Epidemiology of AAA Screening Basic Concepts for Ultrasound - HD

Introduction

My name is Mark Lockhart.

I'm professor of radiology at the University of Alabama at Birmingham.

My topic will be epidemiology of screening of abdominal aortic aneurysm with ultrasound.

The objectives of this presentation are to discuss the concepts associated with screening for abdominal aortic aneurysm to describe the nine basic principles of screening and to address the role of ultrasound for aortic aneurysm screening.

First of all, I will not be discussing the politics of screening, privatization of screening or ethics of screening.

These are the references that I'll be focusing on.

Why Do We Screen?

So why do we screen? As one of my epidemiology professors said, the purpose of screening for a cancer is to reduce the morbidity and mortality from that cancer among the person screened.

Now that's a very big picture view of it.

And there are operational goals for screening, which include the application of a relatively simple, inexpensive test to a large number of persons in order to classify them as likely or unlikely to have the cancer that is the object of the screen.

Now, that's a mouthful, but it really is a detailed concept.

Now, what are some specific concepts that you need to know regarding screening, the timeline of the disease process, epidemiological concepts, and then again, the nine basic principles of screening and how your test would apply.

Timeline of Disease

So this is a big picture description of a timeline of disease.

At some point, all of us are born And everything is Some cells at some time decide to take a left turn and go the wrong way.

And this is the onset of a disease process.

After a while, the cells, weakness disease, cancer cells, whatever, will be large enough that the abnormality is detectable.

Now, that doesn't mean we found it, but that it can be found at some point.

For whatever reason, an imaging test may be performed that we detect the disease.

And this is the time of diagnosis.

Hopefully that is before the actual clinical symptoms.

This is one of my partners, who pretended to pass out for the picture, to sort of demonstrate a ruptured aneurysm in someone, collapsing.

And this would be the development of signs or symptoms.

And then at some point, all of us will die, and that's the end point.

For most of these.

Biases in Screening

Now, when we're evaluating this process, there are smoke and mirrors that affect some of the research trials regarding epidemiology of screening.

Most of these relate to a bias.

So a bias is anything that's a symptomatic, systematic deviation from truth.

What affects the results of your study that make it different than truth?

Two of the biggest ones are lead time bias.

In this case, you find the disease earlier, but you die at the same time.

So really it forces you to live with your disease longer or at least your knowledge of it.

Length bias is the concept that if you only screen occasionally, you'll more likely find slow growing or indolent disease and it will make your results look better.

This is an example using our timeline.

So clinical disease is from the development of signs until your death.

If we do some screening study, we will now diagnose it before those clinical signs.

But if that endpoint death is at the exact same time, which we hope it's not, then we really haven't done a lot of benefit for the patient, but we made it appear as if they live longer.

That's lead time bias.

Now, length bias is shown here.

We have nine years of an interval and all of these arrows are the development of the disease, let's say a cancer, and then the time from the development of that cancer until the patient died.

So you can imagine if you screened one time during this period, you're going to only find some of the disease, and those are the ones in yellow.

Notice that on average, the yellow ones are longer, slower processes than if you, had everybody.

So in this, the mean survival would be approximately four years.

Now, what if instead of a single screen we screened every two years, now we've picked up a lot more of the disease and we've only got a few aggressive fast growing.

The mean survival on this would be three years.

So by screening more, it actually makes it look like the patients had a worse survival, and the reason is the aggressive short cancers would've been missed by the less frequent screening.

The problem is, as you screen more frequently, you find fewer cases, and so you're paying a lot more to find those few cases.

Also, the few cases that you find are the more aggressive ones, so potentially they may not do as well.

Nine Principles of Screening

So let's go to the nine principles of screening.

The disease must be important.

It should have a prevalent, detectable, preclinical disease.

The disease has little pseudo disease, and we'll go over each one of these.

The test should be accurate for preclinical disease.

The test should detect the disease before a critical point so that you can make a difference in the patient.

The test should be safe, it should be cheap and widely available.

Some treatment needs to exist because if you find it and there's nothing to do about it, you really haven't helped the patient.

And then hopefully if treatment is performed before symptoms, it will be more effective than if you wait till the symptoms.

Application to Ultrasound for Abdominal Aortic Aneurysm Screening

Disease Importance

So we're gonna start looking at each one of these and see how they apply to ultrasound for abdominal aortic aneurysms.

So first of all, disease importance.

Do the mortality and mobil morbidity of abdominal aortic aneurysm justify screening?

Well, if you wait until the patient has a ruptured aneurysm and they go immediately to surgery, the mortality is approximately 60 to 80% it.

If you screen the person and you find it and you're able to undergo an elective aortic aneurysm repair, the mortality is three to 5%.

So you can see this is a significant disease unlike screening for something like brown eyes where it really wouldn't make a difference, screening for this can make a significant difference.

