What your breath could reveal about your health

I have a tendency to assume the worst,

and once in a while,
this habit plays tricks on me.

For example, if I feel
unexpected pain in my body

that I have not experienced before
and that I cannot attribute,

then all of a sudden, my mind
might turn a tense back into heart disease

or calf muscle pain
into deep vein thrombosis.

But so far, I haven't been diagnosed
with any deadly or incurable disease.

Sometimes things just hurt
for no clear reason.

But not everyone is as lucky as me.

Every year, more than
50 million people die worldwide.

Especially in high-income
economies like ours,

a large fraction of deaths is caused
by slowly progressing diseases:

heart disease, chronic lung disease,
cancer, Alzheimer's, diabetes,

just to name a few.

Now, humanity has made tremendous progress
in diagnosing and treating many of these.

But we are at a stage
where further advancement in health

cannot be achieved only
by developing new treatments.

And this becomes evident
when we look at one aspect

that many of these
diseases have in common:

the probability for successful treatment

strongly depends on
when treatment is started.

But a disease is typically only detected
once symptoms occur.

The problem here is that, in fact,
many diseases can remain asymptomatic,

hence undetected,
for a long period of time.

Because of this, there is
a persisting need for new ways

of detecting disease at early stage,

way before any symptoms occur.

In health care, this is called screening.

And as defined by
the World Health Organization,

screening is "the presumptive
identification of unrecognized disease

in an apparently healthy person,

by means of tests ... that can be applied
rapidly and easily ..."

That's a long definition,
so let me repeat it:

identification of unrecognized disease

in an apparently healthy person

by means of tests that can be applied
both rapidly and easily.

And I want to put special emphasis
on the words "rapidly" and "easily"

because many of the existing
screening methods

are exactly the opposite.

And those of you
who have undergone colonoscopy

as part of a screening program
for colorectal cancer

will know what I mean.

Obviously, there's a variety
of medical tools available

to perform screening tests.

This ranges from imaging techniques
such as radiography

or magnetic resonance imaging

to the analysis of blood or tissue.

We have all had such tests.

But there's one medium
that for long has been overlooked:

a medium that is easily accessible,

basically nondepletable,

and it holds tremendous promise
for medical analysis.

And that is our breath.

Human breath is essentially
composed of five components:

nitrogen, oxygen, carbon dioxide,
water and argon.

But besides these five, there are
hundreds of other components

that are present in very low quantity.

These are called volatile
organic compounds,

and we release hundreds,
even thousands of them

every time we exhale.

The analysis of these volatile
organic compounds in our breath

is called breath analysis.

In fact, I believe that many of you
have already experienced breath analysis.

Imagine: you're driving home
late at night,

when suddenly, there's
a friendly police officer

who asks you kindly but firmly

to pull over and blow
into a device like this one.

This is an alcohol breath tester

that is used to measure
the ethanol concentration in your breath

and determine whether driving
in your condition is a clever idea.

Now, I'd say my driving was pretty good,

but let me check.

(Beep)

0.0, so nothing
to worry about, all fine.

(Laughter)

Now imagine a device like this one,

that does not only measure
alcohol levels in your breath,

but that detects diseases
like the ones I've shown you

and potentially many more.

The concept of correlating
the smell of a person's breath

with certain medical conditions,

in fact, dates back to Ancient Greece.

But only recently, research efforts
on breath analysis have skyrocketed,

and what once was a dream
is now becoming reality.

And let me pull up this list again
that I showed you earlier.

For the majority of diseases listed here,

there's substantial scientific evidence

suggesting that the disease
could be detected by breath analysis.

But how does it work, exactly?

The essential part is a sensor device

that detects the volatile
organic compounds in our breath.

Simply put: when exposed
to a breath sample,

the sensor outputs a complex signature

that results from the mixture of volatile
organic compounds that we exhale.

Now, this signature represents
a fingerprint of your metabolism,

your microbiome

and the biochemical processes
that occur in your body.

If you have a disease,

your organism will change,

and so will the composition
of your exhaled breath.

And then the only thing that is left to do
is to correlate a certain signature

with the presence or absence
of certain medical conditions.

The technology promises
several undeniable benefits.

Firstly, the sensor can be miniaturized

and integrated into small,
handheld devices

like this alcohol breath tester.

This would allow the test to be used
in many different settings

and even at home,

so that a visit at the doctor's office

is not needed each time
a test shall be performed.

Secondly, breath analysis is noninvasive

and can be as simple as blowing
into an alcohol breath tester.

