Good afternoon. My name is Laura Fayad, and I'm sorry that I cannot be with you in person. We're going to be discussing bone marrow signal on MRI. When should we call it abnormal? This is a pediatric patient with heel pain. Notice the nodular signal and the calcaneus. Should we call this abnormal? Well, it appears that with development, there are a variety of bone marrow patterns that are normal in children, and these consist of confluent signal, patchy signal, as well as nodular signal in the tarsal bones. Similarly, in the wrist, we can see confluent signal along the thysis or nodular signal in the carpal bones, and this is all normal for development. So when assessing the bone marrow, age matters. The bone marrow is a dynamic organ and one of the largest organs in our body. It is composed of hematopoietically active red marrow and yellow marrow, red marrow containing more water and yellow marrow containing more fat. There's a predictable pattern of development from childhood to adulthood, with red marrow being converted to fatty marrow first in the epiphysis, then diaphysis, and then lastly in the metaphysis. So when evaluating the bone marrow, the Spineco T1, or Fast Spineco is what we use nowadays, is the most important sequence. Why? Because fatty marrow is bright on T1 and there is excellent contrast resolution between a marrow replacing abnormality against normal fatty marrow, in this case, an osteosarcoma. So notice that it's dark on T1 and well-defined against the background of normal fatty marrow. Red marrow, on the other hand, is ill-defined and brighter than skeletal muscle on a T1-weighted sequence. These are children, and so you see normal signal in the metadiaphysis, which is red marrow, and fatty marrow in the epiphysis. And then this is converted to fatty marrow in the diaphysis in the adult. In the pelvis and axial skeleton in general, we can have red marrow, and again, notice that it's brighter than skeletal muscle on a T1-weighted sequence. And again, in the spine, we'll see red marrow, but skeletal muscle may not be visible in the field of view. We can also use the adjacent disc, so red marrow is brighter than the adjacent disc on a T1-weighted FASB. So in this talk, we've already discussed that age matters, location in the bone marrow matters, and we've discussed signal on T1-weighted imaging. If we're not sure what we're dealing with, we can always get help from marrow-specific sequences because at the end of the day, we want to rule out bone marrow disorders. The main three are reconversion, replacement, and depletion. Red marrow reconversion is what you'll encounter most commonly, and most commonly, this is due to anemia and patients who smoke or are obese. Red marrow reconversion occurs in the reverse order of normal development, so here we're talking about reconverting fatty marrow to red marrow, first in the metaphysis, then in the metadiaphysis, and then lastly, rarely in the epiphysis. So this is what red marrow reconversion may look like in the metadiaphysis. It's ill-defined, brighter than skeletal muscle, and in this case, we're starting to get some epiphyseal subcapital marrow reconversion here. Notice the normal fatty marrow in the epiphysis that's preserved in contrast to this patient with myeloma, in which the normal fatty marrow of the epiphysis is gone. So anytime you see abnormal bone marrow extending all the way to the end of bone, just be vigilant that you're not dealing with another marrow-replacing disorder. Of course, when you have severe chronic anemia, you can get some involvement of the epiphysis. As we see in this case, it can be very patchy or it can be more confluent. The key here is, of course, to look at the clinical history and confirm that this really is just related to anemia. However, always be vigilant, and any time I see abnormal bone marrow signal in the epiphysis, I'm careful to rule out other marrow disorders and consider further... Because we don't want to miss a case like this. In contrast here, we have abnormal marrow signal, but this time the marrow is very dark and darker than skeletal muscle on T1, and this is the patient with diffuse metastatic disease. So let's talk about marrow replacement then. So here we're talking about metastatic disease and anything that replaces normal fatty or red marrow. So other malignancies, such as myeloma, leukemia, lymphoma, and then anything else, even primary bone needles. And as we said, the FASP and ECHO T1-weighted sequence is very important for identifying marrow replacement against a background of normal fatty marrow. It does so happen that we will encounter signal abnormalities in the bone marrow that we're not sure about. So here we have a signal in the thoracic spine, there's a pathologic fracture potentially, and another signal as you see here in L2. Well, if the signal abnormality denoted by the question mark was the only area of interest on our scan, then we would be hard-pressed to say that this is metastatic disease. And for this, we would turn to one of our marrow-specific sequences, so namely chemical shift imaging. This is a fast one-minute scan, it's T1-weighted, again, in-phase and a post-phase gradient ECHO. And the premise here is that protons attached to lipid precess with very similar yet slightly different frequencies compared to those attached to water. And so if we have a voxel that contains both lipid and water, such as red marrow interspersed with fatty marrow, or bone marrow edema interspersed with fatty marrow, then we would get a drop in signal on the out-of-phase compared with in-phase. But if we have a marrow-replacing process, so a tumor that is replacing normal marrow, we only have water, and then we would get no drop in signal on the out-of-phase compared with the in-phase. So looking at this patient, we have no drop in the out-of-phase compared with in-phase in the areas denoted by the arrows. So these are marrow-replacing areas compared to the other regions that do show a drop, and those are all areas of red marrow interspersed with fatty marrow. This was a patient who was young and had unsuspected metastatic breast cancer. Now there's often this question about whether chemical shift imaging works with marrow-infiltrated disorders like multiple myeloma. Well, here you have little dots in the bone marrow that are rather small but appear to be marrow-replacing. Well, we can see on the out-of-phase that there is a black line around this little dot, which is a fat-water interface, meaning fatty marrow on the outside and water on the inside. So we know that this is marrow-replacement based on this appearance, and this was a patient indeed with multiple myeloma. Another test we can use that is also fast and non-contrast is diffusion-weighted imaging to sort out what's going on with the bone marrow. So as we