Ultrasound and MRI in the Early Diagnosis of Joint Damage in RA

Online CME Program

Hollis G. Potter, MD

Chairman, Division of Magnetic Resonance Imaging
Attending Radiologist, Hospital for Special Surgery
Professor of Radiology, Weill College of Medicine of Cornell University

Theodore R. Fields, MD, FACP

Attending Physician, Hospital for Special Surgery
Professor of Clinical Medicine, Weill Cornell Medical College

Note: CME credit is no longer available for this program.


Theodore R. Fields, MD

Welcome to our program on the use of ultrasound and MRI in the early diagnosis of joint damage in rheumatoid arthritis.

Our first speaker is Dr. Ron Adler, Director of Ultrasound at Hospital for Special Surgery. He will be speaking about ultrasound applications in the diagnosis of early damage in rheumatoid arthritis.

Our second speaker is Dr. Hollis Potter, Director of our MRI program at HSS, who will be speaking about MRI applications in early rheumatoid arthritis.

Following them, Dr. Sergio Schwartzman, a rheumatologist and an attending physician here, will be discussing the clinical implications of the early diagnosis of joint damage and rheumatoid arthritis with the imaging modalities discussed earlier, as well as the overall approach to early rheumatoid arthritis.

The program modules can be viewed in any order.


Ultrasound in the Early Diagnosis of Joint Damage and Synovitis in Rheumatoid Arthritis

Ronald Adler, MD, PhD

It is my pleasure to have an opportunity to speak to you about applications of ultrasound and detection of early disease in patients with rheumatoid arthritis. I would like to structure this talk to first go over some of the advantages of using ultrasound in the musculoskeletal system. Because I know there is quite a varied experience among rheumatologists in general in terms of their use of ultrasound, I would like to go over some technical considerations, just to bring us all up to date in terms of basic knowledge level. Then we can go ahead and discuss the areas of interest, which are really those synovial line structures that we see in the musculoskeletal system, bursa, tendon sheath and articular structures as well.

The advantages, I am sure many of you are familiar with who have used ultrasound in the past, are that it is very available and it is a relatively inexpensive modality to look at the musculoskeletal system. We often perform these studies as target examinations, where the clinician has a specific area of interest that he is concerned about and because of that ease of applicability, we often find ourselves getting referrals directly from physician offices.

One of the real benefits of ultrasound is, of course, that it is real time so that it allows us to perform provocative maneuvers while we are doing the examination and really accentuate a pathologic process as well as provide direct visualization of needle placement during a therapeutic intervention. Of course, it does not involve ionizing radiation, as is true with some of the other modalities currently being spoken about, and there are no artifacts from indwelling hardware, certainly as far as assessing the overlying soft tissues. And this becomes very helpful in patients where there is an indwelling fixation device or indwelling total joint replacement. Of course, it can be used in those patients who are not compatible with other types of imaging modalities, such as magnetic resonance imaging.

To review a little terminology, when we refer to echogenicity in ultrasound, we are talking about image brightness, and that has to do with the number of scatters that reflect the ultrasound beam within the soft tissues. We can see the structures as either being hyperechoic which means they are bright, such as tendons and bones, or we speak of them as being hypoechoic, such as normal-appearing muscle or fluid collections. And if the material or soft tissue is homogeneously dark, then we refer to that as anechoic, and that would be a simple fluid collection.

Most of the soft tissue detail or morphology that we see when we look at ultrasound images is really due to what are known as specular reflectors. Those are the large scattering surfaces that reflect ultrasound and really show us the anatomic inflammation. If you are used to looking at ultrasound images, you notice that inherent graininess within the image. That is known as speckle and that has to do with the small scatters within the soft tissues that are smaller than our ability to resolve, and they just appear as an inherent graininess, or speckle, as I have referred to them.

Gray scale sonography is the ultrasound that we use that gives us morphological detail about soft tissues. This is formed using a pulse echo technique where we have a transducer sending out a short mechanical burst, and it is really the reflected waves off of the soft tissue interfaces, or specular reflectors, that give us the soft tissue detail.

The image is then formed by the reception of these echos to the transducer. We see that on our ultrasound screen as a gray scale image. Now, since we are looking at small part ultrasound, in other words, we are looking at small joints of the hands and fingers for instance, it is usually helpful to use the highest frequency transducer that we have available to us. And some of the current generations of scanners do have transducers as high as 15 megahertz. There are certain specialized features which are now available in some of the current generation of scanners, such as standard field of new imaging and tissue harmonics, as well as compound imaging. All of those features together allow us to improve our depiction of soft tissues within the musculoskeletal system.

