Ovarian tumors: prediction of the probability of malignancy by using patient's age and tumor morphologic features with a logistic model

Gynecol Obstet Invest. 1994;38(2):140-4. doi: 10.1159/000292467.

Abstract

An attempt was made to predict the probability of malignancy of a given ovarian tumor in a certain patient by using the age and simple morphologic features of the tumor. A cohort of 959 patients with ovarian tumors was analysed retrospectively according to the patient's age and tumor characteristics such as greatest diameter, consistency, bilaterality and diagnosis as malignant (271 patients) or benign (688 patients). All variables were entered unconditionally in a logistic regression. The presence of solid/multilocular elements has a 9.6-fold increased risk of malignancy, where a bilateral tumor has a 2.8-fold increase. Significant increase in risk of malignancy was observed in ages under 20 and over 40 years, as well as in tumors with a diameter larger than 9 cm. All variables were highly significant associated with the discrimination between benign and malignant. A formula including all variables has been developed so that the probability of malignancy can be estimated by a scientific calculator. In conclusion, simple, easily determined by ultrasound and reproducible criteria such as patient's age, tumor size, consistency and bilaterality were assembled in a logistic model in order to predict the probability of malignancy for a given ovarian tumor, in an individual patient.

MeSH terms

  • Adult
  • Age Factors
  • Aged
  • Female
  • Forecasting
  • Humans
  • Logistic Models
  • Middle Aged
  • Ovarian Neoplasms / diagnosis*
  • Ovarian Neoplasms / diagnostic imaging
  • Ovarian Neoplasms / pathology
  • Probability
  • Retrospective Studies
  • Ultrasonography