July 19, 2025 • 16 mins
🎙️ Episode 80: Genomic Forecasting of Postoperative Outcomes in Congenital Cardiac Surgery
🧬 In this episode of PaperCast Base by Base, we explore the use of artificial intelligence–driven genome sequencing to predict clinical outcomes following surgery for congenital heart defects.
🔍 Study Highlights:
The authors conducted a prospective observational cohort study of 2,253 patients from the Pediatric Cardiac Genomics Consortium, integrating whole exome sequencing with pre- and post-operative clinical variables to identify damaging genotypes in chromatin-modifying and cilia-related genes.
They applied an AI-based interpretation tool to pinpoint pathogenic variants and employed Bayesian networks to model conditional dependencies between genetic, phenotypic, and surgical complexity variables.
Damaging genotypes in specific pathways were found to increase risks of mortality, cardiac arrest, and prolonged mechanical ventilation, particularly in high-risk surgical categories and certain CHD phenotypes.
Conversely, the absence of these genotypes corresponded to reduced probabilities of adverse post-operative events.
🧠 Conclusion:
These findings underscore the potential of rapid genomic screening to inform preoperative risk stratification and guide targeted interventions to improve outcomes in congenital heart disease.
📖 Reference:
Watkins WS, Hernandez EJ, Miller TA, et al. Genome sequencing is critical for forecasting outcomes following congenital cardiac surgery. Nat Commun. 2025;16:6365. doi:10.1038/s41467-025-61625-0
📜 License:
This episode is based on an open-access article published under the Creative Commons Attribution 4.0 International License (CC BY 4.0) – https://creativecommons.org/licenses/by/4.0/
On PaperCast Base by Base you’ll discover the latest in genomics, functional genomics, structural genomics, and proteomics.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:14):
Welcome to base by base, the AERcast that brings genomics to you
wherever you are. OK, picture this.
A newborn, just tiny, having come through this incredibly
complex life saving heart surgery.
Their recovery is obviously critical.
But what if, right then we couldpredict, like with really
startling accuracy, there's specific risk for serious
(00:36):
problems, things like cardiac arrest or needing a ventilator
for a long time. What if looking into their own
genetic code could give doctors a kind of personalized forecast,
letting them step in before anything bad happens?
It almost sounds like science fiction, doesn't it?
And it's not anymore. It's it's really becoming a
reality. Yeah, exactly.
And this isn't just about, you know, understanding the disease
in general. It's really about looking into
(00:58):
that individual's unique geneticblueprint to predict their
specific path through it all, especially after a huge event
like major heart surgery. It's genuinely about giving
clinicians that foresight, letting them tailor care before
there's even a hint of trouble. So today we're diving into some
really groundbreaking work from the Pediatric Cardiac Genomics
Consortium. This has been a massive
collaborative effort and it's seriously pushed forward our
(01:20):
understanding of congenital heart defects and maybe more
importantly, how they affect patient outcomes.
Their latest work is quite something.
OK, so let's set the scene a bit.
Why is this research so crucial?We're talking about congenital
heart defects, CHD. These are a huge global health
issue. In the US alone, over 40,000
babies are born with them each year.
That's about one in every 100 births.
(01:42):
And these conditions are often life threatening, ranging from,
well, relatively mild to incredibly complex, needing
really intricate surgery. That's right.
And the genetics behind CHD, that's been a huge focus for
researchers for a long time. We've learned quite a bit, like
previous big sequencing studies showed that the genetic causes
can be pretty different between syndromic cases, where the heart
(02:03):
defect is part of a bigger syndrome, and sporadic ones
where it's isolated. The PCGC, the consortium we
mentioned, they specifically highlighted that dominant forms
of CHD often link back to damaging variants in genes that
modify chromatin. That's basically how DNA gets
packaged, whereas reassiforms often involve damaging variants
in cilia genes. Cilia are these tiny hair like
(02:25):
bits involved in cell signaling.One in 100 newborns, that
statistic is just staggering. But even with all the medical
progress predicting how a specific child will do after
complex surgery, that's been really, really tough, hasn't it?
