A nomogram model based on clinical markers for predicting malignancy of ovarian tumors.

OBJECTIVE

The aim of this study was to build a nomogram based on clinical markers for predicting the malignancy of ovarian tumors (OTs).

METHOD

A total of 1,268 patients diagnosed with OTs that were surgically removed between October 2017 and May 2019 were enrolled. Clinical markers such as post-menopausal status, body mass index (BMI), serum human epididymis protein 4 (HE4) value, cancer antigen 125 (CA125) value, Risk of Ovarian Malignancy Algorithm (ROMA) index, course of disease, patient-generated subjective global assessment (PG-SGA) score, ascites, and locations and features of masses were recorded and analyzed ( 0.05). Significant variables were further selected using multivariate logistic regression analysis and were included in the decision curve analysis (DCA) used to assess the value of the nomogram model for predicting OT malignancy.

RESULT

The significant variables included post-menopausal status, BMI, HE4 value, CA125 value, ROMA index, course of disease, PG-SGA score, ascites, and features and locations of masses ( 0.05). The ROMA index, BMI (≥ 26), unclear/blurred mass boundary (on magnetic resonance imaging [MRI]/computed tomography [CT]), mass detection (on MRI/CT), and mass size and features (on type B ultrasound [BUS]) were screened out for multivariate logistic regression analysis to assess the value of the nomogram model for predicting OT malignant risk ( 0.05). The DCA revealed that the net benefit of the nomogram's calculation model was superior to that of the CA125 value, HE4 value, and ROMA index for predicting OT malignancy.

CONCLUSION

We successfully tailored a nomogram model based on selected clinical markers which showed superior prognostic predictive accuracy compared with the use of the CA125, HE4, or ROMA index (that combines both HE and CA125 values) for predicting the malignancy of OT patients.