In addition to the morphologic information that we get from soft tissues on gray scale sonography, doppler, as many of you know, plays a very important role in helping characterize the inflammatory component. If we are looking at active inflammation, we really need to have some functional assessment of the degree of inflammatory response, and doppler techniques to help us do that.

We are all familiar with the doppler equation that you see in the lower left hand side of the screen. The doppler shift that occurs when an Insonating beam bounces off of a moving target and we can certainly display that information as a spectrum in our image, but what we perhaps are more used to that is actually more useful for us as clinicians to display it as color information. Now we generally have two options in how we display this information: either directly as color doppler, which encodes an estimate of the main frequency shift in the image, or we can display it as power doppler, information which is actually more useful for us. I will speak more about that in a few moments.

The problem with color doppler is that it is only good for looking at anatomic information and blood flow in large vessels such as in the carotid arteries. It is subject to a number of problems, such as random noise, which becomes an issue when we begin to turn up our color gain if we wish to increase our flow sensitivity. It is angle dependent and it is also subject to aliasing. Whereas power doppler, on the other hand, is really an estimate of the integrated power spectrum and allows us to take the low power information, which is the noise that we generate when we turn up our color gain, into the low power portion of the spectrum, and we can effectively eliminate that in our power doppler image.

It is not subject to aliasing and it is relatively angle insensitive. So these are all features that we find useful. But because of what power does and its treatment of noise is really what gives us that additional flow sensitivity And here is an example. In the upper left hand corner you see a transverse image of a muscle in which what I have done is basically turn up the color gain. So it begins to write noise throughout the image, and you notice that all possible colors are in the image, but even though the noise, as you see in these images, is of relatively low amplitude.

On the other hand, if you look at the power doppler image in the lower left hand corner, which uses the identical doppler parameters, we have basically excluded the low power component of the image. Notice that you begin to see anatomic information appear -- the gray scale component, the muscle and in addition we begin to see multiple intramuscular vessels. So we can retain all of that morphologic information and basically exclude the noise.

Effectively what that does for us is it increases our dynamic range for looking at inflammation. Of course, inflammation is really the name of the game here. We have shown in the past that a variety of inflammatory processes really can manifest themselves as hyperemia on power doppler imaging and we, as well as a number of other authors, have shown that we expect that changes in power doppler correlate very nicely with clinical improvement following therapeutic intervention. And that has been shown in the literature.

There are ways of quantifying the power signal which are available, though not currently in the many of the machines that are used in the United States, but these techniques are available. Unfortunately I don't have time to speak of them today because that is a presentation in and of itself.

Let's review the types of soft tissues that get involved in early rheumatoid arthritis. We will begin with the common synovial- lined structures situated about the musculoskeletal system, namely the bursa. We think of bursitis as abnormal distention by fluid or synovial debris of the bursa, but we can't tell based on morphologic baseline morphology whether or not these correspond to simply reactive fluid collections or whether that represents active inflammation.

Power doppler can be very helpful in characterizing those fluid collections further in terms of assessing whether or not there is active inflammation going on. Of course, ultrasound provides a very convenient method to guide aspiration and therapeutic intervention. Extended field of view imaging can be very helpful because it allows us to assess the full extent of an abnormality, something that we really didn't have available to us a few years ago. For example, this is a patient with rheumatoid arthritis, a transverse image over the humerus, and you see a large complex fluid collection above the humeral head containing nodular soft tissue and synovial debris corresponding to a very abnormal subacromial subdeltoid bursa. This is something we really wouldn't be able to appreciate on a single small field of view ultrasound image. So, extended view imaging has really been very helpful.

Now this is an example of one of the earlier cases that we used to illustrate the fact that power doppler may give us information regarding therapeutic response. On the left hand side, we have a patient with a painful shoulder. This is a transverse image over the humerus. You'll notice on this image, although it is somewhat old -- it does illustrate the point that there is marked hyperemia in the deltoid muscle. The thin black line that you see below the hyperemic deltoid muscle corresponds to fluid in the subdeltoid bursa, which we aspirated that day, and it had a white cell count of well over 100,000.

The image on the right hand side is, again, basically at the same anatomic level, three weeks following incision and drainage; the patient is clinically much better. They re-aspirated fluid here, the white cell count was less than 10,000, and you'll notice there is a marked diminution in the amount of power doppler signal seen in the overlying deltoid muscle. Again, this illustrates very nicely that there is a marked therapeutic response.

