Imaging Issues in Soft Tissue Sarcomas

SLIDES & TRANSCRIPTS
Monday, June 17

Imaging Issues in Soft Tissue Sarcomas

Murali Sundaram, MD

Slide 1:

Good morning, ladies and gentlemen. While we are setting up, I would just like to take the opportunity to thank the organizers for including imaging in this program, on the evaluation of soft tissue sarcomas.

As you all probably know, until the advent of magnetic resonance imaging a couple of years ago, involvement of radiologists and imaging in the field of soft tissue sarcomas was sporadic at best.

Since its inception, magnetic resonance imaging has shown excellent contrast and spatial resolution of lesions which, augmented by imaging in multiple planes, has resulted in it being the most widely used imaging modality for imaging of soft tissue masses in general and soft tissue sarcomas in particular.

About 15 years ago, there emerged some feasibility clinical papers using F18-fluorodeoxyglucose and positron emission tomography to attempt to grade soft tissue sarcomas.

Over the past five years, there has been studies from about half a dozen centers in this country and Europe, which have looked at positron emission tomography for grading of tumors prior to biopsy, and using the same modality for evaluating response to neoadjuvant therapy and soft tissue recurrence.

The next 18 minutes or so, I would like to limit my remarks to the strengths and limitation of magnetic resonance imaging and positron emission tomography, and draw a few conclusions.

Slide 2:

Before doing that, I would like to very quickly update you on some technical advances.
In the United States, currently about 50 to 60 percent of CT scanners are multi-detector scanners. These are fast scanners that provide greater detail and resolution. They do not have a direct impact on the soft tissue mass, and I do not think will challenge MRI's preeminent role, but they certainly hasten staging.

It is certainly quite possible to do a non-contrast chest CT scan in about a minute or two, as long as it probably takes to do a PA and a lateral chest radiograph.

In the field of magnetic resonance imaging, there is a move toward higher field strength magnets, which result in improved contrast resolution, stunning anatomic detail; but I don't think it particularly corrects some of the shortcomings that we have with the present field strength magnets.

I think the most interesting advance has been combined PET and CT scanning, which I think will have a significant impact on the way we manage patients with soft tissue sarcomas, in that it would be possible to grade tumors by imaging and, at the same setting, systemically stage them and also use CT of the thorax, which is basically considered the better imaging modality for evaluating the lungs

Slide 3:

The ideal imaging modality is one that would permit accurate detection in adults and children, in the neck, trunk, retroperitoneum, pelvis appendicular skeleton, and then go on to locally stage the lesion, permit the interpreter to provide a plausible histological diagnosis, and then use the same modalities for systemic staging, response to neoadjuvant therapy, and to evaluate for recurrence.

As you know, there isn't a single modality that would do all of this, so we have tended to mix and match.

Slide 4:

From its very early days, one of the things that magnetic resonance imaging showed us was that it could detect very small lesions. This is an intermediate-grade vascular tumor, and this is an image from nearly 12 years ago of a fairly large patient, a lot of adipose tissue, who had fibromatosis.

A second lesion was noticed. It was ignored. While following up for the fibromatosis, we noticed it had grown, and this was a case of multiple fibromatosis.

This has been one of the reliant features of modern imaging, in that it permits detection of very small lesions.

Slide 5:

The second strength of MR imaging has been in staging,

Slide 6:

and has been of great benefit to the orthopedic surgeon, as it has been to treating medical oncologists.

Large tumors: MR imaging very reliably measures size in the absence of edema. The tumor has replaced the vastus intermedius, the vastus lateralis, has partially invaded the vastus medialis, displaced the rectus femoris. The femoral vessels, the sciatic nerve are uninvolved.

The tumor sits on the femur. The cortex has not been destroyed, the medullary canal has not been invaded. The staging is complete.

Slide 7:

Another example would be this tumor, which is in the lower lateral aspect of the deltoid, confined to the deltoid, clearly shown on two axial images. Size, local staging complete and this -- it has been repeatedly shown -- is something that has been very accurately achieved by magnetic resonance imaging.

A synovial sarcoma that insinuates between paired bones, in contact with the cortices of both bones but no accompanying invasion of the bone, and the dimensions are again, very clearly depicted.

Slide 8:

It is a characterization that we have made some progress, but not enough. By this I mean that, by looking at a lesion, permitting the interpreter to offer a reasonable diagnosis, first of all, to separate the benign from the malignant and, if one believes it is malignant, what sort of malignancy it is.

We have largely relied on morphology, location, signal features of the mass as to how the mass has behaved in the compartment in which it arose.

Slide 9:

Our strengths have really been in the detection of benign lesions -- lipomas, hemangiomas, elastofibroma dorsi, and lesions of that nature -- because of their signal characteristics.

Slide 10:

Sometimes we can provide a surprisingly accurate benign diagnosis when a malignancy is suspected, as in this 20 centimeter retroperitoneal tumor, confluent with the left ilium, showing a low signal on all sequences, suggesting a collagen-rich hypocellular mass which we felt was a desmoid, and unlikely to represent a sarcoma. This was ultimately diagnosed as a collagenous fibroma.

