Well I think it's time to start. Welcome everyone to one of the best sessions of this meeting, adrenal imaging. My name is Irina Benkos and I'm an adrenal endocrinologist at Mayo Clinic in Rochester and I am absolutely pumped to hear this session because this is all I do this the whole day and having Dr. Blake here with us to present the whole one hour on distinguishing different type of tumors using imaging is a big treat for me and hopefully for us as well for you as well. So I would like to introduce Dr. Blake. Dr. Blake is associate professor of radiology at Mass General Hospital. He is a social professor belonging to the Harvard Medical School. He published a lot on adrenal imaging and I'm suspecting we'll have a lot of images to look at in the next one hour. So at the end of the sessions please use the microphones that you see at the both sides of the room to ask your question. Please briefly introduce yourself and keep your questions short. Thank you. Thanks so much Dr. Benkos for that kind introduction and thank you to ACE for inviting me to speak to you today. It's a real honor to be presented to you today at the annual meeting. I have no relevant financial disclosures. I am a member of the Society of Abdominal Radiology disease focus panels for adrenal and neuroendocrine tumors. So by way of outline I'll start with just some background and then go over the relevant anatomy and then really show how by using the different major modalities of CT, MRI and PET we can differentiate between adenomas and non-adenomatous lesions. I'll also show you on these modalities the imaging features of adrenal disorders and then go over some of the imaging and strategy guidelines. So there's no need to convince this audience of the importance of the adrenals. Despite their small size they play a critical role in physiology and endocrinology. They also are critical in oncology because many primary malignancies metastasize to the adrenals. Indeed it's said that it's the commonest site of metastatic disease per unit weight of any organ. So it's very important in staging. And then radiology has produced this dilemma for all of you that with our improved high resolution imaging we can now detect a lot of small adrenal nodules on studies that were not designed to pick up an adrenal lesion. We see these so-called incidentalomas that are done on exams with an indication for a non-adrenal related reason. And this has led to a number of white papers and guidelines from the radiology and endocrine societies and we'll come back to this later. But as to start with the normal anatomy, these as you know are paired retroperitoneal organs that lie superior and anterior to the kidneys. And they have a body and two limbs so that they have a typically an inverted Y or inverted V shape on both axial and coronal images. And the limbs should really be about three to four millimeters and not thicker than five millimeters. And usually they are separate bilateral organs but occasionally they can be fused across the midline. This is more common in neonates because it's often associated with congenital anomalies. But occasionally we see this in adult patients too where you see fusion across the midline. And when it's like this we call them horseshoe or butterfly adrenals. And when my radiology colleagues need to be convinced about the importance of the adrenals I point out what happens to the different organs when the body is in shock. This is a patient who has a GI bleed and we see that a lot of the other organs have reduced enhancement but the body prioritizes enhancement to the adrenals and they really are avidly enhancing. And the reason why they can avidly enhance like like this is that they have a very strong arterial supply. They have a triple arterial supply from the phrenic artery and I don't know if I'm to stop but also a direct supply from the aorta and also a branch from the renal artery. And so this is why they can enhance so so avidly. And in terms of their venous drainage there is a direct drainage to the IVC on the right side and a more circuitous drainage to the left renal vein and then to the IVC on the left. And this is meant to confer some protection to the left from pressure changes in the IVC whereas chest barotrauma can be directly transmitted to the right adrenal and cause more hemorrhage on the right side usually than the left side. And also with abdominal trauma the right adrenal is situated between the liver and the vertebral column so it tends to be more susceptible to injury in abdominal trauma also. But when we encounter an adrenal mass or nodule the main question clinically is is it a hyper-functioning or non-hyper-functioning lesion and this generally is your domain that you characterize it via clinical means or biochemical means. And the main goal of imaging in terms of classification is to differentiate the benign leave-alone lesions such as non-hyper-functioning tumors, myelolipomas, hemorrhage and cyst from masses that warrant treatment such as a metastasis, an adrenal cortical carcinoma or a pheochromocytoma. And really the reason why we see so many nodules in the adrenals is because of the prevalence of adrenal adenomas. It's felt to be about five to six percent of the general population have these lesions and in younger patients it's less less than two percent but in patients over 70 it's estimated as as 10 percent. So this is why we see so many nodules really and they tend to be small and well-defined and usually of homogenous low attenuation on CT. They can occasionally as in this example hemorrhage or undergo cystic change. Their cells are very similar to the clear cells of the zona fasciculata of the adrenal cortex as it is an adrenal cortical tumor. But occasionally they can have some unusual features but are usually more homogenous. Whereas metastases tend to be more heterogeneous and irregular and larger in general. But there is an overlap on purely morphologic features. And the reason why we're able to characterize adenomas in on CT is that they can tend to contain a large amount of intracellular plasmic lipid. And lipid manifests as low attenuation on CT. And this is why we are able to recognize adenomas. Because the first original seminal paper really was back in 91 which showed that if you have an adrenal