TEDxBoston 2011
Michelle Borkin: Can astronomers help doctors?
Filmed:
Readability: 4.6
190,528 views
How do you measure a nebula? With a brain scan. In this talk, TED Fellow Michelle Borkin shows why collaboration between doctors and astronomers can lead to surprising discoveries.
Michelle Borkin - Physicist
Michelle Borkin is a PhD candidate in applied physics. She works with the Astronomical Medicine Project and interdisciplinary 3D visualization techniques. Full bio
Michelle Borkin is a PhD candidate in applied physics. She works with the Astronomical Medicine Project and interdisciplinary 3D visualization techniques. Full bio
Double-click the English transcript below to play the video.
00:10
I believe
0
310
1546
00:11
that we can both unravel
the mysteries of the universe
the mysteries of the universe
1
1880
4530
00:16
and save human lives at the same time
2
6434
3014
00:19
through interdisciplinary research.
3
9472
2669
00:22
And I'm going to share with you today
just one story, my story,
just one story, my story,
4
12165
4532
00:26
that has crossed these paths.
5
16721
1908
00:28
We start the in supernova
remnant Cassiopeia A.
remnant Cassiopeia A.
6
18653
3429
00:32
It's one of the youngest ones
in our galaxy, about 330 years old.
in our galaxy, about 330 years old.
7
22106
3701
00:35
An astronomy colleague
approached me one day,
approached me one day,
8
25831
2466
00:38
and she had over eight years
of magnificent data,
of magnificent data,
9
28321
3323
00:41
just trying to understand
the 3-D structure of this nebula,
the 3-D structure of this nebula,
10
31668
3291
00:44
the supernova remnant.
11
34983
1514
00:46
But she had no way to look at it.
12
36521
1773
00:49
So I looked at the data with her
and said, "I think I can help you."
and said, "I think I can help you."
13
39078
4599
00:53
And although -- and this is all real data
14
43701
2025
00:55
you're seeing on the screen above me --
15
45750
2443
00:58
this is the Hollywood rendering version,
16
48217
2065
01:00
but the rough draft I made with her
looks something more like this.
looks something more like this.
17
50306
4448
01:04
And she was able to make novel discoveries
about how supernovas explode
about how supernovas explode
18
54778
3857
01:08
and how shells explode within it,
19
58659
2557
01:11
using a piece of software developed
at Brigham and Women's Hospital
at Brigham and Women's Hospital
20
61240
3889
01:15
here in Boston,
21
65153
1309
01:16
called 3D Slicer.
22
66486
1822
01:18
It was originally developed
for looking at patients' brain scans,
for looking at patients' brain scans,
23
68332
4214
01:22
doing surgical planning
24
72570
1454
01:24
and doing 3-D renderings of anatomy.
25
74048
2630
01:26
Who knew our solution was lurking
just across the river?
just across the river?
26
76702
4726
01:31
Now, people don't believe me
27
81452
2503
01:33
when I tell them that astronomy
and medical imaging --
and medical imaging --
28
83979
2867
01:36
these two seemingly different fields --
29
86870
1898
01:38
are really similar.
30
88792
1155
01:39
So we're going to play a little game
I like to call "Which is which?"
I like to call "Which is which?"
31
89971
3274
01:43
I play this with new doctors
and astronomers I work with.
and astronomers I work with.
32
93269
2690
01:45
I'm going to show you
two images on the screen.
two images on the screen.
33
95983
2229
01:48
One of them is biomedical
and one of them is astronomical,
and one of them is astronomical,
34
98236
3274
01:51
and you have to pick them
correctly in your head.
correctly in your head.
35
101534
3323
01:54
So here is the first set.
36
104881
1617
01:56
And again, one of these is biomedical
and one is astronomical.
and one is astronomical.
37
106522
3649
02:00
I'll give you a second
to make your little vote mentally.
to make your little vote mentally.
38
110195
3647
02:07
So it turns out the one on the left
is some of the raw data
is some of the raw data
39
117533
3327
02:10
of the supernova remnant
we were just looking at,
we were just looking at,
40
120884
2337
02:13
and on the right, we have
an angiogram of a patient's heart
an angiogram of a patient's heart
41
123245
2863
02:16
and coronary arteries.
42
126132
1165
02:17
OK, we're going to try another one.
43
127321
1676
02:19
Now, this one is much closer
to my daily bread and butter.
to my daily bread and butter.
44
129021
3330
02:22
Tell me which is which.
45
132375
2298
02:24
And one of these is literally
millimeters across,
millimeters across,
46
134697
2627
02:27
and the other is billions of miles.
