PURPOSE: The TNM staging system in lung cancer, comprising tumor size and extent, nodal involvement and location, and presence of metastases, stratifies cases into prognostic and treatment groups.The system is limited by its inability to include demographic, pathologic and molecular factors, which also impact prognosis.Our study utilizes the ensemble algorithm of clustering of cancer data (EACCD) to create a system that is supplemental, more robust, and more adaptable than the TNM system.METHODS: EACCD is a computer-based predictive system for cancer that preserves the TNM system, but also incorporates additional host features and tumor prognostic factors.EACCD is a two-step clustering method.In the first step, a dissimilarity measure is attained by a partitioning method, and in the second step, the dissimilarity is used with a hierarchical clustering algorithm to obtain clusters of patients.Data were obtained from the Surveillance, Epidemiology, and End Results Program of the National Cancer Institute, USA for the years 1988-1998.Analysis considered tumor histology and grade, and gender, in addition to AJCC stage.RESULTS: EACCD produced a hierarchical, ordered sequence of distinct stratified cohorts of cancer cases.In contrast to the four stages reported from the TNM data, seven groups of patients resulted from EACCD analysis with higher group numbers associated with greater mortality.Group 1 stratified with Stage Ⅰ, Group 5 stratified with Stage Ⅲ, and Groups 6 and 7 stratified with Stage Ⅳ.Stage Ⅰ cancer cases also partitioned to Groups 2 and 3, with decreasing survival, indicating that additional relevant parameters were associated with increased relative biologic aggressive tumor behavior.Intermediate Group 4 consisted of Stage Ⅰ, Ⅱ, and Ⅲ cancer cases, suggesting that localized biologically aggressive cancers may have the same survival as more indolent advanced staged cancers.CONCLUSION: The ensemble clustering algorithm EACCD can be used to adapt the TNM system and combine additional host features and tumor markers to further stratify survival groups.The seven EACCD groups represented a more powerful stratification of lung cancer patients than the TNM staging system.The technique may identify new prognostic factors and determine their influence on existing survival parameters.