MRI Safety in First Trimester
Is there a need for a consensus statement from experts representing major societies such as ISMRM, RSNA, SPR, Obstetrics and Gynecology (e.g. Society of Maternal-Fetal Medicine, ISUOG), about the safety of fetal MRI, particularly in the first trimester for research purposes as no specific FDA guidelines exist?
- Some sites using as early as 12 weeks.
- Do we need more nonhuman primate studies?
- Do we need clear treatment interventions that would alter clinical management to pursue this? (hypothesis-based approach)
- Can we justify technical development (hypothesis-generating) activities?
- Do we first need to define normal?
Is it time for a large prospective T2* clinical trial? Do we need a "Healthy Placenta and Fetal Development" study after Healthy Brain and Child Development (HBCD)?
- Need large cohort to understand the normal variance across GA as well as racial, ethnic, environmental, socioeconomic effects, body habitus, etc
- What are the variations associated with clinical fetal disorders not associated with placental dysfunction?
- What variance arises from different scanner vendors, field strengths, acquisition strategies?
- What is the sensitivity and specificity for poor placental function?
- If T2* too low (or too high) what treatments should be considered? What is the clinical relevance?
- Is low T2* actionable or just an indication to monitor more closely?
- Eventually, we will need push button acquisition with the rapid calculation of T2*
- Are only a few slices used to calculate a whole placental average enough or do we need to pay attention to heterogeneity?
- What is test-retest reliability?
- Do we need phantoms to standardize across sites?
- The most common indication for placental MRI currently
- The most accepted reason for MRI
- Needs to be done in the first trimester
- Need MRI texture analysis vs GA (see MRI vs US below)
- Consider contrast-enhanced imaging using ferumoxytol or other
MRI vs US
- MRI will always be second-line due to expense
- MRI expense is driven by equipment and facility needs that are inescapable for 1.5 and 3T. Low field magnets decrease costs but come with lower SNR, but may be overcome by nanoparticle contrast agents
- Scan costs also depending on scanning time. If the acquisition is fast, costs will go down
- Two roles of MRI
Exploring mechanism with indirect impact on patient care
A diagnostic tool with a direct impact on patient care
- Ultrasound texture analysis shows high success in detecting Accreta
Need to distinguish and encourage both:
- Hypothesis based research with fixed protocols
- Hypothesis generating protocols that aim to generate new methodologies
NIH funding to continue both efforts are needed
Do we need novel technology to bring placental MRI to under-resourced areas?
- Flexible body arrays that allow placental imaging with low field permanent magnets may make placental imaging more feasible in low resourced settings
- Will low field (with less T2*) result in less sensitivity?
- The benefit of low field is fewer artifacts
- Many under-resourced areas have MRIs but no expertise.
- Is the better approach to educate so they can do acquisition with data transmitted to experienced sites for analysis and interpretation?
Need for sharing of Data, Protocols, Analysis Tools, Clinical Data Elements to enable AI
- Larger databases would enable machine learning
- Data harmonization a challenge
- Can all tools be centralized so all can have access
- HPP generated lots of data, needs more sharing, needs common data elements, and common protocols. Can these be made available to all?
- The difficulty is that the sample size for adverse pregnancies is too low in the general population – need to focus on high risk during development and validation
- For machine learning - hard to train with normal data due to variability – focusing on specific clinical phenotype is helpful, need balanced data set with large numbers of abnormals as well as normals
Combining MRI and Blood tests
- Potential of combining blood tests and MRI to better phenotype the placenta and see the impact of gene expression. Could the combination inform better use and interpretation of blood tests?
- Transcriptomics screen for developing MRI signature that something is wrong. Liquid sample first and then MRI? Especially 1st trimester
- Potential of AI to combine MRI, transcriptomics etc but need balanced data set with N and abN
- Fetal Blood Oxymetry. Umbilical cord measures of T2 to provide information on fetal blood oximetry and placental oxygen exchange are desired but currently challenges in spatial resolution in early gestation and motion
- Need to develop methods to distinguish intervillous flow and fetal vascular flow?
- Need to improve placental structural Imaging, standardization, and registration of functional imaging to anatomy
- Need to better manage regional variations in signal due to bowel, SNR, etc
- Need to better understand the interaction between fetal heart and placental exchange
What is the best validation?
- Although standards exist, challenging to accurately assess placental pathology given heterogeneity
- Placental pathology separated in time from imaging
- Is the better outcome metric the health of the neonate (ie pregnancy outcome)?
- Still need to fully assess placental location and anatomy as well as function
- Need AI methods to rapidly assess placental location and morphology
EPI BOLD with Hyperoxia
- Is maternal hyperoxia useful to understanding placental oxygen transport and placental oxygen metabolism?
- Although oxygen will not be a long-term treatment, is it useful to better understand the placental response to clinically applied oxygen?
- Oxygen disrupts metabolism and physiology - is this information clinically relevant?
- Potential for understanding oxygen exchange but challenges in the accuracy (due to motion) and separating maternal from fetal blood flow
Arterial Spin Labelling (ASL)
- What is it measuring?
- Does it matter if detects abnormalities and is a good screen?
- It may not be accurate but ok if differentiates healthy vs placenta at risk
- Low MTT - vsASL may be useful
- PCASL, PASL delayed transit and slow so inaccurate
- SNR low due to slow flow but blood volume high
- More development needed with a focus on vsASL
- Phase-contrast MRI of the intervillous space is a potential alternative for identifying placental blood flow abnormalities.
- Potential in understanding and characterizing placental flow
- IND exists for FGR
- Can't use as a bolus but can use for low TE angiogram, increased SNR for 4D flow, need to give outside the magnet by iv (currently in use for pulmonary emboli)
- Very low doses needed (⅛ what is needed for Gd or ~ 2mL vs the 17mL indicated for clinical therapy)
- Fetal exposure is unclear, but may be irrelevant since already used at much higher doses as iron supplement
Gd - liposomal particles
- No transport into fetus
- Blood pool agent
- More accurate than Ferumoxytol in accreta (data still preliminary)
- IND exists for amyloid