Concanavalin A-peroxidase labeling in cervical exfoliative cytopathology. I. Labeling of normal squamous cells and the detection of cancer.

The lectin binding capacity of the cell surface of normal flattened exfoliated epithelial cells of the uterine cervix was investigated looking for differences between specimens from normal and cancer patients. The method used was a modified concanavalin A-horseradish peroxidase (Con A-HRP) labeling procedure. Both normal and cancer specimens contain labeled as well as unlabeled usual flattened cells. There is a distinct difference between the labeling intensity of labeled and that of unlabeled cells. Quantification of the labeling results has been achieved using a light microscope equipped with a computerized video system. Apparently healthy persons, having a percentage of labeled flattened cells between 54 and 94% (mean = 73%, SD = 10%, N = 40), were totally discriminated by this method from the cancer patients. These patients with a histologically confirmed squamous cell carcinoma, showed a labeling percentage between 10 and 22% (mean = 15%, SD = 4%, N = 10). Hormonal factors, such as phase of cycle and pill use, appeared to have no significant influence. Statistical analysis revealed that at least 99% of all healthy persons will have a labeling percentage above 45%, while at most 1% of the cancer patients will show a labeling percentage above 30%. When choosing the labeling percentage of 45% as critical value, the Con A-HRP labeling might serve as an additional detection method for cancer of the uterine cervix. Moreover, as it is based on the abundantly present normal cells, and not on the often scarce abnormal cells, the method is not liable to sampling and screening errors.