 So, today's session, as I already told you, the first session will be by Dr. Alok Jadusar, who is the Program Director for Pediatric Neuro Radiology Fellowships and he is a radiologist at Annan Robert, Durich Children's Hospital, Chicago. He is Assistant Professor at Northern Western University, Freeberg School of Medicine. So, let us start with Sir's session for today. Hello everyone. It's an absolute pleasure to be speaking at this MRI conference today and I would like to thank Dr. Bhatkar and the organizing team for the invitation. Today I'm going to talk about Pediatric Posteria Fossa Tumors. It's something which we encounter fairly often in our clinical practice, however, it's also a topic that's been evolving very rapidly over the last few years. So, for this talk, I'm going to talk about the classic imaging features of the Posteria Fossa Tumors. I'll provide an overview of the new molecular classification and I'll also discuss how these molecular subgrouping impacts the clinical management as well as the imaging protocols and then I'll briefly touch upon some future directions. To help you have a system tumor for the needing cause of cancer related mortality in children, these are the most common toilet tumors in children, second only to leukemia and lymphoma and they are found for 20 to 25% of all pediatric cancer. Of the CNS tumors, the Posteria Fossa Tumors is a particularly important subgroup which accounts for about 50% of all intracranial pediatric tumors. So, the four big Posteria Fossa Tumors, which we always talk about are measures of last two miles, very better pyrocytic astrocytome miles, Brinsen Pio miles and Epidemo miles and these would be the focus of my talk today. There are certain less common Posteria Fossa Tumors such as ATRT and Choroid Plexus Tumors which should be considered in the differential diagnosis in the appropriate scenarios and I will discuss them briefly. Then there are certain other tumors which are seen only in the setting of genetic tumor. Let's start with Metroloblastoma. Metroloblastoma is the most common malignant brain tumor of childhood. It's also the most common Posteria Fossa Tumor accounting for about 40% of Posteria Fossa Tumors. It's a WHO grade 4 tumor and it belongs to the category of embryology. The age for Metroloblastoma is 4 to 10 years, although it can be seen from instances to a hotel's court. About two-thirds of these tumors are located in the midline involving the court ventricle, although they are centered more towards the group near the word mess. Although one-third of these patients, one-third of these tumors can also be eccentric under the CT-angle system or along the cerebellar peduncle and we'll talk about them in a bit. The emerging appearance of Posteria Fossa Metroloblastoma on the non-contrast head CT we see a hyperdense mat centered in the court ventricle which on MRI demonstrate diffusion restriction, ISO to hyper intensity signal and mild but diffuse post-contrast enhancement. The most important emerging features that help you distinguish Metroloblastoma from other Posteria Fossa Tumor is the higher hyperdensity on CT and the diffusion restriction on MRI. Both of these are features of high singularity and high nuclear sacrifice ratio which is characteristic of these high-grade tumors. On a MOSFET cross-copy we see the expected high-polline low NAAP and a small lipid lactate peak such as finger high-grade tumors. This was an intermediate CE MOSFET cross-copy. However, at a low TEMRS, presence of Torine P has been described as a more specific finding of Metroloblastomas. On the protrusion imaging we see diffusion increase protrusion throughout the tumor. So this Metroloblastoma was a classic subtype based on histopathology. The traditional Metroloblastomas have been classified on the basis of ecology with the classic subtype being most common accounting for about 70% of cases. The second most common is this more plastic or modular which accounts for about 15% of cases although it's the most common subtype in general. And the last subtype, anaplastic or large-cell Metroloblastomas accounts for about 10 to 15% of cases but it has the most common. More recently based on advances in methylation techniques and genetic profiling of these tumors four very distinct subgroups of Metroloblastomas have been described. These are wingless or the rinse which is the least common accounting for about 10% of cases. It's seen in slightly older children with a pK of 10 to 12 years are typically corresponds to the classic histopathological subtype and it has the best problem of the four Metroloblastomas subgroup. Next is the sonic hedgehog which accounts for about 30% of these tumors. This subgroup has a bi-modulate distribution and it's seen in infants and then another pK is seen in adolescents and adults. Histopathological rate corresponds to the anaplastic variant of the Metroloblastoma and it carries an intermediate problem of this. Further the problem of the sonic hedgehog group is dependent on the status of T53 mutation with the T53 gene mutation corresponding to a worst problem of this compared to the wild type. Then we have the group 3 Metroloblastoma which accounts for about 25% of cases typically seen in infants and young children and it has the worst problem of all the four subgroups and finally group 4 which is the most common subgroup accounting for 35% of cases. It's seen in slightly older children with a pK of 10 years and it carries an intermediate. The importance of these molecular subgroups has significantly increased in the recent years and now these are being used for risk stratification and treatment planning. This is also reflected in the most recent WHO classification