This is a plain English summary of an original research article
If fetal brain abnormality is suspected on a pregnancy ultrasound, following this with in-utero MRI (iuMRI) improves diagnostic accuracy. This sequence could allow more informed discussions and decision-making around whether to continue with or terminate a pregnancy.
The NIHR funded study included 565 women of 18 weeks’ pregnancy or more who received ultrasound followed by iuMRI. Diagnoses were confirmed either by postnatal imaging of the baby or at post-mortem examination.
Overall iuMRI gave the correct diagnosis for 93% of scans compared to only 68% of ultrasounds. The accuracy of ultrasound declined above 24 weeks of pregnancy, whereas iuMRI performed well at all times. Clinicians reported that iuMRI scan results caused them to modify their future care planning in over 20% of cases.
Cost effectiveness was not assessed and iuMRI scans are not as widely available as ultrasound. Therefore, resource issues are an important consideration when judging whether or how to implement iuMRI into the diagnostic pathway.
Why was this study needed?
Fetal brain abnormalities occur in about three out of 1000 pregnancies (0.3%). Serious brain abnormalities can have significant impact on the child’s outlook for a normal life.
When a brain abnormality is detected by ultrasound scan, decisions may need to be made whether to terminate or proceed with the pregnancy, depending on the likely prognosis.
Previous studies have suggested that iuMRI following ultrasound may improve the accuracy of diagnosis but it has remained unclear whether iuMRI adds value or has an impact on pregnancy or new born baby outcomes.
Accurate diagnosis in utero is important to help clinical staff assess the prognosis for the foetus and provide parents with the necessary information to make an informed choice and to plan for the next steps of care which may include termination for babies with severe life threatening diagnoses.
What did this study do?
Pregnant women (aged 16 years or over) who had an ultrasound scan from 18 weeks onwards which showed a fetal brain abnormality were recruited from 16 UK centres. Eligible participants received an iuMRI within 14 days of ultrasound.
The accuracy of both ultrasound and MRI were compared against the confirmed diagnostic outcome. This came from postnatal brain imaging in the baby (up to six months of age) or alternatively post-mortem findings if the pregnancy was terminated or the baby stillborn.
Of 903 eligible participants, completed scans and confirmed outcomes were available for 570 babies (565 mothers).
The short timeframe between ultrasound and iuMRI should ensure iuMRI didn’t have an advantage, as brain abnormalities are easier to detect in larger foetuses. However, radiologists interpreting iuMRI had already seen the ultrasound results, potentially leading to confirmation bias and potentially increasing the apparent accuracy of iuMRI.
Because an entry criterion for this trial was a positive ultrasound scan, negative ultrasound results were not followed up to determine if they were truly negative. It was therefore not possible to compare sensitivities or specificities of each test.
What did it find?
- The overall diagnostic accuracy, that is the proportion of scans giving the correct diagnosis, was significantly higher for iuMRI (93%) than ultrasound (68%) (difference 25%, 95% confidence interval [CI] 21 to 29).
- The findings of ultrasound and iuMRI scans were both correct in 68% of cases (385 of 570) and both incorrect 7% of times (39/370). Incorrect interpretations of ultrasound scans were corrected by iuMRI in 25% of cases (144/570) while an incorrect iuMRI followed a correct ultrasound in less than 1% of cases (2/370).
- Ultrasound accuracy was better for foetuses aged 18 to 24 weeks, and declined with increasing age above 24 weeks. iuMRI was just as accurate throughout pregnancy age. Performing iuMRI after ultrasound improved diagnostic accuracy for brain abnormalities by 23% (95% CI 18 to 27) for foetuses between 18 and 24 weeks of pregnancy, and by 29% (95% CI 23 to 36) for foetuses 24 weeks and older.
- The main diagnosis was reported with confidence at 82% (465/570) of ultrasound scans and 95% (544/570) of iuMRI scans. These high confidence diagnoses were more likely to be incorrect for ultrasound (22%, 124/570) than for iuMRI (6%, 32/570).
- Based on clinicians’ reports, iuMRI was estimated to provided additional diagnostic information and change the prognosis for at least 20% of cases.
What does current guidance say on this issue?
NICE guidelines, updated in 2016, recommend that pregnant women have an ultrasound scan to screen for fetal anomalies between 18 weeks and 20 weeks plus six days of pregnancy. They say that women should be informed of the limitations of ultrasound and have the results explained to them to enable informed decision making about whether to proceed with the pregnancy or not.
What are the implications?
The findings suggest that iuMRI would be a useful addition to the current diagnostic and prognostic process for foetal brain abnormalities.
This study did not include an economic analysis and MRI is more expensive than ultrasound. Such data would be needed to inform implementation.
The specialist nature of iuMRI means that access to this investigation may be an issue in some parts of the country. If implementing this into the diagnostic pathway, training issues and ensuring adequate service to avoid any potential delays would need consideration. This is especially important for pregnancies around the legal limit for termination of 24 weeks.
Citation and Funding
Griffiths PD, Bradburn M, Campbell MJ, et al. Use of MRI in the diagnosis of fetal brain abnormalities in utero (MERIDIAN): a multicentre, prospective cohort study. Lancet 2017; 389(10068):538-46.
This project was funded by the National Institute for Health Research Health Technology Assessment programme (09/06/01).
NICE. Antenatal care for uncomplicated pregnancies. CG62. London: National Institute for Health and Care Excellence; 2016.
Produced by the University of Southampton and Bazian on behalf of NIHR through the NIHR Dissemination Centre