PURPOSE

Mediastinal restaging after induction therapy for non-small-cell lung cancer remains a difficult and controversial issue. The goal of this prospective study was to compare the performance of integrated positron emission tomography (PET)--computed tomography (CT) and remediastinoscopy in the evaluation of mediastinal lymph node metastasis after induction chemotherapy.

PATIENTS AND METHODS

Thirty consecutive stage IIIA-N2 non-small-cell lung cancer patients surgically treated at our institution were entered onto this prospective study. N2 disease was proven by cervical mediastinoscopy, at which a mean number of 3.8 lymph node levels were biopsied. After completion of induction chemotherapy, the mediastinum was reassessed by integrated PET-CT and remediastinoscopy. All patients underwent thoracotomy with attempted complete resection and systematic nodal dissection.

RESULTS

PET-CT showed no evidence of nodal disease (N0) in 13 patients, Hilar nodal disease (N1) disease in three patients, and residual mediastinal disease (N2) in 14 patients. Remediastinoscopy was positive in only five patients. The preinduction involved lymph node level could be accurately re-evaluated in 18 patients. This was not the case in the other 12 because of extensive fibrosis and adhesions. In 17 patients, persistent N2 disease was found at thoracotomy. The sensitivity, specificity, and accuracy of PET-CT were 77%, 92%, and 83%, respectively. These parameters for remediastinoscopy were 29%, 100%, and 60%, respectively. Sensitivity (P < .0001) and accuracy (P = .012) were significantly better for PET-CT.

CONCLUSION

After a thorough staging mediastinoscopy, postinduction remediastinoscopy had a disappointing sensitivity because of adhesions and fibrosis. Integrated PET-CT yielded a better result than that obtained in previous studies with side-by-side PET and CT images.