know, diffusion-weighted imaging is a method to assess the Brownian motion of water within tissues, and we're trying to separate out tissues that restrict the motion of water versus those that show no restriction. So if we expose tissues to diffusion-sensitizing gradients denoted by a B value, 5400-800, then a B800 image will be more diffusion-sensitive than a B50, and tissues that restrict will hold on to their signal on a B800 relative to a B50 image, whereas tissues that do not restrict diffusion will lose signal on a B800 image compared with B800. So if we look at this patient here who has an abnormality in the posterior right iliac bone and the B50, 400, and 800 image, the 800 in particular, you see that this signal abnormality is retained on this late B value image in the posterior iliac bone as well as here in the more anterior iliac bone where we don't see it as well on the T2, as well as areas in the sacrum, and these are all areas of metastatic disease as seen by DWI due to their restricted diffusion. We talked about children and normal development of the bone marrow. This is a patient who has diffuse marrow signal that seems to be quite dark. It does involve the epiphysis, which can happen in children early on, but the marrow is quite dark and darker than skeletal muscle on T1-weighted imaging. But furthermore, when we look at the diffusion, we see all this increased signal on the high B value image. This means it's abnormal. It's retaining its signals, restricting, and this is a patient with leukemia because normal marrow would be of low signal on this B800 image. Finally, marrow depletion disorders can occur due to radiation, aplastic anemia, viruses, medication, and anorexia. If you look at this patient here, you would think you were looking at a T1-weighted sequence, but in fact this is STIR. The bone marrow is not fatty but full of water, and this is the true T1-weighted sequence without fat suppression. There's no water here. Furthermore, you can see that the subcutaneous fat has also been depleted, and this is a patient with anorexia nervosa. So in summary, we talked about how age matters, the location in the bone marrow, particularly the epiphyseal signal is important. What the bone marrow signal is doing on T1-weighted imaging is also important for sorting out what we're dealing with. And when you're not sure, get help from marrow-specific sequences, particularly in-and-out-of-phase and diffusion-weighted imaging. Thank you very much. What I'm going to talk about is to continue our refresher course. It's really a great pleasure and honor for me to be here, and I certainly thank Laura and the RSNA for inviting me. And I'm going to talk about malignant soft-tissue lesions simulating benign masses, how to stay out of trouble. And I'm very proud to actually do my clinical work at a military institution, and I hope I don't leave you feeling like this after we're done. I've chosen two topics for this, and while I put how to stay out of trouble, one of my colleagues when I was just walking with her said, you should say, this is the scary part of radiology. And I actually would agree with that. So we're going to talk about how we differentiate high water content masses at cysts, or could it be a myxoid neoplasm, and then hematoma versus hemorrhagic neoplasm. When we see a high water content mass, I think our real job is often to say, is this just a ganglion, a synovial cyst, or bursa, or do I need to entertain the possibility of a myxoid neoplasm? And in my experience, myxoid liposarcoma is the one that can fool us most frequently. Now, this is a colleague of mine that I work with at Walter Reed, and she was kind enough to let me take a picture of this large mass, dorsal to a wrist. And because she's a captive audience, we have a beautiful sonogram showing that this mass is anechoic with some post-acoustic enhancement, and we can see a few vessels around it. We all know this is a ganglion. It's got typical intrinsic appearances, and it's in the right location. And several other examples on MR, and I would just highlight two things about these ganglions. One, they are frequently multilocular, as are bursa, as are synovial cysts, which neoplasms are not in general. And I would point to the enhancement, which should be around the rim and septae of variable thickness, but not significant nodularity. Another very common cyst we see is the popliteal cyst, here on CT, with low attenuation, a small septation, and that typical neck communicating with the joint. And we see the same thing on MR, with low signal on T1, and that neck going between the two typical tendons, the medial head of the gastrocnemius and semimembranosus, with a popliteal or Baker cyst. And while we shouldn't have given contrast, we didn't need to give contrast, this is acceptable contrast enhancement in any cyst, in any organ system, and I always like it when concepts cross over organ systems, because then I don't have to remember and memorize a bunch of stuff. Thin, delicate, non-nodular septae and periphery. We see the popliteal cysts much more apparent on our water-sensitive sequences, with high signal and that typical neck going back toward the joint. This patient is a 70-year-old with rheumatoid arthritis, and she was told both by the radiologist and the orthopedic surgeon that she had a popliteal cyst. Well, does that make sense? I think we got two things right, the all-inclusive we. One, this is popliteal, that I will give you. And it is certainly high water content as well. But if we look, it clearly is not really in the right location for a popliteal cyst, it is too deep, it is not superficial, and it is not far enough medial. And in addition, we have some areas of fat within the lesion, which we very nicely see on the sagittal T1-weighted images, and we should not have fat within an area of a popliteal cyst. If you still wanted to stick with popliteal cyst, let's give contrast. Well, is that acceptable contrast for a cyst? And we certainly know that's not true, because there are nodular areas of enhancement. This is a solid but high water content mass, and this is what myxoid liposarcomas look like. And in fact, 90% of them have a diagnostic appearance on imaging. A high water content mass, but has a small component of fat. That is the most common appearance, and in fact, the popliteal region is the most common location for myxoid liposarcoma. But I really want to emphasize that about 5 to 10% of these, in our experience, don't have intrinsic fat. And those are the ones I don't want you to miss, and to call a cyst or an area of a ganglion. Because I defy you to look at this case and tell me that doesn't kind of look like a cyst. I would point out that it's not multilocular, and if I was to ask you, have you ever seen a synovial cyst or bursa or ganglion here at the knee? And you