Regression in thin malignant melanoma. Microscopic diagnosis and prognostic importance

Forty-eight malignant melanomas of the extremities, 1 mm or less in maximal thickness, were studied to better define microscopic criteria of regression in thin melanomas. Eleven tumors (23%) exhibited definite regression in the form of one or more segmental areas (defects) where the invasive component was replaced by mononuclear cell infiltrate and fibrosis. Thirteen other tumors (27%) had diffuse, nonsegmental changes classified as probable regression. Nineteen lesions (40%) lacked regression, although 14 of these contained focal evidence of host response. Five melanomas (17% of lesions less than 0.75 mm thick) were equivocal for regression. There were no recurrences or metastases. The histologic diagnosis of regression in thin melanomas requires subjective judgments, but segmental defects represent a potentially reproducible criterion. Their width can be measured, and the proportion of the melanoma that has undergone regression can be estimated. The preponderance of data from the literature, supported by this study, indicates that regression has no prognostic importance in the vast majority of thin melanomas. There are observations, however, to suggest that in rare cases, regression may negate the prognostic value of microstaging of a thin melanoma. To date, the type and extent of such regressive changes have not been adequately defined.