Now this is a patient with rheumatoid arthritis. This was one of the earlier studies that we performed looking at therapeutic response. On the image on the left hand side we have a transverse image over the lateral recess of the suprapatellar pouch. The patient is markedly symptomatic at this point, and what we see is a markedly thickened synovial membrane demonstrated by marked hyperemia with the power doppler.

Following intra-articular steroid injection, the patient was asked to come back a week later. On the image on your right hand side, we have a power doppler image over the same anatomic area. You'll notice that there is virtually no flow in this area, corresponding very nicely with the patient's clinical improvement, although there is still thickened synovial membrane. So, gray scale information doesn't always parallel clinical response, whereas power doppler, in our experience at least, seems to do a better job of really following their response to clinical improvement.

Now as is true with bursa -- when we have an abnormal bursa, we see distention of the bursa by fluid and soft tissue -- the same holds true in the case of tenosynovitis, where we are looking at abnormal distention of the tendon sheath by the fluid or soft tissue. The image on the upper right hand side is an extended field of view image of a patient with rheumatoid arthritis who had clinically a puffiness on the dorsal aspect of his wrist. What you see is a localized fluid collection immediately below the extensor pollicis longus tendon. If we turn our transducer 90 degrees and take an extended field of view over the wrist at the level of the fluid collection, we see the extensor pollicis longus tendon, labeled on the left hand side of the image, surrounded by a large fluid collection containing some nodular debris along the periphery as well as some fluid. To the right of that we see both the extensor carpi radialis brevis and longus tendons again with overlying fluid distention of the corresponding tendon sheaths.

While this gives us morphologic information as to the extent of abnormality, it really doesn't allow us to assess the disease activity, and again power doppler is very helpful in terms of looking at disease activity in these patients. On the side, we have a gray scale image over the bicipital groove in a patient who has anterior shoulder pain. We see the long head of the biceps tendon, which is somewhat homogenous surrounded by some abnormal fluid or soft tissue. Since we know the glenohumeral joint is oftentimes in direct continuity with the bicipital tendon sheath, it is difficult to say whether that is normal or abnormal fluid, but we would suspect it based on the fact that clinically the symptoms are along the anterior aspect of the shoulder. But if we look at the image on the right hand side and we see there is marked hyperemia surrounding the synovial membrane of the tendon sheath, we know that this is probably the site of abnormality and, in fact, we injected this patient and he had substantial relief within 24 hours.

Now these are transverse images, on the upper left hand and right hand side of a patient with a swollen wrist and rheumatoid arthritis. This is a transverse image over the common extensor tendon sheath and the tendons themselves are abnormal. You can see they are hypoechoic, and it is a little bit difficult to say what is actually going on around the tendons, whether it is abnormal soft tissue. But again, morphologically, we see that there is an abnormality, but it is very difficult to assess the activity of disease.

Then we look on the right hand side, and we see there is marked hyperemia in the synovial sheath surrounding these tendons. And if we look at the corresponding longitudinal image in the bottom, we see this hyperemia extends over the entire course of the common extensor tendon sheath, so we can see a marked inflammatory response correlating very nicely with the patient's acute symptoms.

Now, the same features we applied to bursa, we can apply to other synovialized structures such as joints. A joint is typically seen on ultrasound as a separation of two smoothly marginated cortical surfaces. If you look at the right hand image, we are looking at the MCP-joint, and we see a smoothly marginated surface on the right side corresponding to the metacarpal head. And we may see a thin band of hypoechoic soft tissue, which is pointed out. That is actually articular cartilage. The adjacent bone that we see slightly to the left of the image is a proximal phalanx. And we typically see an echogenic area intervening between these structures corresponding to some conglomeration of the joint capsule as well as intervening subcapsular fat. This is what we see when we look at a normal joint. Now these principles hold regardless of where we look in the musculoskeletal system.

So now on the left-hand side of the image we have a volar aspect of the wrist. This is a normal radioscaphotrapezial joint, whereas on the right side we have a normal appearing radiolunate capitate joint. So we see nice smoothly marginated articular surfaces with echogenic soft tissue overlying them. That is really what we see when we look at normal structure.

When we see distention of the joint by hypoechoic fluid or soft tissue, then we know there is intraarticular pathology. It can be difficult on ultrasound basically to characterize this either as complex fluid, proliferative synovium, or potentially even iatrogenic material as was true in this case following an intraarticular injection with steroids and anesthetic.