Especially with so many different types of CHD.
What are the main hurdles? Why has it been so hard?
Well, it's complicated on multiple levels.
First, just the sheer range of severity in the heart defects
(02:49):
themselves, that's a huge variable.
Then you've got these incrediblycomplex surgeries needed for
survival, each adding its own set of risks and factors.
And on top of that, there's thishigh degree of genetic
heterogeneity, meaning lots and lots of different genes or
different variations in those genes can lead to similar heart
conditions. So trying to pinpoint specific
(03:10):
genetic risk factors amidst all that noise, it's been.
Yeah. Like finding needles in
countless haystacks. That makes sense.
So faced with all that complexity, how did this
research team actually manage tomake progress?
It sounds like they had a reallyclever sort of two prong
strategy, yeah. How do they start untangling
things, first on the clinical side and then with the genetics?
They did. It was quite innovative.
(03:31):
First the clinical picture. They needed to simplify how they
classify the different CHD typesinstead of getting bogged down
in the thousands of really specific descriptions from the
filer system, which is super detailed but maybe too detailed
for this kind of large scale analysis.
Like a giant map with too many tiny roads.
They condense them. They used a machine learning
model called XG Boost to automatically group all those
(03:53):
descriptions into just 5 clear clinically relevant categories
based on anatomy. So you had left ventricular
outflow tract obstructions, LVO basically blockages,
heterotaxilaterality defects, HTX organs in the wrong place,
atrioventricular canal defects, ABC problems in the heart
center, connotruncal defects, CTD issues with the big arteries
(04:14):
leaving the heart, and then justa category for other defects.
OT. You can kind of think of XG
Boost as a smart digital map maker grouping similar areas
together. OK.
Streamlining the clinical side into 5 manageable groups using
machine learning, That makes a lot of sense.
Then the genetics, which I imagine is the even bigger, more
complex data set. How did they tackle that?
Yeah, that's where AI really stepped up.
(04:34):
They used an AI tool for interpreting genomes called
GMGEM. This tool was absolutely vital
because it helped cut through that problem of hyalilic and
locus heterogeneity. We talked about, you know, many
different genetic variants in many different places causing
issues. GEM assigned what they called a
damaging genotype score. A score of 1.0 was the highest,
meaning really high confidence that a specific genetic variant
(04:57):
is actually harmful. And for this study, they set a
pretty strict threshold, only looking at variance with a score
of one point O or higher to be really sure about the potential
impact. Got it.
High confidence genetic varianceidentified by AI.
So now they have these clearer clinical groups and these high
confidence genetic insights. How do they actually link those
two complex pieces together to predict patient outcomes?
(05:19):
That seems like the magic step. Right, connecting the dots.
For that, they used probabilistic graphical models,
specifically Bayesian networks. Now, these aren't just your
standard statistical models. Think of them more like
incredibly sophisticated detective boards.
They can map out really complex webs of cause and effect, even
relationships that aren't director obvious.
(05:40):
And what's really crucial, especially when you're dealing
with rare outcomes where you might not have huge numbers of
patients, is that these models can tell you how certain they
are about their predictions. They give you uncertainty
estimates. That's a huge advantage.
Wow. OK, Bayesian networks to model
all that complexity and uncertainty.
That's impressive. And who are the patients in the
study? Where did all this incredibly
(06:01):
detailed data come from? The data came from the Pediatric
Cardiac Genomics Consortium itself.
They had a prospective observational cohort of 2200 and
53 CHD patients. And this isn't just any data
set. It's one of the world's largest
collections that integrates genetic data, detailed phenotype
information, and clinical outcomes for these kids.
So they had the exome sequencing, the detailed CHD
(06:23):
classifications we talked about,and crucially, data on how these
patients did after surgery. OK, so with this powerful
approach and this amazing data set, what did they actually
find? What's the big news about
genomics predicting surgical outcomes?