from 2016 which now divides Metroloblastoma along according to genically defined subgroups as well as histologically defined subgroups. For an individual case the owners are on the pathologist to take into account all the histological immunohistochemical and genetic features and come up with the unifying diagnosis as part of the process. Imaging implications of these subgroups these subgroups tend to have slightly different imaging appearance and also clinical behavior in terms of metastatic spread and recurrence. So the wingless subtype like I said has the best diagnosis and these new opiumes are typically located eccentrically either in the CP angle or in the cerebellar potential region. These are typically diffusely enhancing lesions and metastatic spread is there in these tumors. Tonic hedgehogs like I said corresponds to that there's no plastic variant and accordingly it is seen more peripherally along the cerebellar hemisphere may emerge the duro along the centurium. These are typically enhancing tumors and metastasis is relatively less common in these tumors and recurrence is also uncommon for these tumors but when it occurs it's more likely to be in the local surgical bed. For group 3 and group 4 tumors both of these tumors occur in the midline the big difference is that the group 3 tumors are typically enhancing but the group 4 tumors are non-enhancing. Metastatic disease is very common with both of these subgroups at presentation and when they're recurrent the recurrence is almost always at a distance side rather than in the local surgical bed. Two different patients with different subgroups of mesozoblastoma the patient on the left side has a heterogeneously hyperdense mass centered in the right cerebellar peduncle with factory diffusion restriction and post contrast enhancement also on the T2 weighted images. So based on the location this is most likely a wing subtype which it turned out to be. The tumor on the right side is a lobular hyperdense mass which is along the lateral aspect of the cerebellar hemisphere and it's everything particular. The mass amounts to diffusion restriction, low signal on T2 weighted images and post contrast enhancement. Morphologically this is a plastic subtype and it belongs to the molecular group of chronic hedgehog T53 by subtype. These are two midline of the tumor. So the tumor on the left is a diffusion restricting T2 hyperdense mass involving the post ventricles. The mass also denotes diffuse post contrast enhancement. Also there is some modular enhancement along the cerebellar folia as we can see on this T1 post contrast as well as the post contrast T2 flare imaging. So this is a group 3 mesoblastoma with associated leptomanageal spread at the time of diagnosis. This tumor on the right of the screen is again a diffusion restricting T2 hyperdense mass centered in the post ventricles. But this tumor shows minimal tumor enhancement and this is the corresponding group 4. Certain advanced imaging techniques have been used to try and predict these molecules subgroup. This study was able to distinguish between chronic hedgehog and group 3 flash group 4 tumors using low TE and not spectroscopy. What they found was that chronic hedgehog typically have a prominent hole in peak and a prominent lipid lactate peak but lack of a torian peak while group 3 and group 4 tumors readily demonstrate presence of a torian peak with relatively small lipid lactate peak. Also there are multiple attempts being made to use machine learning and artificial intelligence techniques to distinguish between different molecules of groups of mesoblastomas as well as to predict or respond to treatment. One such study which used radiomic feature extraction technique was able to distinguish between chronic hedgehog and group 4 mesoblastoma. Cs of dissemination like I've said is fairly common in mesoblastoma is seen in about 11 to 43% of the patients. It can be intrapanial and or intrapanial. This is most common in group 3 and group 4 and red in the other two molecules as well. The matter is that metastasis to bone and liver has been described but is a disseminated metastatic disease. The patient on the left side has a small diffusion restricting and enhancing force ventricular mass but extensive leptomanin gel metastatic disease along the cerebellum as well as throughout the spinal cord including the fecal factor may have. Again this is one of the common imaging appearance of group 3 mesoblastoma that the primary tumor can be relatively small even in the presence of extensive metastatic disease. The patient on the right side has a size ventricular tumor which only denotes mild and patchy enhancement but there is extensive metastatic disease towards the cerebellum folia and also surface of the spinal cord. In addition there is this enhancing E2 hyperintense intramedular lesion between the lower spinal cord response to a drop metastasis within the central canal of this spinal cord. As we said group 4 tumors can often be non-enhancing or minimally enhancing and that is also reflected in the metastatic disease. Oftentimes the mess can be non-enhancing for example in this patient it's very difficult to identify the metastatic disease on this post-confrost even weighted imaging. However we can already see this modular mess along the cerebellum folia on the decretion weighted and the post-confrost weighted images. It's very important to obtain the clear imaging post-confrost for these tumor patients and it's also important to look at the diffusion weighted and the post-confrost images very carefully not just in the resection bed but elsewhere to look for the combination. Another technique that can be helpful in problem solving could be