paused for a millisecond, you have told me actually what you need to do next. Now if you Googled high water content masses around the knee, you will see some people have made whole careers about writing articles on those, including myself, in fact, if you Googled that. But this is not the expected location. Here is one on CT, and you can see how low attenuation this mass is. So if you see something that looks like a cyst but anything bothers you, don't be a hero, right? Give contrast to a sonogram, prove it's not a cyst mimicker. And the classic, again, is myxoid liposarcoma. So here is pre- and post-contrast. Is that a cyst? I'm pretty impressed with the diffuseness of the enhancement. This is a solid mass, and here another patient with myxoid liposarcoma. Not only this is hypoechoic but not anechoic, and we can see vessels within it also. So another example of myxoid liposarcoma. And there's such high water content in these, this is why they can simulate cysts. And this is an intermediate-grade malignancy. So if we call this a bursa or a cyst and leave it alone for two to three years, that can have significant medical legal implications and certain dramatic implications for our patients as well. So this is something we don't want to miss, that 5% or 10% that can simulate in an area of a cyst. So the next and only other big topic is the differentiation of hemorrhagic neoplasm from hematoma. Obviously, also very important to guide treatment and prognosis. And in my experience, the most common neoplasms to simulate hematomas are undifferentiated pleomorphic sarcoma and synovial sarcoma. And this is in chart form what I look for to make this distinction. And I will show you these features in several examples as we go through this section of the lecture. Now this is a patient who is older who fell and is on Coumadin. That's a helpful history, and you can see the mass. Now hematomas frequently will have curvilinear areas of high signal representing methemoglobin. But unlike the brain, we can't follow the evolution with the itty-bitty baby doo-doo thing that we were all taught as residents because most of us do not have a blood-brain barrier in our buttock. Maybe you do, but no, I'm just kidding. And so it's not going to go through those kinds of sequences, but we can see the area of hemorrhage quite nicely. I would also point to the prominent surrounding edema. And the more edema I see around a mass in the soft tissues, particularly when it's deep, the more likely I think it is not neoplastic, but it's inflammatory, traumatic, or a reactive process. And I'll explain why that is in a little bit. I would also point out the thin rim of enhancement without any focal nodular areas of enhancement. And the wall around a hematoma will often contain some hemocytin. We often, as musculoskeletal imagers, don't get gradient echo sequences. But look at your protocol images. Some people call it cheating. I personally call it looking at all your information. And if you look at that, many times those are gradient sequences, and you can see a little blooming from the hemocytin deposition. Here in this case, sometimes the location is quite helpful for hematoma. And we see this mass, and we can see it intimately related to the tensor fascia. And we can see the high signal intensity within this morel lavalier lesion or chronic hematoma. I would again point out the thickness of the wall. The wall around a hematoma is often much thicker than the pseudocapsular on a neoplasm, and often will contain hemocytin, which you will not see in the vast majority of hemorrhagic neoplasms. Here the surgeon is delivering this bouncing baby hematoma. And here you can see the resection with the wall. And notice the areas of hemocytin within that thick wall also as we've opened this up. Now this patient, on the other hand, presented with a spontaneous hematoma. And I will tell you, when I get that history, I get the heebie-jeebies. Because I do a little medical legal work, and my real fear is we're dealing with a hemorrhagic neoplasm when I get that history. So this is a CT post-contrast. And in fact, 95% of the volume of this is hematoma. But what the heck is a focal nodular area of enhancement doing in a hematoma? It should not be there. And on the sagittal T1, we can see the large area of blood with high signal, and something that's behaving very differently on the MR. After contrast, we can see peripheral and septal enhancement in the hematoma part of the lesion. And clearly, the solid portion of this malignant neoplasm is enhancing much differently, and with an area of diffuse enhancement. Now on the T2-weighted image, we see the mass is very high signal. And we can see fluid levels in it also, clearly representing hemorrhage. Notice that there is little to no edema around this mass. I think one of the hardest things in radiology is to take inference from something that's not there. Because most hematomas, for a long period of time, will have edema around them. This does not, and that makes me very, very nervous. And there is the solid portion of the mass, and that's what we must biopsy to get the correct diagnosis. Here it is with a matched growth specimen showing you the solid malignancy, undifferentiated pleomorphic sarcoma, the area of the hematoma, and that pseudocapsule that holds the hemorrhage in, so it doesn't get out, to cause the areas of the edema in the surrounding muscle. So that's why we see that appearance pathologically. Now this is a case I was involved with as a medical legal. I usually call myself a pseudo-expert. This is a 26-year-old runner who developed a mass. This was called a hematoma. And certainly large parts of it are blood. You can see it's high signal in both T1 and T2. And perhaps on the pre- and post-contrast, it's a little hard to pick up the nodular enhancement, and I think subtraction images can really highlight that nodular area of enhancement posteriorly and inferiorly within this mass, and you can see it on these subtractions as well. The surgeon went anteriorly and superiorly and evacuated this hematoma, and the pathology showed hematoma and no malignancy. And why did that happen? Because we didn't biopsy the correct area in this mass. This is five years later with, quotes, unquote, recurrent hematoma, now involving multiple compartments, and you can see the pulmonary metastases, unfortunately, in this case. Just to show you why you get the pseudocapsule, you can probably remember back to pathology, blue is bad, that is the tumor where the stars are. And you get areas of pseudocapsule around tumors because the muscle dies around the tumor and undergoes fibrosis, and that's what you're seeing between the arrows and the muscle labeled with an M right there in this particular example. So I think most cases you can tell hematoma from hemorrhagic neoplasm, and