WEBVTT
Kind: captions
Language: en-GB

1
00:00:00.000 --> 00:00:03.580
In the early days of photography,
all you had were shades of grey.

2
00:00:03.580 --> 00:00:05.950
Every colour on the
spectrum, from red to violet,

3
00:00:05.950 --> 00:00:08.890
was squashed down by the
photographic process.

4
00:00:08.890 --> 00:00:12.440
And actually, early photographic film
didn't even manage that.

5
00:00:12.440 --> 00:00:15.470
It used chemicals that were
mostly sensitive to blue light.

6
00:00:15.470 --> 00:00:18.970
So if something was bright red,
the chemicals would barely react to it.

7
00:00:18.970 --> 00:00:20.689
It'd just look black on camera.

8
00:00:20.689 --> 00:00:22.230
Which is why dark rooms,

9
00:00:22.230 --> 00:00:24.440
where the photo-sensitive
film was developed,

10
00:00:24.440 --> 00:00:28.420
could be lit with red light
without ruining the pictures.

11
00:00:28.420 --> 00:00:32.510
It turned out that film like that
would also react to x-rays.

12
00:00:32.510 --> 00:00:35.130
Now, visible light can pass
through some solid objects

13
00:00:35.130 --> 00:00:37.360
as anyone who's ever held
their hand up to a light

14
00:00:37.360 --> 00:00:38.710
and looked at their veins can tell you.

15
00:00:38.710 --> 00:00:41.743
But x-rays, with a higher
energy and shorter wavelength

16
00:00:41.743 --> 00:00:44.600
pass through objects a lot more easily.

17
00:00:44.600 --> 00:00:47.340
So the picture doesn't get scattered
into a blurry mess.

18
00:00:47.340 --> 00:00:50.110
So an x-ray photo takes
those short wavelengths

19
00:00:50.110 --> 00:00:54.040
and translates them into
something that our eyes can see.

20
00:00:54.040 --> 00:00:55.870
Unfortunately that higher energy

21
00:00:55.870 --> 00:00:58.360
also means that x-rays
are ionising radiation.

22
00:00:58.360 --> 00:01:00.927
They come with a risk of
cancer and radiation poisoning.

23
00:01:00.927 --> 00:01:06.000
Before that was known, fluoroscopes,
which used a constant beam of x-rays,

24
00:01:06.000 --> 00:01:07.430
seemed like a great idea.

25
00:01:07.430 --> 00:01:10.000
Now they are very carefully controlled.

26
00:01:10.000 --> 00:01:12.620
But x-rays aren't just one wavelength.

27
00:01:12.620 --> 00:01:16.400
They're also on a spectrum,
from soft x-rays to hard x-rays.

28
00:01:16.400 --> 00:01:18.513
And there's a team, here in
Christchurch, New Zealand,

29
00:01:18.513 --> 00:01:22.700
who are doing something that,
in hindsight, seems really obvious.

30
00:01:22.700 --> 00:01:25.780
What if you take x-rays at
several different frequencies

31
00:01:25.780 --> 00:01:29.250
and then translate all those frequencies
to different colours?

32
00:01:29.250 --> 00:01:33.140
- Part of the light spectrum
is the x-ray spectrum.

33
00:01:33.140 --> 00:01:35.760
And in there, we've got
a range of energies,

34
00:01:35.760 --> 00:01:38.480
sometimes they're called
soft and hard x-rays.

35
00:01:38.480 --> 00:01:42.980
But the typical medical imaging device
doesn't use that information.

36
00:01:42.980 --> 00:01:45.270
Sometimes it's used in airport security.

37
00:01:45.270 --> 00:01:48.140
They typically have a
couple of detectors, layers,

38
00:01:48.140 --> 00:01:49.560
on top of each other.

39
00:01:49.560 --> 00:01:52.000
And they get those crude signals out

40
00:01:52.000 --> 00:01:55.460
and then present it to the eye
as two different colours.

41
00:01:55.460 --> 00:01:57.820
But we're actually getting
eight different colours

42
00:01:57.820 --> 00:02:00.520
and we're doing a lot
of computer processing

43
00:02:00.520 --> 00:02:03.570
before we get a 3D image
of what the materials are.

44
00:02:03.570 --> 00:02:08.230
So we actually measure,
that's bone, that's water, that's fat.

45
00:02:08.230 --> 00:02:13.280
We started thinking about it--
myself, in about 2002.

46
00:02:13.280 --> 00:02:17.220
The designers of the detector
started 10 years before that.

47
00:02:17.220 --> 00:02:18.960
It came out of high energy physics.

48
00:02:18.970 --> 00:02:22.820
It came out of CERN's research
and trying to find the Higgs Boson.

49
00:02:22.820 --> 00:02:26.040
Equally importantly, they
found a much better way

50
00:02:26.040 --> 00:02:29.410
of imaging inside humans
than they had before.

51
00:02:29.410 --> 00:02:33.344
Each photon that comes in,
we actually measure the energy of it.

52
00:02:33.344 --> 00:02:37.380
Photons come in so fast,
it has to do it very, very quickly.

53
00:02:37.380 --> 00:02:41.670
The pixels are 110 microns square,
so the size of a hair's breadth,

54
00:02:41.670 --> 00:02:45.590
and each one of them is
about three to four thousand transistors.

55
00:02:45.590 --> 00:02:47.690
They're doing a lot of processing.

