Why is Fetal Soft Tissue Assessment Important? - SD

Introduction

My name is Dr. Wesley Lee. I'm from the Division of Fetal Imaging at William Beaumont Hospital and Royal Oak, Michigan.

Today I will be talking about the importance of fetal soft tissue assessment.

Today's Topic: Why is Fetal Soft Tissue Assessment Important?

By the end of this presentation, I would like you to describe current limitations associated with the use of weight estimation alone for evaluating fetal soft tissue development, and to explain why fetal limb volume measurements may help us to improve fetal weight estimation and nutritional assessment.

And finally, to state how conventional sonographic growth and fractional limb volume parameters are related to birth weight and infant percent body fat.

Qualitative and Quantitative Applications of 3D Ultrasound

During the symposium, we've been talking about the qualitative uses of three dimensional ultrasonography, and we can see some amazing pictures, which range from visualization of a 14 week fetus with the umbilical cord or the appearance of the cardiovascular system with the surrounding 3D anatomy.

On the right you can see bilateral cleft lip. And on the bottom left, there's an example of angiographic views of the outflow tracks of the heart using spatial temporal image correlation. The middle one shows a dilated renal system, and the far right bottom image shows an acardiac twin.

I think that's all amazing in terms of what 3D ultrasound can show us, but there's also some quantitative applications for this technology as well that I'd like to further explain.

Fetal Growth and Nutrition

During this lecture, Doctors Scott and Usher from Montreal once wrote a paper called Fetal Malnutrition, its Incidents, causes and Effects. And they said in the Gray Journal, 1966, that fetal growth is a function of both seed and soil. It is dependent upon the growth potential of the fetus and the availability of intrauterine nutrition, and has brought a sense to fulfill this potential.

The result of these two factors is a wide distribution of birth size at any one gestational age, and a wide variation in the state of nutrition at birth.

According to the March of Dimes, one in every 12 newborns in the United States are delivered with a low birth weight, less than 2,500 grams. And these low birth weight infants are higher risk for perinatal death, developmental delay, learning disabilities, cerebral palsy, and hearing loss.

On the other end of the spectrum, there's been some information from the National Center for Vital Statistics showing the incidences of fetal macrosomia at the varying menstrual ages, which range from 37 weeks or 42 weeks. You can see that at the 50th percentile, that this ranges from 3,100 grams to about 3,500 grams. And then the macrosomic babies being either at the 90th percentile or 95th percentile range from approximately 3,800 grams to about 4,100 grams.

These over nourished babies are also at increased risks related to cesarean delivery, shoulder dystocia, fractured clavicles and brachial plexus injury.

Traditional Approaches to Fetal Growth

Traditionally, the approaches to fetal growth have been based upon a classification system that was first described in the 1960s by Dr. Chenko and colleagues in Denver, Colorado. And essentially, this was a statistical definition of babies, actually infants that were small for gestational age with birth weight less than 10 percentile. And the ones that were large for gestational age had birth weights greater than 90th percentile.

So this information was extrapolated to a fetal population. And for instance, this paper by Dr. Brenner and colleagues determined the weight percentiles as a standard for fetal growth. And they took about 31,000 liveborn infants delivered from 21 to 44 weeks gestation, and also included 400 3430 fetuses at eight to 20 menstrual weeks that were aborted. And they developed these standards as you see in the graph on the right. The national Center for Vital Statistics and the work of Dr. Alexander summarized a US National reference for fetal growth using liveborn data from 1991.

And you can see from this graph that graphs the weight in grams on the Y axis and the menstrual age on the X axis, you can see the weight percentiles with the red line being the 50 percentile. And then you have the 10th and 90th percentile as shown by the white lines and the fifth and 95th percentiles.

Now, this is all interesting, but I'd like to show you a comparison that they made in that paper, comparing it to the large normal reference ranges from Cleveland, North Carolina, from California, from St. Louis and Denver. And the point being is that we have a similar type of graph with a dark line showing the 10th percentile line of the national standard that Dr. Alexander published, as you can see with it from the arrow.

And then if you take any one of these other 10th percentile cutoffs for any of these other large studies, you can see that there's agreement for this boundary up to about 30 weeks or roughly 31 weeks after which time there is some divergence between the national standard that I just described with the other series.

So if you find a fetus and you use a weight calculation and a percentile chart, say for instance, 34 weeks or 33 weeks, you can see that if the dot resides above or below this line or below this line, the point is you can really interpret the information differently depending on which population standard is actually used to interpret your fetal biometry.