We sometimes have the ability on gray scale imaging to distinguish what is fluid from what is soft tissue. This is a patient with rheumatoid arthritis with elbow pain. Looking at a longitudinal image of the posterior recess of the elbow, we see a thickened synovial membrane corresponding to hypoechoic soft tissue. And if you look centrally, you see more of a hypoechoic collection, which is well marginated, which corresponds to complex fluid within the posterior recess.

But, of course, this is morphologic information. We really can't tell the activity of disease until we turn on power doppler, and you can see that there is marked hyperemia here so we know that we are dealing with very active inflammation. This is true again regardless of where we look in the musculoskeletal system. This is another patient with rheumatoid arthritis and, if you look at the image on the left hand side, that is the volar aspect of the radioscaphoid joint. There is distention of the anterior capsule, as indicated by the small yellow arrows, and if you look at the overlying tendon at the flexor carpi radialis tendon, demonstrating a hypoechoic nodular tendon sheath. But, again, this is morphologic information and it is really not until we look at the right hand image when we turn on power doppler that we can see we are really dealing with active inflammatory process due to fairly pronounced hyperemia here.

Sometimes it can be really difficult to distinguish what is abnormal fluid from soft tissue. As we see, fluid can be complex and can have multiple internal echos. And if there is a lot of nodular soft tissue in there, we may not be able to easily compress it to show that it is fluid. Power doppler can be very helpful. This is another patient with rheumatoid arthritis, who had a painful elbow, and the image on your left hand side is a gray scale image over the posterior recess of the elbow. There is all this hypoechoic soft tissue distending the posterior capsule. This patient was sent to us following an attempted aspiration by one of my clinical colleagues, and it came out as a dry tap. We can see why that is. If we look at the image on the right hand side, there is marked hyperemia throughout the soft tissue and inflammatory panus, and that helps us understand very easily why he got a dry tap.

When we have loss of normal cortical contour, those are things that we see in inflammatory erosions. We may see increased blood flow at the site of erosion. We may see adjacent abnormal soft tissue corresponding to inflammatory panus.

Now this was an interesting case, because this was a patient who was sent to us with abnormality in the wrist on plain radiographs but the MCPs were read as normal. If we look at these images, these are the second, third, fourth and fifth metacarpal phalangeal joints along the dorsal aspect. We can in fact see normal morphology in the second and fifth digits, but if we look carefully in the third and fourth metacarpal heads, there are clearly areas of scalloped erosions and adjacent hypoechoic soft tissue. So these are, in fact, inflammatory erosions that were not completely apparent on the plain radiographs of this patient.

This is another patient who had abnormal soft tissue over the dorsal aspect of the wrist. This is the radiolunate capitate joint. You'll notice on the image on the left- hand side, there is distention of the dorsal recess by very heterogeneous hyperechoic soft tissue. And if you look carefully, there is a white arrow pointing out a small surface erosion in the dorsal aspect of the lunate. Again, if we look at the corresponding functional image, if you want to think of it that way, with the power doppler imaging, there is marked hyperemia correlating very nicely with the fact that this is an inflammatory process. And if you look at our erosion in the dorsal aspect of the lunate, that hyperemic soft tissue goes directly into the erosion, which is something we would expect to see with active disease.

So what is the role of imaging in patients with rheumatoid arthritis? Of course we wish to document both the presence and extent of disease. More importantly, we really want to assess disease activity and then we want to follow therapeutic response. So the question that we have to ask ourselves is what should be the role of conventional radiographs, MR and ultrasound in assessing these patients.

This is an example of the patient, where I hope we are going to begin to address some of these issues, where there is clearly abnormality in the wrist. There is extensive subchondral erosion and sclerosis. There is some distraction of the distal radial ulnar joint, so we are really dealing with relatively advanced disease here. And if we look at the corresponding sagittal proton density imaging on MR, we can see some of these morphologic features very nicely. On the left we have the portions of the radial lunocapitate joint and a bit of the radioscaphoid joint on the right. We see distention of the dorsal recesses of both these joints on the proton density imaging, and these are the corresponding ultrasound images on the same patient. If we look above, we have the dorsolunate capitate joint, and we have the volar aspect of the joint on the image below. We see distention of both the volar and dorsal recesses by abnormal soft tissue. If you look at the scaphoid, the cortical surface is quite irregular, corresponding very nicely to the fact that this is an erosive process. And if you look at the corresponding MR images, which is really where we begin to look at functional information in terms of activity disease, on the left hand side we have a coronal image, which is a proton density image, and then we have the fat suppressed T1 weighted image following gadolinium administration. We see that synovial enhancement, which really gives us information about disease activity in a global sense, and in ultrasound we have corresponding abnormality. If we look at the radioscaphoid joint on the upper left hand side in the corresponding power doppler image, we see marked hyperemia in those soft tissues and the same thing in the dorsal lunocapitate joint on the corresponding power doppler image.