it can direct you where to biopsy for solid areas. But there are cases, again, where that biopsy can be directed by what you're seeing on imaging, but most cases I think you can be pretty confident about what you're dealing with. So what I've tried to do is to share with you two areas that I think can fool us into thinking that a lesion is benign when it really is malignant, the high water content mass. If anything bothers you about it, and usually it's not the right location, you've never seen a cyst there, don't be a hero. Give contrast, do a sonogram, make it easy. Because the delay in diagnosis can have significant implications for us and our patients. And then we talked about hematoma and hemorrhagic neoplasm. In hemorrhagic neoplasm, we look for solid nodular regions, prominent surrounding edema or a lack of prominent surrounding edema, and no evidence of hemocytin within the wall. Now we as radiologists, we have a lot of good to give to our patients, but I've tried to share with you how we can stay away from things getting bad, and certainly we never want things to get ugly as radiologists either. So I thank you so much for your attention. Today I'm going to talk about a few orthopedic oncology malignancy mimics. We're going to preview four cases, and you're going to make a mental note of which of the one of the four is actually a true malignancy. And we'll talk as we go through the cases about some principles of trying to distinguish between a mimic and a true malignancy. Note that all of the mimics shown here, the three additional cases, were all thought to be malignancies and were presented at our interdisciplinary orthopedic oncology conference. So if you think it's a malignancy, don't worry, you're not the first one to be fooled. All right, here we go. First case. This is at presentation and one month later. Second case. Third case. And fourth case. All right, I hope you made your decision. We're going to start with the fourth case, which is not our malignancy. So here on these images, you see a T2 hyper intense lesion. And note that even though it looks like it is all fluid signal intensity, I'm going to hold off calling it cystic until I actually look at the post contrast images and see that it is mostly non-enhancing except for a very thin peripheral rim, which is okay for a cyst. And we will see in a little bit why it is dangerous to call a lesion cystic too early. And this cyst is adjacent to the acromioclavicular joint. And in addition to that, we see that there is a full thickness rotator cuff tear and associated severe rotator cuff arthropathy. There's been remodeling of the undersurface of the acromion and the AC joint is also severely degenerated. We see the disc is torn and both the superior and inferior ligaments are torn. The only thing that we don't see here is a tiny neck connecting the AC joint and this cystic structure, but we presume that it is present and that this is a chronic rotator cuff tear leading to an acromioclavicular joint cyst. On to the third case, and I was wondering if any of you guys thought this was the malignancy. Well, it is not. I'm going to show you some additional images that will help to illustrate the point. All right, so on these additional images, we see that there's a geographic region with sharp, non-anatomic borders. And what I mean by that is that it doesn't follow the anatomy of any single muscle compartment. You see here it involves the peroneal muscles, the soleus and the lateral gastrocnemius. And even though it's relatively large, there's no mass effect. In fact, if I took this area away, the signal amyotomy away, the muscles are essentially maintained. The planes have not moved. The fascia are still there. It's almost as if the muscles themselves are not disturbed and there is no mass effect. Now, this geographic area, which has somewhat surpiginous borders, if I use these terms and I applied them to the femoral head, you would probably immediately think of osteonecrosis. Similarly, this is not osteonecrosis, obviously, because it's in the muscle, but it is necrosis or death of the muscle and specifically due to pressure, a pressure effect. The history is that the patient was unconscious for many hours before coming to the hospital due to alcohol. He was lying down on his side. And, you know, here is the ground here and the muscle gets trapped between the fibula and the ground and ends up being necrotic. We see a geographic area of muscle abnormality. It's not actually a mass because it involves more than one muscle compartment in a very quite large area, but there's no actual mass effect. And for those reasons, it's unlikely to be something like a mass or hematoma. It also doesn't follow any specific nerve distribution. And so it's not neurogenic. And once again, if we applied some of these terms that I've used to the bone, you would recognize that as osteonecrosis. Myonecrosis, of course, is due to disrupted muscle blood flow. It could be due to exercise trauma and compartment syndrome and then also to more systemic diseases such as diabetes, vasculitis, radiation, toxins or infection. Diabetes is a common cause of myonecrosis, but it's more diffuse and tends to involve entire muscles and muscle muscle groups rather than more focally, as we saw in the pressure myonecrosis case. And the pattern of involvement is more homogeneous rather than having that serpiginous appearance. And then there's also often more soft tissue with DEMA. Something that is the same, though, is that there is minimal mass effect. So thinking about these things, there is a 37 year old. Here's the case of a 37 year old, which came in with type 2 diabetes and presented with a new onset left hip mass after a prolonged hospitalization. We see this geographic serpiginous area immediately adjacent to the greater trochanter within the gluteus maximus muscle. There's maybe a tiny bit of U1 hyper intensity, predominantly T2 hyper intense. And the question is, is the ideology of what looks like myonecrosis here due to pressure or diabetes following the principles that I just told you, since it is so focal and it is immediately adjacent to the greater trochanter and the patient has a history which would support him being in a supine position for a long time. This is most likely pressure myonecrosis. And also to drive the point home, the patient didn't have any treatment for this mass in particular, and he had a CT two months later and showed that the mass is completely gone so that this is myonecrosis. All right, what about number two? Well, the T1 is the key, as it often is. And this area that looks like fluid is also T1 hyper intense. Now there's this blobby thing in the middle, which looks like maybe on the post contrast images, there's some enhancement, but be wary of saying that without comparing to the T1 pre