FDG-PET, CT, MRI for diagnosis of local residual or recurrent nasopharyngeal carcinoma, which one is the best? A systematic review.

PURPOSE To perform a systematic review to compare FDG-PET, CT, and MRI imaging for diagnosis of local residual or recurrent nasopharyngeal carcinoma. MATERIALS AND METHODS MEDLINE, EMBASE, the CBMdisc databases and some other databases were searched for relevant original articles published from January 1990 to June 2007. Inclusion criteria were as follows: Articles were reported in English or Chinese; FDG-PET, CT, or MRI was used to detect local residual or recurrent nasopharyngeal carcinoma; histopathologic analysis and/or close clinical and imaging follow-up for at least 6 months were the reference standard. Two reviewers independently extracted data. A software called "Meta-DiSc" was used to obtain pooled estimates of sensitivity, specificity, diagnostic odds ratio (DOR), summary receiver operating characteristic (SROC) curves, and the Q* index. RESULTS Twenty-one articles fulfilled all inclusion criteria. The pooled sensitivity estimates for PET (95%) were significantly higher than CT (76%) (P<0.001) and MRI (78%) (P<0.001). The pooled specificity estimates for PET (90%) were significantly higher than CT (59%) (P<0.001) and MRI (76%) (P<0.001). The pooled DOR estimates for PET (96.51) were significantly higher than CT (7.01) (P<0.001) and MRI (8.68) (P<0.001). SROC curve for FDG-PET showed better diagnostic accuracy than CT and MRI. The Q* index for PET (0.92) was significantly higher than CT (0.72) (P<0.001) and MRI (0.76) (P<0.01). For PET, the sensitivity and diagnostic OR for using qualitative analysis were significantly higher than using both qualitative and quantitative analyses (P<0.01). For CT, the sensitivity, specificity, diagnostic OR, and the Q* index for dual-section helical and multi-section helical were all significantly higher than nonhelical and single-section helical (P<0.01). And the sensitivity for 'section thickness <5 mm' was significantly lower than ' =5 mm' (P<0.01), while the specificity was significantly higher (P<0.01). For MRI, there were no significant differences found between magnetic field strength <1.5 and > or =1.5 T (P>0.05). CONCLUSION FDG-PET was the best modality for diagnosis of local residual or recurrent nasopharyngeal carcinoma. The type of analysis for PET imaging and the section thickness for CT would affect the diagnostic results. Dual-section helical and multi-section helical CT were better than nonhelical and single-section helical CT.