lesion that measures less than zero Hounsfield units on unenhanced CT this was typical of a adrenal adenoma with 100 percent specificity and a sensitivity 47 percent. And this is what the study kind of showed. We take a region of interest encompassing about two-thirds of the adrenal nodule and we measure the density. And in this case it was minus five Hounsfield units. And you can also do this using histogram analysis. But I think in most places we still use the region of interest method. And it's because of this lipid, as I said, that is producing low density. That's why it's a very low density lesion usually. And but the sensitivity was quite low in terms of only 47 percent. So this led to a number of other studies and then a meta-analysis was done that produced a new threshold of less than 10 Hounsfield units. Less than or equal to 10 Hounsfield units. And this kept the specificity pretty high with 98 percent and greatly increased the sensitivity to 71 percent. And but it is important to see that that specificity has dropped a tiny bit. And it's because of the rare occurrence of well differentiated adrenal cortical carcinomas can have been reported as showing less than 10 Hounsfield units as have metastases from clear cell renal cell cancer and from hepatocellular cancer. So it's not a like a hundred percent specific. And you need to be aware if those primary tumors are present in the patient. And then the other problem is that some 30 percent of adenomas are so-called lipid poor. They're over 10 Hounsfield units on non-contrast CT. And therefore they cannot be characterized on a non-contrast CT. And indeed most routine studies that are done in America are done with contrast. Over 80 percent are done with contrast. Although as you may have heard due to the shutdown in Shanghai there's a shortage of IV contrast. And we may see this percentage go down significantly over the coming months if the shutdown continues. And with contrast enhanced CT it's generally not the dynamic portal venous phase that we are looking at. It's more the washout of contrast. And this is a so-called de-enhancement. So when you inject contrast it enhances up to an attenuation we'll call the dynamic attenuation at the portal venous phase. And then to measure that you measure the attenuation at the dynamic phase minus the non-contrast which it's enhanced from. And then the de-enhancement or washout is that dynamic attenuation minus the delayed attenuation. So that's usually dropping and we see a washout. And the absolute percentage washout is washout over wash-in or de-enhancement over enhancement times 100 to get a percentage. And then if we don't know the non-contrast density and we can use the relative percentage washout and this turns out to be to just assign a zero to the non-contrast figure in this formula which isn't always true but it still works pretty well if you don't have the non-contrast. And a waterfall is like an extreme example of washout. And the in general we should always use the absolute percentage washout if it's available because if on the relative percentage washout the lipid poor adenomas are penalized because their higher value is not taken into account because by definition lipid poor is over 10 ounce filled units. And the thresholds that has been found at 15 minutes that has helped to distinguish adenomas from non-adenomas is 60% for absolute percentage washout and 40% for relative percentage washout. And this is a graphical demonstration of what happens after you give contrast so time is on the x-axis and attenuation value is on the y-axis and most adrenal lesions enhance fairly briskly into the portal venous phase which is generally between 60 and 75 seconds. And after that adenomas tend to wash out quite briskly and this is because they've got a rich capillary network which allowed them to do that. Whereas non-adenomas like malignancies tend to leak into the interstitial space and hang around longer so it has a slower washout. And the longer you wait in general the more separation there is between these two graphs. So the general standard now has been 15 minutes but over the years people have particularly used 10 minutes in the past or even shorter times but in general the longer you wait the greater the separation and that's why that is generally the standard at the moment. So if you remember at 15 minutes 60% for absolute percentage washout and 40% for relative. And adenomas are greater than this they wash out a lot. And this is how an adenoma that's read the book appears. It's low density on non-contrast less than 10. It enhances but not that high in enhancement generally. Generally it's about between 40 and 80 is usually the general enhancement figure. And then undelayed it enhances quite dramatically and this leads to high relative and absolute percentage washouts. And then the nice thing about washout is that lipid poor adenomas behave just like lipid rich adenomas in terms of washout. So it's not dependent on the lipid content. But you do have to be aware though that there are caveats with washout and any lesion that tends to be hypervascular tends to have high washout or may have high washout. And these include again paracellular carcinoma and renal cell carcinoma are very hypervascular tumors and their metastases can be hypervascular too. And pheochromocytomas are very hypervascular and up to over a third of pheochromocytomas will wash out just like an adenoma. And so that's one thing to keep in mind is that a lot of the studies that did washout excluded pheochromocytoma from the population. And they will, we can't do that in real practice with incidental lesions. So washout may not be quite as good in real practice as in these kind of prior studies that were really comparing adenomas versus metastases. But one thing just to keep in mind is that if you have a very high non-contrast density that should trigger your awareness that it might be, it's not typical of an adenoma to be that dense to start with. If you think about 70% of them are less than 10 Hounsfield units and the remaining 30%, not many would be up in the 40s or higher. So that would make you suspicious no matter what the washout showed. And if it's very low density, it's generally almost always a benign lesion, less than zero Hounsfield units. And