TED2013
Jack Andraka: A promising test for pancreatic cancer ... from a teenager
Filmed:
Readability: 4.7
4,827,010 views
Over 85 percent of all pancreatic cancers are diagnosed late, when someone has less than two percent chance of survival. How could this be? Jack Andraka talks about how he developed a promising early detection test for pancreatic cancer that's super cheap, effective and non-invasive -- all before his 16th birthday.
Jack Andraka - Cancer detector inventor
A paper on carbon nanotubes, a biology lecture on antibodies and a flash of insight led 15-year-old Jack Andraka to design a cheaper, more sensitive cancer detector. Full bio
A paper on carbon nanotubes, a biology lecture on antibodies and a flash of insight led 15-year-old Jack Andraka to design a cheaper, more sensitive cancer detector. Full bio
Double-click the English transcript below to play the video.
00:12
Have you ever experienced a moment in your life
0
726
3709
00:16
that was so painful and confusing
1
4435
3506
00:19
that all you wanted to do
2
7941
2039
00:21
was learn as much as you could to make sense of it all?
3
9980
3770
00:25
When I was 13, a close family friend
4
13750
2950
00:28
who was like an uncle to me
5
16700
1722
00:30
passed away from pancreatic cancer.
6
18422
2765
00:33
When the disease hit so close to home,
7
21187
2331
00:35
I knew I needed to learn more,
8
23518
2106
00:37
so I went online to find answers.
9
25624
2846
00:40
Using the Internet, I found a variety of statistics
10
28470
3090
00:43
on pancreatic cancer,
11
31560
1678
00:45
and what I had found shocked me.
12
33238
2648
00:47
Over 85 percent of all pancreatic cancers
13
35886
3832
00:51
are diagnosed late,
14
39718
1576
00:53
when someone has less than a two percent chance of survival.
15
41294
4326
00:57
Why are we so bad at detecting pancreatic cancer?
16
45620
3948
01:01
The reason? Today's current modern medicine
17
49568
3585
01:05
is a 60-year-old technique.
18
53153
2255
01:07
That's older than my dad.
19
55408
2505
01:09
(Laughter)
20
57913
3546
01:13
But also, it's extremely expensive,
21
61459
2443
01:15
costing 800 dollars per test,
22
63902
3253
01:19
and it's grossly inaccurate,
23
67155
2752
01:21
missing 30 percent of all pancreatic cancers.
24
69907
3417
01:25
Your doctor would have to be ridiculously suspicious
25
73324
2922
01:28
that you have the cancer in order to give you this test.
26
76246
3485
01:31
Learning this, I knew there had to be a better way.
27
79731
3745
01:35
So I set up a scientific criteria
28
83476
2501
01:37
as to what a sensor would have to look like
29
85977
1986
01:39
in order to effectively diagnose pancreatic cancer.
30
87963
3417
01:43
The sensor would have to be inexpensive, rapid,
31
91380
3736
01:47
simple, sensitive, selective,
32
95116
3551
01:50
and minimally invasive.
33
98667
3281
01:53
Now, there's a reason why this test
34
101948
2031
01:55
hasn't been updated in over six decades,
35
103979
3129
01:59
and that's because, when we're looking for pancreatic cancer,
36
107108
3183
02:02
we're looking at your bloodstream,
37
110291
1594
02:03
which is already abundant in all these tons and tons of protein,
38
111885
4849
02:08
and you're looking for this miniscule difference
39
116734
2135
02:10
in this tiny amount of protein,
40
118869
1831
02:12
just this one protein.
41
120700
1286
02:13
That's next to impossible.
42
121986
1926
02:15
However, undeterred due to my teenage optimism --
43
123912
3574
02:19
(Applause) —
44
127486
5576
02:25
I went online to a teenager's two best friends,
45
133062
3798
02:28
Google and Wikipedia.
46
136860
1704
02:30
I got everything for my homework from those two sources.
47
138564
4218
02:34
And what I had found was an article
48
142782
2548
02:37
that listed a database of over 8,000 different proteins
49
145330
3794
02:41
that are found when you have pancreatic cancer.
