We investigated the diagnostic accuracy of FDG-PET in the detection of recurrent lung cancer.
Methods: Thirty-nine lesions in 38 patients with clinically suspected recurrent or residual lung cancer were studied with PET. All PET images were visually interpreted in conjunction with thoracic CT or MRI. Semiquantitative analysis using standardized uptake values (SUVs) was also performed in 25 lesions. FDG-PET diagnoses were correlated with pathological diagnoses and clinical outcome.
Results: The sensitivity and specificity of FDG-PET for detecting recurrent tumors were 100% (26/26) and 61.5% (8/13), respectively. The difference in mean SUV between recurrent tumors and noncancerous lesions was statistically significant [11.2 +/- 5.7 (n = 16) vs. 3.5 +/- 1.8 (n = 9), p < 0.0001]. False-positive results showed relatively lower SUVs than true-positives and also demonstrated increased uptake in a curvilinear rather than nodular shape.
Conclusion: FDG-PET is useful for detecting recurrent lung cancer after treatment. False-positive diagnoses might be reduced by analysis of uptake shape and serial changes in SUV, but further study is needed.