“Seeing is believing” is true for parents who first see their baby on an ultrasound scan. It often makes their pregnancy more real, especially for fathers-to-be. “Seeing is also knowing”, at least to some extent, if all physical structures are present in your baby.
When the ultrasound technique made its way to the obstetrical domain in the fifties, it disclosed the pregnant belly and for the first time allowed us to observe the developing foetus in its own environment. Ultrasound technology has developed over time, we have seen improvements in the quality of registration and in the clarity of the images as well as the addition of three and four dimensional registration for better visualisation. The more you are able to see, the more you can know. Factors that play a role in the quality of registration and in the ability to trace abnormalities in fetal development include the skills of the sonographer, the mother’s size, the baby’s age and its position in the womb.
Last week, I picked up a position as a clinical midwife at a university teaching hospital in which I will be trained as a sonographer. I am very excited about this opportunity and I look forward to adding an extra dimension to my blog by crossing the border between the visible and the invisible in my contributions.
Today I would like to introduce you to an innovative research project called iFind. iFind (Intelligent Foetal Imaging and Diagnosis) is a multi-disciplinary project that runs in the UK and combines the expertise of paediatricians and specialists in the fields of software, robotics and imaging in an effort to improve the quality of antenatal ultrasound scans. In short, the seven year’s project aims at developing an automated scanning procedure, in which a robot performs the scan with multiple probes. Simultaneously, the researchers try to build an electronic ‘atlas’ of normal fetal development out of data obtained through ultrasound and MRI scans, which would allow for an automated analysis of the scans made by the robot.
MRI (Magnetic Resonance Imaging) relies on a magnetic field, radio waves and field gradients instead of on sound waves (as is the case with ultrasound). MRI allows to see through tissue and to better diagnose and monitor the fetus. In this way MRI can provide more detailed images and data for the atlas. MRI has already a place in obstetrics, for instance to better follow up congenital fetal conditions. It is however a very expensive technique, cannot be used for all groups and is not always perceived as a comfortable procedure, which makes it unsuitable as a standard alternative to prenatal ultrasound examination. The use of reference data obtained through MRI scans can however help in optimizing ultrasound scans, as iFind proposes. On top of that, the project has already led to an improvement in MRI imaging through the development of algorithms that ‘correct’ the movement of the fetus.
Some of the MRI footage obtained in the course of the iFind project was published on YouTube. The stunning images can be seen here:
The aim of the iFind project is to provide high quality scans and analyses for everyone. Currently, the quality of the examination depends on the expertise and skills available at a centre. Where there is less expertise or experience, detection rates can be significantly lower than in referral centres. And even there experts may misinterpret scans and miss certain abnormalities that are hard to see in a small, often moving foetus.
In december, a prepublication appeared online in which researchers related to the iFind project propose a new method to automatically detect and localise standard fetal structures, which is demonstrated in this video:
I am excited to follow up on new trends in the field, especially those that aim to improve the quality of prenatal care. It will probably take some time before innovations find their way to practice because of the validation of techniques. Apart from the technical challenges to be overcome, we might also have some bridges to cross to make the technique universally available. Particularly when you bear in mind that in some developing countries ultrasound technique often doesn’t even have a common place in obstetrics practice.
Baumgartner, C.F. et al. (2016). Real-Time Detection and Localisation of Fetal Standard Scan Planes in 2D Freehand Ultrasound. arXiv:1612.05601v1 [cs.CV] 16 Dec 2016. See: https://arxiv.org/pdf/1612.05601.pdf