Objective:

To explore the value of logistic regression model derived from International Ovarian Tumor Analysis (IOTA) group simple rules in differentiation of benign and malignant ovarian tumors.

Methods:

A total of 221 ovarian tumors with pathologic results were retrospectively analyzed. Two ultrasound physicians who blinded to the pathologic results summarized the ultrasound characteristics of each tumor. The probability of malignancy was calculated using the logistic regression model. The diagnostic efficacy was obtained

by the comparison of malignancy probability to the pathologic results.

Results:

A total of 221 ovarian tumors included 168 benign tumors

37 malignant tumors and 16 tumors of borderline malignancy. When borderline tumors were classified as malignant ones and malignant probability of 30% was set as cutoff point

the sensitivity

specificity

positive predictive value

negative predictive value and diagnostic accuracy were 73.6%

100%

100%

92.3% and 93.7%

respectively.

Conclusion:

The logistic regression model derived from IOTA simple rules is a reliable tool for diagnosis of ovarian tumors

but the diagnostic efficacy for borderline tumors remains to be improved

and assistance from experienced radiologists may be needed.