47
137348
2566
02:35
So, it turns out the one on the left
is a confocal microscopy image
is a confocal microscopy image
48
145881
3546
02:39
of a human cornea,
49
149451
1156
02:40
and on the right, we have
a radio telescope image
a radio telescope image
50
150631
3074
02:43
of the star-forming region NGC-1333.
51
153729
2620
02:46
Now, aside from the fact
that these images look similar
that these images look similar
52
156373
2733
02:49
and that doctors trying to find
a tumor in a patient's brain
a tumor in a patient's brain
53
159130
2923
02:52
or a young star forming is similar,
54
162077
1746
02:53
the way the data comes
from the machine or the telescope
from the machine or the telescope
55
163847
2743
02:56
is remarkably similar.
56
166614
1602
02:58
Here's an MRI scanner.
57
168240
2480
03:00
And if you've never seen
the raw data of a patient's brain,
the raw data of a patient's brain,
58
170744
2987
03:03
this is what it looks like.
59
173755
1330
03:05
When the MRI scanner
is acquiring the data,
is acquiring the data,
60
175109
2527
03:07
it goes in slices.
61
177660
1285
03:08
So you can see
the patient's nose, their eyes;
the patient's nose, their eyes;
62
178969
2746
03:11
it kind of progresses
towards the middle of the head;
towards the middle of the head;
63
181739
2690
03:14
you can start to see the cortex,
64
184453
2089
03:16
and it steps through
to the back of the brain.
to the back of the brain.
65
186566
3035
03:19
Now, believe it or not,
66
189625
1317
03:20
telescopes, and particularly radio
telescopes, operate in a similar manner.
telescopes, operate in a similar manner.
67
190966
4619
03:25
If we were to look at the raw data
from these telescopes ...
from these telescopes ...
68
195609
3068
03:28
We're going to look
at a nebula called M16.
at a nebula called M16.
69
198701
3255
03:31
We start with this radio telescope
at the front of the nebula,
at the front of the nebula,
70
201980
3228
03:35
stepping back towards
the middle of the nebula,
the middle of the nebula,
71
205232
2324
03:37
just like the middle
of the patient's brain --
of the patient's brain --
72
207580
2343
03:39
those bright regions
are where young stars are forming --
are where young stars are forming --
73
209947
3050
03:43
all the way to the back of the nebula,
74
213021
1868
03:44
just like the back of the patient's head.
75
214913
2529
03:47
Now, although the doctors
are able to then take this data
are able to then take this data
76
217466
4193
03:51
and look at it in 3-D
and do surgical planning,
and do surgical planning,
77
221683
2688
03:54
this is cutting-edge,
78
224395
1632
03:56
just about as good as you get
with any astronomer,
with any astronomer,
79
226051
2384
03:58
and this is what they have to look at
to understand the 3-D structure
to understand the 3-D structure
80
228459
5056
04:03
and velocity's momentum in our universe.
81
233539
2974
04:06
But we can do better.
82
236537
1685
04:08
So, you might recognize
this nebula more like this:
this nebula more like this:
83
238246
3345
04:11
the famous Hubble image of the Pillars
of Creation or the Eagle Nebula.
of Creation or the Eagle Nebula.
84
241615
4127
04:15
And, I'm going to fade this
out onto a radio image,
out onto a radio image,
85
245766
3033
04:18
it's a false color in the background,
86
248823
2184
04:21
and fade away the Hubble image
you're used to.
you're used to.
87
251031
2720
04:24
But we don't need to just look
at this in 3-D, we can look at it in 2-D,
at this in 3-D, we can look at it in 2-D,
88
254329
3616
04:27
and here I'm using a radiology
tool kit called OsiriX.
tool kit called OsiriX.
89
257969
4034
04:32
When I showed this
to astronomer Marc Pound,
to astronomer Marc Pound,
90
262027
2372
04:34
whose data this is,
91
264423
1325
04:35
he was amazed, because he had
been trying so hard
been trying so hard
92
265772
3359
04:39
to study the impact
of a young group of stars.
of a young group of stars.
93
269155
2817
04:41
And he had this theory
94
271996
1183
04:43
that there's this wind crashing
and tossing the pillars over,
and tossing the pillars over,
95
273203
3144
04:46
and it took him months to prove this
with conventional visualization.
with conventional visualization.
96
276371
3633
04:50
But in one shot, you can
see the shock wave of wind
see the shock wave of wind
97
280028
3628
04:53
blasting through across
to the left-hand side of the screen.
to the left-hand side of the screen.