Such simplicity and ease of use
would reduce patient burden

and provide an incentive
for broad adoption of the technology.

And thirdly, the technology is so flexible

that the same device could be used

to detect a broad range
of medical conditions.

Breath analysis could be used to screen
for multiple diseases at the same time.

Nowadays, each disease typically requires
a different medical tool

to perform a screening test.

But this means you can only find
what you're looking for.

With all of these features,
breath analysis is predestined

to deliver what many traditional
screening tests are lacking.

And most importantly,

all of these features should
eventually provide us

with a platform for medical analysis

that can operate at attractively
low cost per test.

On the contrary, existing medical tools

often lead to rather high cost per test.

Then, in order to keep costs down,

the number of tests
needs to be restricted,

and this means (a) that the tests
can only be performed

on a narrow part of the population,
for example, the high-risk population;

and (b) that the number of tests
per person needs to be kept at a minimum.

But wouldn't it actually be beneficial

if the test was performed
on a larger group of people,

and more often and over a longer period
of time for each individual?

Especially the latter would give access
to something very valuable

that is called longitudinal data.

Longitudinal data is a data set
that tracks the same patient

over the course of many months or years.

Nowadays, medical decisions
are often based on a limited data set,

where only a glimpse
of a patient's medical history

is available for decision-making.

In such a case,

abnormalities are typically detected

by comparing a patient's health profile

to the average health profile
of a reference population.

Longitudinal data would
open up a new dimension

and allow abnormalities to be detected

based on a patient's own medical history.

This will pave the way
for personalized treatment.

Sounds pretty great, right?

Now you will certainly have a question
that is something like,

"If the technology is as great as he says,
then why aren't we using it today?"

And the only answer I can give you is:

not everything is as easy as it sounds.

There are technical
challenges, for example.

There's the need for
extremely reliable sensors

that can detect mixtures
of volatile organic compounds

with sufficient reproducibility.

And another technical challenge is this:

How do you sample a person's breath
in a very defined manner

so that the sampling process itself

does not alter the result of the analysis?

And there's the need for data.

Breath analysis needs
to be validated in clinical trials,

and enough data needs to be collected

so that individual conditions
can be measured against baselines.

Breath analysis can only succeed

if a large enough data set
can be generated

and made available for broad use.

If breath analysis
holds up to its promises,

this is a technology
that could truly aid us

to transform our health care system --

transform it from a reactive system

where treatment is triggered
by symptoms of disease

to a proactive system,

where disease detection,
diagnosis and treatment

can happen at early stage,

way before any symptoms occur.

Now this brings me to my last point,
and it's a fundamental one.

What exactly is a disease?

Imagine that breath analysis
can be commercialized as I describe it,

and early detection becomes routine.

A problem that remains
is, in fact, a problem

that any screening activity has to face

because, for many diseases,

it is often impossible to predict
with sufficient certainty

whether the disease
would ever cause any symptoms

or put a person's life at risk.

This is called overdiagnosis,

and it leads to a dilemma.

If a disease is identified,

you could decide not to treat it

because there's a certain probability
that you would never suffer from it.

But how much would you suffer

just from knowing that you have
a potentially deadly disease?

And wouldn't you actually regret
that the disease was detected

in the first place?

Your second option
is to undergo early treatment

with the hope for curing it.

But often, this would not come
without side effects.

To be precise:

the bigger problem is not overdiagnosis,

it's overtreatment,

because not every disease
has to be treated immediately

just because a treatment is available.

The increasing adoption
of routine screening

will raise the question:

What do we call a disease
that can rationalize treatment,

and what is just an abnormality
that should not be a source of concern?

My hopes are that routine screening
using breath analysis

can provide enough data and insight

so that at some point,
we'll be able to break this dilemma

and predict with sufficient certainty

whether and when to treat at early stage.

Our breath and the mixture of volatile
organic compounds that we exhale

hold tremendous amounts of information
on our physiological condition.

With what we know today,
we have only scratched the surface.

As we collect more and more data
and breath profiles across the population,

including all varieties of gender,
age, origin and lifestyle,

the power of breath analysis
should increase.

And eventually, breath analysis
should provide us with a powerful tool

not only to proactively detect
specific diseases

but to predict
and ultimately prevent them.

And this should be enough motivation

to embrace the opportunities
and challenges

that breath analysis can provide,

even for people that are not
part-time hypochondriacs like me.

Thank you.

(Applause)