 I'll be continuing from where we left in the part one of MRA brain tumor imaging and update. In the first part, we talked about how DTI and MRA spectroscopy are useful as add-ons to the structural imaging in brain tumor imaging. Today we'll look at perfusion, ASL imaging and also functional bold imaging. Let's start off with MRA perfusion first. It is today considered as the biological marker of malignancy grading and prognosis in brain tumors. It measures the degree of tumor angiogenesis and capillary permeability in the given tumor and that is considered as hallmark in differentiating benign tumors from malignancy tumors and in malignancy tumors, what kind of grading that you are dealing with. Increase RC building, typical cutoff is 1.75, is almost always associated with higher tumor grades. Anything lesser than that is considered closer to benign. How does MR perfusion work? We must understand the concept of perfusion, how it is different from simple enhancement that we see on non-perfusion imaging. So simple enhancement which was traditionally used for over 30 years to look at the pattern of enhancement in brain tumors indicates only disruption of blood-brain barrier in the given tumor and at the margins of the tumor. It does not necessarily correlate with tumor grade. MR perfusion principle is based on new angiogenesis. It assesses the number and severity of leaky tortuous vessels in the given tumor. This is associated with increased blood flow in the given tumor and increased blood volume. It also depicts the tumor grading very precisely based on new angiogenesis and presence of leaky new vascularity. How does MR perfusion perform? So what we do is we inject rapid intravenous gadolinium, approximately 15 ml of it at a rate of 5 ml per second and then you obtain a time series of GRE EPI-based intubated images covering entire brain and then looking at the area of interest and this is how sequential set of GRE EPI-based images, 32 star images come. This is the area of tumor. The reservation is not going to be great but this particular portion of brain will pick up contrast early and wash out early. So washing out is seen here. So you get a set of about 32 sets of images which are then post-processed to look at the perfusion of given SOI. What are the current applications of perfusion in brain tumor imaging? First and foremost it differentiates tumor from non-tumor conditions like tumor factor demagnation or benign regions which simulate tumors. It differentiates types of tumors. For example glioma versus lefoma versus metastasis. In the glioma it allows us to grade the tumor more accurately. Also we know that most of tumors will have multiple grading within the tumor. So which one to pick up for biopsy and which is likely to give a higher grade can be assessed using MR perfusion. In post-operative or post-relief therapy situation it differentiates recurrence from post treatment changes. Also it assesses response to therapy very very accurately. It differentiates true progression from pseudo-progression and it differentiates true response from pseudo-response and we'll see some examples. Optimum cutoff values for grading hydrate tumors and differentiating them from low grade tumors is today based on RCBB and also upon Colleen creative ratio and Colleen NA ratio. The typical cutoff of RCBB is 1.75 to differentiate high grade from low grade. That of Colleen creative is 1.56 and that of Colleen NA is 1.60. Based on this MR perfusion has much more sensitivity that is about 95 percent. Quite good specificity that is of 57.5 percent and a very very low error rate compared to MR spectroscopy. And here are the classic examples. If you look at this post-conferenced non-fat-sector tumor-related images, this is why barely enhances. But on MR perfusion it's RCBB was 2.1 and on biopsy this turned out to be an acoustic astrocytromark. One more example on structural imaging there is hardly any enhancement in this left tumor. On MR perfusion it showed perfusivity of 4.1 and on biopsy it turned out to be GBM. Here are two different patients showing ring enhancing regions. The first patient shows ring enhancing region in the midbrain. The other one shows two ring enhancing regions in left front lobe. On morphological imaging it will be virtually impossible to decide what we are dealing with. So look at MR perfusion. This midbrain region is hypoperfused whereas these left front lobe regions are markedly hyperperfused. The first one turned out to be Tiberkruma. The second one turned out to be metastasis from serial lung. Here is another patient having right parietal lobe glioma. The structural enhancement pattern is quite heterogeneous you don't know whether you are dealing with hygrids or low grade tumor. On MR perfusion the perfusion was low to the tune of 1.2. Patient decided to go against the medical advice of biopsy, came back after 22 months and if we look at the imaging now the region and increased MR perfusion in size cystic necrosis has increased enhancement and heterogeneity has further increased MR perfusion showed a CBV of 2.8. So we know we are dealing with a tumor which is low grade to begin with and has now turned hygrids. There is another analysis that we what we have to do when we analyze MR perfusion and that is called as mean grove analysis. So it's basically drop in the signal on it to start set of images that we are looking at. So when we inject contrast after about 18 seconds when the contrast reaches brain if you are dealing with hygrid tumor the CBV in the tumor bed will show steep fall and quick wash out and then the severity of