WEBVTT

00:00:29.592 --> 00:00:30.998
I'm a doctor.

00:00:31.458 --> 00:00:35.632
And one of the hardest things
that I have to do in my profession

00:00:35.632 --> 00:00:39.713
is when I have to tell a patient,
"You have cancer."

00:00:40.573 --> 00:00:44.996
Raise your hand if someone close to you
has been diagnosed with cancer.

00:00:47.329 --> 00:00:50.362
Cancer is something
that touches all of us.

00:00:51.182 --> 00:00:52.312
I've personally lost

00:00:52.312 --> 00:00:56.364
both of my grandmothers
and three of my aunts to cancer.

00:00:57.294 --> 00:01:01.580
The most recent member of my family
to be diagnosed with cancer

00:01:01.580 --> 00:01:06.160
is my wonderful mother-in-law,
Norma, seen here with my husband.

00:01:06.800 --> 00:01:09.635
Norma was visiting us
in Baltimore last year,

00:01:09.785 --> 00:01:12.823
and she was feeling
something wasn't quite right.

00:01:13.563 --> 00:01:18.691
A couple of weeks later,
life changed completely for Norma:

00:01:18.691 --> 00:01:22.701
She was diagnosed
with advanced ovarian cancer.

00:01:24.051 --> 00:01:25.540
Norma's not alone.

00:01:26.720 --> 00:01:27.719
This very day,

00:01:27.719 --> 00:01:33.589
4,828 people will be diagnosed
with cancer in the United States.

00:01:33.759 --> 00:01:35.890
And if we look globally,

00:01:35.890 --> 00:01:39.769
17 million people will be diagnosed
with cancer this year,

00:01:40.075 --> 00:01:43.710
and 9.6 million of them will die.

00:01:45.391 --> 00:01:47.133
How can we change this?

00:01:48.083 --> 00:01:51.438
One of the most important ways
of improving survival

00:01:51.438 --> 00:01:54.168
is detecting cancer earlier.

00:01:54.408 --> 00:01:58.550
The reason for that
is that the later you pick up a cancer,

00:01:58.680 --> 00:02:01.522
the smaller your chances of curing it.

00:02:01.992 --> 00:02:04.022
If we look at colon cancer,

00:02:04.022 --> 00:02:08.399
it starts in the bowel,
then it moves to the lymph nodes,

00:02:08.399 --> 00:02:12.826
and ultimately it metastasizes
to the lungs and the liver.

00:02:13.556 --> 00:02:16.384
If your patients are diagnosed

00:02:16.384 --> 00:02:19.776
when the cancer has spread
to the lungs and the liver,

00:02:20.226 --> 00:02:23.284
they have a 14% chance of surviving.

00:02:23.934 --> 00:02:26.398
But if you could catch it
just a little bit earlier,

00:02:26.398 --> 00:02:28.328
when it was still in the lymph nodes,

00:02:28.328 --> 00:02:32.560
their survival goes from 14% to 71%.

00:02:33.016 --> 00:02:37.640
And if you could pick it up even earlier,
when it was still in the colon,

00:02:37.640 --> 00:02:41.265
your patients have
a 90% chance of surviving.

00:02:42.235 --> 00:02:44.860
So how can we detect cancer earlier?

00:02:45.420 --> 00:02:48.330
Well, I'd like to share some research
that our group is doing,

00:02:48.330 --> 00:02:50.605
which I think is really exciting.

00:02:51.225 --> 00:02:57.031
We know that tumors
release tumor DNA into the blood.

00:02:58.261 --> 00:03:02.549
Imagine if we could detect
that cancer DNA -

00:03:02.549 --> 00:03:06.338
we could potentially
detect cancers earlier.

00:03:07.078 --> 00:03:09.208
What a simple concept -

00:03:10.068 --> 00:03:14.005
detecting tumor DNA in the blood.

00:03:14.605 --> 00:03:18.014
But of course, science and life
is never that simple.