98
283680
2965
04:57
Now, I don't think myself
or any of my collaborators
or any of my collaborators
99
287388
3096
05:00
would've anticipated
how far this has gone,
how far this has gone,
100
290508
2905
05:03
and by sharing the medical
technology with astronomy
technology with astronomy
101
293437
2499
05:05
and astronomy with medical,
102
295960
1354
05:07
we've been able to find
new stars and supernova remnants,
new stars and supernova remnants,
103
297338
2893
05:10
and revolutionize how you do
heart diagnostics
heart diagnostics
104
300255
3203
05:13
and look at data for different patients
and organize it and data-mine it.
and organize it and data-mine it.
105
303482
4622
05:18
I don't have time to show you
all these great projects,
all these great projects,
106
308128
2613
05:20
but I'll show you one of them.
107
310765
1445
05:22
This is a collaboration
I've been working on,
I've been working on,
108
312234
2180
05:24
called The Multiscale
Hemodynamics Project.
Hemodynamics Project.
109
314438
2222
05:26
I'm working with doctors
at Brigham and Women's Hospital.
at Brigham and Women's Hospital.
110
316684
2713
05:29
Now, what this represents is a novel way
of doing heart disease diagnostics.
of doing heart disease diagnostics.
111
319421
4222
05:33
And instead of the conventional
invasive angiography,
invasive angiography,
112
323667
4053
05:37
this is just a CT scan.
113
327744
1775
05:39
What you see here
are the coronary arteries.
are the coronary arteries.
114
329543
2092
05:41
So you have your heart, and the arteries
wrap around the outside.
wrap around the outside.
115
331659
3244
05:44
These are the arteries
you worry about getting blocked
you worry about getting blocked
116
334927
2731
05:47
and giving you a heart attack
117
337682
1418
05:49
and killing you.
118
339124
1154
05:50
So it's really important
that we look at them.
that we look at them.
119
340302
2273
05:52
Now, this is a CT scan of a patient
120
342599
2941
05:55
with a blood-flow simulation --
that's the coloring up there.
that's the coloring up there.
121
345564
4155
05:59
That simulation was originally developed
for studying the structure of DNA,
for studying the structure of DNA,
122
349743
4872
06:04
and then the visualization was done
with a tool kit called VisIt,
with a tool kit called VisIt,
123
354639
3606
06:08
originally developed
for physics simulations.
for physics simulations.
124
358269
2861
06:11
Interdisciplinary.
125
361154
1535
06:12
My assignment was to try and come up with
a new way of looking at this
a new way of looking at this
126
362713
3747
06:16
to make it optimal
for the doctors and hospital:
for the doctors and hospital:
127
366484
2685
06:19
How can we make it the most efficient
for them for a diagnosis?
for them for a diagnosis?
128
369193
3564
06:22
And I came up with this image.
129
372781
1912
06:25
It's 2-D;
130
375038
1159
06:26
I took the whole artery and collapsed
everything into a 2-D plane.
everything into a 2-D plane.
131
376221
4099
06:30
I got some very quizzical looks
132
380783
1525
06:32
when I showed this
to the doctors originally.
to the doctors originally.
133
382332
2165
06:34
But I was inspired to do this
representation from my astronomy work,
representation from my astronomy work,
134
384521
3391
06:37
where we've been using these
tree diagrams along the bottom
tree diagrams along the bottom
135
387936
3277
06:41
to understand the structure of nebulae.
136
391237
2116
06:43
Well, we were inspired in that work
137
393377
1673
06:45
from the bioinformatics
and genome community,
and genome community,
138
395074
2365
06:47
where they use these tree diagrams
to understand their gene expression data.
to understand their gene expression data.
139
397463
3635
06:51
They were inspired
by the evolutionary biologists,
by the evolutionary biologists,
140
401122
2838
06:53
who use these tree diagrams to understand
how species evolve and are related,
how species evolve and are related,
141
403984
3894
06:57
the first of which was drawn
by Sir Charles Darwin.
by Sir Charles Darwin.
142
407902
3038
07:00
Here's an example
from his "Origin of the Species."
from his "Origin of the Species."
143
410964
2835
07:03
So, straight from Darwin,
through biology, physics, astronomy,
through biology, physics, astronomy,
144
413823
4375
07:08
back to medical imaging.
145
418222
1342
07:10
Interdisciplinary.
146
420036
1200
07:11
One may say, "Well, is this
2-D representation better?"
2-D representation better?"
147
421260
4472
07:15
I did a study at Harvard Medical School
to answer just that question.
to answer just that question.