drop and wash out is what we assess using mean grove analysis. Here is an example we know that we are dealing with a hygrid glioma in the right frontal lobe exchanging in the measurement here if you look at the mean grove analysis the drop is very steep recovery is also very fast but it is to the tune of about 60% from the baseline. So when the baseline recovery is low and incomplete we know that there is lot of contrast leakage through the leaky new masculinity that these hygrid gliomas have. So this further confirms that we are dealing with the hygrid glioma compare that with this yogurt glioma in the right perishable region the pickup of contrast is not so steep and wash out is almost 80% 85%. So yogurt gliomas will have more wash out hygrid gliomas will have less wash out. Here is another example of this left middle cerebellar peduncular tumor it's dark on T2 it is shown intense homogenous enhancement we know we are dealing with a lymphoma. If you look at perfusivity this patient had a CBV of 4.2 if you look at mean grove analysis the pickup of contrast is very steep and quick wash out also is very steep and quick and the recovery is almost to the tune of 120% and this is considered to be very very typical of lymphoma. So lymphoma shows approximately 120% baseline recovery which is considered typical. So let's try to differentiate using mean grove analysis glioma from metastasis from lymphoma. Our CBV is very very high in glioma but is almost equally high in metastasis as well as lymphoma. But mean baseline recovery is much below baseline in glioma. It is below baseline but not as low as glioma in metastasis and it overshoots baseline in lymphoma. This happens because of angiocentric grove pattern and widening of perivascular spaces that are associated with lymphoma. Differentiating radiation necrosis versus tumor recurrence is also very clearly defined and has specific protocols when you are talking about MR perfusion. Here are two borrowed examples of differentiating radiation necrosis from tumor recurrence using MR perfusion. Here is a dirty looking region in this left parietal lobe glioma patient post radiotherapy. But if you look at MR perfusion, the perfusion is less than 0.6. So this is such a radiation necrosis. Based on morphological imaging it is very difficult to guess that. This is another patient having two innocuous looking regions in left frontal and parietal lobes in the peri-ventricular region. But if you look at MR perfusion, these are intensely red. So you know you are looking at metastasis along peri-ventricular region in this patient with left parietal lobe glioma which was treated surgically and after giving radiotherapy. So the tumor bed is quiet but there is spread along peri-ventricular white metatars. Here is our own example of bed necrosis in the left temporal parietal tumor. If you look at this region, it looks quite similar to this region. But if you look at perfusion of this region, there is hardly any perfusivity. So we know we are dealing with radiation necrosis here. The radiation oncologist must have overshot his area of interest. In the tumor bed however, there is intense uptake and early washout which is indicative of tumor recurrence. So there is tumor recurrence in the tumor bed in the temporal lobe and radiation necrosis in the parietal lobe. So that is about usefulness of MR perfusion in brain tumors. Let us move on to ASL imaging or artificial spin leveling and how it is slowly but definitely replacing contrast enhanced perfusion. Technically what we do is we do not inject any contrast. That is how it is called as galvanum-free MR perfusion. It uses magnetically labeled arterial blood water protons to use as endogenous stressor. So what is done is you label the protons going through carotid and posterior circulation by applying a tag. Then you stop the tag and get images without tag minus out the tagged images from non-tagged images and you actually get a CBV. This is how normal ASL-CBF map looks and this is how ASL without contrast looks. So here the patient with metastasis in left posterior parietal parietal region, ASL showing 4.2 and DCS map showing 3.6 of our CBV. We know we're dealing with a high-grade tumor. Let us move on to bold imaging or functional imaging as it is called and what is its role in brain tumor imaging. So patients undergoing surgery, the surgeon would like to know the location of tumor and how near or how away it is from delocant cortex and also the surgeon would like to know cerebral dominance, especially in left-handed patients. So there are several articles. This is the one which is followed most often. What it talks about is location of tumor with the delocant cortex and the outcome. So in the lesion, if the tumor is away from delocant cortex by movement 2 centimeters, chance of that patient getting post-resuscitation deficit is virtually zero. If the lesion is located between 1 to 2 centimeters of delocant cortex, one-third of these patients will show deficit. If the lesion is located less than 1 centimeter from delocant cortex, almost half the patients will have post-operative delogical deficit. Here are our overall examples. This is a patient with left frontal tumor. The foot motor function, the primary area, is located over 3.5-4 centimeters away from the tumor. The hand area is again located about 3.5 centimeters away from the tumor and speech area is quite away. So chance of this patient getting post-surgical deficit is very, very low. That is about usefulness of functional MRI or bold imaging in tumor imaging. Let's now look at certain upcoming things like fusion