So this really gives us a way of initially documenting the extent of disease, the amount of inflammation and also seeing that correspondingly in the ultrasound image. Using a relatively inexpensive modality, we can follow those patients as a function of time in response to therapy.

In summary, I hope have shown you that ultrasound, performed using state-of-the-art scanners as we do here at HSS, really enables an exquisite depiction of soft tissues and joints affected in early rheumatoid arthritis. Power doppler information really is a strong adjunct to gray scale imaging, because it allows us to assess activity of inflammation in addition to the morphologic information we get from conventional gray scale imaging and magnetic resonance imaging. And ultrasound should play a very strong complementary role in initially assessing these patients and following their therapeutic response after intervention. Thank you very much for your attention.


MRI Applications to Early Rheumatoid Arthritis

Hollis Potter, MD

My charge this morning is to discuss the use of advanced imaging techniques, specifically magnetic resonance imaging, in the detection of rheumatoid arthritis. Traditionally, we are all familiar with the ACR criteria for the detection of rheumatoid arthritis. What is important to remember is that only the last criterion demonstrated in this slide uses x-ray; thus, imaging is not necessary to establish the diagnosis of rheumatoid arthritis.

We are familiar as well with the early radiographic changes that are seen and described on plain radiographs, including the presence of soft tissue swelling demonstrated on the image on the left, and periarticular osteoporosis, demonstrated on the image on the right.

As the disease progresses, we see marginal erosions, which are of course the hallmark of the disease, demonstrated in the carpus in the upper left image, additionally in the distal radioulnar joint and radiocarpal joint on the lower left, and the interphalangeal joints in the image on the right.

With progression of erosions, there is collapse and ankylosis with deformity that is so characteristically seen on clinical examination. This is what we are trying to prevent. Early diagnosis is of utmost importance in establishing and treating the patient with rheumatoid arthritis. Joint subluxation is also found in the late stages of the disease, and again, what we are trying to target is early treatment of rheumatoid arthritis -- before the patients present to the clinician with this degree of deformity and radiographic change.

Standardized radiographic assessment includes the modified Sharp score, which scores areas of erosion (17 areas, ranging from discrete interruption of the cortex to frank collapse), as well as joint space narrowing, ranging from focal narrowing to frank ankylosis.

Upon review of the MRI literature, very interesting things have been found, including 48 patients studied prospectively with early arthritis, in whom it was found that the inclusion of MRI increased their baseline sensitivity for detection of the disease from 77 to 96% and accuracy from 83 to 94% compared with clinical diagnosis.

Michael Ostergaard, who has done quite a bit of the work in MR assessment of inflammatory arthritis, discovered that new erosions were visualized, at least one, and up to five years earlier by MRI compared with conventional radiographs. The same group of authors studied 32 patients with active rheumatoid arthritis with 12-month follow-up and the erosion progression rate correlated with synovial membrane volume. MRI was more sensitive than radiographs in detecting erosions.

In a nice 6-year prospective study of 31 patients, McQueen studied MR and radiographs using the modified Sharp score, which we are all familiar with and use in our clinical practices. The baseline bone marrow edema score was predictive of the 6-year total Sharp score, and neither epitope status nor swollen or tendon joint count was predictive of radiographic outcome.

Now this is one of the first studies that suggested that an MRI performed at the first presentation of RA can be used to predict radiographic damage, allowing for implementation of disease modifying therapy and, indeed, this is where we want to be in terms of early diagnosis of the disease.

We also want to study things on a cellular level. This is where the disease starts and, unfortunately, we are far too often presented with treating the end-stage or the effect of the disease, rather than the underlying cause. So in order to attack it, we have to be able to visualize, on a cellular level, the synovium and the early processes prior to development of bone disease. So why MRI? MRI has superior soft tissue contrast compared with radiographs and, compared with CT, and allows us to directly visualize not just the osseous erosions, but also the synovium, cartilage and the neovascularity that is associated with inflammatory process. On this axial MR image in a patient with rheumatoid arthritis, you can perceive that heavy synovial load in the suprapatellar pouch, the heavy fronds of hypertrophic synovium that are now visualized on MRI, but are not seen on standardized conventional imaging techniques.

