The long-term objective is to develop a technique for the noninvasive detection of early lung carcinoma using volatile organic constituents in expired air. Preliminary results published by our group (IITRI/Michael Reese Hospital) have identified several volatile organic compounds that appear to be associated with lung cancer. Accordingly, the primary objective of the proposed three-year study is to extend these observations and to explore the diagnostic potential of such compounds using statistical approaches. A study population of 325 subjects are proposed for screening over this three year period. These subjects will include 100 lung cancer patients, and 75 subjects in each of three propensity score matched control groups (normal smoking controls, noncancer lung diseases and nonpulmonary cancers). Four samples (2 fasting and 2 two-hour postprandial) will be collected from each subject. Two samples will be analyzed immediately and the two duplicate samples will be used for backup and quality assurance audits. The collected expired air will be collected, processed, and analyzed by techniques previously developed in this laboratory. Subject screening will be conducted in two phases: an estimation phase and a validation phase. In the estimation phase, variables (peaks) will be screened both individually and in combination for statistical significance using one-way analysis of variance (after appropriate transformation of the data) and by stepwise discriminant function analysis. In the validation phase, those variables that were statistically significant in the estimation phase will be reexamined in a new sample of patients to determine if this previous significance was due to chance. If a variable (i.e., a GC peak representing a compound) or linear combination of variables is significant in both the estimation and validation phases, we will conclude that it is significantly related to the presence of lung cancer. We anticipate that the results of this study will advance preliminary data which will permit the early detection of lung cancer in a smoking population and ultimately result in the development of an inexpensive, routine, noninvasive laboratory procedure.
| O'Neill, H J; Gordon, S M; O'Neill, M H et al. (1988) A computerized classification technique for screening for the presence of breath biomarkers in lung cancer. Clin Chem 34:1613-8 |