then one of the advances in CT is multispectral CT or dual energy CT where we use different energies to assess lesions. And with these dual energies, we're able to produce virtual unenhanced images from a post-contrast image. So this allows us to do a contrast enhanced scan and still get a virtual unenhanced image. And it's been found that these measurements are pretty similar. So we're able to use it in an adrenal CT protocol to give us non-contrast and washout. But some reporters have shown that the densities are a little higher than the true non-contrast value. And this means that if it does measure less than 10 Hounsfield units on a virtual non-contrast image, you can be very confident that it is an adenoma, but you might miss some adenomas because of this higher density, kind of falsely higher density. And then there's a new way of doing multispectral CT, which is detector based. And this has an even greater differential between the true non-contrast and virtual non-contrast. So there has been a new threshold for detector based spectral CT to be 20 Hounsfield units rather than 10 Hounsfield units. But this is still pretty early work. But it is a big advantage to be able to acquire the virtual non-contrast without actually acquiring a true non-contrast study. So in terms of CT in an imaging strategy, it's certainly the most commonly used imaging modality that incidentally picks up these adrenal nodules and also serves to characterize and evaluate these adrenal lesions and is a part of all adrenal algorithms. The non-contrast uniform attenuation, less than 10 Hounsfield units or less than or equal to 10 Hounsfield units is widely accepted as indicative of an adrenal adenoma. The washout is independent of lipid content and is particularly useful for recognizing lipid poor adenomas. It can be useful for staging adrenal malignancies. The multispectral CT advantage of being able to acquire a virtual non-contrast is very appealing, but as I said, it may require its own updated threshold value. And then the big issue with CT is its attendant associated radiation dose to the patient. It is, however, generally an easier scan for the patients than MRI, for example. It's more available. It's quicker and less prone to motion artifact. And it uses a wider and shorter bore, so there's less claustrophobia for patients than what they might experience with MRI. But turning to MRI, this uses magnetic fields and the resonance of protons, and it gives us a lot of tissue characterizing ability. But in general, a lot of them have an overlap and we don't routinely get washout with MRI, even though it is a feasible proposition. But what we generally use to characterize adrenal lesions on MRI is chemical shift imaging. And this is because lipid protons precess or resonate at a slower frequency than water protons. So if you acquire the signal at a time when they're in phase, the two signals will combine and you'll get a higher signal. But if you acquire the signal when the fat protons are 180 degrees opposite phase to the water protons, they will subtract and you'll get a drop in signal. So it's this drop in signal that we're looking for on out of phase images when we assess adrenals on MRI. And there are formulae that you'll see in papers that essentially are trying to quantify this drop on the out of phase image. But I think most people in clinical practice just use visual assessment because it's been reported as being as effective as these quantitative measures. And these thresholds need to be adjusted for different field strengths of the magnet, for example, 1.5 versus 3 Tesla. But it is true that MRI has been reported to be able to characterize additional adenomas when CT densities are indeterminate, particularly in the 10 to 20 Hounsfield unit range, MRI can identify more adrenal adenomas. And this is what we like to see with an adrenal adenoma, you have a, you have a, on the in phase, a nodule which is similar in signal to spleen, spleen is our internal reference because it doesn't contain any lipid. And then on the outer phase, you see how black the adrenal nodule has become, that's because it contains a lot of lipid and it drops signal because of all that lipid, and it's much blacker than the spleen which shouldn't really change. We don't use the liver as the internal reference because liver can undergo fatty change, and this, for example in this patient there is fatty change in the liver. And this is a patient who doesn't show any drop of signal in out of phase imaging, so this is indeterminate, and unfortunately this turned out to be metastases with PET positivity. And this was a larger mass that again didn't show any change relative to spleen between the in and out of phase, so we call this an indeterminate lesion. It was removed laparoscopically and the pathology came back as a adrenal adenoma. But a few months later, the patient returned with metastases in the liver, and on biopsy these turned out to be metastases, and this turned out to be a well-differentiated adrenal cortical carcinoma that had been initially mischaracterized as an adenoma by the pathologist. So this just highlights how difficult it is for pathologists to differentiate adrenal cortical carcinomas from adenomas, and we can't really, biopsy isn't indicated to make that particular distinction. So that's the, unfortunately in that case. So in terms of MRI and an imaging strategy, MRI is generally preferable to CT in young patients. For example, in children, in adolescents, in pregnant women, and in adults less than 40 years of age, again because of the lack of radiation dose. And these groups also have an increased likelihood of malignancy if they have nodules. It occasionally will detect incidental adrenal nodules on MRIs for other indications, just like in CT. It can be an option for follow-up if you just want to follow for size of a lesion. It's also useful in that kind of category of a nodule which has a CT density between 10 and 20, as I said. And unfortunately there are some absolute and relative contraindications to MRI relating to devices and foreign bodies near vital structures. And in general, it is a longer exam for the patient with a higher risk of artifact, particularly motion artifact. It is more expensive. It is noisier than CT, and there is an