50
149124
2487
02:43
So I decided to go and make it my new mission
51
151611
3644
02:47
to go through all these proteins and see which ones
52
155255
2893
02:50
could serve as a biomarker for pancreatic cancer.
53
158148
2523
02:52
And to make it a bit simpler for myself,
54
160671
2438
02:55
I decided to map out a scientific criteria. And here it is.
55
163109
4248
02:59
Essentially first, the protein would have to be found
56
167357
2432
03:01
in all pancreatic cancers at high levels in the bloodstream
57
169789
3175
03:04
in the earliest stages, but also only in cancer.
58
172964
4373
03:09
And so I'm just plugging and chugging through this gargantuan task,
59
177337
2990
03:12
and finally, on the 4,000th try,
60
180327
3322
03:15
when I'm close to losing my sanity,
61
183649
1779
03:17
I find the protein.
62
185428
1890
03:19
And the name of the protein I'd located
63
187318
2669
03:21
was called mesothelin,
64
189987
1392
03:23
and it's just your ordinary, run-of-the-mill type protein,
65
191379
2730
03:26
unless of course you have pancreatic,
66
194109
1708
03:27
ovarian or lung cancer,
67
195817
1729
03:29
in which case it's found at these very high levels in your bloodstream.
68
197546
3025
03:32
But also the key is
69
200571
1808
03:34
that it's found in the earliest stages of the disease,
70
202379
2956
03:37
when someone has close to 100 percent chance
71
205335
2203
03:39
of survival.
72
207538
1274
03:40
So now that I'd found a reliable protein I could detect,
73
208812
3661
03:44
I then shifted my focus to actually detecting that protein,
74
212473
2773
03:47
and, thus, pancreatic cancer.
75
215246
2454
03:49
Now, my breakthrough came in a very unlikely place,
76
217700
3369
03:53
possibly the most unlikely place for innovation:
77
221069
2641
03:55
my high school biology class,
78
223710
2224
03:57
the absolute stifler of innovation.
79
225934
2521
04:00
(Laughter) (Applause)
80
228455
4924
04:05
And I had snuck in this article on these things called
81
233379
3181
04:08
carbon nanotubes, and that's just a long, thin pipe of carbon
82
236560
3598
04:12
that's an atom thick
83
240158
1254
04:13
and one 50 thousandth the diameter of your hair.
84
241412
2695
04:16
And despite their extremely small sizes,
85
244107
2390
04:18
they have these incredible properties.
86
246497
1834
04:20
They're kind of like the superheroes of material science.
87
248331
2591
04:22
And while I was sneakily reading this article
88
250922
2185
04:25
under my desk in my biology class,
89
253107
2095
04:27
we were supposed to be paying attention
90
255202
1599
04:28
to these other kind of cool molecules called antibodies.
91
256801
3709
04:32
And these are pretty cool because they only react
92
260510
2101
04:34
with one specific protein,
93
262611
1783
04:36
but they're not nearly as interesting as carbon nanotubes.
94
264394
2600
04:38
And so then, I was sitting in class,
95
266994
3107
04:42
and suddenly it hit me:
96
270101
2538
04:44
I could combine what I was reading about,
97
272639
2651
04:47
carbon nanotubes,
98
275290
1449
04:48
with what I was supposed to be thinking about, antibodies.
99
276739
2936
04:51
Essentially, I could weave a bunch of these antibodies
100
279675
2765
04:54
into a network of carbon nanotubes
101
282440
1988
04:56
such that you have a network
102
284428
1593
04:58
that only reacts with one protein,
103
286021
2430
05:00
but also, due to the properties of these nanotubes,
104
288451
3320
05:03
it would change its electrical properties
105
291771
1952
05:05
based on the amount of protein present.
106
293723
2508
05:08
However, there's a catch.
107
296231
2083
05:10
These networks of carbon nanotubes are extremely flimsy,
108
298314
3400
05:13
and since they're so delicate, they need to be supported.
109
301714
3489
05:17
So that's why I chose to use paper.