148
425756
3913
07:19
And it turns out, if you present
the image on the left to a doctor,
the image on the left to a doctor,
149
429693
4687
07:24
on average, they find about 39%
of the high-risk regions
of the high-risk regions
150
434404
4421
07:28
that could explode or block
your heart and kill you.
your heart and kill you.
151
438849
2901
07:31
On the right, we can do a little better,
152
441774
2183
07:33
and they're able to find 62% of these
high-risk, dangerous regions.
high-risk, dangerous regions.
153
443981
4390
07:38
But we can do even better,
154
448395
2418
07:40
simply by changing the colors.
155
450837
3089
07:43
The rainbow color map is a sin most
doctors and astronomers and physicists
doctors and astronomers and physicists
156
453950
3621
07:47
are guilty of using. (Laughs)
157
457595
2566
07:50
And it doesn't focus the best
qualities of your visual system.
qualities of your visual system.
158
460185
4599
07:54
The human system can see
brightness variation, contrast ...
brightness variation, contrast ...
159
464808
3614
07:58
not really good at that whole
"green-yellow-blue" thing.
"green-yellow-blue" thing.
160
468446
3052
08:01
But now, if you look in the shades of red
161
471522
2457
08:04
and highlight the regions
that are most diseased with dark red,
that are most diseased with dark red,
162
474003
3157
08:07
now doctors can find 91%
of the high-risk regions,
of the high-risk regions,
163
477184
3728
08:10
simply by changing the colors.
164
480936
1727
08:12
(Applause)
165
482687
3279
08:15
And I would have never known
the importance of color
the importance of color
166
485990
3043
08:19
if it was not for my computer science
and visualization collaborators
and visualization collaborators
167
489057
3317
08:22
showing this to me.
168
492398
1155
08:23
So again: interdisciplinary collaboration.
169
493577
2132
08:25
How do you even get
a collaboration like this?
a collaboration like this?
170
495733
2680
08:28
In the case of astronomical medicine,
171
498437
2050
08:30
it started with a Harvard Astronomy
professor, Alyssa Goodman,
professor, Alyssa Goodman,
172
500511
3087
08:33
serendipitously meeting a computer
scientist and imaging specialist
scientist and imaging specialist
173
503622
4413
08:38
from Brigham and Woman’s Hospital,
174
508059
1682
08:39
and their recruitment
of a very adventurous,
of a very adventurous,
175
509765
2139
08:41
open-minded, young student. (Laughter)
176
511928
1938
08:43
And from there, it has exploded:
we've pulled in cardiologists
we've pulled in cardiologists
177
513890
4007
08:47
and computer scientists and radiologists
178
517921
2582
08:50
and astronomers, physicists,
chemists, computational physicists --
chemists, computational physicists --
179
520527
4334
08:54
I mean, we've brought
so many people together.
so many people together.
180
524885
2191
08:57
And it's been enlightening
181
527100
1623
08:58
to share domains and information
across borders.
across borders.
182
528747
3888
09:02
And we're still going.
183
532659
1182
09:03
And although most of the people
up on the screen are from Harvard
up on the screen are from Harvard
184
533865
3209
09:07
or Harvard Med,
185
537098
1151
09:08
now we cross different institutions
and continents to work together.
and continents to work together.
186
538273
3428
09:11
All I can say is,
it has just been wonderful.
it has just been wonderful.
187
541725
2624
09:14
We're continuing to make new discoveries.
188
544373
2201
09:16
And I just urge you:
189
546598
1897
09:18
attend conferences not in your own domain,
190
548519
2981
09:21
read books and journals not
in your own discipline,
in your own discipline,
191
551524
2919
09:24
watch TED talks and come
to events like this
to events like this
192
554467
2642
09:27
and say hi to the neighbor
sitting next to you,
sitting next to you,
193
557133
2754
09:29
because you really never know
194
559911
1497
09:31
where your next great idea
is going to come from.
is going to come from.
195
561432
2996
09:34
Thank you.
196
564452
1199
09:35
(Applause)
197
565675
2914
ABOUT THE SPEAKER
Michelle Borkin - PhysicistMichelle Borkin is a PhD candidate in applied physics. She works with the Astronomical Medicine Project and interdisciplinary 3D visualization techniques.
Why you should listen
Michelle Borkin’s 3-D imaging work uses tools from astronomy to help doctors visualize patients’ hearts. In 2008, she began PhD work at the Harvard School of Engineering and Applied Science. She is a 2009 TED Fellow, and has been granted a National Science Foundation research fellowship.
Michelle Borkin | Speaker | TED.com