OBSERVER VARIATION IN THE MEASUREMENT OF BRESLOW DEPTH AND CLARK LEVEL IN THIN CUTANEOUS MALIGNANT-MELANOMA

We have assessed the degree of observer variation of both Breslow depth and Clark's level in a series of 50 thin malignant melanomas. Our findings are similar to those of previous international studies in that Breslow depth is the more reproducible measure. Significant intra- and inter-observer variation exists and in some cases it was up to +/- 0.86 mm. Even small differences will potentially affect patient management at our centre and this was analysed using kappa statistics. Good agreement was found between observers and this could be improved by comparing the mean of two or more measurements. This removes larger errors, but smaller observer errors and differences in subjective interpretation of the deepest malignant cell mean that agreement will never be more than 90 per cent. This is high compared with studies of observer variation in other pathological conditions, e.g., dysplasia of the cervix, but where surgical management is potentially disfiguring it is not high enough. We conclude that Breslow depth and Clark's level should not be the sole basis of wide excision protocols.