness of image generation for medical images has been investigated. This study aimed to generate short-term follow-up chest CT images using a latent diffusion model in patients with COVID-19.Materials and methodsWe retrospectively enrolled 505 patients with COVID-19 for whom the clinical parameters (patient background, clinical symptoms, and blood test results) upon admission were available and chest CT imaging was performed. Subject datasets (n = 505) were allocated for training (n = 403), and the remaining (n = 102) were reserved for evaluation. The image underwent variational autoencoder (VAE) encoding, resulting in latent vectors. The information consisting of initial clinical parameters and radiomic features were formatted as a table data encoder. Initial and follow-up latent vectors and the initial table data encoders were utilized for training the diffusion model. The evaluation data were used to generate prognostic images. Then, similarity of the prognostic images (generated images) and the follow-up images (real images) was evaluated by zero-mean normalized cross-correlation (ZNCC), peak signal-to-noise ratio (PSNR), and structural similarity (SSIM). Visual assessment was also performed using a numerical rating scale.ResultsPrognostic chest CT images were generated using the diffusion model. Image similarity showed reasonable values of 0.973 +/- 0.028 for the ZNCC, 24.48 +/- 3.46 for the PSNR, and 0.844 +/- 0.075 for the SSIM. Visual evaluation of the images by two pulmonologists and one radiologist yielded a reasonable mean score.ConclusionsThe similarity and validity of generated predictive images for the course of COVID-19-associated pneumonia using a diffusion model were reasonable. The generation of prognostic images may suggest potential utility for early prediction of the clinical course in COVID-19-associated pneumonia and other respiratory diseases.

引用

页码：622 / 633

页数：12

共 39 条

[1] Combating COVID-19 Using Generative Adversarial Networks and Artificial Intelligence for Medical Images: Scoping Review [J].