diffusion weighted imaging in the spine. For example in this patient there is this modular metastatic focus along the aspect of the thoracic cord which is seen more complicatedly on the diffusion weighted images. Similarly there are two additional modules along the cord ishiana which has been slightly better. Another technique which we have started using routinely in all of the drop metastasis MRF protocol is a 3D high-resolution T2 sequence like a KISS-TSR space. For example in this case we see these two modular foci along the ventral and the topical metastatic which we can also see on these coronal reconstructed images and I think one of the advantages of this technique is that it is 3D and you can get nice axial and coronal reconstructions and personally I have found coronal reconstructions to be very helpful in many of these cases. So these two modules are almost impossible to identify upon the post-confrost T1 weighted images and we can painfully see them on the diffusion weighted images as well. Moving on to next post-confrost tumor which is a cerebellar phylocitic astrocytoma this is the second most common post-confrost tumor accounting for 30-35% of cases. These are benign astrocytic tumors that go at your grade 1 and they typically have an excellent prognosis with resection only without the need for chemotherapy or radiation. The PAH is 5-13 years equally common in male and female 50% of these occur in midline location and 50% are hemispheric. The typical classic imaging appearance of a JPA which is assisted in accounting with a module is seen in about 50% of cases. In about 40-45% cases there is a larger solid component with a central necrosis or cystic area and less than 10% of these tumors can be completely solid and non-necrophic. In sample of a typical JPA we see a mixed solid and cystic mass in the right cerebellar hemisphere with associated base with active hydrocephalus. The fluid component does not suppress completely on flares such as the proteinaceous composition and the solid component is hyper intense on T2 and lack diffusion restriction which are both useful feature in distinguishing it from higher grade tumor. The solid component also demonstrate decreased post-confrost enhancement. This is the second morphological type of JPA where we have a predominantly solid mass with a central necrotic area against the solid masses. Hyper intense on T2 does not demonstrate diffusion restriction as there is diffuse and having post-confrost enhancement. This is a midline posture for the mass which is mixed solid and cystic. The solid portion is hyper intense on T2 does not demonstrate diffusion restriction. I put them on CT and has diffuse post-confrost enhancement. If you just look at T2 and the post-confrost imaging this can be very difficult to distinguish from a major blast trauma however the diffusion rate is and the non-confrost CT images can be very helpful in making that distinction. Brief word about the molecular alterations that can be seen in JPAs. This is the micro-conactivated protein kinase or the MAPK pathway which is probably one of the most important on-prosinic pathway in human cancer. It is implicated in pediatric low grade tumors some of the adult tumors as well as in many cancer predispositions like NF1, tuberculosis, and hematoma syndrome. As far as JPAs are concerned the most important genetic alteration is seen in DPF molecules. So there are two types of B-Ref mutations identified. One is the B-Ref fusion which is seen in about 70 to 80 percent of the co-flip of the JPAs and some authors now equate the presence of B-Ref mutation to the diagnosis of JPA and this is typically associated with better prognosis. Another mutation is the B-Ref M618 point mutation which is seen in only a small percent of the co-flip of the JPA, maybe a slightly higher percent of the super-tentorial JPA like those involving the optic pathway and the hypothalamus. And also this mutation is seen in various other glial and clio-neuromus tumors. This is associated with work prognosis. The important of B-Ref testing is that there are a lot of molecules along the pathway which are targets for chemo-therapeutic agents. For a cerebral JPA, it is sometimes less important because the tumor is restricted completely and may not need chemotherapy but for no great tumors in non-receptible locations like brain stem or discoteltalysis, the importance of B-Ref testing cannot be emphasized enough. Moving on to the next foot-tumor type which is the brain stem clio-neuromus. This is the third most common tumor, accounting for 50 to 30 percent of cases. This is not one tumor entity but it's a heterogeneous group of atrocytic tumors of course, histological subtype and grade is predicted by their location within the brain. So briefly let's talk about the various anatomical locations in the brain stem where the tumor can occur. So the brain stem is divided into mid-brain, pons and medula with the pons being separated from the mid-brain by the pontomism to palyx alkyl and from the medula by the pontum, medula is alkyl. Each of these regions is divided into a ventral or a basal portion and a dorsal portion called tegment and the tegmentum forms the floor of the cerebral aqueduct and the ventricle while the root is formed by the tecton and the cerebral aqueduct. So the most important brain stem tumor is the one that arrives in the basal or the ventral aspect of the palm and is typically called PGR diffused intrinsic point in glioma. These are high-grade tumors who are very aggressive and have an extremely poor prognosis. The second tumor type in the brain stem could involve the mid-brain either in the tecton which are typically grade one tumors, they do not