Abstract:OBJECTIVE:To establish a Parotid Imaging Reporting and Data System (PI-RADS) for CT diagnosis of the parotid gland neoplasms and to investigate the clinical applicable value and feasibility of PI-RADS.METHODS:Patients who had been diagnosed with primary parotid gland neoplasms and had received surgical treatments in Peking University School and Hospital of Stomatology during the period of January 2013 to December 2016 were included in this study. The diagnoses were confirmed by the postoperative pathological examinations in all the patients. The CT imaging data of all patients were retrospectively reviewed and analyzed by two readers in consensus. Imaging characteristics related to the parotid neoplasms were extracted and quantified. Based on comprehensive analysis of the imaging characteristics, the probabilities of the benign and malignant neoplasms were evaluated and classified into six grades, PI-RADS 1-6 (PI-RADS 1: normal parotid gland; PI-RADS 2: confidently benign lesions; PI-RADS 3: probably benign lesions without confirmed evidence of malignancy; PI-RADS 4: suspected malignancy without sufficient evidence of malignancy; PI-RADS 5: confidently malignant lesions; PI-RADS 6: lesions with confirmed pathological evidence of malignancy).RESULTS:A total of 897 patients with 1 003 parotid lesions were included. The lesions included 905 benign and 98 malignant lesions. The proportions of the malignancies in PI-RADS 2, PI-RADS 3, PI-RADS 4 and PI-RADS 5 according to the two readers in consensus were 0.4%, 5.7%, 35.5% and 96.7% respectively. The overall Cohen's Kappa test showed medium consistency between the two independent researchers (κ=0.614, P<0.001, 95%CI: 0.569-0.695). Pearson Chi-square test showed that the proportions of malignancies increased with the diagnostic PI-RADS grades (Cochran-Armitage trend test, Z=-15.579, P<0.001). The results of Pearson Chi-square tests showed significant differences between the grades [PI-RADS 2 and 3 (χ²=12.048, P=0.001); PI-RADS 3 and 4 (χ²=75.231, P<0.001); PI-RADS 4 and 5 (χ²=32.266, P<0.001)].CONCLUSION:PI-RADS can be used to evaluate the risk of malignancy and will be helpful to improve the imaging diagnosis and clinical treatment of parotid gland neoplasms.

Analysis on the CT findings of benign and malignant parotid tumors

Study on detection of tumor markers in serum in diagnosis of parotid tumor

Study on the CT Imaging Findings and Differential Value of Benign and Malignant Parotid Tumors

Quantitative Analysis and Pathological Basis of Signal Intensity on T2-Weighted MR Images in Benign and Malignant Parotid Tumors.

Computed tomography and magnetic resonance imaging findings of lymphoepithelial carcinoma in parotid gland: a report of 6 cases

Value of CT in the diagnosis of parotid neoplasm

Differentiating benign and malignant parotid gland tumors using CT images and machine learning algorithms

Analysis of CT findings in parotid adenolymphomas

Preoperative CT-based intra- and peri-tumoral radiomic models for differentiating benign and malignant tumors of the parotid gland: a two-center study

Machine learning-assisted diagnosis of parotid tumor by using contrast-enhanced CT imaging features

Combined Imaging Diagnosis of Parotid Masses

Contrast-Enhanced Ultrasound and Strain Elastography for Differentiating Benign and Malignant Parotid Tumors

CT features and analysis for misdiagnosis of parotid tuberculosis

Using Deep Learning to Distinguish Malignant from Benign Parotid Tumors on Plain Computed Tomography Images

Efficacy of conventional whole-body <sup>18</sup>F-FDG PET/CT in the incidental findings of parotid masses

Efficacy of Computed Tomography Features in the Differentiation of Basal Cell Adenoma and Warthin Tumor in the Parotid Gland.

[Parotid CT Imaging Reporting and Data System: A Preliminary Study].

22 Cases with Pathologic Correlation Adenoma of the Parotid Gland: A Report of CT and Ultrasound Features of Basal Cell

A retrospective analysis of 350 cases of parotid gland tumors and the value of ultrasound in the diagnosis of benign and malignant parotid gland tumors

The efficacy of clinicopathological and imaging features in differentiating benign from malignant parotid micro- and minitumors: A retrospective analysis

CT-based radiomics analysis of different machine learning models for differentiating benign and malignant parotid tumors