contrast images. And you can see that the areas that look enhancing are actually intrinsically T1 hyper intense. So, and then also on the sagittal image, you can see that there's extensive edema involving a very long length of surrounding muscle, which is something that's not so typical for a malignancy. Oh yeah, by the way, this one is not the malignancy. Hematomas can, and it is actually a hematoma, and hematomas can enhance around the rim, but just not centrally. So this was a traumatic hematoma. The patient was on Plavix for cardiac stent tackled by a 15 year old daughter while playing a game of Thanksgiving touch football. Of course, the question always comes in here whether this is a true hematoma or it's due to a malignancy that bled. We'll address this a little bit more in the next case. In this particular person, he was starting to have signs of compartment syndrome and a lot of pain. It's not typical for us to aspirate these, but they didn't want to take him to surgery because of all the cardiac issues. So we aspirated, it got 50 milliliters of blood, and he actually did quite well. Here it's a few companion cases. This is of course the classic tumor mimic where we see an extensive area of muscle edema in sort of two zones. The central zone has an area of low signal intensity along the edge, which is the beginnings of ossification. And this is myositis ossificans. And of course, radiographs are very important here, except especially radiographs over time. You can see that the second radiograph here, the patient 19 days later has developed a lot more ossification. And just sort of a side note, the rate of ossification is often proportional to the speed of what you would expect for healing fractures. So this patient, also the history of trauma is often not remembered. These are often young people who are probably very active and may have what they don't consider to be trauma, but enough to cause myositis ossificans. There's also other schools of thought that think that trauma is not even actually an etiology of this condition. But despite the developing room of ossification here, this lesion was biopsied and it was myositis ossificans. Here's a last companion case where we can see that there is an intrinsically T1 hyper intense mass without much surrounding edema. And then one month later, it's developing a hemocynarin rim. This is a hematoma on Coumadin. And finally, here's an ultrasound case just demonstrating what a myotendous junction tear in the rectus femoris looks like. This also can look like a mass, but the history is key here. This was a patient who had acute anterior leg pain after kicking a ball and it was just a hematoma. So this last one was a malignancy. The thing that was confusing about this is he was a 15 year old boy with a questionable history of trauma. He was doing some air kicks in bed when this thing developed, but he did have a history of osteosarcoma in the left pubis. So it was decided that we would re-image later to see if the area, it was actually hematoma. And one month later, we can see that it's slightly decreased in size, but it hasn't decreased enough. In addition, if we look back at the original study, we can see that the thick of the rind, which some thought maybe that's just debris because it almost looks like it's layering. It's not actually in the dependent position. Now this can happen that you don't catch it in the dependent position. If you don't wait long enough, you don't wait long enough for the hemocyturin to settle, but even so this line is not straight, it's more curvy. And so it is quite suspicious. On top of that, the follow-up MRI, we can see unlike that case of myositis pacificans where there was extensive edema around the mass, there's really little surrounding edema enhancement or sort of traumatic inflammatory changes around this mass. So it is suspicious. So it was biopsy. And of course we targeted that rind instead of the central low density portion, which was thought to be hemorrhage. And this is a recurrent osteosarcoma with acute hemorrhage. So yes, tumors can bleed if the patient has a history of cancer or the hematoma does not resolve sufficiently, then biopsy it and be wary of that rind. Make sure you target it for a biopsy. And remember that layering debris really has to be in the dependent position in order to be debris. If it's not there, don't rule it out. Better to follow up to be sure. Thank you so much for your attention. I really appreciate it. So today I'm going to be talking about MSK pitfalls. Don't touch lesions on PET-CT. So we'll just review some musculoskeletal pitfalls on PET-CT in a case-based format. And if you're like me, you're just, when a talk starts, when a musical starts, anything starts, you're like, when is this going to be over? So there are nine cases, so you can kind of pace yourself. All right, case one. This is a 73-year-old woman, metastatic adenocarcinoma of the pancreas, metastatic to the ovary, has been on single agent gemcitabine and presents for restaging. And as this spinning MIP can show you, there's a lot of areas of uptake in the shoulders. We kind of filtered these out at the hips, but the one I want you to focus on is these little things in the chest wall that look like they're in the posterior chest wall. When we look at the axial images, we see that these localize to the posterior chest wall, deep to the serratus muscle, inferior to the scapula. And we can see that these localize to these little soft tissue masses deep to the scapula. So this is a pretty pathognomonic entity, but the patient had good insurance, so we got additional imaging. This was not questioned on PET. It was called what it was, but then the medical oncologist didn't feel comfortable. This is a T1-weighted image, T2-weighted image in post-contrast, and a T1 coronal. And you can see this is a fatty lesion with some soft tissue density intensity on the T1-weighted images. And then the teaching is that this entity doesn't enhance, but we've imaged enough of these to see that they do have some mild enhancement. The key is the fat and the location. And this is elastofibroma dorsi, and you can read about it here. It's a pretty common pitfall on PET CT. They can have increased uptake, and that's why they can become problematic for us. Usually they're unilateral. When they're bilateral, it becomes a little bit easier of a call. Case two. This is a 64-year-old man with multiple myeloma, had a stem cell transplantation three months ago. There's a bunch of abnormalities here. This is the injection site. There's this big mass of myeloma in the scapula, and then this is the area that I want to focus on. And you can see the huge mass in the scapula right here, but I want you to pay attention to the lymph nodes and the muscle. And as you can see, this localizes the scapula and the