Serum estradiol were examined by radioimmunoassay in 80 patients with breast disease; 13 normal controls, 18 mastopathies, 33 breast cancers, 7 oophorectomized, and 9 adrenalectomized. Estradiol extraction method was modified by urea combing, and obtained accurate results by removing of lipids substance disturbing of radioimmunoassay. Premenopausal mastopathies have a relative high values than that of normals, but within normal limits. Postmenopausals have lower values than that of premenopausals. Breast cancers have various values most of within normal limits. Postmenopausal breast cancers also showed similar results, but nore various than that of premenopausals. The patients after hormonal surgery have low values. Hormone therapy for advanced and recurrent breast cancer is selected by hormone dependency of tumor, but the success chance is more expectably by agreement with hormone concentration in serum. Estradiol assay is useful as an indicator for surgical hormone therapy.