Comparison of (18)F-FDG PET/CT, MRI and SPECT in the diagnosis of local residual/recurrent nasopharyngeal carcinoma: A meta-analysis

• Three modalities were compared for the detection of local residual/recurrent NPC.
• SPECT is superior to MRI for the detection of local residual/recurrent NPC.
• PET/PEC–CT is superior to MRI for the detection of local residual/recurrent NPC.
• The addition of CT to PET cannot significantly improve the diagnostic accuracy.
• 201TI-SPECT and MIBI-SPECT have the same diagnostic accuracy.

SummaryThe objective of this study was to assess the overall diagnostic value of MRI, SPECT and (18)F-FDG PET/CT in detecting local NPC residual/recurrence with a meta-analysis. We performed a systematic review with meta-analyses to compare the diagnostic performance of nuclear magnetic resonance Imaging (MRI), single photon emission computed tomography (SPECT) and 18-fluoro-2-deoxyglucose positron emission tomography ((18)F-FDG PET/CT) as imaging modalities for the detection of local residual or recurrent nasopharyngeal carcinoma (NPC). MEDLINE, EMBASE and publisher databases were searched in December 2014. Methodological quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool. Pooled estimation and subgroup analysis data were obtained by statistical analysis. Seventeen studies were included in the meta-analysis. The pooled sensitivity estimates for (18)F-FDG PET/CT (90%) and SPECT (85%) were not significantly higher than MRI (77%) (p = 0.096 and 0.164, respectively). The pooled specificity estimates for (18)F-FDG PET/CT (93%) and SPECT (81%) were significantly higher than MRI (76%) (p = 0.033 and 0.042, respectively). The pooled DOR (Diagnostic odds ratio) estimates for (18)F-FDG PET/CT (73.27) were significantly higher than MRI (12.09) (p = 0.019) while the pooled DOR estimates for SPECT (78.69) were not significantly higher than MRI (12.09) (p = 0.872). For (18)F-FDG PET/CT, there were no significant differences between PET–CT and PET on all of the variables including sensitivity, specificity, PLR (Positive likelihood ratio), NLR (Negative likelihood ratio) and DOR (P > 0.05). For SPECT, there were no significant differences between 201TI-SPECT and MIBI-SPECT on all of the variables including sensitivity, specificity, PLR, NLR and DOR (P > 0.05). Both (18)F-FDG PET/CT and SPECT are very accurate for the detection of local residual or recurrent NPC, they are superior to MRI in distinguishing recurrent NPC from fibrosis or scar tissue after RT in irradiated fields with distortion of normal architecture. For (18)F-FDG PET/CT, the diagnostic accuracy PET/CT was not significantly different than that of PET alone. For SPECT, 201TI-SPECT and MIBI-SPECT have the same diagnostic accuracy.