Assessment of parotid masses: which MR pulse sequences are optimal?

The objective of this paper is to determine which MR pulse sequences are optimal for delineation of lesion and predicting pathologic nature of lesion with signal intensity. A prospective study was performed in 53 parotid masses (39 benign and 14 malignant lesions) in 53 patients. Signal intensity of lesion was visually assessed and lesion/parotid contrast-to-noise ratios were measured. On visual assessment, detection sensitivity was 100% for nonenhanced nonfat-suppressed T1-weighted images, 91% for nonfat-suppressed fast spin-echo (FSE) T2-weighted images, 83% for gadolinium-enhanced fat-suppressed T1-weighted images, and 75% for fat-suppressed FSE T2-weighted images. The highest contrast-to-noise ratios were obtained with nonenhanced T1-weighted images. Hypointensity of lesion relative to the parotid gland on nonfat-suppressed FSE T2-weighted images was seen in 11 of 14 malignancies, 12 of 15 Warthin tumors, and two of 18 pleomorphic adenomas. Cystic portion of hyperintensity on nonenhanced T1-weighted images was solely seen in benign tumors (n = 11). Thus, the highest accuracy (81%) (79% sensitivity and 82% specificity) for predicting malignancy was obtained with a criterion of hypointensity on nonfat-suppressed FSE T2-weighted images plus absence of cystic portion of hyperintensity on nonenhanced T1-weighted images. Nonenhanced T1-weighted images combined with nonfat-suppressed FSE T2-weighted images is optimal for delineation of lesion and prediction of pathologic nature of parotid masses.