Seung Baek Hong1,2, Sang Hyun Choi3, Kyung Won Kim1, Seong Ho Park1, So Yeon Kim1, So Jung Lee1, Seung Soo Lee1, Jae Ho Byun1, Moon-Gyu Lee1. 1. Department of Radiology and the Research Institute of Radiology, University of Ulsan College of Medicine, Asan Medical Center, 88 Olympic-Ro 43-Gil, Songpa-Gu, Seoul, 05505, South Korea. 2. Department of Radiology, Pusan National University of Hospital, Busan, South Korea. 3. Department of Radiology and the Research Institute of Radiology, University of Ulsan College of Medicine, Asan Medical Center, 88 Olympic-Ro 43-Gil, Songpa-Gu, Seoul, 05505, South Korea. edwardchoi83@gmail.com.
Abstract
OBJECTIVES: To systematically determine the diagnostic accuracy of 18-fluorodeoxyglucose positron emission tomography/magnetic resonance imaging ([18F]FDG-PET/MRI) for the detection of liver metastases and evaluate the sources of heterogeneity in the reported results. METHODS: PubMed and EMBASE databases were searched up until December 31, 2017, to identify original research studies reporting the diagnostic performance (Se and Sp) of PET/MRI for liver metastases, in comparison with PET/CT. Study quality was assessed using QUADAS-2. The summary Se and Sp of the studies were estimated using hierarchical modeling methods. To determine causes of study heterogeneity, the presence of a threshold effect was analyzed, and meta-regression analysis was performed. RESULTS: Of 546 articles screened, eight suitable articles were identified, with seven for per-lesion analysis, and four for per-patient analysis. The meta-analytic summary Se and Sp for per-patient-based analysis were 99.2% (95% CI, 31.4-100.0%, I2 = 89.4%) and 98.6% (95% CI, 84.0-99.9%, I2 = 0.0%), respectively, while for per-lesion-based analysis they were 95.4% (95% CI, 78.3-99.2%, I2 = 99.7%) and 99.3% (95% CI, 93.8-99.9%, I2 = 96.5%). PET/MRI showed higher Se (95.4% vs. 68.3%) and Sp (99.3% vs. 95.8%) than PET/CT. Meta-regression analysis showed five significant factors affecting study heterogeneity: study subject characteristics, study design, MRI technique (DWI, HBP after injection of liver-specific contrast media), imaging review method, and reference standard. CONCLUSION: The diagnostic accuracy of [18F]FDG-PET/MRI for liver metastasis was high overall, but substantial heterogeneity was found. Further randomized controlled studies or prospective studies are needed to investigate the role of PET/MRI in liver metastasis in comparison with PET/CT. KEY POINTS: • [ 18 F]FDG-PET/MRI has high meta-analytic Se and Sp for the diagnosis of liver metastasis. • PET/MRI using DWI and HBP images significantly increased diagnostic accuracy. • Study heterogeneity was associated with subject characteristics, study design, MRI technique, image review method, and reference standard.
OBJECTIVES: To systematically determine the diagnostic accuracy of 18-fluorodeoxyglucose positron emission tomography/magnetic resonance imaging ([18F]FDG-PET/MRI) for the detection of liver metastases and evaluate the sources of heterogeneity in the reported results. METHODS: PubMed and EMBASE databases were searched up until December 31, 2017, to identify original research studies reporting the diagnostic performance (Se and Sp) of PET/MRI for liver metastases, in comparison with PET/CT. Study quality was assessed using QUADAS-2. The summary Se and Sp of the studies were estimated using hierarchical modeling methods. To determine causes of study heterogeneity, the presence of a threshold effect was analyzed, and meta-regression analysis was performed. RESULTS: Of 546 articles screened, eight suitable articles were identified, with seven for per-lesion analysis, and four for per-patient analysis. The meta-analytic summary Se and Sp for per-patient-based analysis were 99.2% (95% CI, 31.4-100.0%, I2 = 89.4%) and 98.6% (95% CI, 84.0-99.9%, I2 = 0.0%), respectively, while for per-lesion-based analysis they were 95.4% (95% CI, 78.3-99.2%, I2 = 99.7%) and 99.3% (95% CI, 93.8-99.9%, I2 = 96.5%). PET/MRI showed higher Se (95.4% vs. 68.3%) and Sp (99.3% vs. 95.8%) than PET/CT. Meta-regression analysis showed five significant factors affecting study heterogeneity: study subject characteristics, study design, MRI technique (DWI, HBP after injection of liver-specific contrast media), imaging review method, and reference standard. CONCLUSION: The diagnostic accuracy of [18F]FDG-PET/MRI for liver metastasis was high overall, but substantial heterogeneity was found. Further randomized controlled studies or prospective studies are needed to investigate the role of PET/MRI in liver metastasis in comparison with PET/CT. KEY POINTS: • [ 18 F]FDG-PET/MRI has high meta-analytic Se and Sp for the diagnosis of liver metastasis. • PET/MRI using DWI and HBP images significantly increased diagnostic accuracy. • Study heterogeneity was associated with subject characteristics, study design, MRI technique, image review method, and reference standard.