To optimise a fast sequence for MR-sialography and to compare a parallel and non-parallel acquisition technique. Additionally, the effect of oral stimulation regarding the image quality was evaluated. All examinations were performed by using a 1.5-T superconducting system. After developing a sufficient sequence for MR-sialography, a single-shot turbo-spin-echo sequence (ss-TSE) with an acquisition time of 2.8 sec was used in transverse and oblique sagittal orientation in 27 healthy volunteers. All images were performed with and without parallel imaging technique. The assessment of the ductal system of the submandibular and parotid gland was performed using a 1 to 5 visual scale for each side separately. Images were evaluated by four independent experienced radiologists. For statistical evaluation, an ANOVA with post-hoc comparisons was used with an overall two-tailed significance level of P =.05. For evaluation of interobserver variability, an intraclass correlation was computed and correlation >.08 was determined to indicate a high correlation. All parts of salivary excretal ducts could be visualised in all volunteers, with an overall rating for all ducts of 2.26 (SD +/- 1.09). Between the four observers a high correlation could be obtained with an intraclass correlation of 0.9475. A significant influence regarding the slice angulations could not be obtained (p = 0.74). In all healthy volunteers the visibility of excretory ducts improved significantly after oral application of a Sialogogum (p < 0.001; eta (2) = 0.049). The use of a parallel imaging technique did not lead to an improvement of visualisation, showing a significant loss of image quality compared to an acquisition technique without parallel imaging (p < 0.001; eta (2) = 0.013). The optimised ss-TSE MR-sialography seems to be a fast and sufficient technique for visualisation of excretory ducts of the main salivary glands, with no elaborate post-processing needed. To improve results of MR-sialography, it is reasonable to use an oral Sialogogum.
Read full abstract