Other advantages exist in the setting of longitudinal examinations, as there is ionizing radiation. MRI can obtain direct multiplanar acquisitions and does not require post-processing reformations. MRI can differentiate between joint inflammation that is potentially reversible and joint destruction, as we look through the continuum between synovitis and eventual panus formation.

This is a patient with rheumatoid arthritis, initial presentation. You can see the large focus of synovial load on the sagittal image on the left in the suprapatellar pouch. You can appreciate the early marginal erosion in the tibia seen on the image on the right. The nice thing about MR is that because the soft tissue contrast is so profound, we can differentiate different patterns and distinguish different patterns of synovitis. This is a patient with a degenerative pattern of synovitis, not typical of an inflammatory arthropathy of the rheumatoid type. We do see some hypertrophic synovium in the suprapatellar pouch, but without the conspicuity or without the same morphology as the synovium seen in the inflammatory process shown in the earlier example.

In the same patient, we can see delamination of cartilage, causing a mechanical irritant as a result of traumatic cartilage injury. In other joints, MRI has also been proved efficacious in detecting the degree of synovial load. This is a patient with rheumatoid arthritis with a painful elbow, and you can appreciate not only profound joint effusion, but also the large degree of panus formation and hypertrophic synovium seen in all the joint recesses of the elbow joint. This is causing small areas of marginal erosion that were not seen on an x-ray, demarcated by the arrows on the images shown before you.

This is a patient with plain radiographs. You can see on the left that this patient had rheumatoid arthritis and presented largely with shoulder pain. Clinical concern was exactly where the process was primarily located. You can see the large rice bodies and particulate hypertrophic pannus, seen in the subdeltoid bursa. While the patient does have a small joint effusion, the large volume of disease is really confined to the bursa, helping to guide clinical management in this particular patient.

This is an axial image in the same patient and here you can appreciate that large hypertrophic rice bodies in the distended bursa. You can also appreciate the marginal erosions seen en face in the humeral head. This is an arthroscopic picture of the same patient, showing you these rice bodies that are so characteristic on both MR examination and at the time of arthroscopic synovectomy.

It also is quite helpful in the cervical spine, where of course, disease can have its most serious manifestations. This patient has subaxial spondylosis in addition to inflammatory changes, and unfortunately it is also manifest as an abnormal signal in the cord at the subaxial level. In a more characteristic C1-2 involvement, the arrow on the left is showing you translation of the C1-2 articulation. These are performed in both extension and flexion and can be very efficacious in assessing the degree of dynamic instability in patients with atlantoaxial instability and high cervical disease. Unfortunately, in this particular patient, there is also subaxial stenosis with superimposed degenerative changes, leading to profound stenosis and cord compression identified on the image on the right.

This is a 64-year-old patient with diffuse finger swelling who was known to have carpal disease. Here you can see on the axial images, on the top, the really diffuse hypertrophic synovium distending the flexor and extensor sheaths and causing quite a bit of synovitis, with particulate matter and some marginal erosion at the proximal interphalangeal joint.

Ultimately, what we want to visualize not only the synovium, but also the cartilage. A unique thing about MRI, as opposed to x-ray, and as opposed to CT scan, is that we directly and noninvasively evaluate cartilage. The thin cartilage, however, of the wrists and the fingers presents us with a diagnostic challenge, and that is where the type of MR technique is really important.

A wrist coil and a high field strength are necessary for adequate signal to noise ratio. We are looking at very thin cartilage, and because of that, the dedicated extremity units are completely unsuitable in detection of early marginal erosion of cartilage or an evaluation of synovium that precedes bone erosion seen in this process. Standardized pulse sequences are, of course, very helpful, not only for serial studies for interval comparison but also for multicenter trials.

What I wanted to spend some time on is the use of gadolinium. Gadolinium is an MR contrast agent that develops a magnetic moment and behaves like a "magnet" due to unpaired electron spins. With normal dosages and scan times, it primarily acts to shorten an MR parameter known as T1 relaxation time. We may double the recommended dose for MR, and that is completely safe.

Relative contraindications are prior idiosyncratic reactions which have been described using gadolinium, nursing mothers (as it is secreted into breast milk), and hemolytic anemia, as it has been shown to promote extravascular hemolysis in ex-vivo models. It is not toxic and that is the nice thing as compared to standardized x-ray and CT contrast agents. It is eliminated by GFR, and it is removed by dialysis and can be used in patients with renal insufficiency.

