0:00

So I want to walk you through some of the primary factors

0:04

and how each of the factors impact both the image

0:08

quality and the radiation dose.

0:11

Among them,

0:13

one of the most fundamental primary factor is tube current.

0:17

The unit of tube current is mA, milliampere.

0:21

Ampere was a French

0:23

physicist who did a lot of fundamental work on tube current.

0:26

That's why the unit is mA.

0:30

mA is called milliampere, one-thousandth of an ampere.

0:34

Typical mAs we used in CT, if you look in

0:37

any of the patient images, you can see the mA value

0:41

displayed on which each image is. It's usually 10

0:45

mA, 100 mA, 200 mA, and so forth. That's the range we

0:49

are talking about.

0:50

So what is so important about tube current?

0:53

So if

0:55

you imagine the X-ray tube, it's like a light bulb.

0:59

When the X-ray tube is turned on, there is a filament which heats up.

1:03

When it heated up, the electrons are released.

1:07

It's called thermionic emission.

1:09

These electrons coming out from the cathode end, the filament of the X-ray

1:13

tube, now move towards the anode,

1:17

where there is a high melting point material is used, so this

1:21

electron will bombard on the anode, creating an

1:24

X-ray, such as Bremsstrahlung X-ray or characteristic X-ray.

1:29

Therefore, the tube current is in a way

1:32

directly influences the amount of X-ray produced in an X-ray tube.

1:38

Along with the tube current, we also commonly use what is called a tube

1:41

current-time product, mAs,

1:44

and that's purely multiplying the tube current with the scan time

1:48

per rotation. When I say scan time, this is the

1:52

time for the X-ray tube to go around the patient in one

1:56

rotation. It's not the total study time or the total

2:00

CT scanner time, it's just per rotation.

2:03

mAs is the tube current multiplied by the scan time per

2:07

rotation.

2:09

The factor which influence the image quality is the image

2:13

noise,

2:15

and they go in opposite direction.

2:17

So the image noises decreases with increase in tube current.

2:23

And the fundamental relationship is this one.

2:26

The increase in tube current increases radiation

2:30

dose linearly,

2:32

which means higher the tube current, higher the patient

2:36

dose.

2:37

And that principle works very straightforward with tube current.

2:42

For example, in this one, in this panel of images, you

2:46

see how the tube current can impact the image noise or image

2:49

quality.

2:51

There are four panels of images here, the same images of the abdominal

2:55

images

2:56

you have acquired at 200 mAs. That is the mA time,

3:00

the second 200 mAs. This is done at 200 mAs.

3:04

You can see the object in the kidney here.

3:07

You can also see some low contrast objects in the liver.

3:11

However, if you decrease the tube mAs to 150,

3:16

you begin to see, then notice the image becoming slightly grainier.

3:20

If you go down further to 100 or 50 mAs,

3:24

it becomes quite grainy. The image noise increases because there are

3:28

less amount of X-rays participating in image generation,

3:32

therefore, the image is noisy. And you can still see this

3:36

object,

3:37

even though it is kind of like a diffusing, is merging with the background.

3:41

But the object you were seeing in the liver is gone.

3:44

It is almost you are not able to see.

3:46

Therefore, in this panel, the renal cyst is

3:50

observed at all four radiation levels,

3:54

whereas the small vessels observed in these liver

3:57

images is gone.

4:00

This begs to the question is, which mAs should I use?

4:04

The answer is the mAs should be determined based on what you

4:08

are looking in the image.

4:11

If you're looking only for the large object here, conspicuous

4:15

object, you don't need a 200 mAs. You can use what is called

4:19

as the low mAs, therefore, resulting in lower radiation dose to the

4:23

patient. If you're looking at a very fine object, very

4:26

low subtle contrast object, then you may have to increase the tube

4:30

current higher.

4:32

I'll show you more example this when I do the cardiac CT, where

4:36

in one instance, we use very low techniques such as calcium scoring,

4:40

and we use very high technique for CT angiography.

4:44

And for the principle here is this is what the image noise can

4:47

obscure low contrast object.

4:51

So to summarize this tube current concept,

4:55

the tube current indicates the quantity of

4:59

X-rays produced.

5:00

The take home message of this one is tube current is an indicator of

5:04

the quantity of X-rays.

5:07

It is the amount of X-ray produced in X-ray tube, and the

5:10

relationship of patient dose varies linearly with tube

5:14

current. This is one of the fundamental principle behind the tube current.

5:19

On the other hand, if you decrease the tube current

5:23

or change the tube current by 50%,

5:27

that will decreases the radiation dose by 50%, but

5:31

will increase the image noise by square root of two.

5:35

This principle applies for the conventional way we

5:39

image reconstruct, where it depends on the Poisson statistics

5:43

and the number of photons contributing to the image.

5:46

So for simplicity purpose,

5:48

increase the tube current, image looks very nice.

5:51

The radiation dose is also high to the patient.

5:54

If you decrease the tube current, the radiation dose to the patient is low.

5:58

However, image becomes very grainier, and it depends on the decision

6:02

factor which of these images is sufficient for diagnosis

6:06

is the way to go towards optimization of this CT

6:09

protocol.So here are shown three different panels of

6:13

abdomen image. These are simulated images.

6:16

As you can see on the CT images, look at all the information available

6:20

here. This information is available on the scanner

6:24

depending on the way the PACS is set up.

6:26

In this particular image, it tells what kV was used on this

6:30

image, what mA was used, and what was

6:34

the scan time. This is a 0.5-second rotation time.

6:38

The number of detectors used is 32 detector times

6:42

one, one millimeter detector, and a table speed of 10 millimeter per

6:46

rotation table speed. And you also have in here

6:50

the window width and window level, which we talked

6:53

earlier when I talked about the fundamental display, the

6:57

impact of Hounsfield unit on the display.

7:00

Window level is 50, means it's slightly higher than the water, and then the

7:04

window covers all the way 150 to

7:07

-150, all the images, all the objects in this window is displayed

7:11

here.

7:13

Now, in these three panels,

7:16

this is done at three different mA,

7:19

48 mA, 69 mA, and 184.

7:24

In order to understand the variation of the tube current, the way

7:28

we physicists do is we keep all the other factors same.

7:31

So you have everything else is same, we only change the tube current.

7:35

By doing so, we see that the DCT radiation dose

7:39

index, which I'm going to discuss more detail about the radiation dose index

7:43

later, but the CTDI volume, let's for our case

7:47

right now, look at the relative variation.

7:50

Four milligray, six milligray, and 17 milligray.

7:55

So the question is, do we need this image to diagnose something

7:59

in the patient image?

8:01

If not, we can get away with this mA or this mA.

8:04

That's what the principle of dose standardization or optimization

8:08

comes into picture.

8:10

So regrouping the same principle,

8:14

decreasing the tube current reduces dose linearly.

8:18

Shown here is the tube current setting,

8:21

correspondingly the effective dose setting.

8:25

Again, the effective dose values I'm going to explain in more detail when we

8:29

talk about radiation dose for CT radiation dose, the symmetry, and

8:33

so forth. But for relative comparison, showing here is like

8:36

if this increases linearly,

8:39

50 to 400, eight times it increases

8:43

to 10 millisievert. That's a straightforward linearity.