Slide 11:

Our problems are with a different type of lesion, but we are still able to characterize a fair number of sarcomas -- well-differentiated liposarcoma; de-differentiated liposarcoma, which almost shows a biphasic appearance; fat and non-fatty lesion confluent, myxoid liposarcomas, when one identifies wisps of fat within a lesion that has a very low signal; tumor mass within the saphenous vein, which would suggest a leiomyosarcoma.

Slide 12:

Our problems lie with lesions like this. It is one similar to the case I showed you, low signal on T1, bright on T2, in our jargon, long on T1, long on T2.

Could this be benign? It is possible. A neural lesion could look like this; a very acellular desmoid could look like this. Could it be malignant? It certainly could be malignant.

Could one predict what the malignancy is? I think it would be very difficult to do so. Could one determine what grade this might be if this is malignant, and the answer is no.

Slide 13:

Here are a few other lesions that highlight the problem, dark on T1, bright on T2. Size is easy to determine. Anatomic staging is accurate confluent with the biceps tendon, brachioradialis, separate from the radius -- but what is it?

That is the problem. This turned out to be a synovial sarcoma.

Here is a lesion that is entirely superficial, less than three centimeters. Again, it is a problem in trying to separate the benign from the malignant. This was interpreted as a fibroushistiocytoma.

Slide 14:

Necrosis in the large tumors can be determined and, in a tumor like this, there will not be much difficulty in predicting that it would be a high-grade sarcoma: it is large, it has breached the compartment in which it arose and, using gadolinium, one sees a large amount of necrosis within the tumor, and this is something that can be estimated prior to treatment.

Slide 15:

That has been one of the shortcomings of magnetic resonance imaging -- and that is in trying to predict the histology of the lesion that we encounter, and it has been estimated that we only do this to the tune of 25 to 40 percent.

This figure could perhaps be improved if you were to say that we were reasonable in suggesting the diagnosis of a non-specified sarcoma in several instances; but we certainly don't do very well in trying to predict its subtype.

Slide 16:

Post-neoadjuvant therapy, MRI has played a useful role in evaluating size. When the patient is being treated, MR imaging is able to determine certainly whether there has been a decrease in size; but, as we know a decrease in size doesn't necessarily correspond with a high degree of necrosis, and that would require using dynamic studying. In almost all of these patients, they would be restaged prior to surgery, at which time one could determine size, necrosis and the margins of the lesion.

Slide 17:

Here is an example. This is a girl who first was seen with a pelvic rhabdomyosarcoma when she was six. She returned last Christmas at the age of 22 with a mass in her forearm.

Because she had a previous rhabdomyosarcoma, it was felt that this could be a second rhabdomyosarcoma or, alternatively, a second primary malignancy, and a synovial sarcoma was suggested in the differential diagnosis. The lesion was biopsied and it turned out to be a rhabdomyosarcoma.

The patient was given neoadjuvant therapy; and this is an example of where one uses the dynamic technique to show that there is no evidence of residual tumor.

Precontrast scan showed a little, faint signal. This is the dynamic study and what one is looking for is immediate enhancement, which one doesn't see, and that is delayed imaging, which looks very similar to the precontrast, and we felt fairly confident about this patient not having any residual tumor, and that is how this patient has been managed.

Slide 18:

Otherwise, we rely largely on morphology for determining recurrence. This is a patient who was referred in with a known diagnosis of what was advertised as a malignant fibroushistiocytoma, and it shows a solid, superficial lesion, external to the lateral retinaculum, consistent with a recurrence.

Slide 19:

Just for purposes of continuity, this is the same case that I showed you earlier which is a neurofibrosarcoma, the lesion in the deltoid. It was removed, this shows post-operative inflammatory changes. There is nothing that is mass-like to suggest a recurrence. One year later, the abnormal signal has disappeared and there is just a local, focal abnormality, and there is no evidence of recurrence.

Slide 20:

When we do identify a mass following surgery, as in this case -- which is a case lent to me by Dr. Kransdorf, who is in the audience -- the tumor was removed.

At three-month follow-up, there was a mass -- low signal, bright signal. Is this a recurrence or is it something that is related to the recent surgery?

A gadolinium study shows that this is a seroma, because it does not enhance. That is something that we rely on, and I think our orthopedic surgeons accept that.

Slide 21:

A similar-looking lesion, where an excisional biopsy had been performed and then the patient was sent in with a diagnosis of an intermediate grade leiomyosarcoma. The patient was restaged. Mass: dark on T1, bright on T2, as the great majority are; some subcutaneous hemorrhage.

This time, using gadolinium, the mass enhances, so one could predict that this is a residual tumor, and that is what this turned out to be.

Slide 22:

Now that largely is what MRI imaging offers in the detection, staging, and evaluation of soft tissue sarcomas. For tumor necrosis, we need to use dynamic MR studies as well as post-processing methods to calculate this.

Slide 23:

Now, positron emission tomography, which largely has been with F18- fluorodeoxyglucose, differs from magnetic resonance imaging in that it measures glucose metabolism, and, thereby, predicts the biological activity of the mass.