What really jumps out? What they found is essentially A
genetic risk map. Imagine knowing pretty clearly
that certain genetic red flags, specifically these damaging
(06:45):
variants in chromatin modifying genes, don't just increase the
chance of having a defect like LVO, but also mean a 1.6 fold
higher likelihood of bad outcomes after surgery.
It's more than just an association, it's like a
powerful early warning. And within that LVO group, for
kids with hypoplastic left heartsyndrome specifically, that
enrichment for damaging chromatin variants was even
higher, 1.9 fold. It confirms older ideas, but
(07:09):
adds much more specific detail for certain subtypes.
So chromatin genes flagged risksfor LVO outcomes?
What about those cilia related genes you mentioned for the
recessive forms? How do they connect to outcomes?
Yeah. For patients with heterotaxy,
the HTX group, they found a really striking 2.6 fold
enrichment for damaging recessive or bilelic cilia
related genes. And drilling down even further
(07:30):
within that group, specific mutations and something called
the FO XJ1 pathway showed a massive 6.9 fold enrichment.
So again, it reinforces what we kind of knew, but adds critical
detail about how genetics links to outcomes in specific patient
groups. And the impact wasn't subtle,
right? These damaging genotypes really
ramped up the risk of serious problems after surgery.
Absolutely, they significantly increase the probability.
(07:53):
For example, having one of thosedamaging de Novo chromatin
variants, meaning it's new in the child not inherited, meant a
1.8 fold higher chance of mortality.
That's huge. Also a 1.7 fold increase in the
risk of cardiac arrest and a 1.6fold jump in needing prolonged
mechanical ventilation. Similarly, those damaging
(08:14):
recessive or bilelic cilia genotypes, they increase
mortality risk by 1.4 fold, cardiac arrest by 1.5 fold and
eating that long term ventilation by 1.4 fold.
These are really substantial increases in risk.
We'll wait until you hear this part.
What I found equally fascinatingwas that the absence of these
damaging genotypes was also really informative.
It actually lowered the risk. That's a key finding too, isn't
(08:35):
it? It absolutely is.
It's the other side of the coin.For instance, if a patient
didn't have a damaging de Novo chromatin genotype, their
relative risk for mortality dropped significantly down to
.55 five, and if they didn't have a damaging recessive or
bilelic cilia genotype, the mortality risk was 0.72.
So knowing what genetic risks aren't present is also
(08:57):
incredibly powerful. It can offer reassurance, maybe
guide less aggressive monitoringor intervention where
appropriate. And these genetic impacts
weren't uniform across the board, right?
They're amplified in certain situations.
Tell us about that. Yes, context mattered a lot.
Take the highest risk surgeries,the really complex ones
classified as Stat 4 or Stat 5. Among patients undergoing those
procedures, those who unfortunately died were 1.8
(09:19):
times more likely to carry a damaging chromatin variant and
1.7 times more likely to have a damaging recessive cilia
genotype. And looking at specific defect
types among LVO patients, havingthat damaging de Novo chromatin
variant boosted the mortality likelihood 2.3 fold.
Or for HDX patients who had a cardiac arrest, having a
damaging recessive cilia genotype made that event 3.3
(09:41):
times more likely. It gives us really granular risk
insight. What about?
Kids who had other issues besides the heart defect, things
like extracardiac anomalies or ECAS.
How did genetics factor in there?
Right Ecas themselves, as you might expect, had a big impact.
They increase the probability ofmortality 2.8 fold on their own
and prolonged ventilation one point sevenfold.
(10:02):
But the study found a crucial interaction with genetics
damaging. De Novo chromatin genotypes made
it 2.5 times more likely for a patient to have both an ECA and
unfortunately die after surgery,and 2.4 times more likely to
have both an ECA and need prolonged ventilation.
It really shows this kind of synergistic risk when you have
both factors present. A double whammy, essentially.
(10:24):
And what about the cilia genotypes in patients with ECAS?
Similar story there, a damaging recessive or bilelic cilia
genotype increased mortality risk 1.5 fold in patients who
also had an ECA. But here's a really striking
one. For HTX patients, the ones with
laterality defects who also had an ECA, having a damaging cilia
genotype increased their probability of needing prolonged
ventilation by a huge 4 point O fold.
