3D turbo spin-echo sequence with motion-sensitized driven-equilibrium preparation for detection of brain metastases on 3T MR imaging

Abstract

Background and purpose: MSDE preparation is a technique for black-blood imaging. Our purpose was to evaluate the usefulness of a 3D TSE sequence with MSDE preparation in detecting brain metastases by comparing it with conventional sequences.

Materials and methods: Postcontrast images of 227 patients who were suspected of having brain metastasis were prospectively obtained by using 3 T1-weighted 3D sequences: a gradient-echo sequence (MPRAGE), TSE-noMSDE, and TSE-MSDE. The number of visualized blood vessels and the lesion-to-normal CNR were compared among the 3 sequences. An observer test involving 9 radiologists was performed, and their diagnostic performance by using TSE-MSDE, MPRAGE, and combined TSE-MSDE and MPRAGE was compared by means of an FOM as an index of diagnostic performance derived by the JAFROC analysis, sensitivity, FP/case, and reading time.

Results: TSE-MSDE resulted in significantly better vessel suppression than the other 2 methods. TSE with and without MSDE resulted in significantly higher CNRs than MPRAGE. In the observer test, significantly higher sensitivity and FOM as well as significantly shorter reading time were achieved by TSE-MSDE compared with MPRAGE, but FP/case was significantly higher with TSE-MSDE. Combined TSE-MSDE/MPRAGE resulted in significantly higher sensitivity and FOM and similar FP/case and reading time compared with MPRAGE alone.

Conclusions: With blood vessel suppression and increased CNR, TSE-MSDE improves radiologists' performances in detecting brain metastases compared with MPRAGE, but it may increase FP results. Combined with MPRAGE, TSE-MSDE achieves high diagnostic performance while maintaining a low FP rate.

Fig 1.

Diagram of a TSE-MSDE used in this study. The MSDE preparation consists of a 90° excitation pulse, two 180° Malcolm Levitt refocusing pulses with each pulse sandwiched by bipolar motion-sensitizing gradients, and a −90° flip back pulse. The motion-sensitized gradients are inserted in all 3 (x, y, and z) directions. All refocusing pulses are implemented as composite pulses (90x-180y-90x). Bipolar gradients are used to reduce eddy current effects. An additional bipolar gradient is inserted in front of the sequence for further eddy current compensation.

Fig 2.

The distribution of the 50 metastatic lesions used in the observer study according to lesion diameter, which ranged from 2 to 36 mm, with an average of 6.8 mm.

Fig 3.

Number of blood vessels visualized in a single section of a postcontrast image obtained by using TSE-MSDE sequences at different VENC settings in 2 healthy volunteers. Lower VENC settings were used in earlier scans for volunteer 1, whereas the imaging order was reversed for volunteer 2.

Fig 4.

Postcontrast TSE-MSDE images for a volunteer (volunteer 1) obtained at 6 different VENC settings that are indicated in millimeters per second. Note that fewer blood vessels are visualized at lower VENC settings.

Fig 5.

Comparison of blood vessel numbers in the MPRAGE, TSE-noMSDE, and TSE-MSDE. A, The results for cases in which postcontrast images were obtained in the order of MPRAGE, TSE-noMSDE, and then TSE-MSDE (order 1, n = 25). B, The results for the reverse order (order 2, n = 25). Significant differences are found among the 3 imaging methods for both orders (P < .001, respectively).

Fig 6.

Comparison of CNRs for metastatic lesions for MPRAGE, TSE-noMSDE, and TSE-MSDE. A and B, Results for order 1 (n = 25) and order 2 (n = 25), respectively. Significant differences are found between MPRAGE and TSE-noMSDE and between MPRAGE and TSE-MSDE for both orders (P < .05, respectively), whereas no significant difference is found between TSE-noMSDE and TSE-MSDE for either imaging order.

Fig 7.

A brain metastasis imaged with TSE-MSDE (A) and MPRAGE (B). Note that the lesion is clearly visualized in TSE-MSDE (arrow), whereas it is poorly visualized in MPRAGE (arrow). Also note the well-suppressed blood vessels in TSE-MSDE. In the observer study, this lesion was overlooked by all 4 residents and by 2 of the 5 board-certified radiologists on MPRAGE, while it was correctly detected by 2 residents and all 5 board-certified radiologists on TSE-MSDE.

Fig 8.

An enhancing blood vessel mimicking a metastatic tumor. TSE-MSDE shows a nodulelike enhancement (arrow) in the right frontal lobe (A), which is not visualized in the next section at 2 mm below (arrow) (B). C and D, In MPRAGE, an enhancing blood vessel (arrows) running across the 2 contiguous sections is visualized in the corresponding area, revealing that the nodulelike enhancement in TSE-MSDE is a partially faded blood vessel.