Natural History of the Disease

Next, do we understand the natural history of the disease?

So is it well understood?

Is our understanding supportive of an adequate detectable preclinical phase?

And is there a low occurrence of pseudo disease?

So I'll discuss the first one.

First abdominal aortic aneurysm is prevalent enough, and we do understand it well enough that screening may work.

We understand that 5% of men age 65 have the disease, and then if you wait till later, approximately 10% of men age 79 will have the disease.

Also, we know that there are risk factors such as tobacco use and collagen vascular disease, which can increase the risk.

Further, knowledge of the natural history needs to be, an understanding of the clinical, situation such as the onset of symptoms or signs till death.

That's the clinical phase.

The preclinical phase is when the disease exists but have not yet developed symptoms or signs.

And then the sort of golden part that we really are focused on is the detectable preclinical phase where that there is disease.

Our study can find it, they haven't developed symptoms and we can treat them.

So this is an example.

Here's the clinical phase from signs till death.

And the preclinical is all the way from those first cells having some problem until the clinical signs, but we can't detect those individual cells.

So the key part for us is this point from a detectable disease until the signs.

Pseudo Disease

Now, what is pseudo disease?

Pseudos disease means there is a finding and there's two types.

Type one means there is a finding, but there's no disease progression.

It's a benign process that just looks like an aggressive or a problematic disease.

Type two pseudo disease is a little more difficult.

There is say a cancer say in the thyroid and potentially that thyroid cancer may never cause problems for that patient.

But if you do a biopsy and you find cells, you can call it a cancer.

However, it may not affect the patient in any way relating that to abdominal aortic aneurysm.

Not all patients with an abdominal aortic aneurysm die from rupture.

In fact, a lot of people never know they have it and they die of a totally different process.

Trauma disease, never know they had it.

Test Characteristics

So switching from the disease to the test, we're gonna talk about, test characteristics and how to apply it to the disease.

So when we're evaluating a test, we want to know disease, test specific.

Is it sensitive? Is it specific? And is it cost effective?

Now, again, my epidemiology professor always said, talk about the denominator first.

So sensitivity means of the patients with disease, what proportion will have a true positive test?

So the other way to think about this is a highly sensitive test.

You will reduce the number of false negatives in a specific, in a specificity you look at of the patients without disease.

What proportion test negative.

So in this case, a highly specific test will lower your amount of false positives depending on what threshold value you choose to call it disease that will increase or decrease your sensitivity and will have the opposite effect on the specificity.

Now from a program standpoint, we wanna look at the predictive value positive and predictive value negative.

So predictive value positive is of all the positives, what proportion are true positives?

This depends greatly on the prevalence of the disease, unlike sensitivity or specificity.

And then likewise, predictive value negative would be of the patients with a negative test.

What proportion test with a true negative?

Cost Effectiveness

Something that you'll hear about is, incremental cost effectiveness ratio.

This is cost effectiveness analysis.

And what they try to do is from a broad viewpoint, look at how does that screen compare to other screens or no screen at all in terms of, improving the quality of life and health of the patient.

The way we calculate this is we'll take the more expensive test, which is also usually the most effective, subtract it from the lesser and we will get a ratio.

So say the more expensive test is $40,000 and the least is $20,000, but the life quality adjusted life years is 5.2 versus 5.0.

Well, actually I said that a little wrong because this is actual life years that were affected.

Say the patient after correction of an abdominal oral aneurysm had a stroke or a heart attack or some other life limiting, effect, where that their quality of life is only half as good as a purely healthy patient, well then we would have to reduce this to 2.6 and 2.5.

And now our cost per quality, quality adjusted life year goes up.

Now there has been a trial that looked at this for screening of abdominal aortic aneurysm and they found that the cost for one quality adjusted life year was approximately $71,000.

This is actually a pretty good value and would in most situations be considered cost effective.

Detecting the Critical Point

The next thing is whether or not the test can detect a critical point.

Is it a slow enough disease where you have a chance to make a difference?

And ultrasound for AAA screening actually works pretty well.

So if you have 10 years follow up after a single screen in patients greater than or equal to 65 years old, three centimeters predicts 68% of all patients that will either need subsequent repair or have death from the aneurysm.

A 2.5 centimeter threshold will identify 91% of those patients.

But obviously you're picking up a lot more patients that will never have problems.

So here's a more detailed list by size and you can see that anything under three centimeters, your likelihood of rupture or acute therapy goes significantly down.

And then by the time you're greater than five centimeters, there is a very high likelihood, that you will have problems from your aneurysm.

Safety of the Test

So safety of the test is important.

This is key because the test is occurring now, the potential negative outcome is occurring years later and we discount for every year by about 5%.

So you can understand if you had one person die from your test.

Now that is going to outweigh the benefits to so many people 10 to 15 years from now.