(10:47):
It really underscores how combining genetic information
with other clinical factors gives you a much sharper picture
of the overall risk. It's just incredible the level
of detail they can pull out now.It really does feel like giving
doctors a kind of genetic crystal ball.
So stepping back, what's the bigpicture here?
What does this all mean for patient care going forward?
It feels like this really cements the role of genome
(11:09):
sequencing in the clinic. I think it does.
This study being the largest yetto really connect specific
genotypes to actual surgical outcomes in CHD.
It provides pretty definitive evidence.
It shows that genomic data is truly crucial for predicting
severe post op problems, especially when you're looking
at higher risk surgeries, specific types of CHD, or
(11:29):
patients who also have extracardiac anomalies.
We're moving beyond just diagnosing the defect itself to
actually forecasting that individual patients likely
journey through recovery. And the practical application,
how does this change what doctors can actually do for
these kids? It's more than just knowing the
risk, right? Absolutely.
The power lies in being able to act on that knowledge.
(11:50):
Quantifying the risk means doctors can potentially use
specific therapies earlier, maybe even before the surgery,
to try and head off those predicted complications.
So for example, if pre op sequencing flags a damaging
cilia genotype in a patient, particularly one with
heterotaxy, the clinical team immediately knows OK, this child
(12:10):
is at much higher risk for breathing problems after
surgery. That knowledge could trigger
immediate more aggressive airwayclearance, maybe using specific
nucleic drugs, adjusting ventilation strategies, even
avoiding certain anesthetics that are known to mess with
ciliary function. It allows that shift from just
reacting to problems to proactively trying to prevent
them based on the child's own genetics.
(12:32):
So it's genuinely moving towardsprevention and truly
personalized care planned even before the operation starts.
That feels like a massive step. It really is.
And think about it, as the cost of whole genome sequencing keeps
falling and the turn around times get faster, this kind of
detailed genomic information will become routinely available
before surgery. That enables true precision risk
(12:52):
stratification. It allows for interventions
tailored to the individual childthat could just dramatically
improve their outcomes. This kind of research is paving
the way for whole genome sequencing to become a standard
part of care for Critically I'llnewborns, offering this
unprecedented predictive insight.
It sounds incredibly promising, but you know every major study
has its limitations or things toconsider.
(13:14):
What did the authors acknowledgein terms of limitations for this
work? That's always important to
consider. Yeah, they were upfront about a
few things. First, the PCGC cohort, while
huge, wasn't technically an inception cohort.
That means it might slightly underrepresent very early death
that happened before enrollment.So their complication rates
might be conservative kind of lower bounds.
(13:35):
They also noted they didn't haveenough statistical power to
really dig into the genomic impact on less severe types of
CHD. The focus was understandably on
the higher risk cases. And, you know, while these large
clinical registries are amazing resources, there can always be
some variability in data quality, even with careful
checks. Finally, actually replicating
this exact study elsewhere is tough right now, simply because
(13:58):
the PCGC data set is so uniquelydeep and comprehensive.
However, they do point out that their Bayesian modeling approach
is actually pretty robust even when data is limited.
And crucially, their main findings align really well with
what's already known from biology and clinical experience.
That adds a lot of confidence. OK, so this deep dive really
shows us that bringing together genome sequencing and AI gives
(14:19):
us this incredible new power to forecast outcomes after these
major heart surgeries for kids. It feels like we're truly moving
into an era where personalized genetic insights can actively
shape clinical decisions and treatments.
Absolutely connecting it all back.
Knowing a patient's specific genetic risks allows for that
proactive tailored care. It could make a huge difference
(14:42):
in their post operative journey,whether that's by identifying
those heightened risks so the team can prepare, or maybe
confirming a lower risk profile,which can be immensely
reassuring for families and might guide less intensive
follow up. Looking forward then, what do
you see as the single biggest way this could transform
precision medicine and pediatriccardiac care?
How could this really change lives?
