Background: The accuracy of (18)F-2-deoxy-fluoro-D-glucose positron emission tomography (FDG-PET) for diagnosing nodal status in patients with lung cancer was initially reported as excellent, but, with increasing experience, the problem of false-positive and false-negative assessments has been observed.
Purpose: To evaluate the accuracy of FDG-PET for diagnosing nodal status in lung cancer patients with pathologically proven N2 lymph nodes and compare it with that of computed tomography (CT).
Material and methods: Nineteen pN2 patients (13 males and six females) with primary lung cancer undergoing preoperative CT and FDG-PET were investigated. Lymph nodes were considered to be positive when uptake higher than the surrounding mediastinal level was visually observed. Slight symmetrical mediastinal uptake was considered to be negative, representing benign physiological accumulation. Radiological and pathological correlation was investigated, and the association between FDG accumulation and the size of metastatic lymph nodes and metastatic lesions was evaluated.
Results: Of the 19 patients, nodal stage determined by using FDG-PET was cN0 in four (21%) cases, cN1 in three (16%), cN2 in nine (47%), and cN3 in three (16%). On CT, nodal stage was cN0 in three (16%) cases, cN1 in seven (37%), cN2 in eight (42%), and cN3 in one (5%). Thus, FDG-PET provided correct N-staging in 47%, under-staging in 37%, and overstaging in 16%. CT staging was correct in 42%, underestimated in 53%, and overestimated in 5%. The maximum area of metastatic foci was 15.8 +/-21.3 mm(2) (mean +/- SD) in false-negative nodes and 75.0+/-56.3 mm(2) in true-positive nodes (P<0.0001).
Conclusion: Diagnostic accuracy of FDG-PET (47%) was low and similar to that of CT (42%). The possibility of false-negative as well as false-positive findings should be recognized in interpreting PET images. Micrometastasis appeared to be the greatest cause of false-negative findings.