Objective: To assess the diagnostic performance of Positron Emission Tomography using fluorodeoxyglucose (FDG-PET) in comparison to conventional imaging modalities in the assessment of patients with cervical and ovarian cancer.
Methods: Studies published between 1966 and 2003 were identified using an OVID search of the MEDLINE database. Inclusion criteria were use of a dedicated scanner, resolution specified, >/=12 human subjects, clinical follow-up >/=6 months or histopathology as reference standard, and sufficient data provided to construct a two-by-two table. Two reviewers independently abstracted data regarding sensitivity and specificity of PET.
Results: 25 studies (15 cervical cancer, 10 ovarian cancer) met inclusion criteria for full text review. For cervical cancer, pooled sensitivity and specificity of PET for aortic node metastasis are 0.84 (95% CI 0.68-0.94) and 0.95 (0.89-0.98). Pooled sensitivity and specificity for detection of pelvic node metastasis are: PET, 0.79 (0.65-0.90) and 0.99 (0.96-0.99); MRI, 0.72 (0.53-0.87) and 0.96 (0.92-0.98). Pooled sensitivity for CT is 0.47 (0.21-0.73) (pooled specificity not available). Pooled sensitivity and specificity of PET for recurrent cervical cancer with clinical suspicion are 0.96 (0.87-0.99) and 0.81 (0.58-0.94). For ovarian cancer, pooled sensitivity and specificity to detect recurrence with clinical suspicion are: PET, 0.90 (0.82-0.95) and 0.86 (0.67-0.96); conventional imaging, 0.68 (0.49-0.83) and 0.58 (0.33-0.80); CA-125, 0.81 (0.62-0.92) and 0.83 (0.58-0.96). When conventional imaging and CA-125 are negative, pooled sensitivity and specificity of PET are 0.54 (0.39-0.69) and 0.73 (0.56-0.87), respectively. When CA-125 is rising and conventional imaging is negative, the pooled sensitivity and specificity of PET are 0.96 (0.88-0.99) and 0.80 (0.44-0.97).
Conclusions: There is good evidence that PET is useful for the pre-treatment detection of retroperitoneal nodal metastasis in cervical cancer. There is fair evidence that PET is useful for the detection of recurrent cervical cancer. PET is less useful for the detection of microscopic residual ovarian cancer but has fair sensitivity to detect recurrence in the setting of a rising CA-125 and negative conventional imaging studies. Available studies are limited by low numbers of patients and wide confidence intervals.