To the Editor.—
In the April Archives (1980;116:422-426), Callen and Headington claimed that their study "once again demonstrated a higher incidence of internal malignancy in patients with Bowen's disease." They suggested that the article by Andersen et al1 underestimated the true incidence of internal cancer. However, if we read the article by Andersen et al carefully, we become aware of their sound approach. That is, they used incidence rates of internal malignant neoplasms among the general population by age, sex, and year for comparison, they employed the Danish Cancer Registry, death certificates, and questionnaires to identify cancer cases, and, most important, they included only patients who had internal malignant neoplasms after their Bowen's disease (which we consider the most appropriate way of studying the relationship of internal malignant neoplasms to Bowen's disease).It is not entirely suitable to calculate the incidence of internal malignant neoplasms in the study of