MR imaging of salivary duct carcinoma

Ken Motoori, Yumiko Iida, Yuichirou Nagai, Seiji Yamamoto, Takuya Ueda, Hiroyuki Funatsu, Hisao Ito, Okamoto Yoshitaka

Research output: Contribution to journalArticlepeer-review

53 Citations (Scopus)


BACKGROUND AND PURPOSE: Salivary duct carcinoma (SDC) is regarded as a high-grade malignancy in the current classification of salivary gland neoplasms. The aim of our study was to describe the MR imaging features of SDC. METHODS: Nine patients with SDC underwent MR imaging study. The apparent diffusion coefficient (ADC) values of SDCs were measured from diffusion-weighted images. Time-signal intensity curves (TICs) of the tumors on dynamic MR images were plotted, and washout ratios were also calculated. TICs were divided into four types: type A, curve peaks < 120 seconds after administration of contrast material with high washout ratio (≥30%); type B, curve peaks <120 seconds with low washout ratio (<30%); type C, curve peaks >120 seconds; type D, nonenhanced. We correlated the MR findings of SDC with the pathologic findings. RESULTS: All tumors had ill-defined margins and showed low to moderately high signal intensity for contralateral parotid gland on T2-weighted images. The average of the ADC values of the SDCs was 1.16 ± 0.14 [SD] × 10-3 mm2/s. Seven of nine (78%) tumors had type B enhancement. On the other hand, six of nine (67%) tumors with rich fibrotic tissue also had type C enhancement. CONCLUSION: The findings of ill-defined margin, early enhancement with low washout ratio (type B), and low ADC value (1.22 × 10-3mm2/s) were useful for suggesting malignant salivary gland tumors. Although it was reported that type C enhancement was specific for pleomorphic adenoma, SDC frequently has type C-enhanced focus.

Original languageEnglish
Pages (from-to)1201-1206
Number of pages6
JournalAmerican Journal of Neuroradiology
Issue number5
Publication statusPublished - 2005


Dive into the research topics of 'MR imaging of salivary duct carcinoma'. Together they form a unique fingerprint.

Cite this