For me, the most transformative aspect is that potential shift
(15:04):
from a reactive wait and see approach to a proactive know and
prepare strategy for every single child facing these
surgeries. It means optimizing everything
from how they prepare for surgery to how closely they're
watched afterwards, potentially stopping complications before
they even start, or at least lessening their impact.
This level of personalized care truly based on a child's unique
(15:27):
genetic makeup, it really has a potential to redefine outcomes
for these vulnerable kids and lead to many more healthy
futures. This episode was based on an
Open Access article under the CCBY 4.0 license.
You can find a direct link to the paper and the license in our
episode description. If you enjoy this analysis, the
best way to support Base by Baseis to subscribe or follow in
(15:48):
your favorite podcast app and leave us a five star rating.
It only takes a few seconds but makes a huge difference in
helping others discover the show.
Thanks for listening and join usnext time as we explore more
science base by Base.
Welcome to base by base, the AERcast that brings genomics to you
wherever you are. OK, picture this.
A newborn, just tiny, having come through this incredibly
complex life saving heart surgery.
Their recovery is obviously critical.
But what if, right then we couldpredict, like with really
startling accuracy, there's specific risk for serious
(00:36):
problems, things like cardiac arrest or needing a ventilator
for a long time. What if looking into their own
genetic code could give doctors a kind of personalized forecast,
letting them step in before anything bad happens?
It almost sounds like science fiction, doesn't it?
And it's not anymore. It's it's really becoming a
reality. Yeah, exactly.
And this isn't just about, you know, understanding the disease
in general. It's really about looking into
(00:58):
that individual's unique geneticblueprint to predict their
specific path through it all, especially after a huge event
like major heart surgery. It's genuinely about giving
clinicians that foresight, letting them tailor care before
there's even a hint of trouble. So today we're diving into some
really groundbreaking work from the Pediatric Cardiac Genomics
Consortium. This has been a massive
collaborative effort and it's seriously pushed forward our
(01:20):
understanding of congenital heart defects and maybe more
importantly, how they affect patient outcomes.
Their latest work is quite something.
OK, so let's set the scene a bit.
Why is this research so crucial?We're talking about congenital
heart defects, CHD. These are a huge global health
issue. In the US alone, over 40,000
babies are born with them each year.
That's about one in every 100 births.
(01:42):
And these conditions are often life threatening, ranging from,
well, relatively mild to incredibly complex, needing
really intricate surgery. That's right.
And the genetics behind CHD, that's been a huge focus for
researchers for a long time. We've learned quite a bit, like
previous big sequencing studies showed that the genetic causes
can be pretty different between syndromic cases, where the heart
(02:03):
defect is part of a bigger syndrome, and sporadic ones
where it's isolated. The PCGC, the consortium we
mentioned, they specifically highlighted that dominant forms
of CHD often link back to damaging variants in genes that
modify chromatin. That's basically how DNA gets
packaged, whereas reassiforms often involve damaging variants
in cilia genes. Cilia are these tiny hair like
(02:25):
bits involved in cell signaling.One in 100 newborns, that
statistic is just staggering. But even with all the medical
progress predicting how a specific child will do after
complex surgery, that's been really, really tough, hasn't it?
Especially with so many different types of CHD.
What are the main hurdles? Why has it been so hard?
Well, it's complicated on multiple levels.
First, just the sheer range of severity in the heart defects
(02:49):
themselves, that's a huge variable.
Then you've got these incrediblycomplex surgeries needed for
survival, each adding its own set of risks and factors.
And on top of that, there's thishigh degree of genetic
heterogeneity, meaning lots and lots of different genes or
different variations in those genes can lead to similar heart
conditions. So trying to pinpoint specific
(03:10):
genetic risk factors amidst all that noise, it's been.
Yeah. Like finding needles in
countless haystacks. That makes sense.
So faced with all that complexity, how did this
research team actually manage tomake progress?
It sounds like they had a reallyclever sort of two prong
strategy, yeah. How do they start untangling
things, first on the clinical side and then with the genetics?
They did. It was quite innovative.
