Maternal and fetal incidental findings on antenatal magnetic resonance imaging

Magnetic resonance imaging (MRI) is an increasingly utilized modality in obstetric practice for the further characterization of both fetal and placental abnormalities, as well as for abdominal assessment in women with suspected pathology. It is also used in some circumstances where pathology is suspected but ultrasound (US) images may be suboptimal, such as in cases of raised body mass index (BMI) or oligohydramnios [1]. With the increased diagnostic capability of MRI as a result of improved signal-to-noise ratio and enhanced image quality [2] associated with 3-tesla (T) imaging systems and the use of motion correction post processing techniques [3], unexpected abnormalities, both in the mother and the fetus, are commonly identified. Additionally, with large research studies using MRI, such as those in our own institution including iFIND (Intelligent Fetal Imaging and Diagnosis) [4], the Placenta Imaging Project [5] and the Developing Human Connectome Project [6], unexpected abnormalities may also be detected in otherwise low-risk pregnancies.

Previous studies in adults undergoing imaging have shown that unexpected findings are common, occurring in more than a third of adults undergoing cardiac MRI [7]; although most are unlikely to be of clinical significance, they often need further evaluation to reassure both the clinician and the patient. This may result in anxiety for women and increased health care costs. However, in some circumstances, identifying an unexpected anomaly may also alter the course of a patient’s treatment.

There are a number of studies in the literature regarding incidental findings in adult and paediatric patients, and guidelines have been published with regard to the management of such cases [8‐13], but data regarding antenatal MRI, where there is the possibility of unexpected findings in both the mother and fetus, are more limited [14].

This study aims to characterize the incidence of both maternal and fetal incidental findings in a large sample of pregnancies to facilitate enhanced counseling of women before MR imaging as well as to help inform the creation of care pathways for follow-up when abnormalities are identified.

A retrospective study was undertaken of all women who had an MRI during pregnancy at the Centre for the Developing Brain at St.. Thomas’ Hospital in London, a tertiary referral service and research centre, between January 2015 and December 2019. Incidental findings were identified from the clinical reports (generated routinely for both clinically indicated and research fetal imaging within the department).

Participants included all pregnant women who underwent a scan for clinical indications who also signed written consent forms that their data could be used for research purposes, those with known fetal/maternal conditions who were undergoing imaging as part of research studies and pregnant women who were participating as healthy controls in research. All women undergoing imaging received information sheets in advance of their MR examination highlighting the possibility of incidental findings being identified. This was then discussed again with the women at the time written consent was obtained just before the scans. All women undergoing imaging were also informed that the department has a standard operating procedure for the follow-up of all incidental findings.

All MRIs were performed on either a 1.5-T (Ingenia; Philips, Best, the Netherlands; or Area XMR; Siemens Healthineers, Erlangen, Germany) or 3-T (Philips Achieva) imaging system. Although the exact imaging protocol varied depending on the research study, for all examinations, an initial T2-weighted image was performed to ensure full coverage of the entire uterus, placenta and cervix. This may or may not include all maternal pelvis structures. However, the views obtained of maternal organs could be in unorthodox planes. All studies should include the fetal head and brain. Most also include the fetal body. Some late gestational age brain studies may not have dedicated fetal body imaging, but the fetal body would have been covered by the initial whole uterus scan. When repeat imaging was undertaken during the same pregnancy, all of the scans were included, but each abnormality was counted once only.

Maternal incidental findings were defined as previously unidentified abnormalities of the maternal structures (including the maternal cervix) and fetal abnormalities were defined as previously unidentified abnormalities of the fetus, amniotic fluid volume, placenta or umbilical cord. In accordance with a previous study by Abdullah et al. [14], incidental findings were categorized as level I, II or III. Level I findings had little to no clinical significance and were deemed as not requiring any further treatment or evaluation [11]. These have to be discussed with the mother and therefore may still cause anxiety. Level II findings had unknown clinical significance and were thought to be clinically relevant during pregnancy and level III findings were of high clinical significance and required urgent follow-up/care planning [14]. Categorisation of these significance levels was decided by a consensus of the multidisciplinary team of authors of this manuscript including radiologists, obstetricians and gynaecologists, and paediatricians.

A specialist perinatal radiologist reviewed all images and the majority were double reported; when a registrar or fellow reviewed the images, these were reviewed and signed off by one of three senior perinatal radiology consultants. Scans with incidental findings were all double reported.

All MRIs during 2015–2019 were included to assess maternal incidental findings. To assess fetal incidental findings, only the MRIs from healthy women, with US fetal anomaly scans reported as normal, who had volunteered for research studies during this period were assessed. In pregnancies with a fetus with a known structural abnormality, whether scanned for clinical or research indications, it was believed that an additional fetal abnormality could be part of an underlying syndrome rather than a true unexpected incidental finding.

Follow-up data were collected when available in order to assess the impact of any incidental findings on care pathways.