00:03:19.304 --> 00:03:22.406
And for great endeavors,

00:03:22.406 --> 00:03:25.948
there are always multiple challenges
that you need to overcome.

00:03:26.113 --> 00:03:30.899
And whether you're climbing Everest
or trying to diagnose cancer earlier,

00:03:30.899 --> 00:03:33.406
you have to overcome these challenges.

00:03:34.116 --> 00:03:39.085
The first challenge that we faced
in trying to diagnose cancer earlier

00:03:39.085 --> 00:03:40.946
was a technical one.

00:03:40.946 --> 00:03:46.567
And that is that the amount
of tumor DNA in the blood is tiny:

00:03:46.823 --> 00:03:53.653
one to five mutant fragments
among a sea of 10,000 normal fragments.

00:03:55.933 --> 00:03:59.020
I'm part of a team
at Johns Hopkins University

00:03:59.020 --> 00:04:02.940
led by Bert Vogelstein, Ken Kinsler,
and Nick Papadopoulos,

00:04:03.042 --> 00:04:06.539
and our goal is to detect cancers earlier.

00:04:07.889 --> 00:04:10.881
The team has worked on this
for many, many years

00:04:11.071 --> 00:04:14.036
and ultimately came up
with a novel technique

00:04:14.036 --> 00:04:16.498
which we call "safe sequencing."

00:04:16.998 --> 00:04:20.082
And using this technique,
what we do is we take DNA

00:04:20.082 --> 00:04:24.929
and we attach a unique identifier
or barcode to the DNA

00:04:24.929 --> 00:04:26.618
before it's sequenced.

00:04:26.968 --> 00:04:28.963
Using this technology,

00:04:28.963 --> 00:04:34.827
we're now able to identify
a single mutant template in the blood

00:04:34.827 --> 00:04:38.183
when it's surrounded
by 10,000 normal templates.

00:04:38.426 --> 00:04:40.618
So we've overcome our first challenge.

00:04:41.418 --> 00:04:43.808
But can it actually detect cancer?

00:04:44.408 --> 00:04:48.617
And we looked at 220 patients
with pancreatic cancer.

00:04:49.397 --> 00:04:53.248
Now, pancreatic cancer
is a devastating disease.

00:04:53.488 --> 00:04:57.209
It has the worst survival of any cancer,

00:04:57.439 --> 00:05:02.523
with just over 8% of patients
diagnosed with it ultimately surviving.

00:05:03.163 --> 00:05:08.573
Imagine if we could identify
pancreatic cancer earlier -

00:05:08.713 --> 00:05:12.415
what an incredible difference
this could make to those patients.

00:05:13.149 --> 00:05:14.367
So did it work?

00:05:14.877 --> 00:05:16.595
The answer is yes.

00:05:16.595 --> 00:05:21.201
We were able to identify tumor DNA
in the bloodstream

00:05:21.201 --> 00:05:24.869
in 30% of the patients
with pancreatic cancer.

00:05:25.489 --> 00:05:27.946
So we thought, "That's good."

00:05:28.306 --> 00:05:32.375
But the question we asked
is "How can we do better?"

00:05:33.235 --> 00:05:36.060
Another challenge
that we need to overcome.

00:05:36.790 --> 00:05:41.839
So we went back to the drawing board,
and we came up with a novel concept,

00:05:41.839 --> 00:05:48.095
which was to combine tumor DNA
with other tumor markers

00:05:48.095 --> 00:05:51.436
in such a way that we were able
to greatly increase

00:05:51.436 --> 00:05:55.381
the number of patients with
pancreatic cancer that we could identify

00:05:55.381 --> 00:05:58.572
from 30% to 64%,

00:05:59.082 --> 00:06:04.653
while still ensuring that healthy patients
were not misdiagnosed as having cancer.