require any treatment other than CSF conversion or they can arise ventrally in the mid-brain when they can also extend into thalamide. These are also typically grade one but can be more infiltrative grade two to three tumors. When these tumors occur in the brain stem, they are more common along the dorsal aspect and they can have an exothermic component extending into the overlying extractal space and can involve the cervical plantar part. And finally the least common group of these dorsal point in tumors, these are lower grade and the main importance is in distinguishing these tumors from the more aggressive ventral counterpart. So the DIPG has been renamed in the most recent WHO classification as the diffuse glioma, which has the signature genetic alteration of K-27 mutation within the HPG source. DIPGs account for 60 to 75 percent of all brain stem tumors, they are typically at-strategic tumors. On histopath they are grade two to four however for practical purposes they are all treated as grade four tumors because of the uniformly poor prognosis. The PK is three to ten years and they are equally common in male and female. The typical clinical presentation of these brain stem gliomas is an axiom cranial novelties like in the seven-year-old child and this is also the typical clinical imaging appearance on MRI. So we have this large D2 hyperintensive infiltrating mass involving the pump, the mass extends specifically into the pre-founding system, increasing the basilar artery. On the sagittal images we can see that the mass is centered more ventrally within the pump with mass effect on the floor of the post ventricle. There is no appropriated diffusion restriction. There is no post contrast enhancement in the mass and the mass also extends into the right cerebellum peduncle and can sometimes also extend into the cerebellum. A lot of imaging features have been studied as predictors of prognosis within DIPG. The two imaging features which have been consistently associated with poorer prognosis are presence of necrosis and ring enhancement and diagnosis and extra-contin extension into the cerebellum peduncle. However, vertical extension into mid-plane on medulla is typically not associated with worse prognosis. Then there are certain advanced imaging parameters which have been associated with worse prognosis which includes high relative CBV and K-tran on initial fusion imaging and lower diffusion value based on ADC histogram. Another thing to remember about these tumors is that the imaging appearance can change rapidly after radiation treatment. For example, these tumors can show increasing areas of enhancement and necrosis after radiation and that should not be confused with progressive tumor. The ADC value is typically decreased after radiation and which may be attributed to increasing rates from the radiation. However, on MR perfusion the CBV values can transiently increase after radiation again. It's something useful to know about as it should not be confused with tumor progression. One of the most useful predictor of tumor versus T2 progression is the tumor volume on the T2 weighted images and also you can use MR spectroscopy as a problem solving technique. I briefly talked about genetic mutations in TIPG so K27M is the characteristic mutation which can involve the 3.1 or 3.3 and there are slight clinical differences between these two subsides. The 3.1 mutation is seen in a slightly younger age group for the 6 years and has a longer survival while the 3.3 mutation is seen in slightly older age group and is associated with a problem. Some tumors can involve the medulla or the mid-brain as you already discussed. These are two different patients. The patient on the left is an 11 year old child who presented with a large expansile T2 apparent as a mass involving the medulla with exoperative total extension as well as extension into this cervical spinal cord. There is no diffusion restriction between the tumor and the follic portion the most heterogeneous but avid post-frontal enhancement. So this is a grade 1 JPA however it is in a non-refectible location. Similarly we have another tumor which is centered in the mid-brain involving both the tectum and the tegmentum as hyper-intensive 22 weighted images no diffusion restriction amongst this patchy post-frontal enhancement against this for the lower-grade tumor. Moving on to the last major post-refective sub-type which is the ependymoma. Ependymomas are the least common accounting for 10 to 24 percent of the tumor. These are clear tumors with ependymal differentiation and they are typically WHO grade 2 but they can be grade 3. The peak age is slightly younger than megaloblastomas and JPA seen in 1 to 5 year age group. Most of them are sporadic although they have been associated with NF2 in addition to schwannomal and meningeomal. Based on the morphology the tumors can be divided into mid-floor type, lateral type and roof type. The mid-floor type is the most common which involves the post-ventricle and it is centered in close relation to the dorsal aspect of the brain stem. Lateral sub-type is slightly less common as it is centered in the CP angle system and the roof type is the least common of the three. On imaging these tumors are typically more heterogeneous than megaloblastomas and JPAs. Calcifications are very common in about half the cases and you may also see hemorrhage and cystic change. Another important imaging feature is this plastic appearance of the tumor where it insinuates to the post-ventricular outflow foramina. On the diffusion weighted imaging these tumors have intermediate diffusion between megaloblastomas and JPAs. So this is a 4 year old child with the typical mid-floor type