We have studied whether it is possible to detect lung cancer in high risk patients using information that is present in the buccal mucosa. The goal is to create a simple, inexpensive test—an assay that could be used by GPs and dentists to screen for lung cancer. Recent studies have shown that normal appearing buccal mucosa can be analyzed to predict the presence of not only adjacent oral malignancies but also of distant tumors as well. We believe this effect extends to the lung.
We have developed an automated system for cytometry of quantitatively (Feulgen-thionin) stained specimens of buccal mucosa scrapings. Dubbed Automated Quantitative Cytometry (AQC), the system analyzes nuclear conformation and chromatin texture that reflect subtle DNA distributional changes in buccal cell nuclei. Several thousand cells are analyzed per specimen and the data reduced to a single score that predicts the likelihood of the presence of cancer. No manual pathological review is done as part of the AQC analysis. The system was developed through a field study of buccal specimens collected from 150 confirmed lung cancer patients and 990 high-risk negatives. The two groups were matched for smoking status (mean pack years, 52 versus 54; median pack years, 45 versus 42). An automated analysis system and decision rule was created to separate the two groups.
The cross validated performance on the 1140 field study specimens was 66% sensitivity at 70% specificity. Sensitivity for Stage I lung cancer (which comprised 47 of the 150 cases) was 61%.
Buccal mucosa contains information that separates patients with lung cancer from high risk negatives. Further study is warranted to turn this analysis approach into an early detection test.
Analysis of buccal mucosa may become a regular part of screening for lung cancer.
Roger Kemp, Employee Perceptronix Medical Inc.