So it's key that it be a safe test.

Well, luckily ultrasound is one of the safest imaging modalities that we have.

And so this fits very well as a screening modality.

Availability and Cost

What is the availability and cost compared to other modalities such as C-T-R-M-R-I?

We have wide access and it is a relatively inexpensive test, compared to other modalities.

Existence of Treatment

Switching over to the treatment, is there existence of a therapy?

So for aaas, there are two very good, potential treatments and numerous articles have been written showing that these are effective.

This can be an open surgical repair or an endovascular stenting.

Both are effective.

However, the stenting actually requires lifelong follow up to look for leaks or if for change in the aneurysm, often on an annual basis.

So there is downstream cost associated with stent repair.

Benefit of Early Detection

For early detection, it's important to consider does the early detection benefit the patient.

Now benefit is hard to show and it's hard to prove in an epidemiologic trial.

And I'm gonna go over some examples from mammography to show you that even a good trial can be challenged.

And then there's the question of whether you look at overall mortality or disease specific mortality.

And then it can be very hard to determine whether there's reduced toxicity from treatment of earlier disease.

Now obviously difference of treating ruptured aorta versus non ruptured, is pretty easy, but of early detection of non ruptured versus later detection of non ruptured, that's less, certain.

However, again, improved survival if the surgery is performed in a non-emergent setting definitely suggests that this is a positive, screening, option.

Lessons from Mammography

Now what are the lessons from mammography?

There were seven randomized controlled trials over 450,000 women, multiple countries.

The largest showed, 31% disease specific mortality reduction in over a hundred thousand patients, but only a 1% all cause mortality.

However, when a group evaluated the seven trials, they basically said there were problems with five of the seven trials, threw them out.

And of the true trials that they had left, they said there was no difference.

This had a great effect on mammography for over a decade and shows you that even large prospective randomized trials can be challenged.

Endpoints for Abdominal Aortic Aneurysm

Most trials do use disease specific mortality.

And the reason is, you know, if you're evaluating for an abdominal aortic aneurysm, should you let a traumatic head injury in a car accident in a patient with controlled disease be included as the outcome, doesn't really make sense.

So what are some end, endpoints for abdominal aortic aneurysm?

Well, there's a meta-analysis that showed that there's no significant reduction in all cause mortality.

Again, older patients more likely have cardiac events or stroke unrelated to, the actual aortic aneurysm, however, there was a significant reduction in disease specific mortality in men.

For whatever reason, there was no significant reduction in the disease, specific mortality in women.

And this had an effect on the government's view towards screening of aaa, in the elderly population because, it fit many of the good criteria, meaning good potential in men, four to 8% of the patients had it.

If there was no abdominal aortic aneurysm at age 65, then your likelihood of death after age 65 from the aneurysm itself was very low.

And by screening, they reduced the disease specific mortality by 42%.

And these are people who were offered a screen.

Now if they accepted a screen, 53% disease specific mortality, so obviously very useful, for screening.

And the range was about 18,000 to $72,000 per quality adjusted life year.

Now because of the results in women, initially the government did not have a welcome to Medicare free screen for women for aortic aneurysm.

And it wasn't until, later political pressure, that it was expanded to women.

Follow-Up Screening

Now one of the goals is to, once you find it is to treat it or if it's not yet at the treatable size to watch it, and once it gets to the right size, then you would have to treat it.

So that means we have to follow it up.

This will add costs, but it will also hopefully reduce the rupture rate.

So to maintain a rupture rate of less than 1%.

There was an article this year that showed if the abdominal aortic aneurysm is three to 3.9 centimeters or anything less than four, probably yearly screen is a little bit overdone.

And they felt that it would be adequate to be every two to three years for screening because of the growth rate.

Once a aneurysm reaches four to 4.4 centimeters, the screening should be every one to two years.

And then once you're at 4.5 to 5.4 centimeters, if the patient is not going to surgery, then you would screen them every six months to one year.

Strengths of Ultrasound Screening

So what are the big picture ultrasound screening strengths, that we see?

Ultrasound in the older population has the potential to have a one-time screen and be done.

You've called out a large number of patients who no longer need any additional imaging.

Now the ones you find you will have to continue follow if they don't go to treatment.

It does have low false positive rate.

There is broad access, it's a rapid technique, and there are low cost associated with this.

So based on the principles of screening ultrasound for abdominal AO aneurysm seems to be a very good option.

Conclusion

So in conclusion, screening has potential to improve disease outcomes in the proper setting, and we go back to O'S nine principles.

Many of the benefits of screening are iis read, meaning that there are lead time and length biases that can make them look better than they really are.

However, screening by ultrasound for abdominal, aortic aneurysm fulfills many of the criteria for a good screening program and seems to be having a positive effect.

Thank you for your attention.