COLOR FLOW-DIRECTED DOPPLER STUDIES OF OVARIAN MASSES - COMPUTER-ANALYSIS

OBJECTIVE: To determine if analyzing the entire color Doppler image (C DI) pulse wave of an ovarian mass can improve the ability to predict i ts histopathology. STUDY DESIGN: The CDI of 42 histopathologically dia gnosed adnexal masses were retrospectively analyzed. Using an image an alysis software program, the following parameters were calculated: are a under the curve, compactness, Feret diameter, perimeter, shape facto r and width of the waveform. Using an automated curve-fitting software program, the up and down slopes were processed separately for the opt imal equation and coefficient for each slope. Two computerized neural networks were created, both consisting of an input layer, one hidden l ayer and an output layer of three neurons: benign, borderline and mali gnant. The first network contained two input neurons: pulsatility inde x (PI) and resistance index (RI). The second network contained 10 inpu t neurons consistent with the shape and slope parameters calculated. T he coefficient of determination (R2) was determined for each network. RESULTS: The neural network utilizing RI and PI failed to train (1,397 runs, 67,056 facts, R2 = 0.59, 0.12 and 0.43 for benign, borderline a nd malignant masses, respectively). The network using the 10 calculate d parameters achieved an R2 of 0.96 after 685 runs and 27 facts. CONCL USION: Analyzing the CDI studies of ovarian masses, using the entire p ulse wave, improved the ability to differentiate between their benign, borderline and malignant histopathology.