(03:31):
First the clinical picture. They needed to simplify how they
classify the different CHD typesinstead of getting bogged down
in the thousands of really specific descriptions from the
filer system, which is super detailed but maybe too detailed
for this kind of large scale analysis.
Like a giant map with too many tiny roads.
They condense them. They used a machine learning
model called XG Boost to automatically group all those
(03:53):
descriptions into just 5 clear clinically relevant categories
based on anatomy. So you had left ventricular
outflow tract obstructions, LVO basically blockages,
heterotaxilaterality defects, HTX organs in the wrong place,
atrioventricular canal defects, ABC problems in the heart
center, connotruncal defects, CTD issues with the big arteries
(04:14):
leaving the heart, and then justa category for other defects.
OT. You can kind of think of XG
Boost as a smart digital map maker grouping similar areas
together. OK.
Streamlining the clinical side into 5 manageable groups using
machine learning, That makes a lot of sense.
Then the genetics, which I imagine is the even bigger, more
complex data set. How did they tackle that?
Yeah, that's where AI really stepped up.
(04:34):
They used an AI tool for interpreting genomes called
GMGEM. This tool was absolutely vital
because it helped cut through that problem of hyalilic and
locus heterogeneity. We talked about, you know, many
different genetic variants in many different places causing
issues. GEM assigned what they called a
damaging genotype score. A score of 1.0 was the highest,
meaning really high confidence that a specific genetic variant
(04:57):
is actually harmful. And for this study, they set a
pretty strict threshold, only looking at variance with a score
of one point O or higher to be really sure about the potential
impact. Got it.
High confidence genetic varianceidentified by AI.
So now they have these clearer clinical groups and these high
confidence genetic insights. How do they actually link those
two complex pieces together to predict patient outcomes?
(05:19):
That seems like the magic step. Right, connecting the dots.
For that, they used probabilistic graphical models,
specifically Bayesian networks. Now, these aren't just your
standard statistical models. Think of them more like
incredibly sophisticated detective boards.
They can map out really complex webs of cause and effect, even
relationships that aren't director obvious.
(05:40):
And what's really crucial, especially when you're dealing
with rare outcomes where you might not have huge numbers of
patients, is that these models can tell you how certain they
are about their predictions. They give you uncertainty
estimates. That's a huge advantage.
Wow. OK, Bayesian networks to model
all that complexity and uncertainty.
That's impressive. And who are the patients in the
study? Where did all this incredibly
(06:01):
detailed data come from? The data came from the Pediatric
Cardiac Genomics Consortium itself.
They had a prospective observational cohort of 2200 and
53 CHD patients. And this isn't just any data
set. It's one of the world's largest
collections that integrates genetic data, detailed phenotype
information, and clinical outcomes for these kids.
So they had the exome sequencing, the detailed CHD
(06:23):
classifications we talked about,and crucially, data on how these
patients did after surgery. OK, so with this powerful
approach and this amazing data set, what did they actually
find? What's the big news about
genomics predicting surgical outcomes?
What really jumps out? What they found is essentially A
genetic risk map. Imagine knowing pretty clearly
that certain genetic red flags, specifically these damaging
(06:45):
variants in chromatin modifying genes, don't just increase the
chance of having a defect like LVO, but also mean a 1.6 fold
higher likelihood of bad outcomes after surgery.
It's more than just an association, it's like a
powerful early warning. And within that LVO group, for
kids with hypoplastic left heartsyndrome specifically, that
enrichment for damaging chromatin variants was even
higher, 1.9 fold. It confirms older ideas, but
(07:09):
adds much more specific detail for certain subtypes.
So chromatin genes flagged risksfor LVO outcomes?
What about those cilia related genes you mentioned for the
recessive forms? How do they connect to outcomes?
Yeah. For patients with heterotaxy,
the HTX group, they found a really striking 2.6 fold
enrichment for damaging recessive or bilelic cilia
related genes. And drilling down even further
(07:30):
within that group, specific mutations and something called
the FO XJ1 pathway showed a massive 6.9 fold enrichment.