increased risk of claustrophobia with MR over CT. So turning to PET, or positron emission tomography, this uses the emission of positrons to produce an image. And the standard radio tracer that's the main workhorse is fluorodeoxyglucose, or FDDG. This is widely used in oncology for whole body staging of malignancies. And so we don't use this to specifically address the adrenal, but you get it on every PET scan. The adrenal is essentially included. And it's been shown that the adrenals show less adrenals and benign lesions, so similar or less FDDG activity than liver. But there are some false positive findings in the adrenals. For example, most pheochromocytomas will show increased uptake. Brown fat that you're probably all aware of can show uptake in the periadrenal area. Some 5% of adrenal adenomas will show uptake, and these are considered to likely be functioning adrenal adenomas will show it, and that kind of makes intuitive sense. And then inflammatory and infectious lesions can also show uptake. And then the false negative lesions could include metastases that have undergone hemorrhage or necrosis, small metastases, or metastases from low-grade tumors. But this is what we like to see in an adrenal lesion on PET. Essentially little to no uptake, similar or less than liver. And so this left adrenal nodule. And similarly with lipid poor, also shows little to no uptake on PET. So that's useful, whereas metastases, say for example in this patient who has a lung cancer also has an adrenal metastasis that is very FDDG avid. And this is a patient who has intravascular CNS lymphoma, and the adrenals are fairly normal looking on CT, but on PET they were avidly enhancing, or showed avid uptake rather. And this was due to kind of lymphoma involving the adrenals. So it shows that sometimes PET can be very useful in detecting low-volume disease. And this was a patient who had a left adrenal lesion that measured less than 10 hgf units on CT, but showed intense uptake on FDG. And we thought maybe this is one of these 5% of adrenal adenomas that show uptake. But you do have to carefully co-register these images. And when we co-registered it better, we saw that the low-density lesion didn't really show uptake, whereas it was a more enhancing focal area that did. And we targeted both the low-density area that came back as an adenoma, and the enhancing lesion which showed FDG uptake, came back as a metastasis. So this is an example of a collision tumor with a metastasis occurring in a patient who already had an adrenal adenoma. And then as I said, we sometimes see this uptake around the adrenals with normal looking adrenals. And this is an example of brown fat. And we typically see the typical pattern in the neck where there's uptake in the brown fat areas in the neck, and then extends down the mediastinum and sometimes around the adrenals. And brown fat, as opposed to white fat, has more mitochondria, and these act as engines to burn calories to produce heat, and it's a potential confounder therefore when interpreting PET images. And it's particularly dramatic if the patient happens to have a pheochromocytoma because that'll really activate the brown fat by its catecholamine excess, so you'll see a lot of uptake in that situation. They tend to be patients with increased brown fat uptake. Typically are young women who are thin, and in general, the younger you are, the higher the standardized uptake value. And then we have a number of kind of newer radio tracers such as DOTATATE, which can be useful in terms of paragangliomas and metastases, but there's quite a lot of physiological uptake in the adrenals. So it's not as good as, for example, F18-DOPA for a primary pheochromocytoma, but it is better than FDG-PET in terms of pheochromocytomas. And then you can also combine PET with MRI instead of PET with CT, and this again, we can see a lung cancer with adrenal metastases. So in terms of an imaging strategy, as I said, it's widely used in oncology for whole-body staging of many malignancies and for re-evaluation of tumors. As I said, it's not specifically to evaluate the adrenals that we're doing a PET, but it does give useful adrenal information when it's done. DOTATATE is certainly better than FDG for pheo, but F18-DOPA is considered probably the best of the new agents for pheochromocytomas and paragangliomas. PET, as I said, is now usually combined with CT because they are kind of mutually complementary and CT provides the attenuation correction and speeds up the whole exam for the patients. And then PET-MR has been up to now more research-based, but is slowly entering the clinical sphere. So in terms of imaging features of adrenal disorders, we have Cushing syndrome due to functioning adrenal lesions, which will behave, like I said, low-density on CT and dropping signal and out-of-phase MR. And typically because of the suppressive effect on ACTH, you'll get a very thin contralateral gland. So in both these patients, the left adrenal is very thin. And even if the functioning tumor is kind of exophytic on one side, you can even see that it produces some thinning of the ipsilateral side as well. So that's because it reduces the ACTH and its effect on the gland. Whereas in Cushing's disease, where you have increased ACTH, you get diffuse thickening of the adrenal glands. This would be well over five millimeters thick, these limbs. And then sometimes we see these very bizarre-looking adrenal glands, where you've got diffuse kind of large nodules and thickening, and the nodules are very low in density. And this tends to be adrenocorticotropin-independent macronodular adrenocortical hyperplasia. And you get massively enlarged adrenals with high lipid content. And then in terms of adrenal insufficiency, often, sometimes we can see calcification in the adrenals from prior hemorrhage or infection. Or there are a number of neoplasms, including pheochromocytoma and adrenocortic carcinoma, and myelolipoma that can all calcify. So that's one of the ways we can see, or very thin adrenals can also indicate adrenal insufficiency, less than two millimeters in thickness. And in terms of hemorrhage, CT tends to be very high in attenuation and then decrease in density over time. And MRI similarly goes from a high signal on T1 to having more of a ringed