Positive predictive values of MRI and ultrasound-guided fine-needle aspiration biopsy for isolated parotid swelling in adults: A STARD comparative analysis

Objectives: To evaluate positive predictive values (PPVs) of magnetic resonance imaging (MRI) and ultrasound-guided fine-needle aspiration biopsy (UFNAB) in patients with isolated parotid swelling. Material and methods: Observational study following the STARD guideline, based on a cohort of 212 patients from 18 to 93 years of age, with isolated parotid swelling (malignant: 16.9%; benign: 83.1%), consecutively operated on between 2015 and 2020, after work-up including MRI and UFNAB in an otorhi-nolaryngology department of a university hospital. The main endpoint was PPV for diagnosis of benign tumor, malignant tumor and the most frequent etiology. Secondary endpoints were correlations between PPVs and clinical factors for malignancy, and the impact on PPV of various situations: dynamic analysis on MRI; diagnostic disagreement between MRI and UFNAB; and UFNAB PPV according to MRI diagnosis. Results: PPVs for MRI and UFNAB were respectively 45.4% and 88.8% for malignant tumor, 89.6% and 46.9% for benign tumor, and 88.1% and 85.2% for pleomorphic adenoma (the most frequent etiology). Tumor fixation and history of head and neck radiation therapy PPVs were the only one higher than the MRI one for malignant tumor. MRI PPV did not differ between groups with or without dynamic analysis. PPV for malignant tumor, benign tumor and pleomorphic adenoma on MRI and UFNAB was respectively 42.8% and 33.3%, 42.8% and 100%, and 36.3% and 50% in case of diagnostic discordance. When MRI suggested malignant tumor, UFNAB PPV was 51.8% for malignant tumor, 67.7% for benign tumor, and 37.5% for pleomorphic adenoma; when MRI suggested benign tumor, it was 32.2% for malignant tumor, 91.5% for benign tumor, and 88.5% for pleomorphic adenoma; and, when MRI suggested pleomorphic adenoma, it was 23.5% for malignant tumor, 93.9% for benign tumor, and 92% for pleomorphic adenoma. Conclusion: Systematic association of UFNAB to MRI did not fundamentally improve diagnostic accuracy. UFNAB appeared most valuable in case of history of radiation therapy, in case of tumor fixation, and when MRI diagnosis was uncertain and/or suggested malignant tumor and/or the apparent diffusion coefficient (ADC) was low. The contribution of UFNAB when MRI suggested benign tumor or especially pleomorphic adenoma was more limited.(c) 2022 Elsevier Masson SAS. All rights reserved.