The fact of the matter is that the diagnosis of intrauterine growth restriction is not standardized.

Lack of Consensus on IUGR Diagnosis

There was a recent review of 108 original articles and 11 case reports that were submitted to our white journal, the Journal of Ultrasound and Obstetrics and Gynecology between 1995 and 2007, using the mesh term IUGR. You can see from this graph that many of the articles came from the United Kingdom, United States, Germany, Italy, Sweden, and Netherlands, and the proportions as seen in this pie chart, but we found no consensus for the IUGR diagnosis.

In fact, about a third of them, 34.5% used the criteria for IUGR as the estimated fetal weight less than a 10th percentile, 15% used EFW less than a fifth percentile, 3.4% used EFW less than a third percentile, and 1.7% used a definition of the abdominal circumference less than 10 percentile. And 18.6% used the definition of the AC less than a fifth percentile.

Some of them also advocated using some combination of this criteria with doppler. So the bottom line is that there really is no consensus for IUGR diagnosis from an international perspective.

So today I'd like to talk to you a little bit more about how fetal soft tissue can help.

History and Studies on Fetal Soft Tissue Assessment

There's been a number of investigators throughout the past two to three decades talking about fetal soft tissue. Starting with some description of the fetal circumference by Dr. Dieter in 83, Dr. Felipe Shante and Romero and Hobbins talked about the subcutaneous tissue of the arm and leg. And then Dr. S talked about the fetal thigh and caster conferences.

It goes on and on and on, involving the fetal abdomen, the cheek, the cheek diameter, the fetal buttocks, and even the fetal thigh soft tissue thickness.

So the problem has been though is it's difficult to replicate these measurements and to use them in a clinical manner, but as the technology improves it, there are actually some options that are being considered currently in the white journal.

In the White Journal. In 2003, there was a editorial by Dr. Schwartz and Glan from Colorado, talking about ultrasound and the assessment of fetal growth disorders. Is there a role for subcutaneous measurements?

And then Dr. Fred Batal from Colorado was quoted as saying in another publication, in Animal Studies, regardless of the model being used, fetal growth restriction fetuses are unable to deposit normal fat stores in late gestation. When FGR babies are examined after delivery, they show all the clinical signs of reduced fat stores. Thus, if fat concentration could be estimated fairly precisely during fetal life with ultrasonography of fatty lean ratios in various sites, this might help distinguish the normal small from the FGR fetuses.

Well, one of the studies from Colorado demonstrated a reduction of subcutaneous mass, but not lean mass in normal fetuses in Denver, Colorado. And they were measuring the subcutaneous fat shown by ultrasound around the fetal abdomen, as well as tracing out the fetal fat versus the lean muscle mass in the fetal limb.

It's sort of a tedious process, but it was kind of an interesting study. They went on to work with their Italian colleagues to show differences in fat and lean mass proportions in normal and growth restrictive fetuses, and found that fetuses with growth restriction have reduced subcutaneous fat and lean body mass compared to normal controls. And that subcutaneous fat concentration decreases are proportionally greater when compared to decreases in lean body mass.

There was one review of the literature looking at the prediction of macrosomia with soft tissue by Dr. Shahan. And they used things like ultrasound derived estimate fetal weight, clinical estimates, estimate fetal weight, and soft tissue parameters including the upper arm subcutaneous thickness, the femoral subcutaneous tissue, cheek diameter, and so on and so forth.

And they were able to show some correlation in the prediction with the parameters marked in yellow. But there's other ones that were not significantly helpful using the receiver operator curves. And they stated that the potential explanations for the poor performance of some of the soft tissue includes a different population, possible ethnic variations and distribution of soft tissues. Irreproducible measurements of soft tissue or inability of these proposed models to identify newborns with weights greater than 4,000 grams.

Now, if you look at the fetus, and this is a fetal MRI at 28 weeks up on the upper left, you can see the cross section of limb, and you can see the fat layer around the lean body mass, which consists of the muscle, darker muscle, and the bone. You can see how it delineates or how it appears on the fetal limbs at 28 weeks and 34 and a half weeks.

And then if you look at this view, you can see how you can see the fat tissue and the muscle of the fetus. It's this type of tissue that we want to quantify as an index of intrauterine nourishment or malnourishment.