appearance to finally having a very hypo-intense imaging features because of hemocytarin. And this is a patient who had very high signal in the right adrenal gland on T1. But post-contrast, there was some heterogeneity, which should suggest to you that there might be an underlying mass. But this was read as being just hemorrhage. But on follow-up, you see that the hemorrhage is beginning to resolve. And underlying the hemorrhage is a T2 slightly bright tumor that is enhancing. And this turned out to be a pheochromocytoma that had hemorrhaged. And we sometimes see adrenal infarction where you see lack of enhancement post-contrast on MR. And this is sometimes seen with thromboembolic disease, particularly in patients who've got antiphospholipid syndrome. You can also see it in various connective tissue disorders. And we've seen a number of cases associated with COVID-19, which can portend a poorer prognosis. But all of these infarctions are often complicated by hemorrhage and are often combined. So turning to tumors from the medulla, pheochromocytoma, as you all know, are a catecholamine-producing tumors from chromatin tissue. It doesn't obey all the 10% rules, but about 10% are bilateral, 10% malignant, and 10% are extra-adrenal, and about 10% calcify. There are various syndromes that are associated. And then, as I said, 18F-DOPA is probably the best for pheochromocytoma at the moment. And traditionally, it was considered to be a light bulb bright lesion on T2-weighted images. So on T2-weighted images, fluid is bright, so the CSF is bright, the gallbladder is bright. And this was an example of a very bright pheo. But in general, this only occurs about 20% of the time. Most of the time, it's slightly bright or may not be bright at all because of all the degeneration and hemorrhage that can occur. But a feature that is more like light bulb bright is its vascularity. Most of the time, it's really quite vascular. Here we see on CT and on MR. Researchers have shown that if you have a very avidly enhancing lesion, if it's over 130 hf units on the portal venous phase, that's very specific for pheochromocytoma. And other people have used even higher figures to get more specificity. But Mohammed and Al suggested that actually, we don't really want... You guys will diagnose the pheochromocytomas with great specificity. We need to kind of raise your awareness of that this could be a pheo. So really anything over 85 hf units is generally more likely pheo than adenoma. The higher it is on the dynamic phase, the more likely it is to be a pheochromocytoma. And because as I said, most adenomas won't enhance up to this level. So that's a useful information. And again, this is a patient who has a left adrenal mass. It's not particularly bright on T2, but it is very hypervascular on MR as well. So because pheochromocytomas can undergo so many different degenerations as we've seen it, they can hemorrhage, they can produce cystic change, they can be fibrotic. We called it an imaging chameleon because a chameleon is this lizard that can change color to kind of better camouflage itself. So pheos are often difficult to diagnose on imaging as they are clinically. But this was for an example, a patient who had what appeared to be a brisk washout. But if you look at the dynamic density, it's quite high. And then if you look at the non-contrast density, it's also quite high. So all of these should trigger alarm bells to say this shouldn't be considered to be just an adenoma slam dunk. This is more likely to be a pheochromocytoma. And as I said, over a third of pheochromocytomas will mimic adenoma in terms of washout. And then in terms of containing fat, this is a very rare phenomenon that pheos can do. It's generally considered to be less than 1% in most studies, but a recent study showed a higher percentage perhaps because they just looked for a little foci of fat. But if the lesion is homogenous and non-avidly enhancing, it's very unlikely to be a pheochromocytoma. But it can mimic other adrenal lesions. We'll see that adrenal cortical carcinoma has propensity to involve the venous structures and the pheos can do the same thing. About 10% are malignant and it's hard for a pathologist to determine that until they actually metastasize. And this is one that did metastasize. It was a big left adrenal mass that has a lot of cystic change and calcifications and then liver metastases and bone metastases. And as I said, you can have extra adrenal paragangliomas anywhere from the skull base to the bladder base. And these tend to be quite vascular also. In terms of hyperaldosteronism, this is caused by an adenoma in about 40% of cases and hyperplasia in most other cases. And these can behave like we've discussed already, but can sometimes be very small. And sometimes we need adrenal venous sampling to kind of be sure that we're lateralizing to the right side. So turning to adrenal cortical carcinoma, this is very rare. It's one to two per million, but it's what we all want to avoid missing. And sometimes it can have clinical and biochemical manifestations. With imaging, it often is large, irregular, and heterogeneous. It often has this involvement of vessels, evidence of invasion, local metastases. It's quite a lethal cancer. But at some point, it is starting as a smaller lesion and then quickly gets to that four centimeter lesion. But we rarely catch it at a smaller lesion. However, as we've kind of seen, when they're well differentiated, they can behave a bit like adenomas. They've been reported as having occasionally a marked washout pattern, although usually it will behave more like a malignancy in general. And usually it's very heterogeneous, as both these lesions are. Both these lesions are involving the venous structures, so that's typical of adrenal cortical carcinoma. Their symptoms are huge, and it's hard to tell where they're coming from. But looking at coronal and sagittal images, you can kind of see its influence on the surrounding structures. And in terms of the general size criteria, lesions greater than four are generally considered to have a higher risk of malignancy. But this is a rather crude measure for risk of malignancy, as there are quite a few benign lesions that could be greater than four centimeters, too. So you kind of have to use