A malignancy risk prediction model of parotid masses using ultrasound image characteristics and clinical features
10.3760/cma.j.cn131148-20210120-00048
- VernacularTitle:腮腺肿物超声图像特征、临床特征及恶性风险预测模型
- Author:
Yanping HE
1
;
Weijun HUANG
;
Bowen ZHENG
;
Weiguo CHEN
;
Genggeng QIN
Author Information
1. 南方医科大学南方医院放射科，广州 510515
- Keywords:
Ultrasonography;
Parotid mass;
Diagnosis;
Nomogram
- From:
Chinese Journal of Ultrasonography
2021;30(7):609-614
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To construct and evaluate a parotid mass malignancy risk model based on ultrasound image characteristics and clinical features of parotid masses.Methods:Ultrasound images and clinical features of 214 patients with parotid masses in the First People′s Hospital of Foshan were retrospectively collected from June 2018 to August 2020. The pathology results were taken as the golden standard. All the clinical features and ultrasound image features were first screened using regression analysis, and then the screened features were used to build a prediction model.Results:Malignant tumors of the parotid gland appeared on ultrasound as hypoechoic solid masses with or without abnormal cervicofacial lymph nodes with poorly defined borders and irregular morphology. Multifactorial analysis showed that facial nerve function, cervicofacial lymph node abnormalities, maximum diameter, morphology and borders of the mass were independent predictors of the risk of malignant parotid masses. A Nomogram prediction model was established using the above 5 indicators, and the results showed a concordance index(C-index) of 0.896 (95% CI=0.834-0.958) for Nomogram. The standard curve showed good agreement between the predictive effect of Nomogram and the actual situation of benign and malignant parotid swellings, with an internally validated C-index of 0.878. Conclusions:Ultrasound is of great value in identifying benign and malignant parotid tumors. The Nomogram model using ultrasound image features and clinical characteristics can assess the biocharacteristics of parotid masses, and the model shows high accuracy in predicting the risk of malignancy of parotid masses.
