The Potential of Low-Field-Strength MRI for Simpler Fetal Scanning

See also the article by Aviles Verdera et al in this issue.

Download as PowerPoint

Since its earliest days, MRI has provided valuable information in fetal medicine (1,2). However, this capability has not been widely used even within well-resourced health care settings. There are possible reasons for this: issues with positioning pregnant women within the limited space available in scanners with conventional bore diameters, and problems with radiofrequency penetration at higher field strength. These factors can require longer times to set up a patient's examination, hence increasing the cost of scanning and making the procedure an unattractive part of a routine clinical pathway. However, low-field-strength, wide-bore scanners are becoming available and may address these issues.

In this issue of Radiology, Aviles Verdera et al (3) aim to determine the baseline performance of low-field-strength MRI in fetal imaging. This is a prospective trial in which 79 pregnant participants between approximately 17 and 39 weeks of gestational age underwent a 20-minute fetal MRI examination on a low-field-strength, wide-bore (80-cm) 0.55-T MRI scanner. The participants included 47 low-risk pregnant participants (control pregnancies) and 32 participants with pregnancy-related abnormalities (21 participants were clinically referred because they were unable to undergo higher field scanning).

The design of the protocol allowed a complete radiologic assessment of fetal anatomic structures and a quantitative assessment of some key parameters in the fetus and placenta. This included two sets of structural, T2-weighted, single-shot, partial Fourier, fast-spin-echo examinations (often known as half-Fourier acquisition single-shot turbo spin-echo, or HASTE), set up to cover the whole uterus and the brain separately (each acquired as three orthogonal stacks); T2* relaxometry performed in the brain and placenta using a gradient-echo single-shot echo-planar imaging sequence; and diffusion-weighted imaging in the placenta performed with seven b values between 0 and 1000 sec/mm². Finally, a cine scan was acquired to assess fetal movement.

Key brain features (biparietal and transcerebellar diameter) were obtained by following previously described protocols and normative curves by two radiologists experienced in fetal radiology and neuroradiology, respectively. Two radiologists independently double-read the first 40 MRI scans. An obstetrician experienced in fetal MRI assessed cervical length and lung volume. These anatomic measures were compared with literature curves obtained at 1.5-T and 3-T MRI or at US. The image quality, percentage of sections with artifacts, and the ability to perform a full clinical radiology report to assess qualitative scores were assessed for all 21 clinically referred participants. The placenta and brain were manually segmented by an experienced fetal MRI physicist, and standard T2* and apparent diffusion coefficient metrics were calculated in the fetal brain.

The authors reported some clinical findings, including incidental findings, with a maximum quality score (score of 3 on an ordinal scale of 1–3) in 13 of 21 clinically referred participants with pregnancy-related abnormalities and a full report possible in 20 of 21.

The standard fetal measurements agreed with published large cross-sectional 1.5-T and 3-T studies and showed a good interclass correlation greater than 0.93 on all fetal MRI scans. The T2* values (177 msec at 30 weeks in the placenta and 282 msec at 30 weeks in the fetal brain, both decreasing with gestational age) were longer than at higher field strengths. A review of the data shows a quadratic increase in relaxation rate with field strength, which would be expected, particularly if T2* is dominated by the local effects of deoxygenated blood (4).

The placental apparent diffusion coefficient and perfusion fraction values were similar to values found at higher field strengths but were scattered. The authors suggested that this may have been due to more conservative placental segmentation. However, it would seem to have required further investigation because many different regions of interest have been considered and different variations with gestational age have been reported (5,6).

The study by Aviles Verdera et al (3) showed that a commercially available, low-field-strength (0.55-T), wide-bore (80-cm) MRI scanner can be used to perform a short (20-minute) fetal imaging protocol that will provide anatomic information and quantitative data related to placental function and fetal compromise. The study protocol was broad and included scans that might be indicated for both fetal and placental complications. The quantitative results were generally robust and repeatable, although a further case-controlled study is required to determine definitively whether the system has sufficient sensitivity to the radiologic changes commonly identified and discriminated using fetal MRI.

It is expected that low-field-strength MRI would be more robust in pregnancy. A larger scanner is likely to be more acceptable to women and, using a survey, the authors found that participants who had undergone both 0.55-T and 3-T MRI scored the 0.55-T scanner as being much more comfortable. This is probably a benefit in fetal MRI because it is usually necessary to place patients away from the supine position to avoid aortocaval compression. If participants are more relaxed, they are likely to be still, allowing for faster scanning and better image quality.

A lower field strength also reduces artifacts from B₀ inhomogeneity in the body. This is an issue in the fetus because echo-planar imaging is often used to freeze fetal motion for quantitative imaging, and although the uteroplacental unit has uniform magnetic susceptibility (7), gas in the adjacent bowel can sometimes cause problematic image artifacts at the edge of the uterus. Reduced fat-water shifts will also improve image quality.

Another advantage of lower field strength is that B₁ is more homogeneous. Areas of signal dropout are often observed when imaging at late gestation or in polyhydramnios at 3 T because of radiofrequency interference in the large, highly conducting pregnant uterus at high field strengths (8). This problem is usually addressed by using region-specific radiofrequency adjustment (amplitude or shape of the radiofrequency field). However, this will not be necessary at 0.55 T, where this type of interference does not occur because the wavelength of radiofrequency in tissue is greater than 1.5 m at 22 MHz; radiofrequency adjustment will be more stable and radiofrequency distributions will be more homogeneous. This reduces the setup time required for scanning and makes scanning more repeatable and robust.

MRI has moved to higher field strengths because it provides increased sensitivity, which generally translates into increased spatial resolution at fetal MRI. The spatial resolution of the fast-spin-echo scans used here (1.5 × 1.5 × 4 mm³) is lower than that typically used for fetal fast-spin-echo scans at higher field strengths (9). The authors did not explain whether this was the minimum spatial resolution achievable. It remains to be seen whether this compromises clinical efficacy, particularly at early gestations. However, it is likely that many fetal MRI examinations can be performed at lower field strength with only certain cases being referred for higher-field-strength MRI.

The specific absorption rate of the radiofrequency field is reduced at a low field strength. The authors reported that the specific absorption rate was about one-ninth of that required to perform similar examinations at 3 T. This is important because heat deposition is a concern in fetal MRI: Fetuses have reduced heat-loss mechanisms compared with adults, and heat can be teratogenic. However, relatively high specific absorption rate single-shot fast-spin-echo examinations need to be used because they are fast enough to freeze most fetal motion.

The authors (3) indicated that the scanner is cheaper for several reasons including quicker scan times (because of increased patient compliance and faster scanner setup and adjustment) and reduced helium consumption and installation issues. They also argued that the reduction in artifacts eliminates the need for correction tools and thus reduces the need for specialist techniques. These issues are likely to be somewhat site-specific, but it is true that a relatively quick scan protocol on a cheaper and more patient-acceptable scanner is more likely to expand the use of MRI in fetal medicine and obstetrics.

This study (3) has shown that 0.55-T wide-bore MRI can produce good-quality data in the fetus and placenta with reduced scanner setup time. Further work will be necessary to establish the relative diagnostic efficacy of this scanner in investigating the range of congenital abnormalities referred for fetal MRI. This is an important step in establishing the sensitivity and efficiency of low-field-strength MRI, which will be important in selecting the right field for a particular scan.