Purpose: The purpose of this study was to assess the comparative benefits and limitations of (18)F-fluorodeoxyglucose (FDG) PET and CT-MRI in documented or suspected recurrence of cervical cancer after primary treatment.
Methods: Three patient groups were enrolled. Group A patients had biopsy-documented recurrent or persistent cervical cancer. Group B patients had suspicion of recurrent tumour on CT-MRI without biopsy proof and were potentially curable. Group C patients were in complete remission after previous definitive treatment for histologically confirmed cervical carcinoma but had elevated serum squamous cell carcinoma antigen (tumour marker) levels despite negative CT-MRI. Clinical management decisions were recorded with CT-MRI alone and with additional FDG PET. Discordances and concordances between CT-MRI and FDG PET results were identified and related to final diagnosis as based on histopathology or follow-up.
Results: A total of 150 patients (ten regions per patient) were eligible for analysis, with 58 in group A, 52 in group B and 40 in group C. For the 149 discordant regions, 126 (84.6%) had final diagnoses. Of these final diagnoses, there was additional benefit from FDG PET over CT-MRI in 73.8% (93/126), with FDG PET correcting false negatives (FNs) on CT-MRI in 74.2% (69/93) and correcting false positives (FPs) on CT-MRI in 25.8% (24/93). Among lesions confirmed by FDG PET, 75.4% (52/69) were extra-pelvic. There was additional benefit of CT-MRI compared with FDG PET in 26.2% (33/126): in nine (27.3%) CT-MRI results were shown to be true positive (TP) whereas FDG PET yielded FN results, while in 24 (72.7%) CT-MRI corrected FP results on FDG PET. Among the nine FNs on FDG PET that were identified by CT-MRI, four were extra-pelvic. Among the FPs on FDG PET that were excluded by CT-MRI, 79.2% (19/24) were extra-pelvic.
Conclusion: For recurrent cervical cancer, the benefits of FDG PET exceed those of CT-MRI owing to the ability of FDG PET to identify extra-pelvic metastases and its higher sensitivity and specificity.