# MRI Signatures of Parotid Tumours Impacting Management Decisions: A Retrospective Study With Radiology and Pathology Correlation

## TL;DR

This study shows that MRI can accurately predict the nature and type of parotid tumors, helping guide treatment decisions.

## Contribution

The paper introduces MRI signatures that predict parotid tumor malignancy and histopathology with high accuracy.

## Key findings

- MRI signatures achieved 94.23% accuracy in distinguishing benign and malignant parotid tumors.
- T2 signal intensity and enhancement patterns help identify low-grade mucoepidermoid carcinoma.
- MRI signatures significantly impact decisions about elective neck dissection.

## Abstract

Fine needle aspiration (FNA) from parotid tumour is inadequate and nondiagnostic in 8% and FNA/biopsy from deep lobe is technically challenging; hence, our first objective was to evaluate MRI findings which best predict the benign and malignant nature of parotid tumour. Our second objective was to develop MRI signatures for parotid tumour histopathologies including grades of carcinoma, to help in decision making regarding elective neck dissection.

Two head and neck radiologists retrospectively evaluated and developed signatures of common benign and malignant parotid tumours using morphology and signal intensity–related variables for 98 patients on MRI available in PACS from 01 January 2016 to 26 December 2022. T1 weighted image (WI), T2WI, short tau inversion recovery, diffusion WI/apparent diffusion coefficient and postcontrast T1WI sequences were evaluated. The developed MRI signatures were then validated by a blinded third radiologist.

Sensitivity, specificity, accuracy, positive and negative predictive values using MRI signatures were 92.31%, 100%, 94.23%, 100% and 81.25%, respectively, for benign and malignant nature of parotid tumours with a highly significant p‐value (< 1e‐04). Developed MRI signatures also showed high statistical performance and significant p‐value for parotid tumour histopathologies and grades of mucoepidermoid carcinoma (MEC). T2 signal intensity and enhancement patterns can help identify low‐grade MEC, impacting management decisions regarding elective neck dissection.

MRI can predict the benign and malignant nature, parotid tumour histopathologies and grades of MEC when typical signatures are present, impacting management decisions.