lymph nodes and the deltoid. We did get MR on this patient. You can see some linear increased uptake, sorry, linear increased T2 signal. I'm going to be doing that going back and forth between PET and MR. And then on the post-contrast images, as we scroll down, we can see that the, we can see the big mass coming off the scapula, and then some big regional lymph nodes, and then a peripherally enhancing collection in the deltoid muscle. So this is kind of an en vogue diagnosis these days. So big nodes, linear deltoid abnormality. This is pre-COVID. The patient had had a stem cell transplant three months ago. And after a stem cell transplantation, myeloma patients have to get their immune system reconstituted. So they basically redo all their immunizations. And in this case, the patient had a pneumococcal shot three to six months after her stem cell transplantation. And what you're seeing is just a vaccine reaction. Case number three is actually one of the more interesting ones. They're all interesting, of course, but this is one that I fell for. This is September, 2015. This is a 42-year-old woman. In 2010, had melanoma diagnosed on the right knee. It was localized disease, negative central lymph node biopsy. Unfortunately, the patient went on to develop liver and mesenteric metastases. And at some point after this, PET-CT in September, 2015, was started on these agents, dabrafenib and tramitinib. So this is the first scan from June, 2016, after the therapy was started. You can see some things have happened. She's undergone some in-transit metastatic resection, but concerning in this patient, or for the radiologist reading the PET, was all these areas of increased uptake in the lower extremities. This was not read by me. It was called, these were called concerning for mets, and the patient was basically kept on the therapy. When we look at this on the CT images, you can see that the area of uptake localizes to the subcutaneous fat, and there's really not much going on on the CT images. Maybe there's a little rim of high attenuation, but it's centrally fatty. So the therapy was continued, and then the PET CT I was reading was the one from August, 2016. I was like, holy moly, this patient has had metastatic disease that has been untreated, and I started looking at the clinical notes to see if they did not read the PET report. I was kind of having a medical legal moment. Again, the PET and the CT images look very similar to the last time. Intense uptake that is out of proportion to the findings on CT images. So luckily, the medical oncologists were ahead of the curve compared to me, and this was BRAF inhibitor-associated pinaculitis. And once you've seen this, you've seen enough to make the call each time. You have to make sure the patient does have the appropriate therapy. It happens with BRAF inhibitors, which is one of the drugs the patient was on, and the appearance is the same, basically like pinaculitis anywhere else. This time, it's associated with this agent. And the two that have been approved by the FDA are dabrafenib and this one, and these are the two names there. The fourth case is a 54-year-old woman with stage four breast cancer. The patient does have residual active hepatic metastases and has been getting treatment on Taxol for about four months. The finding in this case was this intradural uptake at the lower T-spine, and you can see it right here, and localizes to the lower thoracic spine. And if we look on the sagittal recons, you can see that there's also some area of increased uptake in the upper C-spine, not as concerning because this is more diffuse. This one looks more nodular. And the question was, which is the real lesion, the one in the C-spine, the one in the T-spine, both or neither? And it turns out this is neither. This is variant physiological uptake, maximum at the C4 and T11, T12 level. T11, T12 is where the spinal cord has its lumbar enlargement where all the ganglia come out, and similarly in the upper C-spine. So when you see this characteristic location, don't freak out, it's not a tumor. The next case, a 75-year-old woman, left shoulder melanoma that was resected 12 years ago, had one year of vaccine immunotherapy, local recurrence was resected. Unfortunately, a patient developed brain metastasis, a single brain metastasis this year of the imaging which was resected. So currently NED comes in and we have this increased uptake a little bit lower, you don't see the ribs anymore. This is at maybe the L1, L2 level, and very similar appearance, mild increased uptake in the spinal canal. And there it is on the sagittal. There it is, unfortunately, three months later, which basically declared itself as a leptomeningeal metastasis. Fairly common, unfortunately, in melanoma patients. This could initially be mistaken for the normal variant uptake. It's a little bit lower than typical, but as anyone who's red-spine MRs knows, the CONUS can be higher or lower than average. But I'm not sure if this was strictly a miss just because of the normal variant, but just be aware if it's lower than usual to be on alert. Case number six. This is a 68-year-old woman with newly diagnosed endometrial carcinoma. And I want you to draw your attention to an area of increased uptake in the shoulder and in the lumbar spine. This one ended up just being a degenerative uptake. The L-spine lesion was the area of concern. And you can see that it's localizing to apparently to the spinous process. And this was called the spinous process metastasis, unfortunately. If you look at the sagittal recons or even the oblique MIP, you can see that this is a linear pattern of uptake and it's localizing to the interspinous bursa. And this is a common pitfall when you're just scrolling through the axial images and you see the spinous process from one level and then you don't necessarily notice the microscopic gap in advanced degenerative disease in the interspinous space or Bastrop's disease. So this can be very active and it can be a fake out for you on PET. All right, almost done. So this is a 70-year-old man with newly diagnosed lung cancer, which you can see here. And unfortunately, also newly diagnosed chronic lymphocytic leukemia. This was a staging PET-CT and obviously you can see the big mass in the left lung and the astute PET reader noticed this increased uptake at the shoulder. When you look at it on axial images, you can see that it's localizing to the rotator interval. And the double malignancy was a big enough concern that they raised concern for this for metastatic disease or leukemic cells. As MSK radiologists, we know that rotator interval is a special place for us. This was an MR that was done for further workup of this lesion. You can see that there's effacement of the fat in that area. It enhances, which any inflammatory thing does. And then this was a tumor protocol. We didn't get the nice images, but you can appreciate some thickening of the inferior capsule over here. The additional history may really push this past the goal here. Shoulder pain, limited range of motion. This was adhesive capsulitis. So we went back and looked at a, we published this as a case report and then looked systematically. This is actually fairly common. So your NukeMed colleagues may call you about this. About 0.5% of patients have rotator interval inferior capsule uptake. So about 200 PETs or so, they'll run into it. And you can basically do a good job of diagnosing adhesive capsulitis on PET if you see uptake in these areas. So just something to consider. 