This is a contrast-enhanced angiogram in a normal patient. You can appreciate the detail with which we are able to see the radial artery, ulnar artery, the complete deep arch (this patient happens to not have a complete superficial arch), but also the detail of the digital arteries that we are able to see on a 2-minute study using gadolinium through an intravenous injection. It is important to remember that this is not an arterial injection; this is relatively noninvasive and quite a fast study. Just to show you again the detail. This is a 16-year-old patient who had a vascularized distal radial bone graft. Clinical concern is if we could see the vascular pedicle that had been placed in the scaphoid. To show you the conspicuity with which we see these small vessels, see on the first run that we can actually see the well-vascularized scaphoid, so we are picking up quite small vessels.

When it comes to our ability to depict such small vessels, the utility of contrast enhanced MR angiography in depicting neovascularity in inflammatory arthritis is very interesting. Now when we think about the use of gadolinium salts in rheumatoid arthritis, what largely has been described is not actual depiction of vessels but depiction of inflammatory synovium. Gadolinium is injected and distributed into the interstitial space by a rate dependent on local capillary permeability and perfusion, which, of course, are increased in inflammation. Indeed, histopathologic correlation between synovium inflammation and early enhancement rate has been shown. This requires very fast MR gradients and high field strength imaging, where we are injecting the gadolinium and taking rapid successive images; essentially the whole exam is done in a few minutes.

Michael Ostergaard, again, nicely determined that MR-determined synovial membrane volume correlated very well with fibrin deposition, subsynovial leukocyte infiltration, with a trend correlation to the vascular proliferation, which is, again, one of the hallmarks of the early stages of the disease, where we are trying to find these areas of neovascularity. Synovial membrane also correlated with both blood vessel density and area, based on immunohistochemical assessment of vessels in a nice study discerned by Gaffney et al.

So what are the capabilities? We can do MR angiograms of these small blood vessels, relatively non-invasively, via intravenous injections. We may in fact be able to discern neovascularity associated with inflammation and determine synovium vascularity in regions of inflammation. The hypothesis would then be that serial MRI can track the angiogenesis at an early stage in this inflammatory disease that unfortunately we are catching far too late.

Traditionally what have been used, however, for MRI studies are bone marrow edema, which has been shown in many studies to correlate to synovitis and synovial volume assessment and bone erosion. The latter are confirmed in at least two planes of imaging, and again due to its superior soft tissue contrast and direct multiplanar capabilities, MRI is more sensitive than plain films in detecting these early erosions of bone. Ideally, we want to get it even earlier: we want to directly visualize the cartilage, and in fact we want to look at vascular recruitment on magnetic resonance angiography.

This is a coronal fat suppression technique. This is showing you marginal erosions, manifest as the bone marrow edema pattern and the effusion in the MP joint of a patient with inflammatory arthritis. This is not early disease. This is moderate in the course of the disease, although this is going to disclose the extent to disease to much a greater extent than plain films would. This is a gradient echo technique. It is a nice high contrast technique on MRI, showing you those marginal erosions of the long finger MP joint, as well as multifocal joint effusion indicative of the inflammatory process.

This is a corresponding and complementary cartilage sensitive echo technique. There are techniques developed at this institution for evaluation of cartilage, and here we also visualize the presence of marginal erosions and can also assess directly the effect on the articular cartilage over the metacarpal heads. Unfortunately in this patient, there is more severe involvement of the carpus. You can see quite well-formed erosions, some of which have a low signal intensity or sclerotic rim, indicating some measure of indolence in terms of some of the disease process. Hypertrophic synovium and panus, unfortunately, fills the extensor compartment, with involvement of the MP-joints as well. The image on the right is a cartilage sensitive technique, where we are able to get direct visualization on the intermediate signal or "gray" articular cartilage, to discern exactly how much cartilage is missing from areas of erosion and also quantify the degree of hypertrophic synovium.

Now this is the angiography. Again, a lot of panus formation in the carpus, but what is really interesting is that when we look at the MP-joints in this patient, you can see the areas of synovial contrast enhancement of gadolinium at the second MP-joint. What we are really interested in, beyond that level, are very small vessels. This patient has unfortunately fairly long-standing disease, so I show you this additional image. This is a 29-year-old woman with radiographic evidence of carpal disease. I think all of you looking at these images will say "Well, gosh, I can see all of the marginal erosions in the carpus, I am comfortable that this patient has rheumatoid arthritis based on the radiographic depiction." What was interesting though, when we looked at her on MRI, these are oblique views again, there was significant erosion of the carpus. This is a non-contrast fat suppression technique. This technique is used to evaluate for bone marrow edema, and here we can see a lot of bone marrow edema at the sites of erosion, pretty substantial disease in the carpus. This is with gadolinium. This is several minutes after gadolinium, where we can see enhancement within the hypertrophic synovium, and we can see the marginal erosions, making it very complementary to the technique I showed you just previously.