So again, it reinforces what we kind of knew, but adds critical
detail about how genetics links to outcomes in specific patient
groups. And the impact wasn't subtle,
right? These damaging genotypes really
ramped up the risk of serious problems after surgery.
Absolutely, they significantly increase the probability.
(07:53):
For example, having one of thosedamaging de Novo chromatin
variants, meaning it's new in the child not inherited, meant a
1.8 fold higher chance of mortality.
That's huge. Also a 1.7 fold increase in the
risk of cardiac arrest and a 1.6fold jump in needing prolonged
mechanical ventilation. Similarly, those damaging
(08:14):
recessive or bilelic cilia genotypes, they increase
mortality risk by 1.4 fold, cardiac arrest by 1.5 fold and
eating that long term ventilation by 1.4 fold.
These are really substantial increases in risk.
We'll wait until you hear this part.
What I found equally fascinatingwas that the absence of these
damaging genotypes was also really informative.
It actually lowered the risk. That's a key finding too, isn't
(08:35):
it? It absolutely is.
It's the other side of the coin.For instance, if a patient
didn't have a damaging de Novo chromatin genotype, their
relative risk for mortality dropped significantly down to
.55 five, and if they didn't have a damaging recessive or
bilelic cilia genotype, the mortality risk was 0.72.
So knowing what genetic risks aren't present is also
(08:57):
incredibly powerful. It can offer reassurance, maybe
guide less aggressive monitoringor intervention where
appropriate. And these genetic impacts
weren't uniform across the board, right?
They're amplified in certain situations.
Tell us about that. Yes, context mattered a lot.
Take the highest risk surgeries,the really complex ones
classified as Stat 4 or Stat 5. Among patients undergoing those
procedures, those who unfortunately died were 1.8
(09:19):
times more likely to carry a damaging chromatin variant and
1.7 times more likely to have a damaging recessive cilia
genotype. And looking at specific defect
types among LVO patients, havingthat damaging de Novo chromatin
variant boosted the mortality likelihood 2.3 fold.
Or for HDX patients who had a cardiac arrest, having a
damaging recessive cilia genotype made that event 3.3
(09:41):
times more likely. It gives us really granular risk
insight. What about?
Kids who had other issues besides the heart defect, things
like extracardiac anomalies or ECAS.
How did genetics factor in there?
Right Ecas themselves, as you might expect, had a big impact.
They increase the probability ofmortality 2.8 fold on their own
and prolonged ventilation one point sevenfold.
(10:02):
But the study found a crucial interaction with genetics
damaging. De Novo chromatin genotypes made
it 2.5 times more likely for a patient to have both an ECA and
unfortunately die after surgery,and 2.4 times more likely to
have both an ECA and need prolonged ventilation.
It really shows this kind of synergistic risk when you have
both factors present. A double whammy, essentially.
(10:24):
And what about the cilia genotypes in patients with ECAS?
Similar story there, a damaging recessive or bilelic cilia
genotype increased mortality risk 1.5 fold in patients who
also had an ECA. But here's a really striking
one. For HTX patients, the ones with
laterality defects who also had an ECA, having a damaging cilia
genotype increased their probability of needing prolonged
ventilation by a huge 4 point O fold.
(10:47):
It really underscores how combining genetic information
with other clinical factors gives you a much sharper picture
of the overall risk. It's just incredible the level
of detail they can pull out now.It really does feel like giving
doctors a kind of genetic crystal ball.
So stepping back, what's the bigpicture here?
What does this all mean for patient care going forward?
It feels like this really cements the role of genome
(11:09):
sequencing in the clinic. I think it does.
This study being the largest yetto really connect specific
genotypes to actual surgical outcomes in CHD.
It provides pretty definitive evidence.
It shows that genomic data is truly crucial for predicting
severe post op problems, especially when you're looking
at higher risk surgeries, specific types of CHD, or
(11:29):
patients who also have extracardiac anomalies.
We're moving beyond just diagnosing the defect itself to
actually forecasting that individual patients likely
journey through recovery. And the practical application,
how does this change what doctors can actually do for
these kids? It's more than just knowing the
risk, right? Absolutely.
