COMPUTER-AIDED DIAGNOSIS - AUTOMATIC DETECTION OF MALIGNANT MASSES INDIGITIZED MAMMOGRAMS

A computerized method to automatically detect malignant masses on digi tal mammograms based on bilateral subtraction to identify asymmetries between left and right breast images was developed. After the digitiza tion, in order to align left and right mammograms the breast border an d nipple were automatically detected. Images were corrected to avoid d ifferences in brightness due to the recording procedure. Left and righ t mammograms were subtracted and a threshold was applied to obtain a b inary image with the information of suspicious areas. The suspicious r egions or asymmetries were delimited by a region growing algorithm. Si ze and eccentricity tests were used to eliminate false-positive respon ses and texture features were extracted from suspicious regions to rej ect normal tissue regions. The scheme, tested in 70 pairs of digital m ammograms, achieved a true-positive rate of 71% with an average number of 0.67 false positives per image. Computerized detection was evaluat ed by using free-response operating characteristic analysis (FROC). An area under the AFROC (A(1)) of 0.657 was obtained. Our results show t hat the scheme may be helpful to the radiologists by serving as a seco nd reader in mammographic screening. The low number of false positives indicates that our scheme would not confuse the radiologist by sugges ting normal regions as suspicious. (C) 1998 American Association of Ph ysicists in Medicine. [S0094-2405(98)00406-4].