This study aimed to assess the balanced steady-state free precession (bSSFP) cine sequence applied during fetal MRI to a wide range of clinical cases of suspected esophageal and tracheal obstruction or swallowing diffuculties.

7 human pregnant patients (mean gestational age 25w+4d, range 19w+6d-32w+0d) with fetuses with various clinical presentations of suspected tracheal or esophageal obstruction or swallowing difficulties were scanned using MRI at 1.5T or 3.0T.

6 patients were scanned using an Avanto 1.5 T system (Siemens Healthcare, Germany) with 2 flexible 6-channel body Matrix coils. Cine imaging was performed using a bSSFP sequence based on the Siemens TrueFISP GRE sequence. Sequence parameters: TR/TE 3.93/1.97ms, FA: 49°, 300x300mm FOV, 1.6x1.6mm voxels, 7.0mm slice thickness, 501Hz/Px bandwidth, acceleration factor 2, temporal resolution 0.5-0.7s.

1 patient was scanned using an Ingenia 3.0T MRI scanner (Philips Medical Systems, Netherlands). This sequence was based on a Philips TFE sequence. Sequence parameters: TR/TE 3.2/1.62ms, FA: 45°, 348x384mm FOV, 1.5x1.5mm voxels, 3.0mm slice thickness, acceleration factor 2, temporal resolution 0.4-0.5s.

The highly mobile fetus is a significant challenge in MR imaging. The bSSFP sequence was planned on the latest anatomical image, acquiring 1-5 dynamics to check positioning. This fast (1-2s) acquisition was repeated until the acquisition plane correctly intersected the anatomy of interest. A longer cine was then acquired, typically 200 dynamics. This was usually sufficient to observe at least one swallow event.

Attempted swallows were observed for all fetuses. On average 3 (range 1-6) short cine acquisitions were required to confirm the position of the image plane. In all cases between 1 and 3 full length cine acquisitions were acquired, each lasting 2.5 minutes. Repeat acquisitions were performed when a swallow event was not observed or the event was poorly visualised.

At 1.5T bSSFP banding artefacts were insignificant. At 3.0T banding artefacts were significantly increased, but did not affect the area of interest.

The bSSFP sequence provided relatively poor contrast between tissue types. However, in all cases the bSSFP sequence demonstrated excellent fluid/tissue contrast with fluid appearing bright. This allowed assessment of structures surrounded by amniotic fluid, such as the face, palate, oropharynx, nasopharynx, esophagus and trachea, as well as the movement of these structures during swallowing. In 3 cases the fetus was seen to swallow a large bolus of fluid (Figure 1). In 2 cases a small amount of fluid was seen to pass down the esophagus due to partial obstruction.

Amniotic fluid motion was clearly visible using this sequence, allowing observation of fetal regurgitation. In 3 cases the fetus was seen to fill the oro- and nasopharynxes before fully or partially expressing fluid out of the mouth (Figure 2). In a case where the fetus had a unilateral cleft lip and alveolar palate, fluid is seen entering the nasal cavity through the cleft and being expressed back into the surrounding fluid.

With conventional anatomical imaging of the fetal oropharynx and nasopharynx, clear definition of these structures is limited when not filled with amniotic fluid, with visualisation of the upper oesophagus being rare. For anatomical imaging the bSSFP sequence is poor due to relatively little contrast between tissue types compared to standard single-shot T2 weighted sequences used in fetal imaging.¹ However, the contrast between fluid and tissue is excellent in the bSSFP. The fluid bolus outlines the mouth and esophagus allowing assessment of its patency or visualisation of the extent of obstruction. Regurgitation (n=3) and ease of swallowing is clearly seen with this sequence. The high temporal-resolution sequence freezes motion and is sufficiently fast to show development of jets of fluid expressed from the fetus’s mouth.

At 3.0T the banding artefacts are significantly increased. However we have demonstrated the technical feasibility of acquiring clinically acceptable fetal bSSFP cine images at 3.0T.

The number of positioning attempts depends heavily on fetal movement and operator experience. Where the fetus is highly mobile it is worth persevering. In our experience the fetus may become agitated on beginning a new sequence due to the change in scanner noise, but often settles down, perhaps becoming familiar with the sound, allowing successful acquisition.

Cine acquisition is not currently widely used in fetal MR imaging, perhaps due to relatively poor tissue contrast and potential banding artefacts. The properties of the bSSFP sequence lend themselves, however, to the assessment of tracheal and esophageal structures due to the high fluid/tissue contrast. This study demonstrates that bSSFP sequences can provide additional information for a wide variety of clinical presentations of esophageal or airway obstruction, guiding parental and clinical decisions, and aiding safe delivery.