56
00:02:47.690 --> 00:02:50.310
We've only got a two-dimensional image,
so we've actually got to turn that

57
00:02:50.310 --> 00:02:53.890
into a three-dimensional image.
And that's called computed tomography.

58
00:02:53.890 --> 00:02:55.722
We do a lot of computer processing

59
00:02:55.722 --> 00:02:59.540
and produce 3D slices any way
through the object.

60
00:02:59.540 --> 00:03:01.380
With the resolution we've got,

61
00:03:01.380 --> 00:03:04.420
we've got to solve tens
of billions of equations.

62
00:03:04.420 --> 00:03:06.150
And we've got to try and
do that very quickly.

63
00:03:06.150 --> 00:03:09.420
- The goal is to make it as easy
as possible for people to understand.

64
00:03:09.420 --> 00:03:11.450
For bone, you'd typically use white

65
00:03:11.450 --> 00:03:14.500
because that's what people
think is what bone should be.

66
00:03:14.500 --> 00:03:16.430
But if you're looking at
something like iodine,

67
00:03:16.430 --> 00:03:18.230
a solution of iodine is clear.

68
00:03:18.230 --> 00:03:21.410
So we typically assign red
and most people in medicine

69
00:03:21.410 --> 00:03:24.150
are used to seeing things
assigned to different colours.

70
00:03:24.150 --> 00:03:25.820
It's a combination of what looks good,

71
00:03:25.820 --> 00:03:28.340
what's intuitive to them,
what's easy to understand.

72
00:03:28.340 --> 00:03:31.560
We've got to the point we
can now test things in humans

73
00:03:31.560 --> 00:03:33.480
and we're about to start
our first clinical trial.

74
00:03:33.480 --> 00:03:37.130
The challenges in going
from where we were in 2006

75
00:03:37.130 --> 00:03:40.040
on our first mouse, to being
able to do the first human,

76
00:03:40.040 --> 00:03:42.090
we had to move to the human energy range.

77
00:03:42.090 --> 00:03:44.900
So the x-rays you need
to get through a mouse

78
00:03:44.900 --> 00:03:46.900
are much softer, lower energy

79
00:03:46.900 --> 00:03:49.940
and we need to be able to
do large areas of detectors

80
00:03:49.940 --> 00:03:52.680
so you can actually scan someone
in a reasonable timeframe.

81
00:03:52.680 --> 00:03:54.800
And then there's the data handling issues.

82
00:03:54.800 --> 00:03:58.150
Typically we think of
scanning one litre of volume

83
00:03:58.150 --> 00:04:00.310
giving us about 10 gigabytes of data.

84
00:04:00.310 --> 00:04:04.210
Can you actually scan a large object
and manage that kind of data?

85
00:04:04.210 --> 00:04:07.470
At the moment our scanner
uses about half as many x-rays

86
00:04:07.470 --> 00:04:09.380
as a standard CT.

87
00:04:09.380 --> 00:04:11.740
But if you think of imaging
something like a wrist,

88
00:04:11.740 --> 00:04:14.820
which is not radiosensitive,
it's not a thyroid gland,

89
00:04:14.820 --> 00:04:16.110
it's not a breast gland.

90
00:04:16.110 --> 00:04:19.330
The actual dose from a wrist
scan is very, very low.

91
00:04:19.330 --> 00:04:20.483
Depending on which country you live in,

92
00:04:20.483 --> 00:04:24.320
it's the equivalent of several hours
to a day or two of background radiation.

93
00:04:24.320 --> 00:04:26.948
Most x-ray detectors need
to have a certain amount

94
00:04:26.948 --> 00:04:29.620
of photons before you can measure them.

95
00:04:29.620 --> 00:04:31.360
Otherwise you're in the electronic noise.

96
00:04:31.360 --> 00:04:33.938
Whereas we work at very,
very low count rates

97
00:04:33.938 --> 00:04:38.360
and so we can measure the statistical,
or the quantum noise quite easily.

98
00:04:38.360 --> 00:04:40.560
So we can operate at lower doses.

99
00:04:40.560 --> 00:04:44.590
Our goal is to keep the dose the same
but provide 10 times as much information.

100
00:04:44.590 --> 00:04:47.470
- There are all sorts of different ways
of looking inside the body.

101
00:04:47.470 --> 00:04:49.520
I'm really excited, in the next six months

102
00:04:49.520 --> 00:04:52.760
we're going to scan lots
of interesting bits,

103
00:04:52.760 --> 00:04:55.840
like broken arms and gout in the fingers.

104
00:04:55.840 --> 00:04:57.860
Maybe in a year's time, start looking at

105
00:04:57.860 --> 00:05:02.270
more exciting things like
diseases of blocked arteries.

106
00:05:02.270 --> 00:05:03.370
We can actually say,

107
00:05:03.370 --> 00:05:05.070
not just simply there's
something absorbing,

108
00:05:05.070 --> 00:05:08.800
but whether it is
calcium or iodine or fat,

109
00:05:08.800 --> 00:05:11.760
that's an injury, is there a bit of metal
in there or whatever.

110
00:05:11.760 --> 00:05:13.820
We can answer those questions.

111
00:05:13.820 --> 00:05:19.060
And so we've got a great prospect
of actually seeing a lot of things

112
00:05:19.060 --> 00:05:20.910
that haven't been seen before.

113
00:05:20.910 --> 00:05:24.080
- Thanks to all the team at Mars Bioimaging
and the University of Canterbury.

114
00:05:24.080 --> 00:05:26.610
Pull down the description for more
about them and about their work.