Personalized cancer therapies driven by AI
[Music]
[Music]
let me give you a sneak peek
into the future in the future
everything will be truly personalized in
the future
we will be designing one therapy for
each individual patient
personalization is now becoming a
reality due to the advancements in
artificial intelligence or ai
ai is enabling us to develop cancer
therapy that are specifically tailored
to the individual patient
based on the genetic code or the dna of
the individual
we can use the wealth of information
contained within the dna of the
individual
to tailor the therapy to that specific
cancer patient
but now you may ask yourselves how do we
get from the dna
of an individual cancer patient to a
personalized cancer therapy
using ai
in order to understand this let’s take a
step back and look at cancer therapy
in a traditional cancer therapy approach
such as chemotherapy
we’re using a one-size-fits-all model
these types of therapies target cancer
non-specifically
which is which causes side effects and
is not always effective in fighting off
the cancer
the challenge however is in cancer and
in tumors
tumors each tumor is unique and we can
argue that a more effective therapy can
be developed if we treat the tumor as
such
in a personalized approach we want to
target each patient
uniquely more importantly we want to
target each
tumor uniquely we basically want to look
at the foreign elements that are
displayed on the surface of the tumor
cells
and target these
these foreign elements arise from
mutations that can be caused by for
example
environmental factors such as sun or
smoking
mutations are completely normal events
in the human body and we have developed
many mechanisms to mitigate the
potentially
dangerous actions that these mutations
can have
however in cancer these one or more
mutational events
have not been mitigated and this can
result
in the development of a potentially
malignant tumor
tumor cells are disguising themselves as
normal
cells this means that tumor cells are
putting up a protective shield
allowing them to attract all the
essential nutrients they need to grow
allowing them to ultimately grow out of
control
the immune system cannot recognize the
tumor cells as foreign
our immune system consists of so-called
t-cells and t-cells constantly
survey the body and monitor whether
there’s any danger looking
they are experts in finding foreign
intruders that are trying to invade the
body
such as bacteria and virus virus such as
the coronavirus which i guess you are
all experts in now in light of the
pandemic
in cancer the tumor should in principle
be just as foreign to the human body
as a virus is but somehow it is not
the tumor goes unnoticed by the immune
system the tumor is in disguise
we want to design personalized cancer
therapy that helps the immune system
see the tumor we want to help the immune
system see the foreign elements
on the surface of the tumor cells and we
want to
boost an immune response that makes sure
that we have the right t cells present
and enough of the t cells present to
actually fight off the tumor
in a personalized cancer therapy what we
want to do is we want to develop a
therapy specifically designed
based on the genetic code of the dna of
the individual
patient we want to bring this these
foreign elements that are on the surface
of the tumor we want to bring them out
in the open
and the way we do that in a personalized
cancer therapy approach
is actually by providing the body with
more of this foreign material
exactly as we would do in a vaccine
against coronavirus
so let me illustrate what these foreign
elements actually are
these foreign elements are what we call
neo-epitopes
or new epitopes and epitopes are small
molecules that are displayed on the
surface of our cells
in complex with a specialized immune
molecule called
mhc neo-epitopes
or new epitopes have arisen from
mutations in our dna
and then they are displayed on the
surface of the cells in complex with
these mhc
each cancer patient or each tumor of
each cancer patient
is defined by a unique set of neopotents
and also each patient has a unique set
of mhc
it’s only a subset of the molecules that
are bound to the mhc or displayed on the
surface by mhc
that are actually neopotopes and
identifying these
of uh is a hugely complex task also
the immune system being able to see
these is also
a highly complex task for the immune
system
imagine that these neops are screaming
at the top
of their lungs trying to inform the
immune system that there is danger
the neoprobes are trying to tell the
immune system that is something wrong
with this cell
there’s something wrong with this tumor
cell and that they should immediately
kill
the cell the problem is however
that these this signal that the neotopes
are trying to convey is drowning in
other single
signals provided by the the tumor cell
but also
surrounding the tumor cell that results
in the knee octop signal not being heard
what we want to do in a neoprete-based
personalized
cancer therapy is we want to shift this
balance
and we want to amplify the dangerous
signal
basically allowing the immune system to
be able to
hear the voices of the neopets we want