00:06:06.313 --> 00:06:09.395
One of the really exciting things

00:06:09.395 --> 00:06:11.919
about tumor DNA

00:06:11.919 --> 00:06:16.606
is that the mutations
that are present in pancreatic cancer

00:06:16.606 --> 00:06:21.271
are also present
in multiple other cancers.

00:06:22.051 --> 00:06:23.657
So what does this mean?

00:06:24.727 --> 00:06:27.359
It means that potentially,

00:06:27.359 --> 00:06:33.842
you could identify many different types
of cancers with a single blood test.

00:06:35.672 --> 00:06:37.727
And that's what we try to do.

00:06:37.727 --> 00:06:43.157
So building on this concept,
our group developed a test

00:06:43.157 --> 00:06:47.740
which would screen
for eight common types of cancers:

00:06:47.740 --> 00:06:49.281
cancer of the esophagus,

00:06:49.281 --> 00:06:53.930
cancer of the stomach,
the colon, the pancreas, the liver,

00:06:53.930 --> 00:06:58.752
breast cancer, lung cancer,
and ovarian cancer.

00:06:58.878 --> 00:07:01.355
We called it CancerSEEK.

00:07:02.155 --> 00:07:06.099
And we evaluated it
in just over 1,000 individuals

00:07:06.099 --> 00:07:09.063
who had one of those
eight different types of cancers

00:07:09.063 --> 00:07:12.113
as well as 800 healthy individuals.

00:07:13.043 --> 00:07:14.351
And it worked.

00:07:15.151 --> 00:07:18.714
We were able to identify
every cancer type.

00:07:19.115 --> 00:07:22.200
The number of patients
that we were able to identify

00:07:22.200 --> 00:07:24.960
depended on the type
of cancer that they had.

00:07:25.140 --> 00:07:29.763
So we were able to identify
33% of the patients with breast cancer,

00:07:30.533 --> 00:07:35.757
72% of the patients with cancer
of the pancreas or stomach cancer,

00:07:36.207 --> 00:07:42.208
and 98% of the patients
who had liver cancer or ovarian cancer.

00:07:43.608 --> 00:07:45.058
Overall,

00:07:45.058 --> 00:07:51.867
CancerSEEK identified a median of 70%
of the eight different types of cancers.

00:07:53.787 --> 00:07:56.525
One of the most important findings

00:07:56.525 --> 00:07:57.704
in this study

00:07:58.144 --> 00:08:03.991
is that five of the eight cancers
have no screening test.

00:08:05.001 --> 00:08:10.725
CancerSEEK identified
between 69% to 98%

00:08:10.725 --> 00:08:12.809
of these five cancers.

00:08:13.902 --> 00:08:16.906
(Applause)

00:08:24.177 --> 00:08:27.333
If we look at these eight common cancers,

00:08:28.753 --> 00:08:34.701
they account for 60% of the deaths
due to cancer in the United States.

00:08:35.291 --> 00:08:38.461
Identifying them earlier will save lives.

00:08:41.104 --> 00:08:42.492
We're currently evaluating

00:08:42.492 --> 00:08:47.620
how good CancerSEEK is
at identifying cancer earlier

00:08:47.620 --> 00:08:50.650
in 10,000 healthy individuals

00:08:50.830 --> 00:08:55.361
who have no symptoms
and no history of cancer.

00:08:56.371 --> 00:08:58.998
There will be more challenges ahead.

00:08:59.228 --> 00:09:02.185
But our hope is that in the future,

00:09:02.185 --> 00:09:07.500
we'll have a single blood test
which can identify multiple cancers,

00:09:07.500 --> 00:09:11.767
and we'll save lives
by detecting cancers earlier,

00:09:12.487 --> 00:09:17.336
ultimately helping ensure
people such as Norma

00:09:17.336 --> 00:09:21.661
have the very best chance of surviving
and enjoying life to the full.

00:09:22.221 --> 00:09:23.220
Thank you.

00:09:23.220 --> 00:09:26.218
(Applause)