Present clinicopathologic staging of non-small cell lung cancer is limited in its ability to provide more than a general prognostic estimate in patients with lung cancer who have resectable disease. This study was performed to identify whether the ability to adapt tumor tissue from resectable (stage I to IIIa) non-small cell lung cancer was associated with a poorer prognosis and an increased risk for early tumor recurrence. We attempted to culture a tumor specimen obtained from 90 patients with resectable non-small cell lung cancer. We used a culture medium conditioned by exposure to the lung cancer cell line A549-1, a known producer of autocrine lung cancer growth factors, and provided tumor colony scaffolding using a feeder layer of inactivated fibroblasts, and found these measures improved tumor culture yields. Twenty-two cell lines were obtained, a success rate of 24.4%. Tumor recurrences were more common (79%) among the culture-positive patients than among the culture-negative patients (37.5%; p < 0.002). For all patients, survival at 19 months in the culturepositive patients was 50.0%, compared with 83.6% in the culture-negative patients (p < 0.005). The median survival for the culture-positive patients was 15 months, versus 21.7 months for the culture-negative patients (p < 0.004). The establishment of a culture was a predictor of shortened survival for patients with stage I disease. In patients with stage I disease, survival at 19 months was 54.5% for the culture-positive patients, versus 89% for the culturenegative patients (p < 0.02). On the basis of our findings, the ability to adapt tumor specimens to cell culture appears to be a negative prognostic indicator for surgically resectable non-small cell lung cancer. The results of culture of the primary tumor may be used to more closely estimate the individual patient's risk for recurrence and determine his or her potential need for adjuvant therapies.
Bibliographical noteFunding Information:
This research was supported by grants from the NIH to Jill M. Siegfried, PhD (CA50694), and to Joseph R. Testa, MD (CA45745), and by a contract from the Ncr (EDRN. No N01-CN-15393) awarded to the University of Pittsburgh. Jay D. Hunt, MD, was supported by a National Research Service Award (CA09033) from the NCr. We thank the Pittsburgh Cancer Institute Tissue Bank for assistance in obtaining tumor tissue. We also thank Frank D'Amico, PhD, head of Biostatistics at Dusquenne University, for his statistical analysis support in this work.