Back to Search Start Over

The Performance Comparison of 18F-FDG PET/MRI and 18F-FDG PET/CT for the Identification of Pancreatic Neoplasms.

Authors :
Xing, Haiqun
Ding, Haiyan
Hou, Bo
Hao, Zhixin
Hu, Ya
Zhu, Wenjia
Liang, Sayuan
Feng, Feng
Li, Fang
Zhao, Yupei
Huo, Li
Source :
Molecular Imaging & Biology; Jun2022, Vol. 24 Issue 3, p489-497, 9p
Publication Year :
2022

Abstract

<bold>Purpose: </bold>To determine the optimal imaging tool for clinical evaluation of pancreatic neoplasm by comparing the performance of 18F-FDG PET/MRI and PET/CT.<bold>Procedures: </bold>Patients with suspected pancreatic neoplasms underwent PET/MRI and PET/CT in the same day prior to resection or endoscopic ultrasound-guided fine-needle aspiration. Histology served as the golden standard of lesion classification. Visual assessment on lesion type and lesion malignancy via PET/MRI and PET/CT images was compared. Standard uptake values (SUVs) of PET images from the two scanners were measured and their correlations were further evaluated.<bold>Results: </bold>Thirty-nine patients were included for the final analysis. In visual assessment, we found MRI achieved better performance than CT in differentiating solid and cystic neoplasms, with accuracy of 100% vs. 87%, respectively. In visual malignancy diagnosis, the accuracy of PET/CT was 92.3% for overall lesions and 90.9% for cysts, while the accuracy of PET/MRI was 92.3% and 86.4%, respectively. Besides, semi-quantitative analysis achieved better specificity than visual assessment for both hybrid modalities (100% vs. 87.5% for PET/CT; 100% vs. 81.5% for PET/MR). Furthermore, strong correlation of SUV was found between PET/CT and PET/MRI, with Pearson's correlation coefficients > 0.82.<bold>Conclusions: </bold>In this study, we found PET/MRI and PET/CT, both using 18F-FDG as tracer, had comparable overall performance in identification of pancreatic neoplasms. Interestingly, for patients who had suspected pancreatic neoplasm but invisible FDG uptake, PET/MRI had shown exceptionally better performance, probably because MR images could detect tiny abnormal structures to improve diagnosis. [ABSTRACT FROM AUTHOR]

Details

Language :
English
ISSN :
15361632
Volume :
24
Issue :
3
Database :
Complementary Index
Journal :
Molecular Imaging & Biology
Publication Type :
Academic Journal
Accession number :
156759571
Full Text :
https://doi.org/10.1007/s11307-021-01687-w