Now on the cartilage sensitive imaging, we pick up quite a bit of destructive process due to the presence of hypertrophic panus formation in the midcarpal, radiocarpal, and distal radioulnar joint compartments, as well as in the common carpometacarpal compartment. This is what is interesting: radiographs are completely negative. Indeed, the fat suppression is negative in the MP-joints. In this patient, we are picking up marginal erosions on the cartilage-sensitive technique in the metacarpal heads that were not seen on any other type of imaging. And you can see that area of abnormal signal in the subchondral bone, with its effect upon the cartilage of the dorsal margin of the metacarpal heads of the second and third MP-joints.

This is our angiogram. First one, we have a complete deep arch. We notice she is radial dominant. We see a lot of early contrast enhancement in the first few seconds in that hypertrophic panus formation around the cartilage. Later we see a very small vessel approaching the second MP-joint. On the late arterial volume, we pick up some additional vessels. This is essentially a loop reformation combining all three runs. You can see this area of hypertrophic vascular recruitment following an "arch", if you will, adjacent to the MP-joint.

The color segmentation model is a bit more striking; again you are looking at the neovascularity of the MP-joints. The carpus I know is very striking, but what we really want to focus on, and what was occult on all the other imaging techniques, is the areas of vascular recruitment in the MP-joints. This is where we want to be. We want to be detecting the disease before we are already seeing bone erosions on our fat suppression techniques and clearly, before we see them on the plain films. This is where we need high field strength, superior signal to noise, dedicated coils and expertise in MR angiography.

This is a loop run with all different phases, just to show you. You can appreciate again that "arch" of neovascularity coming off, with recruitment of the metacarpal branch of the second MP-joint.

So where are we? We have both quantitative and qualitative assessment using MRI in rheumatoid arthritis. The "quantitative" is that we can provide an actual number of synovial volume using post-processing assessment on a standardized PC. We can provide cartilage volume. We can actually discern vessels and recruitment in the areas of neovascularity. It can also be qualitative, that is a qualitative assessment of erosions, marrow edema and synovitis.

So we conclude that MRI is more sensitive than radiographs in detection of early erosive disease. It has been well documented that the rate of MR synovial enhancement occurs very early, within the first 90 seconds post-injection, and correlates to synovial blood vessel density. MRI allows us to directly visualize cartilage before there are marginal erosions in the bone. The role of MRI in detecting vascular recruitment and angiogenesis certainly deserves greater study. Down the road, we can postulate that there may be tagged MR agents to vascular endothelial growth factor that will provide a more specific marker for areas of active disease. But there is a need to standardize techniques for multi-center trials, improve inter-observer variability, which has been a previous issue in published literature and would be improved by better spatial resolution and improvement of MR techniques. This is why standardization of high field imaging of patients with inflammatory arthritis with dedicated pulse sequences sensitive for cartilage, and using standarized contrast enhanced MRI angiographic techniques is essential, as is an overall reduction of cost to make these techniques ubiquitous to every patient. I thank you for your attention.



Theodore R. Fields, MD

We hope that these three speakers have helped you to understand the use of ultrasound and MRI in the early diagnosis of joint damage in rheumatoid arthritis. Dr. Schwartzman has discussed how these modalities can influence the decision to treat early RA aggressively. Various factors will determine how widely used these modalities will be in rheumatology practice. Further studies of the prognostic implications of these studies will play a major role. Also important will be the quality of images that can be produced with the less-expensive machines which rheumatologists can purchase for their offices. We will need to evaluate whether the images produced by these machines can be compared with the images of the more expensive and higher-resolution machines at academic centers, and whether the published results from the higher-resolution machines will be reproducible on the lower-resolution equipment. Ultimately, cost will be a major factor in how these modalities are employed in daily rheumatology practice.

To get CME credit, you will need to register, view the video or read the text of the program, answer test questions, and submit your evaluation electronically. A certificate for this program will be sent to you by mail. We invite you to check Rheumatology.HSS.edu for other CME programs available at our website.

We hope that this program has been helpful to you in the evaluation and management of your patients with rheumatoid arthritis.

^ Back to Top
Request an Appointment