The efficacy of 1% toluidine blue in the identification of oral malignancies and potentially malignant oral lesions was evaluated among a group of Asian patients (n = 102) with undiagnosed oral mucosal lesions and conditions (n = 145). The trial, utilising a ready-to-use kit, was controlled by histopathologic evaluation of a total of 87 dye-retained or dye-negative lesions. Eighteen oral carcinomas all retained the dye and there were no false negatives, yielding a test sensitivity of 100%. Eight of 39 oral epithelial dysplasias were toluidine blue-negative, giving a false negative rate of 20.5% and a sensitivity of 79.5% for oral epithelial dysplasias. The specificity of the technique was low (62%). Five dysplastic lesions were detected solely by the kit and this suggests that the method is valuable for surveillance of high-risk subjects in addition to its remarkable sensitivity in the detection of invasive carcinoma.