to provide the neopotopes
with a megaphone in order for the immune
system to be able to hear them better
in order to fight off the tumor
let me try to visualize how we would
design
such a personalized cancer therapy
this you see here is a cancer patient
or more specifically this is the healthy
dna of a cancer patient
this is the dna of the tumor of this
specific cancer patient
what we want to do is we want to
identify these mutations that are unique
to the tumor of this specific patient
and the way in which we do that is
overlay the dna of the tumor with the
dna of the healthy tissue
in order to isolate the mutations
this results in potentially a list of
thousands of mutations
and what we want to do is we want to
based on these thousands of mutations
we want to identify which of those are
actually displayed on the surface of
cells
and and are able to interact with the
immune system in order to inform that
this tumor
is actually a tumor cell so what we want
to do is we want to find the needle in
the haystack
we want to find the neotopes that will
be displayed on the surface
and this is a challenging task in
principle what we want to do is we want
to actually test each of these mutations
in the lab
we want to test whether they are
displayed on the cell surface and we
also want to test whether they are
they have the ability to interact with
the immune system
however as you can imagine doing this
testing all of these in the lab would is
a highly complex task
would take very long time and is
therefore not feasible
so we need other tools to identify these
neops and this is where ai comes into
the picture
an ai algorithm in the context of
neopitot identification is trained on
vast biological data sets that contain
information about
molecules that are displayed on the
source of cells and we can use that
information
to identify our neotopes an example of
such data
is that we take cells from
lots and lots thousands and thousands of
different people
we then strip the surface of these cells
for the for their epitopes
as you can imagine this results in an
extremely long list of epitopes
what we want to do with ai is we want to
look at this really long list of
epitopes
and we want the ai to identify patterns
in these
epitopes that will allow us to look at
similar patterns in neotopes
in order to figure out what is actually
presented on the cell’s surface
what we do with our ai algorithm when we
see when we get a new patient and we
want to see
what neopotopes i are unique to this
patient’s tumor
is that we use our ai and the ai uses
all its past
experience that it’s learned from all of
the other data it’s seen
and it uses that past experience to
figure out for this specific patient
which neo-epitopes should we be bringing
out into into the open
which neopetopes which voices should we
be amplifying
in order for the immune system to be
able to spot the tumor
once we’ve identified our neopotopes
with our ai algorithm
we can go back into the lab and we can
then make more of these
neops synthetically we can then put
these neopotopes
into our personalized cancer therapy and
give this back to the patient
in order to amplify the danger signal
and for the patient’s
own immune system to be able to fight
off the cancer
let me finish off by telling you a story
about tomorrow
based on many of the things that we can
already do today
after many years in the sun mostly
without sun protection
my grandmother harriet has developed
skin cancer
she visits the hospital and the hospital
decides that
a personalized cancer approach would be
the way to go for her
a tumor biopsy and a blood sample is
taken from my grandmother
and this is the material we need to
design personalized therapy specific for
her
once she is waiting for her therapy to
be produced we’ll take a look behind the
scenes
and see how do we actually produce this
personalized cancer therapy
we look at the dna of the healthy tissue
of my grandmother
and we look at the dna of our tumor and
this is what we use to identify the
computational
or computationally identify the
mutational landscape of her tumor
specifically
we can use the blood sample to identify
the mhc the immune molecules that are
specific to her
once we have this information we can
then use ai to extract the neopotopes
that are specific to her tumor
these computational steps can ach can be
achieved within a matter of hours
once we’ve identified identified on
neoprotops
we can go back into the lab synthesize
them
make a personalized cancer therapy
specific for my grandmother
and when she comes back to the hospital
she can be given her therapy
and after some time my grandmother’s own
immune system will be able to fight
off the cancer
just imagine this world of tomorrow a
world
where personalization is the norm and
where each cancer patient can be treated
specifically based on his or her tumor
a world where each cancer patient can be
given a cancer cancer treatment
with no off-target toxicities and with
high effectiveness in fighting off
the in the body that is the cancer
tailor-made personalized cancer therapy
driven by ai can potentially make this
reality
thank you
[Applause]
you