that on a case-by-case basis. We should look for a regular shape is often malignant, and heterogeneity is also often malignant. These are lung cancer metastases that are quite heterogeneous and bilateral. We do just have to be aware with the advent of immunotherapy that you can get effects of the treatment that can be kind of confusing. This is a patient who had normal appearing adrenals and pituitary before starting ipilimumab for melanoma. And then after two months of treatment, the adrenals have got thick, and the pituitary has also got thick. So these were adrenals before, these are the adrenals after treatment. And then after stopping the treatment, they returned to normal. So this is an example of immunotherapy kind of producing pseudo lesions, and those were just kind of thickening. This is a similar kind of case with very nodular appearance of the adrenals, and that again on biopsy turned out to be an immunotherapy effect as opposed to progression. So these are kind of pseudo nodules. So it's something to keep in mind with these new agents. And then ultrasound usually doesn't detect many adrenal lesions, but when it does, it's usually an echogenic lesion such as this, which turns out to be a fat-containing lesion on CT. And this is a myeloid lipoma, which is a relatively uncommon benign tumor, but it represents about 6% of incidental loma lesions, and it's composed both of hematopoietic tissue and adipose cells. And when there's a lot of fat in the lesion, it's virtually diagnostic. So you can get varying amounts of fat and hematopoietic tissue, so it can be more challenging if they're very small amounts of fat, or kind of equal amounts of fat and hematopoietic tissue, you can almost get a more fluid density. But when there's fat, it's virtually diagnostic. On ultrasound, they're echogenic. On MRI, they're T1 bright, but lose signal on fat suppression as opposed to hemorrhage. And then they can hemorrhage when they get big. And as I said, sometimes they can have very small amounts of fat. And this can be tricky because when there are only small amounts of fat, this could also be my lipomatous metaplasia of another tumor, such as adrenal cortical carcinoma or Pheo. And all three of these can calcify too. So this often is a differential. Are we dealing with an unusual myelipoma or a malignancy? But when it's like this, you almost need to see prior showing that it's stable to be able to say this is kind of a myelipoma with little fat. So turning to the society imaging guidelines, well, first of all, I have to say that there is a lack of randomized controlled trials and longitudinal cohort studies to inform these guidelines. But if we look back at how these guidelines have evolved, if you look at the first decade of this century, both the ACE guidelines and William Young were suggesting that all adrenal incidental lomas should be re-imaged quickly and annually, whether they had a benign looking phenotype or not. But then some papers came out saying that, well, there's really very little chance of a malignancy on these incidental patients. And that there is a risk of, by having lots of CT scans, you might cause a cancer by repeating all these scans rather than diagnosing one. So this led to the European Society of Endocrinology basically saying that they advise establishing if adrenal mass is benign or malignant at the time of detection. But if you see that it looks like a benign mass and it meets all the criteria, no further imaging is required, which was a change. And then they did say if the mass is indeterminate on non-contrast CT, you've got a negative hormonal workup, you have different options depending on the patient's clinical context. You could have immediate additional imaging with another modality. You could get interval imaging or go to surgery. So there are major dichotomy. So in terms of no further imaging being needed for benign appearing adrenal masses, this was supported by El-Hassan's kind of meta-analysis in the Annals of Internal Medicine. And also then there was a disease stage clinical review by Dr. Vaida and Dr. Bankos and others, basically also saying that there was no strong evidence to say that we should keep imaging at these benign looking lesions. But, and then there was an official statement actually from ACE also saying that. But in terms of the radiology groups, the ACR also updated their algorithm in an attempt to kind of reduce the amount of imaging. So just like the endocrine groups, they also said that if it looks benign, you don't need to do any more imaging. They also said that mild lipomas and dense calcification, you don't need to do further imaging. As we said, greater than four centimeters is different and it's the case by case kind of appearance. They emphasized if there were priors, if you could see it wasn't growing and was stable, you could consider it benign. But they did say that in masses over two, between two and four centimeters, or if there was some history of malignancy, then you should work up with a non-contrast CT and if necessary, an adrenal washout protocol. And then if it didn't enhance at all, it would be a benign cyst. If it was washing out like an adenoma, then it was likely to be a benign adenoma as long as the primary tumor wasn't a clear cell RCC or HCCC. And then if it was less than the washout, then it's a case by case kind of situation. But the big difference I think that the radiology ACR have made is that they have this, for one to two centimeter, even if it's over 10 Hounsfield units, they still think it's statistically likely benign and they're just saying to get a 12 month follow up adrenal CT. So that's different to most of the other groups and in fact, the most recent one is a paper that's probably gonna be published in JAMA Surgery this year for guidelines for adrenalectomy. And they also agree that we don't need to do a routine follow up of non-functional adrenal nodule with benign imaging characteristics. And everyone says that there's no strong evidence to support further imaging. And that's been endorsed by the SAR. But they, like the other endocrine groups, continues to say that any size nodule over 10 Hounsfield units should be further characterized at the time of diagnosis to kind of risk stratify those patients. So that is a kind of a