A MULTIVARIATE LOGISTIC-REGRESSION ANALYSIS IN PREDICTING MALIGNANCY FOR PATIENTS WITH OVARIAN-TUMORS

Objective. Our objective was to improve the preoperative diagnosis of ovarian malignancy using a multivariate logistic regression analysis o n the basis of demographic, serologic, gray-scale morphological, and D oppler variables. Methods. One hundred seventy-one patients with ovari an tumors (120 benign, 51 malignant including 9 tumors of low malignan t potential) were studied with transvaginal B-mode, color, and pulsed Doppler ultrasonography before surgery. Based on the gray-scale ultras ound imaging, each tumor was classified as a unilocular cyst, multiloc ular cyst, unilocular cyst with solid parts, multilocular cyst with so lid parts, or solid tumor. Intratumoral blood flow velocity waveforms ere recorded on all tumors except unilocular cyst and were evaluated f or resistance index (RI) and peak systolic velocity (PSV). Serum CA 12 5 levels were also measured. Results. Twenty tumors were unilocular cy sts and were all benign. Seventy tumors including all unilocular cysts which showed no hows were all benign. The remaining 101 tumors (50 be nign, 51 malignant including 9 tumors of low malignant potential) pres ented intratumoral blood flows. Univariate and multivarlate logistic r egression analyses were conducted to identify variables predictive of ovarian malignancy in these 101 tumors. The variables included age, me nstrual state, serum CA 125 levels, B-mode classification, RT, and PSV . In univariate analysis, menopause, the positivity of CA 125 (greater than or equal to 35 U/ml), and PSV larger than or equal to 10.4 cm/s were found to be significantly associated with malignant tumors. The P SV value of 10.4 cm/s was the median in benign tumors. Multivariate an alysis showed that serum CA 125 levels (greater than or equal to 35 U/ ml) (P = 0.002) and PSV (greater than or equal to 10.4 cm/s) (P < 0.00 1) were to be independent predictors of malignancy. Conclusion. These results suggest that intratumoral PSV is the strongest means of differ entiating benign from malignant ovarian tumors with suspicious gray-sc ale ultrasonographic findings. (C) 1998 Academic Press.