High-resolution image cytometry on smears of normal oral mucosa: a possible approach for the early detection of laryngopharyngeal cancers.

Abstract

The purpose of this study was to investigate the possibility of identifying laryngopharyngeal cancers by nuclear chromatin texture feature analysis of cell nuclei from mucosal scrapings obtained from clinically and cytologically noncancerous areas of the soft palate in patients with cancer. The collective consisted of 68 controls and 77 cases of laryngopharyngeal carcinomas. After Feulgen staining, 3000 cell nuclei were automatically measured using a high-resolution image analyser (CytoSavant Oncometrics, Vancouver, BC, Canada). Texture features were extracted for calculation of a discriminant function, which allows the two groups to be distinguished. Two parameters allowed the two populations to be distinguished. The classifier reached an overall performance of 72.7% sensitivity, 82.4% specificity, a positive predictive value of 80.5%, a negative predictive value of 75.1%, and an area under the receiver operating characteristics (ROC) curve of 0.7754. Our work shows that subtle changes in the chromatin distribution in cell nuclei from ostensibly normal cells in the vicinity of carcinomas are demonstrable in the oral cavity of patients suffering from laryngopharyngeal cancers. It may be possible to develop this method into a valuable clinical tool to reduce the high rate of delayed diagnosis of oral and laryngopharyngeal cancers.