Fractional Limb Volume

So in it's been a little few years since we originally introduced the concept of fractional limb volume. It's a limb sub volume based on 50% of the long bone diaphysis length.

So if this is the arm, you would take the measurement of the humerus diaphysis, and then the computer software would define a sub volume of limb that's centered in the middle of the bone based on 50% of the long bone length. So this is called fractional arm volume, or FAV and fractional thigh volume, or FTV.

And the normal ranges and the technical considerations have recently been published in the April issue in the White Journal 2009, as shown here.

This is an example of how the 3D volume dataset is used to acquire the fetal limb. These are orthogonal multiplanar images. And you can see I can just magnify the thigh and I can look back and forth, move the plane back and forth to look at the axial views of the soft tissue on the right.

And then the next thing I would do is use the measurement tool, fractional limb volume right there, and take a essentially take a measurement of the femoral diaphysis length from one end to the other end, at which time the computer, and this is 4D view software from GE Healthcare. You can actually take five manual traces of the sub volume that's defined by the computer.

And once this is accomplished over about two minutes or less, then you get the fractional limb volume.

Now, I wanna explain why we go to this, why we were going this route. Because when Dr. Chang FGM Chang from Taiwan originally talked about fetal limb volume and birth weight, he would be describing approximately 10 minutes per limb. This procedure standardizes the measurement, and it takes approximately one and a half to two minutes to accomplish.

Now, eventually this will be automated, but this is a great first start, I think. So here's the fifth one here, and then once this is finished, then essentially you can go, you have the option of going back and correcting a manual trace. And once you finish that, then you can say done, after reviewing your slices to get a tomographic view of all the five slices to make sure that that's what you're happy with.

And the fractional limb volume in this case would be 2.6 milliliters or cubic centimeters.

Technical Considerations for Fractional Limb Volume

Alright? Now, there are some technical considerations for this new parameter. This is an example of a fractional limb volume at 12 weeks. And that's pretty easy because you see a lot of fluid around the limb, around 20 weeks. You see the limb can be sandwiched between the placenta and the body of the uterus, sorry, the body of the fetus.

But you can still see how you can create these tomographic images of these soft tissue slices. Now it's important to understand why we're emphasizing the middle portion of the limb. Because if you take this panel A and you can see the dot over here corresponds to a cutting plane that creates a axial cut of the limb. And you can see the axial limb over here. You can see the borders quite well.

But if you go over to one side, one end or the other end, you can see that the edges of the soft tissue borders become fuzzy. And this is what we want to avoid. So that's why we emphasize the center of the limb.

I think that the use of color filtering such as STIC helps to improve the tissue contrast. So you can see the borders better. And then if you put too much pressure on the maternal abdomen, you can actually obliterate the interface between the soft tissue and the uterine wall, as opposed to letting up pressure. And you can see that the fluid fills the space, and you can see the borders quite well.

Also, another potential pitfall is using an inadequate sweep angle into the screen, so to speak. And you can see how it may look like that you have a sagittal acquisition of the fetal limb, but the sweep doesn't go completely through because the orthogonal plane shows that the limbs actually indeed cut off.

And if the baby moves, then that causes a problem during the acquisition. So if that happens, essentially it's not usable, the information's not usable, and you shouldn't save it and should just simply do another acquisition.

Sometimes it's confusing to know whether the limb is an arm or a leg. But if you look carefully, you can see not only the features such as the elbow, the shoulder, and the chest, sometimes the ribs. There's also a line of demarcation where the deltoid muscle is attached to the upper arm. And you won't see this little notch on a leg.

Alright, so fractional limb volume is a new parameter with new obstetric applications that will should be helpful for us or as an indirect index of fetal soft tissue development and nutritional status. And also can be used for fetal weight estimation and can be used for something called individualized growth assessment.

Fractional Limb Volume for Fetal Nutritional Assessment

So let's start with the use of this parameter for fetal nutritional assessment. This information again is has just been published in the April issue of the journal ultrasound in obstetrics and gynecology. And this work was in conjunction with my colleagues at the perinatology research branch at the National Institutes of Child Health and Human Development.

But essentially, we the ranges, the normal ranges of fractional arm volume and fractional thigh volume. And in that paper, we also talk about the reproducibility of the study using these Bland Altman plots.

So, for instance, for fractional arm volume, intra observer, you can see that if you take blinded observers taking these measurements on the arm, that there's only a 2.2% bias between blinded observers and 95% of the observations or the measurements are within the limits of minus 6% to 10.5%.