70 or 40-year-old woman with stage 3c melanoma, right foot primary, the 3c means nodal metastases. And you can see that the patient has some right inguinal uptake, but there's something also at the left hip. So let's get to the axial images. You can see there's some increased uptake in an external iliac lymph node and some intra-articular uptake at the contralateral side at the left hip. So right now this patient is stage 3c. If we call this a MET, then we're upstaging to four. So it's important to recognize this as kind of a non-malignant uptake. It's very uncommon for metastases to go inside joints. And when we did a retrospective review with a very talented undergrad, we actually found only one case of an intra-articular tumor and it was lymphoma and the patient already had widespread disease everywhere. So if you see this, this is just a chart of all the time course of these. Mean SUV was about five. More common in males. For some reason, it was most commonly left-sided. They can last up to a year and there's no association with symptoms. And the last case, this is a 62-year-old man with chronic renal failure, mycosis fungoides, post-stem cell transplant. Unfortunately, had GVHD, graft-versus-host disease, which was treated with immunosuppression and then developed a fungal respiratory tract infection. And you can see these areas of increased uptake in the bones in the lower extremities and also maybe a proximal hemorrhoid. When we look at the actual images and correct for misregistration, a little bit mentally, you can see that these are actually located on the surface of the bone and not in the medullary cavity. So here's the multiple choice question. Let me simplify the clinical history. Chronic renal failure, fungal respiratory tract infection. This ended up being voriconazole-associated periostitis. Now you've seen it on PET. I'm sure you've all seen it on x-rays and CT. It's basically a fluorosis because of the voriconazole loading dose followed by the maintenance dose. Your kidneys get rid of it usually, but if you have renal failure, you end up with this voriconazole-associated fluorosis, basically. Acknowledgement to Dr. Ginsberg, who came up with the PETFALS term, and Dr. Rorin, whose presentation style I've stolen. Thank you. What I'd like to do in this session is talk about osseous lesions and specifically when to worry. And at the end of this session, what I'd like you to be able to do is to recognize the radiographic features that characterize a biologically active or aggressive intraosseous lesion and be able to identify the appropriate follow-up recommendations that you can make. Now when we analyze lesions, there's a number of things we look at, but that's not what we're going to talk about. We're going to talk about when it's just you and the solitary lesion of bone. When you're in the ER, when you're working at night, when you get a case and the only history is patient fell, rule out, fracture, and you have this radiograph that you're looking at, it's got a bone and that bone has a big hole in it, what do you do next and what do you recommend and what is the appropriate thing and when do you need to worry? And when we have a solitary lesion of bone as an isolated finding with no history, what we have to look at really is the matrix, the periosteal reaction and the margin of the lesion. The matrix really is the key to the histology. We could see an osseous matrix, a fibro-osseous matrix or a cartilaginous matrix, but that doesn't help us with the biological behavior of the lesion, which is really assessed by the periosteal reaction and most importantly, by the margin. And if you think about it, bone provides the structural stability to this skeleton. It is incredibly strong and when we look at the femur, the longest of the bones, it's like a strong column with a thick wall that's filled with fat and as that continues and goes to the articular surface, when it flares, that thins and the structural integrity is supported by the trabeculation of the cancellous bone that replaces it as that cortex thins. And when we put a hole in that bone, it creates alterations in the distribution of the biomechanical forces and the stresses that are generated and the bone responds to that change in the distribution of those forces and stresses and given sufficient time, it will form a sclerotic rind around that hole and that's the basic principle. In bone, we say slow growing is indolent and therefore benign and rapidly growing is aggressive and malignant until proven otherwise, something that is true for bone but not for soft tissue. And we know that from the work of Willem Lodwig who really studied the patterns of osteolytic bone destruction in fibrosarcoma and correlated that with patient survival. And he divided them into three groups and this is what you learned back in the AFIP, the geographic group where there was, he defined as a single large, well-defined hole in bone or a few confluent holes and he noted those patients with fibrosarcoma that had that pattern had the best survival with many survivors. Those with primitive on the other end of the scale which he described as multiple tiny holes which fade imperceptibly from a zone of maximal involvement to a cortex that's quite intact had very few survivors. And those that were moth-eaten which he called many scattered and confluent holes giving the impression of arising from multiple foci was in between those two. And then he studied more patterns of osteolysis and he studied patterns, behaviors of various levels of malignancy and eventually had five groups that we learned about again at the AFIP. Type one was geographic which he subdivided into three types, types one, two, and three or A, B, and C rather. And then of course the primitive and moth-eaten types. And it's interesting, he used those terms geographic for those. Now if you remember back to AFIP was a little different than it was then and he used 14 variables to describe each of those patterns. Well, if you think back to when you were at the AFIP and I was there in the early 1980s, this is how I learned the Leibniz classification which was