Prospective evaluation of logistic regression models for the diagnosis of ovarian cancer

Objective: To test the accuracy of three logistic regression models in diag nosing malignancy in women with adnexal masses. Methods: This was a prospective collaborative study. Women were recruited f rom three hospitals and all assessments were performed at the Gynaecology U ltrasound Unit, King's College Hospital. One hundred women with known adnex al masses were examined preoperatively. The demographic, biochemical, and s onographic data recorded for each patient included age, menopausal status, CA 125 levels, ultrasound morphology, and Doppler blood flow analysis. The diagnosis of malignancy was made for each woman using three logistic regres sion models previously described by Alcazar et al, Tailor et al, and Timmer man et al. Variables used in these models were then combined to form a new model. The results were compared with the final histopathologic diagnosis. Results: Sixty-seven women had benign tumors and 33 had ovarian cancer. Wom en with malignant tumors were older than those with benign masses. There we re significant differences in CA 125 levels, presence of papillary prolifer ations, and ascites between the two groups. The sensitivities and specifici ties achieved respectively by the models were as follows: 45% and 93% with Tailor et al's model, 9% and 99% with Alcazar et al's model, and 73% and 91 % with Timmerman et al's model. There was no significant improvement over t he performance of Timmerman et al's model and the new combined model. Conclusion: All models performed less well than originally reported. Combin ing the models did not lead to a significant improvement in performance. La rger sample sizes that incorporate all types of ovarian tumors are necessar y to design more accurate diagnostic models. (Obstet Gynecol 2000;96:75-80. (C) 2000 by The American College of Obstetricians and Gynecologists.).