But we did this with the arm volume for the same observer and between observers. And we did this all for the thigh volume for the same observer and the different observers. So the reproducibility of the technique appears to be sufficient for clinical use.

We also defined normal ranges for fractional arm volume and fractional thigh volume. And if you refer, if you go to a website called www.obsono.org, you'll see some education materials which explain the rationale for the soft tissue measurements, as well as it provides a online calculator that calculates the percentile or z-score for normal FAV, and even gives you an online demonstration of the technique.

So this is an example of the online calculator where you type in the enter the menstrual age in weeks and days, and then you type in the FAV FTV measurement and say calculate, and it gives you a Z-score or a percentile, and graphs one or more measurements for each parameter.

Correlation of Fetal Growth Parameters with Birth Weight and Infant Body Composition

Well, with that background, I'd like to address the question of how do fetal growth parameters that we use every day in actual birth weight correlate with newborn infant body composition. And I also wanted to see how these, how the neuro biometry, such as biparietal diameter, head circumference, abdominal circumference, femur length, how that also compares with the new soft tissue parameters.

And this while this also was a paper just published in the April issue of the White Journal, at in collaboration with the Perinatology research branch of the NIH, we've been using air displacement plethysmography.

You might have heard about the athletes that are placed into this compartment called the BOD Pod. And where the athletes or where their patients, they sit in this capsule and they displace some certain amount of air to allow a calculation of volume displacement. And then using the well known relationships between density, mass and volume, then it's possible to estimate percent body fat and lean body mass.

The same principles have been applied to the infant, the newborn infant, and this is called the PEA POD system. And you can see a picture of it down below. This is somewhat analogous to underwater hydrostatic weighing. When people go into the tubs of water displace the water, and then they're able to determine the percent body fat.

At any rate, you can see the baby being weighed here, and then the baby goes into the compartment and displaces air, and we get percent we have a two compartment model of fat percent fat and lean body mass.

Now, if you take we took 87 fetus newborn pairs, and we measured them within four days of delivery. We measured the biparietal diameter heads circumference, the humerus and femoral diaphysis length. We calculated estimate fetal weight on a basis of two Hadlock formulas using the AC and fem length or a head AC and fem length.

We looked at abdominal circumference. We added a fractional arm volume, mid thigh circumference, and we actually looked at birth weight as well as fractional thigh volume. And essentially we found that birth weight, which has been long known to correlate well with the nutritional status of the baby, explains about 44.7% of the variance in neonatal body fat on the base of these PEA POD studies.

And but if you look at the other biometric parameters, for instance, if we look at abdominal circumference, the AC only explains 24.8% of the variance in percent body fat of the newborn. And if you look at even weight is only explaining 28% or 30% of the variance in neonatal percent body fat.

Which means that if the goal is to determine which babies are truly malnourished on the basis of percent body fat, then we're not doing a very good job with just the traditional biometric parameters.

Now if you look at fractional, the best parameter is actually fractional thigh volume that essentially has identical results with birth weight. So this is food for thought, especially for those in Europe that use AC alone to make the clinical definition of an intrauterine growth restricted baby.

So infants who are delivered during the late third trimester have percent fat mass that correlate best to actual birth weight and fractional thigh volume as compared to abdominal circumference. And our findings may have clinical ramifications for those who use AC alone, or even EFW prediction models that are based on AC that exclude soft tissue parameters to identify fetuses at risk for growth abnormalities.

Fractional Limb Volume for Fetal Weight Estimation

Alright, so this brings us to the second point, the second major area of how we can use fractional limb volume for fetal weight estimation. And the idea is not only to improve the precision of fetal weight estimation, but also to add a nutritional component, a soft tissue development component to this estimation that will give us an idea of how the baby, how the fetus is nourished.

So to state the problem, generalized fetal nutritional status is commonly evaluated using fetal weight or abdominal circumference. We already know that newborns that are growth restricted or macrosomic are at higher risk for poorer outcome. But fetal weight, unlike birth weight, actual birth weight cannot be directly measured. So there's a certain uncertainty about this estimation.

A recent literature review by Dr. Dudley showed that there was no and this is like 11 different methods of fetal weight estimation concluded that there was no preferred method, and the size of random errors remain a major obstacle to confident use in clinical practice with 95% of the confidence intervals exceeding 14% of birth weight in all studies.

