Purpose: The aim of this study was to evaluate the efficacy of PET imaging and compare it with the performance of CT in mediastinal and hilar lymph node staging in potentially operable non-small cell lung cancer (NSCLC).
Methods: Fifty-nine patients with potentially resectable NSCLC who underwent preoperative PET and CT imaging were enrolled into this prospective study. All patients underwent surgical evaluation by means of mediastinoscopy with mediastinal lymph node sampling (14 patients) or thoracotomy (45 patients).
Results: The prevalence of lymph node metastases was 53%. Overall, the sensitivity, specificity, accuracy, PPV, and NPV of PET were 79%, 76%, 78%, 86%, and 76% for N0 and N1 lymph nodes and 76%, 79%, 80%, 67%, and 83% for N2 lymph nodes, while those values for CT were 66%, 43%, 58%, 68%, and 43% for N0 and N1 stations and 43%, 66%, 54%, 41%, and 66% for N2 lymph nodes, respectively. PET correctly differentiated cases with mediastinal lymph node involvement (N2) from those without such involvement (N0 or N1) in 76% of cases. Statistical analysis of the diagnostic accuracy of nodal involvement showed that PET improves diagnostic accuracy significantly in the detection of both N0 or N1 and N2 status in the individual patient based on analysis, compared with CT (P < 0.01 and P < 0.01, respectively). When preoperative nodal staging was compared with postoperative histopathological staging, 38 (65%) patients were correctly staged, 9 (15%) were overstaged, and 12 (20%) were understaged by PET, while 29 patients (49%) were correctly staged, 13 (22%) were overstaged, and 17 (29%) were understaged by CT.
Conclusion: It has been clearly shown that PET is more accurate than CT for the differentiation of N0 or N1 from N2 disease in patients with NSCLC. However, PET imaging alone does not appear to be sufficient to replace mediastinoscopy for mediastinal staging in patients with lung cancer, especially in geographic regions with high granulomatous or inflammatory mediastinal disease prevalence.