difference between the groups, but the endocrine groups have been fairly consistent in that regard. But in terms of issues, I think all the groups say that one centimeter is the cutoff for assessing adrenals, nodules in the, but there are seven, eight, nine millimeter nodules that look very discreet and very real on our imaging and they don't meet the one centimeter criteria. So what do we do with those things? In terms of that one to two centimeter nodule being, not being kind of assessed at the time of diagnosis, is that a sensible route or not? And then even two to four centimeter nodules that look benign, if you have a 3.8 centimeter nodule and a four centimeter gonna possibly send it to surgery, do you just say, well, that's it, we're just leaving you now, we're not gonna get any more imaging. So there's kind of a few things to think about in terms of size. And then we've got to really think about too, are these nodules meeting the homogenous criteria? Like it needs, and homogeneity is in the, some is in the eye of the beholder. There's a wide range. The margin too needs to kind of, needs to be remembered if you're looking at it. And then this whole thing of the greater than 10 Hounsfield units, in terms of when we further characterize those lesions, or if we do, and then also with dual energy CT, do we need to have a new threshold for that? And then just be aware that the washout figures that have been reported, we're comparing adenomas to metastases. They're not, haven't really been assessed in the incidentaloma population. So that could be a problem because we don't know if it's gonna be a pheochromocytoma or not, for example. And I think that that enhancement information regarding pheochromocytomas is useful and should probably be added into the algorithms to kind of increase the detection of pheochromocytomas. But I do think all these strategies are evolving and are quite complex for both radiologists and for endocrinologists. And I think it would be a useful application of artificial intelligence to have an algorithm-based decision support, both for you ordering these studies and for us interpreting them. And just to finish up with some pseudoadrenal lesions to be aware of. Anything near the adrenal can look like an adrenal lesion. You can even get adrenal pseudonodules by its kind of S-shape, but you should see it on more than one plane. If it's only seen in one plane, it's probably just an adrenal pseudonodule. The celiac ganglion lies right beside the adrenal. Gastric diverticular are quite common and kind of hang down beside the adrenal. And really any vessel that's in the area or any organ can sometimes produce an adrenal-looking lesion. This is an example of a gastric diverticular that you can see the connection with the stomach. And as I said, prior exams can really be helpful to let you know if things are growing or not. But occasionally things that grow, for example here, isn't that they are malignant, it's that there's ectopic ACDH or something driving their growth. Or sometimes hemorrhage can produce increase in size too. So in summary, I think if you can remember that for an adenoma, they generally are less than or equal to 10 HV units of non-contrast CT, and they have greater than 40% relative percentage washout or 60% absolute percentage washout. They tend to show signal drop on out-of-phase MRI. And just remember that pheochromocytoma is a great mimicker, but is usually hypervascular. So in conclusion, I think imaging can help characterize many adrenal lesions, including most instantiolomas. CT and MRI are the main imaging modalities. The societies have some different strategy guidelines, particularly regarding follow-up imaging. I think greater harmonization between imaging and endocrine societies could lead to more consistent and optimized clinical practice. Thank you very much for your attention. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Pre-contrast, post-contrast, Hounsfield units, also about, for example, washout and chemical shift. You know, for the clinicians, for you, of course, you know, it's clear, but for us, especially for endocrinologists, it's a little bit more complicated. At the last slides, you mentioned something about the artificial intelligence and deep learning. What do you think about, you know, the artificial intelligence? Where do we stand and what is happening in the field? Because I think that it will definitely help clinicians to help them to see, you know, how we can actually predict what is going on with the adrenals, especially with the tumors, what we can expect, you know, from that. Yes, well, I think it's an exciting time for medicine in general. And I think if we can apply artificial intelligence wisely, it can be very helpful. We already are lucky that we have some decision support in our hospital that allows us, you know, plug in the features that we see, and then it can churn out, based on the algorithm, what the diagnosis should be and what potentially the follow-up should be. So that's the kind of low-hanging fruit. But I do think that, as I said, all these algorithms are continually being updated and evolving, and I think it would be helpful to have a similar kind of thing for choosing what study to order, what is compliant with algorithms. And, you know, people say that artificial intelligence may be able to do a better job than radiologists. Already, you know, it's doing very well in a number of different areas, and I think patients are said to be very accepting of, say, a chest X-ray being read by artificial intelligence. But in my experience, with a lot of deep learning, it almost needs some human oversight, because it can be brilliant, but it can also be illogical. So you need some human oversight. Thank you so much. It was a really great talk. Thank you. Thank you, please, sir. Dr. Sergev, Charleston, South Carolina. I have a question and more of a suggestion. I've seen recently, the last five years, two patients with benign adrenal lesions that were followed, actually, but by non-endocrinologists for a few years, one actually for about 10 years. And they showed up with livid Cushing syndrome with me. I have a small suggestion, I don't know. Some radiologists, when they read benign versus malignant, because that's what can be differentiated