The power lies in being able to act on that knowledge.
(11:50):
Quantifying the risk means doctors can potentially use
specific therapies earlier, maybe even before the surgery,
to try and head off those predicted complications.
So for example, if pre op sequencing flags a damaging
cilia genotype in a patient, particularly one with
heterotaxy, the clinical team immediately knows OK, this child
(12:10):
is at much higher risk for breathing problems after
surgery. That knowledge could trigger
immediate more aggressive airwayclearance, maybe using specific
nucleic drugs, adjusting ventilation strategies, even
avoiding certain anesthetics that are known to mess with
ciliary function. It allows that shift from just
reacting to problems to proactively trying to prevent
them based on the child's own genetics.
(12:32):
So it's genuinely moving towardsprevention and truly
personalized care planned even before the operation starts.
That feels like a massive step. It really is.
And think about it, as the cost of whole genome sequencing keeps
falling and the turn around times get faster, this kind of
detailed genomic information will become routinely available
before surgery. That enables true precision risk
(12:52):
stratification. It allows for interventions
tailored to the individual childthat could just dramatically
improve their outcomes. This kind of research is paving
the way for whole genome sequencing to become a standard
part of care for Critically I'llnewborns, offering this
unprecedented predictive insight.
It sounds incredibly promising, but you know every major study
has its limitations or things toconsider.
(13:14):
What did the authors acknowledgein terms of limitations for this
work? That's always important to
consider. Yeah, they were upfront about a
few things. First, the PCGC cohort, while
huge, wasn't technically an inception cohort.
That means it might slightly underrepresent very early death
that happened before enrollment.So their complication rates
might be conservative kind of lower bounds.
(13:35):
They also noted they didn't haveenough statistical power to
really dig into the genomic impact on less severe types of
CHD. The focus was understandably on
the higher risk cases. And, you know, while these large
clinical registries are amazing resources, there can always be
some variability in data quality, even with careful
checks. Finally, actually replicating
this exact study elsewhere is tough right now, simply because
(13:58):
the PCGC data set is so uniquelydeep and comprehensive.
However, they do point out that their Bayesian modeling approach
is actually pretty robust even when data is limited.
And crucially, their main findings align really well with
what's already known from biology and clinical experience.
That adds a lot of confidence. OK, so this deep dive really
shows us that bringing together genome sequencing and AI gives
(14:19):
us this incredible new power to forecast outcomes after these
major heart surgeries for kids. It feels like we're truly moving
into an era where personalized genetic insights can actively
shape clinical decisions and treatments.
Absolutely connecting it all back.
Knowing a patient's specific genetic risks allows for that
proactive tailored care. It could make a huge difference
(14:42):
in their post operative journey,whether that's by identifying
those heightened risks so the team can prepare, or maybe
confirming a lower risk profile,which can be immensely
reassuring for families and might guide less intensive
follow up. Looking forward then, what do
you see as the single biggest way this could transform
precision medicine and pediatriccardiac care?
How could this really change lives?
For me, the most transformative aspect is that potential shift
(15:04):
from a reactive wait and see approach to a proactive know and
prepare strategy for every single child facing these
surgeries. It means optimizing everything
from how they prepare for surgery to how closely they're
watched afterwards, potentially stopping complications before
they even start, or at least lessening their impact.
This level of personalized care truly based on a child's unique
(15:27):
genetic makeup, it really has a potential to redefine outcomes
for these vulnerable kids and lead to many more healthy
futures. This episode was based on an
Open Access article under the CCBY 4.0 license.
You can find a direct link to the paper and the license in our
episode description. If you enjoy this analysis, the
best way to support Base by Baseis to subscribe or follow in
(15:48):
your favorite podcast app and leave us a five star rating.
It only takes a few seconds but makes a huge difference in
helping others discover the show.
Thanks for listening and join usnext time as we explore more
science base by Base.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Special Summer Offer: Exclusively on Apple Podcasts, try our Dateline Premium subscription completely free for one month! With Dateline Premium, you get every episode ad-free plus exclusive bonus content.