BACKGROUND Inconsistent information on the prognostic significance of non-small cell lung cancer (NSCLC) isolated tumor cells (ITC) has been reported to date. We sought to evaluate the survival for NSCLC in a group of patients in which the presence of bone marrow isolated tumor cells and their DNA ploidy was assessed. METHODS Seventy patients (58 males [83%]; median age 70 years, range 49-89) with T1-4, N0, M0 clinical staging entered the study; 68 who underwent complete resection, were included in the follow-up. Two patients with clinical stage T2 and T4, N0, M0 were excluded because of pleural carcinosis discovered at thoracotomy. Recruitment ended in 2002. None received neoadjuvant therapy. The rib bone marrow was extracted and assessed for ITC by hematoxylin and eosin (H&E) staining, immunohistochemistry and flow cytometry. The latter was regarded as positive when >10% of cells reacted to pan-cytokeratin antibody MNF116. DNA ploidy was studied by propidium iodide staining. Patient follow-up was with chest X-ray and abdominal US every 6 months, and CT-PET scan every 12 months for at least 5 years after surgery. Causes of death were assessed. RESULTS Rib bone marrow ITC were documented in 17 patients (25%), 6 with DNA euploidy (p stage: I 4; III 2), and 11 with DNA aneuploidy (p stage: I 5; II 4; III 2) while 51 (75%) patients were free of ITC (p stage: I 32; II 8; III 9; IV 2). The median follow-up was 61 months, 21 patients died from causes unrelated to NSCLC and 12 patients died from causes related to tumor relapse. Significant survival differences were observed according to stage, presence of ITC and DNA aneuploidy. In particular free from recurrence survival was significantly reduced in stage IA and IB patients presenting aneuploid ITC (Wilcoxon (Gehan) test p=0.031). CONCLUSIONS The prognostic role of bone marrow ITC seems to be corroborated by DNA ploidy studies. Patients with bone marrow ITC with abnormal DNA content showed a significantly reduced survival particularly in stage I NSCLC.