Evaluation of a similarity anisotropic diffusion denoising approach for improving in vivo CEST-MRI tumor pH imaging.
Magn Reson Med
; 85(6): 3479-3496, 2021 06.
Article
em En
| MEDLINE
| ID: mdl-33496986
ABSTRACT
PURPOSE:
Chemical exchange saturation transfer MRI provides new approaches for investigating tumor microenvironment, including tumor acidosis that plays a key role in tumor progression and resistance to therapy. Following iopamidol injection, the detection of the contrast agent inside the tumor tissue allows measurements of tumor extracellular pH. However, accurate tumor pH quantifications are hampered by the low contrast efficiency of the CEST technique and by the low SNR of the acquired CEST images, hence in a reduced detectability of the injected agent. This work aims to investigate a novel denoising method for improving both tumor pH quantification and accuracy of CEST-MRI pH imaging.METHODS:
An hybrid denoising approach was investigated for CEST-MRI pH imaging based on the combination of the nonlocal mean filter and the anisotropic diffusion tensor method. The denoising approach was tested in simulated and in vitro data and compared with previously reported methods for CEST imaging and with established denoising approaches. Finally, it was validated with in vivo data to improve the accuracy of tumor pH maps.RESULTS:
The proposed method outperforms current denoising methods in CEST contrast quantification and detection of the administered contrast agent at several increasing noise levels with simulated data. In addition, it achieved a better pH quantification in in vitro data and demonstrated a marked improvement in contrast detection and a substantial improvement in tumor pH accuracy in in vivo data.CONCLUSION:
The proposed approach effectively reduces the noise in CEST images and increases the sensitivity detection in CEST-MRI pH imaging.Palavras-chave
Texto completo:
1
Base de dados:
MEDLINE
Assunto principal:
Imageamento por Ressonância Magnética
/
Neoplasias
Limite:
Humans
Idioma:
En
Revista:
Magn Reson Med
Assunto da revista:
DIAGNOSTICO POR IMAGEM
Ano de publicação:
2021
Tipo de documento:
Article
País de afiliação:
Itália