Background: To compare fusion, positron emission tomography-computed tomography (PET-CT) with CT alone in detecting ovarian carcinoma recurrence.
Methods: Fifty-one consecutive patients underwent 53 restaging PET-CT scans with a concurrent diagnostic quality CT scan. Two body imaging radiologists independently assessed the CT's; each then teamed with a nuclear medicine specialist to review the PET-CT's. Two teams conferred for consensus on the presence of disease in the chest, abdomen, and body overall with CT alone and with PET-CT, using a six-point reader confidence metric to determine accuracy and receiver operating characteristic (ROC) curves. Reader agreement was compared using kappa. Recurrence was determined by two gynecologic oncologists reviewing clinical records from time of presentation to at least 13 months (mean 22.7) after imaging.
Results: Recurrence was based on histopathology in 17% (9/53). Seventy-two percent (38/53) cases had recurrence, with two cases showing isolated chest recurrence. PET-CT accuracy exceeded CT for body 92% (49/53) vs. 83% (44/53), chest 96% (51/53) vs. 89% (47/53), and abdomen 91% (48/53) vs. 79% (42/53). ROC curves for PET-CT dominated that for CT alone; this difference was statistically significant for abdomen and for body overall (P < 0.01). Interobserver agreement was better for PET-CT than for CT alone.
Conclusions: PET-CT demonstrates greater accuracy and less interobserver variability than CT alone.