Purpose: This study was designed to retrospectively evaluate the diagnostic yield of FDG PET for the diagnosis of recurrent ovarian cancer.
Methods: Eighty FDG PET scans were performed on 55 patients owing to suspicion of relapse, and 45 FDG PET scans were performed on 31 patients who were clinically disease free. PET results were compared with the results of conventional radiological imaging (CIM) and serum CA 125 levels, and related to pathological findings in 54 cases or clinical follow-up in 71 cases.
Results: CIM correctly identified 49 cases with recurrence [sensitivity (SE) 53.3%] ,and there were 27 true negatives [specificity (SP) 81.8%] However, 43 cases were false negative and six were false positive. The positive predictive value (PPV), negative predictive value (NPV) and accuracy (ACC) of CIM were 89%, 38.6% and 60.8%, respectively. FDG PET correctly detected recurrent disease in 80/92 cases (SE 86.9%, p<0.05) and ruled out relapse in 26/33 cases (SP=78.8%). The PPV, NPV and ACC of PET were 91.9%, 68.4% and 84.8%, respectively. Standardised uptake values did not provide additional diagnostic accuracy compared with visual analysis. The CA 125 results showed an SE of 57.6%, an SP of 93.9% and an ACC of 67.2%. In 23 patients with positive serum CA 125 levels, but negative CIM, FDG PET was positive and relapse was confirmed. Furthermore, FDG PET was positive and relapse was confirmed in 11 patients with negative serum CA 125 levels and CIM.
Conclusion: FDG PET may detect recurrent ovarian cancer earlier than CIM, with higher sensitivity and even higher diagnostic accuracy.