Keratin profiles of normal and malignant oral

Aims-To assess keratin profiles from smears of malignant and contralateral normal oral mucosa as part of the development of a screening procedure for oral cancer based on exfoliative cytology. Methods-Smears were taken from oral cancers (confirmed by biopsy) and from the contralateral site of 20 patients. Using a panel of antikeratin antibodies, the keratins expressed by these cells were identified using a standard immunocytochemical technique (Vectastain) and assessed on a 3 point scale. Results-Using x2 analysis, noticeable differences between the keratin profiles for malignant mucosal smears compared with the contralateral mucosal smears were found. This was particularly evident for the simple epithelial keratins. Conclusion-Individual keratins can be identified in smears from oral cancers. The identification of simple epithelial keratins seem to be the best keratin markers associated with malignancy. Their detection within smears from oral lesions could be valuable in the early diagnosis of oral cancer. (7 Clin Pathol 1993;46:352-356) Department of Dental Surgery G R Ogden S McQueen D M Chisholm Department of Anatomy and Cell Biology, University of Dundee E B Lane Correspondence to: Dr G R Ogden, Department of Dental Surgery, Dental Hospital & School, Park Place, Dundee DD1 4HN, Scotland Accepted for publication 7 October 1992 Worldwide, oral cancer is the 6th most common cancer.' In the United Kingdom about 1900 new cases and 960 deaths occur every year.2 Ideally disease prevention is what is required. But there seems little evidence to suggest that exposure to the two factors most frequently associated with oral cancer (alcohol and tobacco) is decreasing.'4 Even then total reliance on prevention is not possible because all the aetiological factors for oral cancer are not known.5 6 Despite the fact that oral cancer can be cured if treated early enough,7 the 5 year survival (about 35%)8 has not really improved with advances in surgery, radiotherapy, and chemotherapy.9 The main reason is probably the late presentation of these tumours.'° 11 In turn this may be due to: (i) the asymptomatic nature of the early lesion'0; (ii) lack of self examination by patients'2 "; (iii) misdiagnosis by clinician'O-'2; and (iv) the patient's fear." 12 These obstacles have to be overcome if the prognosis is to improve. Recent calls for a screening procedure for oral cancer still rely on subjective assessment by the clinician as to whether the mucosa appears clinically malignant or not.'4 15 Recent advances in the quantitative assessment of oral cytology16-20 may offer a suitable technique for such screening, analogous to the screening procedure based on exfoliative cytology for cervical cancer, a disease which may have a similar incidence to oral cancer.'4 Heavily keratinised oral lesions make it difficult to gain a representative sample of the underlying lesion, but this is offset by their relatively low malignant transformation rateabout 2%.21 Far more worrying are the speckled leucoplakias and erythroplakias which are reported to undergo malignant change more frequently. These lesions are amenable to sampling using exfoliative cytology. Not all general practitioners would biopsy the oral mucosa and they may be more inclined to smear an abnormality of the mucosa. Furthermore, in patients who are under review following treatment of an oral squamous cell carcinoma or in patients who have widespread instability of the oral mucosa, repeat biopsies at every review are not practicable. Such patients would soon fail to attend for review. Exfoliative cytology permits the frequent sampling of such sites with minimal inconvenience to the patient. This forms part of the overall management of our patients treated in Tayside. As well as assessment of nuclear status, exfoliative cytology samples can be assessed for differentiation status. In epithelial cells differential expression of keratin filament proteins is a well established indicator of differentiation, against which many well characterised antibodies are available.222' Keratins are for the most part conserved during malignant transformation when all other identifying criteria of the cell's origin may have been lost. Because cytokeratin profiles are useful in tumour diagnosis24 their identification in oral smears may detect subtle changes in tissues undergoing malignant change. Keratin expression has been studied in normal oral epithelia,25 but keratin antibodies had not previously been applied to smears from oral cancers. Knowledge of the normal keratin profile determined the range of keratins studied. Methods Paired smears were taken from biopsy confirmed oral cancers and from the contralateral mucosal site of 20 patients, using a Cytobrush (MedScand, Colgate Medical Ltd, 352 group.bmj.com on October 13, 2017 Published by http://jcp.bmj.com/ Downloaded from