In order to construct a multivariate model for predicting early recurrence and cancer death for patients with stage I non-small cell lung cancer, 271 consecutive patients (mean age, 63 ± 8 years) who were diagnosed, treated, and followed at one institution were studied. All patients were clinical stage I with head and chest/abdominal computed tomograms and radionuclide bone scans without evidence of metastatic disease. Pathological material after resection was reviewed to verify histological staging. Follow-up documented the time and location of any recurrence, was a median 56 months in duration, and was complete in all cases.
Data recorded included age, sex, smoking history, presenting symptoms, pathological description, and oncoprotein staining for erbB-2 (HER-2/neu), p53, and KI-67 proliferation protein. Immunohistochemistry of oncogene expression was performed on two separate archived paraffin tumor blocks for each patient, with normal lung as control. All analyses were blinded and included Kaplan-Meier survival estimates with Cox proportional hazards regression modeling.
Data, including immunohistochemistry, were complete for all 271 patients. Actual 5-year survival was 63% and actuarial 10-year survival was 58%. Significant univariate predictors (P < 0.05) of early recurrence and cancer-death were: male sex; the presence of symptoms; chest pain; type of cough; hemoptysis; tumor size > 3 cm diameter (T2); poor differentiation; vascular invasion; erbB-2 expression; p53 expression; and a higher KI-67 proliferation index (>5%). An additive oncogene expression curve demonstrated a 5-year survival of 72% for 136 patients without p53 or erbB-2, 58% for 108 patients who expressed either oncogene, and 38% for 27 who expressed both (P <0.001). Multivariate independent predictors of early recurrence and cancer death (P < 0.05) were symptomatic presentation, erbB-2 expression, T2 size, vascular invasion, p53 expression, and poor differentiation.
These data allowed the creation of a multivariate model which quantified the risk of recurrence and cancer death for patients with stage I non-small cell lung cancer. This model, based on complete data from 271 patients, represents the largest analysis of its type in the literature and can form the basis for multi-institutional randomized adjuvant trials for “high risk” patients.
↵1 Supported by NIH Grant CA56749 and a research grant from Abbott Laboratories (Chicago, IL).
↵2 To whom requests for reprints should be addressed, at Division of Thoracic Surgery, Brigham and Women's Hospital, 75 Francis Street, Boston, MA 02115.
- Received July 8, 1994.
- Accepted October 31, 1994.
- ©1995 American Association for Cancer Research.