There is increased glycolysis where there is tumor. It can stage the lungs and identify involved nodes all at the same sitting in about 45 minutes. Morphologic detail is limited, and, therefore, it is limited for local staging.

Slide 24:

Recent studies have suggested -- prospective studies -- that positron emission tomography can separate the high-grade sarcomas from the low-grade, although there might be some difficulty in separating the low-grade from the benign. It has also been used for determining tumor necrosis following neoadjuvant therapy, and for tumor recurrence.

Slide 25:

What is not known with PET for soft tissue sarcomas -- but is felt would be a problem, based on what has been learned with PET in other sites -- is that one might anticipate false positives in pre-biopsy cases if there is inflammation or infection associated with the soft tissue mass. It is also unclear what inflammation or infection would do in the post-surgical circumstance.

Slide 26:

The potential influence of knowing the biological activity of a mass is that it could save a biopsy. If we could reliably predict a mass that is demonstrated on MR as likely being a low-grade sarcoma or a benign lesion, a wide excision could be performed pathology you could confirm that this is a high-grade sarcoma, thereby circumventing the procedure -- doing a biopsy, determining what grade of sarcoma it might be, and then deciding whether it could be removed or receive neoadjuvant therapy.

The biological activity of the mass could influence systemic staging at the same sitting. If it is a high-grade lesion, one can look at the rest of the body and then go on to looking at the lungs, and then it can also be used for determining tumor response and for post-surgical follow-up, and I would like to show you a few examples.

Slide 27:

This is a large, soft tissue sarcoma confluent with the femur, not invading the femur; and this is a color positron emission scan -- the lighter color showing the greater metabolic activity, this area in red showing necrosis -- and this is a high-grade, necrotic sarcoma.

Slide 28:

This is a rhabdomyosarcoma, pre- and post-neoadjuvant therapy. This is the mass the patient presented with and, detected at the same exam, the presence of nodes in the popliteal fossa and in the groin.

Following neoadjuvant therapy, there has been some shrinking of the mass, increasing size of the nodes with confluence to some of the smaller nodes, both in the popliteal fossa and in the groin.

Slide 29:

This is a patient with known neurofibromatosis, who had a large mass in the left thigh. It shows increased metabolic activity with necrosis, consistent with a neurofibrosarcoma, while a neurofibroma on the contralateral thigh shows minimal metabolic activity.

I would submit that it is this sort of lesion that I don't think we at MR could predict would be a neurofibrosarcoma, which I think early experience with positron emission tomography seems to suggest that it could do.

This is a gluteal sarcoma at the same sitting. One can see metastatic nodules in the lungs.

Slide 30:

This is an example with a recurrence posterior to the knee -- increased abnormal activity on positron emission tomography. MR imaging also demonstrates it, but more clearly defines the dimensions of the mass and its relationship to adjacent muscles.

Slide 31:

This is my final image with positron emission tomography, and I think it is this sort of case that has produced a certain amount of enthusiasm for this technique.

This is a large sarcoma which has a standard uptake value measured at 5.3, and within this sarcoma is another nodular abnormality which has an uptake value of 14.2.

Follow neoadjuvant therapy, a large portion of the mass shows good response with decrease in the uptake value, yet the nodular portion shows very poor response, and this is one that would be considered a very poor responder.

Those who are very enthusiastic about this technique have suggested that this is a very, very good predictor of which areas of the tumor have responded and which areas have not responded, recognizing the heterogeneity of soft tissue sarcomas. I think this shows a great deal of promise for predicting tumor response.

Slide 32:

So, how would one put this all together? I would suggest that, for identifying and local staging, magnetic resonance imaging would remain preeminent.

If you want to grade a lesion prior to biopsy, and if it is an intermediate- or high-grade lesion, systemic staging could be done at the same sitting with a CT of the thorax. CT of the thorax is now standard procedure for patients with sarcomas, but one could combine it with a PET and I think we would discover more lesions.

Now, the response to neoadjuvant therapy -- as to how we want to use PET or magnetic resonance imaging -- really depends on the medical oncologist and the orthopedic oncologist.

If you want to determine response to systemic staging -- and the purpose for determining response to systemic staging is to change your chemotherapy if one is not responding -- then PET would be a simple way to do repeated scans.

Anecdotally, if I might say, we have been able to predict tumor necrosis in bone tumors with magnetic resonance imaging, and I can't really say that I have noticed oncologists wanting to measure tumor response while the patient has been on neoadjuvant therapy with the intent of changing treatment if it is turning out to be a poor responder. However, I would say that if one wants to measure response to neoadjuvant therapy, PET lends itself to that process.

Magnetic resonance imaging, too, can be used for the same process; but it is a little more elaborate, but can be done for determining tumor necrosis.

The evaluation of recurrence -- where I work and in most other places -- magnetic resonance imaging has been used. A case could be made for using PET because, at the same setting, you could stage for local and systemic disease.

I would say that an area that has been, I think, not looked at close enough and would be a very inexpensive, very simple, and very easy procedure to repeat frequently, particularly in the appendicular skeleton, is ultrasound; and there really have been no studies at all that have compared these three modalities for evaluating recurrence. Thank you.