PURPOSE: Ovarian epithelial tumors of borderline malignancy belong to a disease entity distnct from that ofbenign and overt malignant tumors, and the younger age of patients and earlier stage at diagnosis compared withinvasive ovarian cancers makes conservative surgery an important issue. The purpose of this study was to evaluatewhether there are any characteristic radiologic findings to suggest the presence of these tumors. MATERIALS AND METHODS: In 20 pathologically proven tumors of ovarian borderline malignancy, the findings of ultrasonography (n= 17) (with color Doppler ultrasonography in two cases), computed tomography (n = 8), or magnetic resonanceimaging (n = 6) were compared with pathologic findings and, were retrospectively reviewed and analyzed. Imageswere evaluated for size, loculation, thickness and regularity of wall and septum, endocystic vegetation, solidportion, calcification, local invasion and distant metastasis. RESULTS: Among 20 unilateral tumors, 18 weremucinous and two were serous; 19 (95%) were at stage I, and one was at stage II (local invasion of uterus, leftpara-aortic lymphadenopathy, and malignant ascites). Tumor size ranged from 10 - 40 cm in the largest diameter(mean, 21 cm). Radiologic findings of thick, irregular wall or septum were notified in 14 tumors (70%), endocysticvegetation in 12 (60%), enhancing solid portion in 11 (55%), and calcification in two (10%). Of 18 mucinoustumors, 17 (94%) were multilocular. One serous tumor was unilocular and one was multilocular. Color Dopplerultrasonography of two mucinous tumors revealed blood flow signals in the wall, septum, solid portion orendocystic vegetation, and resistive indices were 0.5 and 0.4, respectively. CONCLUSION: When ovarian tumors arepresent - especially in young females and at a lower stage -, radiologic findings of a thick, irregular wall orseptum, endocystic vegetation, and a small enhanced solid component, for example, indicate the possibility ofovarian epithelial tumors of borderline malignancy.