110
305203
2016
05:19
Making a cancer sensor out of paper
111
307219
2294
05:21
is about as simple as making chocolate chip cookies,
112
309513
2368
05:23
which I love.
113
311881
3549
05:27
You start with some water, pour in some nanotubes,
114
315430
3345
05:30
add antibodies, mix it up,
115
318775
2414
05:33
take some paper, dip it, dry it,
116
321189
2368
05:35
and you can detect cancer.
117
323557
3432
05:38
(Applause)
118
326989
6507
05:45
Then, suddenly, a thought occurred
119
333496
3234
05:48
that kind of put a blemish on my amazing plan here.
120
336730
4346
05:53
I can't really do cancer research
121
341076
1855
05:54
on my kitchen countertop.
122
342931
1242
05:56
My mom wouldn't really like that.
123
344173
2143
05:58
So instead, I decided to go for a lab.
124
346316
3250
06:01
So I typed up a budget, a materials list,
125
349566
2404
06:03
a timeline, and a procedure,
126
351970
2227
06:06
and I emailed it to 200 different professors
127
354197
2519
06:08
at Johns Hopkins University
128
356716
1689
06:10
and the National Institutes of Health,
129
358405
1704
06:12
essentially anyone that had anything to do with pancreatic cancer.
130
360109
3176
06:15
And I sat back waiting for these positive emails to be pouring in,
131
363285
2949
06:18
saying, "You're a genius!
132
366234
1517
06:19
You're going to save us all!"
133
367751
1582
06:21
And — (Laughter)
134
369333
4289
06:25
Then reality took hold,
135
373622
1321
06:26
and over the course of a month,
136
374943
2414
06:29
I got 199 rejections out of those 200 emails.
137
377357
4575
06:33
One professor even went through my entire procedure,
138
381932
2460
06:36
painstakingly -- I'm not really sure where he got all this time --
139
384392
2973
06:39
and he went through and said why each and every step
140
387365
3984
06:43
was like the worst mistake I could ever make.
141
391349
2871
06:46
Clearly, the professors did not have as high
142
394220
2637
06:48
of an opinion of my work as I did.
143
396857
3122
06:51
However, there was a silver lining.
144
399979
2142
06:54
One professor said, "Maybe I might be able to help you, kid."
145
402121
3011
06:57
So I went in that direction.
146
405132
2488
06:59
(Laughter)
147
407620
3897
07:03
As you can never say no to a kid.
148
411517
2664
07:06
And so then, three months later,
149
414181
2480
07:08
I finally nailed down a harsh deadline with this guy,
150
416661
2876
07:11
and I get into his lab,
151
419537
1364
07:12
I get all excited, and then I sit down,
152
420901
2319
07:15
I start opening my mouth and talking,
153
423220
1747
07:16
and five seconds later he calls in another Ph.D.
154
424967
2633
07:19
Ph.D.'s just flock into this little room,
155
427600
3717
07:23
and they're just firing these questions at me,
156
431317
2215
07:25
and by the end, I kind of felt like I was in a clown car.
157
433532
2482
07:28
There were 20 Ph.D.'s plus me and the professor
158
436014
1951
07:29
crammed into this tiny office space
159
437965
2287
07:32
with them firing these rapid-fire questions at me,
160
440252
3071
07:35
trying to sink my procedure.
161
443323
2338
07:37
How unlikely is that? I mean, pshhh.
162
445661
2469
07:40
(Laughter)
163
448130
4348
07:44
However, subjecting myself to that interrogation,
164
452478
2724
07:47
I answered all of their questions,
165
455202
1458
07:48
and I guessed on quite a few but I got them right,
166
456660
2158
07:50
and I finally landed the lab space I needed.
167
458818
4552
07:55
But it was shortly afterwards that I discovered
168
463370
2193
07:57
my once brilliant procedure
169
465563
1999
07:59
had something like a million holes in it,
170
467562
2027
08:01
and over the course of seven months,
171
469589
1581
08:03
I painstakingly filled each and every one of those holes.