Well, so we have current technical limitations for estimated fetal weight. Now if you look at the pictures on the left top, you can see that the picture of a growth restricted baby at term weighing 2845 grams, and you can see the thigh of a macrosomic baby at term. And intuitively, it's obvious that there's a difference between those two legs.

But the question is, how can we characterize those two appearances just on the basis of abdominal circumference or femur length? And if you can imagine having someone come through the door of the emergency room and then the first thing that happens is that the nurse takes a length of the femur bone and a tape measure on the belly. And then if you're trying to use that information to say whether the person's healthy or not, well, you know, that's a tall order on the base of that limited information.

There's some very interesting theory from Dr. David Barker, known as the Barker Hypothesis, which suggests that low birth weight newborns, such as the one on the left here, as opposed to the normal one, as opposed to the one on the far right being term rubicund and Macrosomic, the one on the left, the low birth weight baby are actually at higher risk for developing disease in later adult life, diabetes, heart disease, high blood pressure, and this is a result of something called fetal programming.

We'll talk more about that in just a little bit. So the current prediction models do not include soft tissue parameters for fetal weight estimation because these soft tissue parameters are difficult to standardize. However, three ultrasound fetal limb volume measurements can be time consuming. And the fetal soft tissue borders are fuzzy at the ends of long bone.

So this is where the fractional limb volume comes in. In the year 2001, we used the fractional limb volume to develop fetal weight estimation models that incorporate 2D that add the 3D volume, fraction thigh volume or arm volume to traditional 2D parameters.

And you can see from this busy table, you can see the original Hadlock models using AC and femur length or AC. And we can use AC and femur length model from Houston as originally described by Dr. Hadlock net 85, or we can use modified a modified Hadlock model that uses the same parameters, but modified for our own population.

We can have a formula which is arm volume alone, thigh volume alone. We can say thigh volume, AC and thigh volume, and arm volume AC. But the point is that if you look at the systematic random error of these estimations of term babies, you can see that the addition the best results came from the combination of thigh volume and AC were the random error or precision was improved from the Hadlock models for 9 to 10%, down to 6%.

And that the proportion of actual birth weights within five or 10% were actually much higher when you included a soft tissue parameter.

Also, there are some instances where you have a ventral wall defect where it's very difficult to get the abdominal circumference. So in that case, fractional limb volume alone may be helpful for helping to estimate fetal weight.

So our long term study objective was to develop new fetal weight prediction models based on conventional biometry with soft tissue parameters throughout mid to late pregnancy. And we studied 271 pregnancies that were scanned using 3D ultrasound within four days of delivery.

And we took the standard biometry and also the FAV FTV and used the GE Voluson platform and software. And these are the original Hadlock formulas using looking at AC femur length and or BPD AC and femur length.

These are the modified formulas based on our own population that are analogous to the original Hadlock models. And one of the models that are currently being validated that seems to be to so far had the best accuracy and precision is the one that's been transformed. Log of birth weight equals some constant. And then using the variables of natural log of biparietal diameter by biparietal diameter square abdominal circumference and thigh volume fraction of thigh volume squared.

And this information has been submitted to one of the major journals since currently undergoing peer review. Almost finished, and it should be published pretty soon.

So we looked at systematic random errors, systematic error or accuracy being defined as predicted birth weight minus actual birth weight over actual birth weight times a hundred, and that the standard deviation of the mean percent differences representing the precision or random error of these estimations.

And just to briefly summarize, if you take the original Hadlock using a model using BPD AC and femur length and compare it to the modified Hadlock, you can see that you can improve the random error from 8.5% to 7.6% in our population in Michigan using the Houston equation, we had a 7% overestimation on the average that went down to 0.3.

And by just simply substituting FTV the fraction thigh volume for femur length, we're able to bring down the random error again to 6.6%. And if you take the look at the proportions of birth weights between five and 10%, you can see that the proportion of babies that are accurately classified increases.

We're currently conducting a prospective validation study in conjunction with an National Institutes of Health. And we hope to have this information in the near future.

Fractional Limb Volume for Individualized Growth Assessment

So the third area that fractional limb volume where it can help is as a growth parameter. And the traditional approach has been used is for instance, if you take a humerus length on the Y axis and the length gestation on the x axis, we have a series of a range of normal values, and we plot an individual compared against a population based standard.

Now, we know from the previous slides I showed you that depends on which population you use as to how you interpret the dot or individual against the population based standard. And that is the traditional approach.