by all imaging modalities, they stay there. And a lot of non-endocrinologists, they do not perform, or they don't refer, for endocrine testing. I would like a suggestion, maybe in the radiological society, when something is read as benign or malignant, when it's benign, some kind of hormonal evaluation suggested, so in order not to miss. Very important, because most of the, I mean, we know most of the adrenal tumors produce cortisol excess, and we can miss a lot of Cushing syndromes. Absolutely, and I think that's why it would be good to have more cooperation between endocrine and radiology societies to, if we're producing an algorithm, it would be your domain to kind of say when it should, that should be assessed, or when is the right time to do those, that biochemical assessment. But I know from personal experience when we were trying to introduce a referral to endocrinology, or the primary care doctors were very, we have to really wordsmith it very carefully. And I was just in our hospital, in a national level, I don't, it's gonna be, it can be sometimes quite challenging. It's easier said than done to get the right wording. And, but we would kind of defer it to you, and that's where artificial intelligence could bring it all together, because you're managing patients, it's not just what imaging you're gonna order, it's what biochemical test you're gonna do. So I agree, and you have that expertise, so I think it should be combined as much as possible. Otherwise we have all these different guidelines from different groups, and it's very hard for any primary care doctor to keep it all straight, or for any of us to keep it straight either. Thank you, a great talk. And I just wanted to comment on that, because I could not agree more that the algorithms from AGR did not contain referral to endocrinology. Thank you. Thank you. Sir? Yeah, Fadil Khair from Connecticut. 20 year old woman, right adrenal tumor one centimeter, 28 density, increased to 140, 140. And the left 0.9, washout was more than 60%, and the left 0.9 centimeters, still 28 density, went up to 140, washout below 60%. Stable over two years, silent for pheochromocytoma. Of course there are silent pheochromocytomas, we know that. Did not show up in frequent workup for two years. What do you make out of this? It's a true case I have, challenging. Yeah, yeah, and I think, as I kind of alluded to, there are, nothing is 100% in life or in radiology, and there are caveats and exceptions. But certainly with younger patients, you've got to, they're less common, so you've got to be careful that you're not missing a malignancy or a pheochromocytoma. But, you know, there are exceptions to all the rules, so it's possible, though it's a little unusual that they might still be adenomas, but you'd certainly be suspicious from the enhancement characteristics. But you said that it's all been negative, and it's. And stable over two years, so far. Yeah, so I think it's probably then a benign. There are other entities that I didn't address today, but ganglioneuromas can show kind of progressive enhancement. Hemangiomas are rare, but can be very hypervascular, but wouldn't usually show washout. And so they, it probably is an unusual, bilateral unusual adenomas, but when you first see that, you would really need to make sure there wasn't a pheo, or possibly an adrenal cortical carcinoma in a young patient with kind of an unknown syndrome or something. Thank you. Thank you. I think we have time for two more questions. Please, go ahead. Hi, thank you, great talk. Hassan Shawal, Winnie, New York. Clarification about the epilimumab adrenal change. Was that only images, changes, or it was associated with, or it could be associated with biochemical adrenal changes like cortisol or endothelial deficiencies? I missed the first word. The epilimumab changes. Oh, you're fine. Is that only imaging change, transient imaging change, only thickening in the adrenal glands for, I don't know, a couple of months, three months, it was not associated or will not be associated with biochemical changes? Is it adrenalitis changes? Yeah, it's felt to be adrenalitis changes. When you get hypophysitis, it's felt that it interferes with function, so it wouldn't really be driving adrenal hypertrophy. And like in one of the cases, there was pituitary increase too. And in the case that we've seen, if you stop the epilimumab, it returns to normal. And I'm not sure about what biochemical profile was associated with them when they became nodular or thick, but it's meant to be kind of an inflammation of those organs. You can get it in the thyroid too. You can get a colitis. There's a lot of inflammatory process that can occur with those agents. Thank you. Please go ahead. Thanks for a great presentation. Really appreciate it. I'm David Brown from Rockville, Maryland. And you showed some examples of what you referred to as sort of fat or plump adrenal glands in an ACTH-dependent disease state. What I'm occasionally running into is a patient who has clinical and biochemical characteristics of ACTH-independent disease, ACTH-independent Cushing disease. And there's a little bit of disagreement on what the adrenal imaging looks like. The adrenals will look basically morphologically normal, but just a little too plump as though there was ACTH drive, but in fact, there isn't. So I'm wondering if you could just offer a little guidance about when to think that morphologically normal looking adrenal glands are just a little too plump. When would you start thinking about some subtle hyperplasia versus just a morphologically normal gland? Yeah, so as I said, most of the limbs, the width of the limbs usually is by three to four millimeters but once it gets over five millimeters, that's beyond the normal range of the adrenal limb width. So anything over five is abnormal. And so that's where I would draw the line. The normal kind of extreme range is two to five millimeters. So once you get over five millimeters, that's what I take as. Okay, thank you. Thank you. And we're out of time, but I do know, sorry, you want to ask a question, but maybe we can leave it after dismissing everyone. Thank you. Thank you.

adrenal imaging

Dr. Benkos

Dr. Blake

adrenal tumors

imaging techniques

CT

MRI

PET

adenomas

non-adenomatous lesions