172
471170
3456
08:06
The result? One small paper sensor
173
474626
3050
08:09
that costs three cents and takes five minutes to run.
174
477676
3150
08:12
This makes it 168 times faster,
175
480826
3585
08:16
over 26,000 times less expensive,
176
484411
2827
08:19
and over 400 times more sensitive
177
487238
2659
08:21
than our current standard for pancreatic cancer detection.
178
489897
3010
08:24
(Applause)
179
492907
9558
08:34
One of the best parts of the sensor, though,
180
502465
2395
08:36
is that it has close to 100 percent accuracy,
181
504860
2597
08:39
and can detect the cancer in the earliest stages
182
507457
2362
08:41
when someone has close to 100 percent chance of survival.
183
509819
3393
08:45
And so in the next two to five years,
184
513212
2264
08:47
this sensor could potentially lift for pancreatic cancer survival rates
185
515476
3265
08:50
from a dismal 5.5 percent
186
518741
2135
08:52
to close to 100 percent,
187
520876
2096
08:54
and it would do similar for ovarian and lung cancer.
188
522972
2891
08:57
But it wouldn't stop there.
189
525863
2342
09:00
By switching out that antibody,
190
528205
1695
09:01
you can look at a different protein,
191
529900
1619
09:03
thus, a different disease,
192
531519
1606
09:05
potentially any disease in the entire world.
193
533125
3705
09:08
So that ranges from heart disease
194
536830
1841
09:10
to malaria, HIV, AIDS,
195
538671
2733
09:13
as well as other forms of cancer -- anything.
196
541404
2729
09:16
And so hopefully one day
197
544133
2191
09:18
we can all have that one extra uncle,
198
546324
2416
09:20
that one mother, that one brother, sister,
199
548740
3084
09:23
we can have that one more family member to love,
200
551824
3026
09:26
and that our hearts will be rid of that one disease burden
201
554850
4856
09:31
that comes from pancreatic, ovarian and lung cancer,
202
559706
2861
09:34
and potentially any disease,
203
562567
2305
09:36
that through the Internet anything is possible.
204
564872
2398
09:39
Theories can be shared,
205
567270
1678
09:40
and you don't have to be a professor
206
568948
1452
09:42
with multiple degrees to have your ideas valued.
207
570400
2660
09:45
It's a neutral space,
208
573060
1402
09:46
where what you look like, age or gender,
209
574462
2860
09:49
it doesn't matter.
210
577322
1174
09:50
It's just your ideas that count.
211
578496
2202
09:52
For me, it's all about looking at the Internet
212
580698
2953
09:55
in an entirely new way
213
583651
2088
09:57
to realize that there's so much more to it
214
585739
2119
09:59
than just posting duck-face pictures of yourself online.
215
587858
4897
10:04
You could be changing the world.
216
592755
3257
10:08
So if a 15-year-old
217
596012
2033
10:10
who didn't even know what a pancreas was
218
598045
3087
10:13
could find a new way to detect pancreatic cancer,
219
601132
3574
10:16
just imagine what you could do.
220
604706
2491
10:19
Thank you.
221
607197
1332
10:20
(Applause)
222
608529
5708
ABOUT THE SPEAKER
Jack Andraka - Cancer detector inventorA paper on carbon nanotubes, a biology lecture on antibodies and a flash of insight led 15-year-old Jack Andraka to design a cheaper, more sensitive cancer detector.
Why you should listen
After Andraka’s proposal to build and test his idea for a pancreatic cancer detector was rejected from 199 labs, the teen landed at Johns Hopkins. There, he built his device using inexpensive strips of filter paper, carbon nanotubes and antibodies sensitive to mesothelin, a protein found in high levels in people with pancreatic cancer. When dipped in blood or urine, the mesothelin adheres to these antibodies and is detectable by predictable changes in the nanotubes’ electrical conductivity.
In preliminary tests, Andraka’s invention has shown 100 percent accuracy. It also finds cancers earlier than current methods, costs a mere 3 cents and earned the high schooler the 2012 Intel Science Fair grand prize.
Jack Andraka | Speaker | TED.com