quite different. And that was because John Madewell who was the chairman and registrar in the 70s thought it was too difficult for Lodwig's classification system for residents to use and he changed it and he's the one that came up with this concept of well-defined with sclerosis, well-defined with no sclerosis or ill-defined and that's the one that we use today. So in reality when we thought we learned the Lodwig classification, we actually learned the Lodwig-Madewell classification. And in addition, in his classic articles from the radiologic clinics of North America in 1981 with Ragsdale and Sweet, he really added two additional patterns of osteolysis or margins, that of a change in margin or change over time and a combination margin as well as the invisible margin and the invisible margin was a big deal back in the 1980s and today with MR it is not uncommon at all. Some of you may remember this slide from your time at the AFIP but it really shows the concept of margin quite well. If we take a look at the interface between the tumor and the host bone and we look at it at higher power, we can see the advancing margin of the tumor and remember there's only two things that remove bone, orthopedic surgeons and osteoclast. And you can see the advancing margin of the tumor and the osteoclast that it stimulates which remove bone. And it's the rate at which that bone is removed that corresponds to its biological behavior. So it can be slowly growing where we'll see sclerotic margin it can be relatively more aggressive and if it stimulates the resection or the removal of bone or osteolysis at the rate at which it grows, it can be well-defined without sclerosis. A faster, more aggressive pattern is ill-defined, more aggressive yet is moth-eaten and more aggressive yet is permeative and the most aggressive is the invisible pattern. And we can use that information to decide what we need for follow-up. Now Maybell also introduced the changing margin to show a change in biological behavior which is nicely demonstrated in this case where we can see an area of sclerosis and along the lateral margin of the distal femur an area of complete destruction. Nicely shown on CT as well with indolent and aggressive. This benign process undergoing malignant transformation in this fibro-osseous lesion with secondary malignant transformation to an undifferentiated high-grade pleomorphic sarcoma. And of course he also introduced the concept of invisible which is nicely demonstrated in this case with a radiograph of the ankle which was interpreted as normal with the corresponding MR showing this tumor growing so rapidly that it's encasing the trabeculae before they can be described. So there's no lucency on the radiograph. And it's interesting to note that studies have shown that 30 to 50% of the cancellous bone has to be lost to be recognized as a lucency on radiographs. In contrast to cortex which even small holes can usually be identified quite readily. So this is what we have now for the assessment of osseous lesions. And just to put it in perspective when Lodwell proposed his system he tested it with a test library with the results of that test library of 223 cases showing increasing grade of malignancy from that geographic 1A, 1B, and 1C to that of moth-eaten and permutative. Interesting to note he had no cases of osteomyelitis or tumor-like lesions, but these were all tumors but there was an increase in grade. And this is from his early studies when he did this. And there was a study that was published in 1994 which showed quite different numbers for benign and malignant in comparison to what Lodwick had. And the reason for this is quite simple because the people who did this study didn't really use the original Lodwick study but they used the Lodwick-Madewell grading system rather than the Lodwick one. And I bring this up because as we talk more about it the Lodwick-Madewell system is really what we've used today and this is the system that, or the chart that Lodwick used for his 14 characteristics to assess it. And you can see the difference in benign, particularly in the 1B and 1C, which we now use for our well-defined nosclerotic and ill-defined geographic lesions. So that brings us to the 21st century. And more recently a modified Lodwick-Madewell grading system has been suggested and it's presented here. And if you notice the numbers are really quite similar to what they really got with the Lodwick-Madewell system. And what's nice about this is that we can characterize things very easily. And we know that those with a geographic sclerotic or geographic well-defined nosclerotic margin are generally benign, the overwhelming majority. And certainly with their sclerosis it's probably well over 95 and in one study 99 plus percent. So when I see a lesion and I have no history and I have a sclerotic margin, I'm very happy to dismiss it. And most often there are things like old non-ossifying fibroids, occasionally atrocious lipomas, sometimes fibrous dysplasia but they're typically benign and it's an easy diagnosis to make. If I'm not 100% sure I may get a follow-up but that's pretty much all I do. When I have a lesion that's geographic with no sclerosis that's a little more problematic because those have a higher incidence of malignancy. While it's still small, it's there. And we also know that metastases and myeloma can also have that appearance. So on these I'm a little more cautious. If I see a younger person and I think it's an enchondroma or I think it's a cyst, I'm happy to get a follow-up. And I usually do it in three months if there are no old films to compare with. If it's an older patient, a mature adult, I may get an MR to look at it. And again, if there's any concern whatsoever, usually an MR or follow-up but those I typically will not just dismiss. Those that have a geographic margin that is ill-defined all get oncology consultation because about half of those will be malignancy. And certainly any of those with a changing margin or permittive or moth-eaten pattern of osteolysis will also get an oncology referral because they certainly are evil until proven otherwise. So in summary, radiographs, although they may not be specific for a diagnosis, they certainly accurately depict tumor growth rate and the biological potential for individual lesions. Often we can get a definitive diagnosis but when we don't, certainly we can use the characters of the lesion to identify its biological potential and therefore we can use that to direct follow-up appropriately. Anyway, the point I'd like to make is that careful attention to detail will accurately predict when bone goes bad. Thank you.

bone marrow MRI

pediatric heel pain

T1-weighted sequences

osteosarcoma

red marrow reconversion

myeloma

metastatic diseases

chemical shift imaging

myxoid liposarcoma

PET-CT pitfalls