There's a more mathematical way of doing this called individualized growth assessment or IGA that has been popularized by Dr. Russell, Dieter, Baylor, and Eva Rovic. This is the Rovic model that relates any parameter P to time administer. T is gestational age of the pregnancy, a coefficient C and the coefficient K and SC is essentially derived from the slope of second trimester growth, assuming that the growth is normal for any type of 2D or 3D parameter during the second trimester of pregnancy.

And K depends on what kind of parameter it is. If it's 1D, it's one, it's 2D, it's two, it's if it's 3D, it's three ideally. And then S is some sort of control mechanism that is still unclear about what controls it, but it helps to slow down the baby.

And essentially what we have here on the right is an example of a fetal femur length and our thigh circumference. And we have these dots here. We have actual data from the fetus before 26 weeks that helps establish C. And then from that information, we use the function to calculate the expected range of statistical range of normal for that given fetus, given the growth velocity during the second trimester using each fetus as its own control.

So in this case, we don't have to keep wondering about trying to scan a patient from India or southeast Asia, or someone in plotting against an inappropriate curve.

So this is this concept of IGA still warrants a further investigation, but this is what we've tried to do is to apply the fractional limb volume to this mathematical formula using the fraction thigh volume and arm volume as seen here.

So this is a noisy slide that essentially says that the Rovic functions fit all the parameter trajectories well, and we got some pretty good results from this approach.

This is just one example of a fetus that was growth restricted at term. And you can see that three graphs here, one showing fractional thigh volume, one on the right showing thigh circumference and estimated fetal weight.

Typically speaking, we would follow this baby using the estimated fetal weight graph, and you can see that that'd be in her normal range. But fractional thigh volume actually showed that there was evidence of growth restriction even as early as roughly 28, maybe 30, 33 weeks here. And the thigh circumference, the traditional 2D parameter also showed no problem.

So you know, this goes back to the concept of what really constitutes growth and the fact that growth is not really simple. When you talk about fetal growth, are you talking about the weight? Are you talking about skeletal growth? Are you talking about soft tissue development? It's a very complex topic.

So individualized growth assessment allows accurate prediction of third trimester sonographic parameters where each fetus acts as its own control, and it may provide earlier detection and improved monitoring of fetal growth abnormalities.

Fetal Origins of Adult Disease: The Barker Hypothesis

Just a quick word about the fetal origins of adult disease, the Barker hypothesis. We know that the development of the embryo depends upon a complex interaction between the environment and genetics.

And Dr. Barker is a medical epidemiologist from the United Kingdom, who pioneered some work. He has a famous book out Mothers, babies and Health in later life, where he developed a concept of fetal programming affecting disease in later adult life.

And the basic idea is that if the fetus is exposed to a hostile intrauterine environment due to chronic placental insufficiency, that there are some adverse environmental factors such as hypoxemia, ischemia, acidosis, and depending upon how that fetus responds to those environmental insults depends on how the baby turns out being small for gestational age, low birth weight, normal or macrosomic.

But the Barker hypothesis, also known as the thrifty phenotype theory, predicts that the one that the small one on the left will be at a higher risk for developing diabetes, heart disease, hypertension, and neurologic problems, and perhaps decreased longevity as a result of these intrauterine insults.

There's also some growing body evidence from many disciplines. But this one talks about the fetal programming of body composition and musculoskeletal development. The fact that there are changes in fat distribution, reduced muscle mass and strength, and low bone mineral content that can be caused by altered stem cell function, cell numbers and resetting of regulatory hormonal axes.

Dr. Beattie from the UK said in the mid 1990s about the assessment of newborn infants. Only when we apply proper measures of neonatal nutritional status rather than simplistic categorization based on birth weight alone, can we hope to determine the most useful methods of identifying and assessing disorders of intrauterine growth and the prediction of subsequent related morbidity and mortality.

Future Research Directions

So future research directions will emphasize the mechanisms underlying the fetal programming of fat, muscle and bone mass and the animal models. Then this needs to be confirmed in human studies. We need to develop effective interventions to optimize early growth and prenatal nutrition as a priority.

And we need long-term followup studies of cohorts with well characterized mothers and their offspring to confirm long-term benefits of early intervention.

Conclusion

I'd like to end with a quotation from Dr. Simpson. There are tiny puny infants with great vitality they never seen. The rest these infants will live for, although their weight is inferior, their sojourn in the womb is longer.

Thank you.