During the study period, 2,569 examinations were undertaken. One hundred and seventy-three women had two scans in the same pregnancy, nine women had three scans and one woman had four. Of the women who had multiple scans in the same pregnancy, 10 women had an additional finding on a subsequent scan (7 cases of a short cervix and 3 cases of mild/moderate hydronephrosis) and 5 had an additional fetal finding (1 case of renal pelvis dilation, 2 matured placenta, 1 fetus was noted to have the cord looped around the neck and 1 had a head circumference >97th centile). When multiple scans were undertaken in the same pregnancy, all scans were included but any incidental findings only counted once. Twenty-eight women were imaged in two separate pregnancies and four women in three separate pregnancies. Of the MRIs undertaken, 1,099 were in healthy volunteers. Imaging details can be seen in Table 1. Scans were reported by 14 different experienced clinicians, including 3 senior radiologists (A.E.T.J., M.R. and J.C.); all other clinicians were perinatal radiology fellows (including authors A.E. and S.A.).

There were 135 fetuses with incidental findings and, of these, 11 had 2 abnormalities. There were, therefore, 146 findings in total and the overall incidental finding rate in fetuses from uncomplicated pregnancies was 12.3% (Table 2).

In the 459 pregnancies with maternal incidental findings, 60 women had 2 additional findings, 1 woman had 3 and 1 woman had 4. There were 524 findings in total. The incidental maternal finding rate in women undergoing a fetal MRI was 17% (Table 3).

Of the 89 women who had a short cervix, less than 25 mm, the median gestational age at MRI was 31 completed weeks (range: 20–38 weeks). Four of these examinations were performed after 37 weeks’ gestation. The gestation at delivery was available for 69 cases, median gestation at delivery was 38⁺⁰ weeks (range: 26–42 weeks). Forty-eight were delivered at term (70%). Of the 21 delivered at less than 37 weeks’ gestation, the median gestation at delivery was 35⁺³ weeks (range: 26⁺⁰–36⁺⁶ weeks). Four of the 21 women were healthy volunteer controls in research studies and 1 woman had a history of depression and was participating in the BIBS (Brain Imaging in Babies) study but otherwise had normal antenatal ultrasound scans. Five women had multiple pregnancies, three of whom were iatrogenically delivered between 36 and 37 weeks’ gestation. One patient had premature rupture of the membranes and was known to be at high risk of preterm delivery before the scan (although the cervical length was unknown before imaging). One woman delivered iatrogenically preterm due to preeclampsia and another due to growth restriction. A further case had a cerclage inserted as a consequence of a short cervix identified on MRI at 21⁺⁵ weeks’ gestation, but subsequently elected to terminate the pregnancy due to a suspected diagnosis of tuberous sclerosis on a repeat MRI at 30⁺⁰ weeks’ gestation. The exact details surrounding delivery, other than gestation, were not available for seven women, although all were imaged for fetal abnormalities and it is possible some may have elected to terminate their pregnancies.

Five women had an open cervix, one with a cord prolapse noted at MRI (Fig. 1). This woman was participating in a study assessing women at high risk of preterm delivery. She was known to have premature rupture of the membranes, underwent imaging at 23⁺⁵ weeks’ gestation and delivered at 24⁺³ weeks’ gestation for maternal chorioamnionitis. A neonatal death occurred shortly after delivery. Of the other four women with an open cervix (gestational age range: 19–30 weeks), one was participating in a study of those at high risk for preterm birth and was known to have ruptured membranes, two were twin pregnancies undergoing clinically indicated MRIs (one for cardiac rhabdomyomas and one for ventriculomegaly in one of the fetuses) and the fourth woman was undergoing a fetal cardiac MRI for left atrial isomerism. The median gestational age at delivery was 24⁺³ weeks (range: 19⁺⁶–36⁺³ weeks).

In addition to women with an open cervix, the other level III maternal finding was a grossly dilated rectum identified on an MRI undertaken at 29⁺⁰ weeks’ gestation for suspected fetal brain anomalies (Fig. 2). This woman complained of nausea, vomiting and abdominal pain and hence emergency review by the colorectal surgeons was organized; however, the symptoms improved and this was not undertaken. The patient’s care was transferred to another trust for delivery due to the fetal diagnosis of rhomboencephalosynapsis with severe ventriculomegaly secondary to aqueductal stenosis, with a reduction in cortical folding, where paediatric neurosurgical input was required.

Of the level III fetal findings, there was one case of suspected polymicrogyria. However, the images were subsequently reconstructed using a different algorithm and the gyration was believed to be appropriate for gestational age. The fetus with grade 1 germinal matrix haemorrhage had an uncomplicated neonatal course (Fig. 3). No follow-up data were available for the fetus with a cerebellar haemorrhage. One fetus had suspected subependymal heterotopia and ventricular asymmetry identified on MRI at 28⁺⁶ weeks’ gestation. A repeat MRI was undertaken at 32⁺⁴ weeks and no abnormalities were identified.

Three suspected cases of congenital pulmonary airway malformation were identified. One case was reviewed in the Fetal Medicine Unit and the lesion was no longer visible on US in the third trimester. Two patients had chest radiography and computed tomography (CT) performed postnatally and no congenital pulmonary airway malformation was identified in either case. This may represent the natural evolution of the condition.

In this study of a large population of pregnant women undergoing fetal MRI, 17% of women had an incidental maternal finding and 12.3% of uncomplicated pregnancies had an incidental fetal finding.

Incidental findings are an inevitable consequence of fetal MRI and units undertaking examinations should ensure adequate locally agreed upon protocols and pathways are in place to optimize outcomes for both mother and child.