Positron emission tomography may underestimate the extent of thoracic disease in lung cancer patients

Objective: Although widely utilised in the staging of lung cancer various studies have questioned whether the accuracy of this staging modality is sufficient to replace conventional invasive staging techniques. We have therefore reviewed our experience in order to assess the accuracy of PET CT as an intrathoracic staging toot for non-small cell lung cancer (NSCLC). Methods: Two hundred patients referred for surgery between June 2006 and January 2008 underwent PET CT followed by staging mediastinoscopy and, if appropriate, resection. Results of scans and histopathology were correlated and analysed. Results: Overall, PET CT correctly staged 99 out of 200 patients (49.5%), under-staged 59 (29.5%), and over-staged 42 (21%). Superior mediastinal nodes were incorrectly staged by PET CT in 35 (19%) of 186 patients undergoing mediastinoscopy: in 15 (8%) mediastinoscopy revealed metastatic disease not detected on PET CT and 20 (11%) had negative histology despite a positive scan. Five (2.5%) resections were benign despite avid FDG uptake, and 6 (3%) were malignant despite a negative scan. PET CT had false positive result of 6.5%, 5.5%, 4.5% and 3.5%, respectively for hilar, station 2, 7 and 5 node groups. The false negative result was 12.5%, 10.5% and 8%, respectively for hilar, intrapulmonary and station 4 nodes. Twelve (6%) of patients were under-staged regarding chest wait and mediastinal invasion, and 10 (5%) patients had metastatic nodules in the lung (T4) not detected by PET CT. Stage I or 11 disease was identified by PET CT in 141 patients of whom 26 (18.4%) had IIIa or higher stage disease. The false positive and negative predictive values for PET CT with respect to N2 or greater status were 17.2% (11.8-24.2) and 48.6% (32.2-65.3), respectively. Conclusions: Our experience would suggest that PET CT alone is not sufficiently accurate to replace mediastinoscopy and other conventional biopsy techniques in the evaluation of NSCLC cases. It may better be viewed as a valuable additional toot with which to inform decision making and to screen for disseminated disease. (C) 2009 European Association for Cardio-Thoracic Surgery. Published by Elsevier B.V. All rights reserved.