of ependymal mass. On CT we have a hypodense mass which is gathered from great calcifications and obstetrics as a specialist. On MR we see a T2 hyper intense mass which extends throughout to bilateral foramina of Plushka, extends dorsally to the foramina of Majanti and extends inferiorly into the typical spinal canal. The mass does not demonstrate any diffusion restriction and there is a mild patchy post-ventral. This is an example of the lateral type of ependymal mass which is centered in these about these angles. Again it is a T2 hyper intense mass without any diffusion restriction and then also diffuse both contrast and have some. The lateral type of megaloblastomas typically have a worse problems compared to a mid-floor type partly because of difficulty in obtaining a complete surgical infection given the presence of important neurovascular structures in that location but also partly because of the more aggressive genetic subtype associated with these tumors. Cs of dissemination can occur with ependymoma but it is much less common compared to ependymoma and if you see Cs spread at the time of diagnosis then you should consider the anaplastic sub. The genetic subgroups of ependymomas are not as well defined as megaloblastomas and JPAs however most recently the intratentorial ependymomas have been divided into two major groups the procephalotide A which is also the pediatric type and more commonly seen in mid-lines and the procephalotide B which is called the adult type with more commonly seen in the T3 angle system and it has a worse problem compared to type A. Also based on genetic studies it has been recognized that the intratentorial, procephalotorial as the final ependymoma are distinct disease entities and they have different genetic mutations and these subtypes of ependymoma are most probably going to be included in the next W.H.O. classification which comes out in 2020. A brief word about some of the less common procephalotumors including ATRT and Coralphexas tumors ATRTs are rare tumors seen in children less than two years of age. These are W.H.O. grade 4 tumors characterized by the presence of ratoid cells and the signature genetic mutation is the loss of I and I1 or the DRG1 gene. These can occur both to crime and intratentorially however half of them are located in the intratentorial location. These are on the differential diagnosis called methyloblastomol because of their hyperdensity or PT and presence of deficient restriction on MRI. However these are typically seen in much younger age youth as we can see this patient is probably three months old. So the young age in association with relatively heterogeneous appearance of the tumor can't help you but just the diagnosis of ATRT although you may not be able to completely distinguish it from methyloblastoma. Coralphexas papilloma these can occur in pediatric as well as the age group. Pediatric tumors are more commonly located in the natural ventricle when the adult tumors tend to be located in the forced ventricle. The key imaging features of Coralphexas papilloma is the morphology of the tumor. They have this lobular from like appearance to the margin. These are typically hyper intent demonstrate post-frontal enhancement and speckle calcification is very common in about 25 years. This is an 18 year old with Coralphexas papilloma. We have a hyper dense mass and it is a forced ventricle with multiple parts. The mass is hyper intense on T2 weighted images. There is not enough restriction. So like I said this is lobular and from like morphology of the tumor should be your clue to distinguish like JPA and ependymal mass in this case. When somebody I talked about the four major for safe of the tumors I talked about how location morphology and decrease restriction can be extremely helpful in distinguishing between these peptides. I talked about the evolving molecular classification and how it applies to imaging and finally I spoke about some diagnostic indicators based on imaging. For future even the rapid advances in the molecular subtyping and the treatment protocols of these tumors, our conventional imaging may not provide adequate information to guide the precision medicine and we need to really rethink the RADFAT approach we have for our image interpretation. It probably needs to be more a RAD molecular approach going forward. The goals of imaging in future should be to predict the molecular subtypes and the important being when the patient goes for surgery the only information surgeon has is imaging and oftentimes surgeon is faced with a dilemma between achieving a complete restriction versus risking neurological deficits especially for tumors such as modular blast tumors which can be closely related to brain steps. Having some idea about the molecular subtypes can help them decide how aggressive they need to be in trying to remove that last bit of tumor. Also molecular subtyping can take a few weeks to few months depending on your access and resources and imaging based predictions can then also be very helpful for treatment planning in terms of radiation another goal of imaging should be to identify quantitative imaging biomarkers which can help predict response to these novel therapies to this end multiple advanced imaging techniques including diffusion or fusion and MRF techniques are being looked at. However the most potential is probably with AI and machine learning techniques that you are using a radiometric extraction or a CNN based deep learning technique and it should be interesting how they plan out in the future. With that I would like to thank you for your attention if you have any questions or if you like a copy of this slide be free to email me. Thank you so much.