imaging. So if you can add PET to MRI, it will give you not only the structural information provided by MRI and functional information provided by MRI. It will also add metabolic facet to it. So a neurosurgeon will get not only structural and functional MRI information, but the surgeon will also get metabolic blood information. So it will give extreme human anatomical information using MRI, superior soft tissue blood characterization by MRI, better temporary resolution by MRI, and comparatively less radiation hazard. Here is a borrowed example of a 44-year-old male who was a known case of left temporal low-grade glioma operated six years back and treated with ADSRP, now has suspected recurrence. What we see on routine structural MRI is the heterogeneously elastic region in the left front, the perimenticular white matter, which is hyperperfused on perfusion imaging, shows high value on SL imaging and also hypermetabolic on PET imaging. On spectroscopy, it is showing elevated choline and presence of lipid lactate. So we know we are dealing with a high-grade recurrence. Also on DTI imaging, there is destruction of the adjoining tracks. So all this information, if provided together, allows neurosurgeon to plan his therapy much more accurately. One more thing that we should be aware of is also sampling bias. So it's our job to tell neurosurgeon where to biopsy from if it's not going to rescind the tumor. For example, again in this borrowed example of right frontal lobe tumor, chance of getting high-grade changes in the given tumor are better in this intensely enhancing area compared to its corpus carousel part, which will show low-grade tumor. Let's now look at some newer trends and the whole focus is now moving from pure structural and functional imaging to molecular imaging because we now know that tumor prognosis depends not only on histology, but also on its molecular microstructure. There are three such important mutations which we need to report. That is IDH, ISD1, 1P192 correlation and MGMT methylation. And these will actually tell you, in addition to the morphological and functional imaging, the eventual outcome in the given patient. Let's understand some key concepts in this. So if the patient is IDH1 mutated, it has better prognosis and these are typically seen in low-global astrocytomase. If there's no mutation, it is called as a wild type of llamas and they have worse prognosis and these are mainly seen in primary GDNs. If patient has 1P192 correlation, which is GDN that has better prognosis compared to 1P192 intact, those patients have worse prognosis and deletions present in majority of oligodendrologmas. If patient has decreased activity of MGMT methylation, they have better prognosis compared to patients who have normal activity, which will have worse prognosis. And this key concept was put forth in this meta-analysis conducted by New England General of Medicine about three years back, which talked about how long is the survival in low-grade llamas versus high-grade llamas. So this talks about temporal survival. So if you look at this graph carefully, low-grade llamas, pandas have coarse cysteine survival. High-grade llamas hardly anybody survives beyond two years. If you look at these patients in this blue line, these started off as low-grade llamas, but they had wild type of LVH mutation. So instead of behaving like low-grade llamas, they have actually behave like high-grade llamas. So this tells us importance of how molecular imaging is going to play a part in outcome for the tumor in future. And the importance was given by the new WHO 2016 classification of brain tumors. So this has molecular imaging has significantly changed the classification of a number of tumor families, and it gives a greater reliance on molecular markers. And this is how today we are supposed to define the tumors. So histological classification is given an importance, but equal importance is today given to molecular information. So classification of tumor is divided into four layers. We finally integrated diagnosis, histological classification, WHO grade, and molecular information. These three were always available. This one has added a new facet, and this has also changed several tumor classifications from 2016. So an anaplastic oligodendro, when you are describing it, you have to call it as infiltrating gamma with oligodendrogal features by microscopy, WHO grade III classification, and IDH1 mutation with whole arm loss of both 1p and 19 new legs. Based on this, several new entities like epitheloid, glioblastoma, diffuse meningomaniacal, leptomeningial, and cellular tumors have come forth as new entities. And this will probably be the way we will have to deal with brain tumors in future. So to conclude with integration of conventional and advanced imaging techniques, we can now provide increasingly detailed information about underlying pathologies. These details will aid in improving our understanding of brain tumors and help in development of new treatment strategies and regimes in future. So as I said in the previous part of this talk, radiology has to be both image-centric and patient-centric. We must understand that we are an important part of the patient management. Inter-speciality communication and coordination is very, very important, and we should be the first part of it. We should aim to change from volume based to value based imaging, from interpretation focus to outcome focus reporting, and the responsibility of training our future generation with not only the advances, but these values lies with us. Thank you again for your attention.