 Dear all, this is Amit Choudhury. I am a lecturer at Tata Memorial Hospital, Mumbai. This presentation on DSC perfusion of brain tumors will include a video documentation of open source software for analysis of perfusion images. So, brain tumors typically have increased vascularity, which is because of new angiogenesis, which is brought about by these tumors by expression of certain factors like a vascular endothelial growth factor in addition to many other such growth factors. This leads to generation of leaky blood vessels or capillaries which lack a blood-brain barrier. As a result, there is contrast enhancement, which is because of leakage of the contrast from the blood vessels into the interstitium of the tumor. And there is also contrast recirculation because of disorganized capillaries, which do not immediately send contrast into the veneals. So, there are multiple MRI perfusion techniques, which are available these days. The perfusion imaging can be done with or without using contrast agent. Perfusion techniques with the use of contrast agents include dynamic contrast enhanced perfusion and dynamic susceptibility contrast perfusion. And those without contrast agents include arterial spin labeling, which is non-contrast based technique which uses tagged blood and blood oxygen level dependent imaging and intravoxial incoherent motion, which is a new kid on the block. So, this is the protein DAC perfusion protocol, which is followed these days, which is based on the ACRS in our perfusion guidelines. Typically, use of either 1.5 tesla or higher field strength magnets recommended for perfusion imaging. This implies a GRE based EPI sequence routinely. And the position plane is usually axial, oblique axial plane, that is the ACPC or anterior commissure, posterior commissure line plane. Acquisition is made in 2D mode. And this involves administration of a double dose of water, that is about 0.2 millimoles per kg at the rate of 4 to 5 ml per second, which is done by means of a power injector. The repetition time is usually 1.5 seconds or lesser and a flip angle routinely employed is higher of about 75 degree with approximately 40 time points, which are routinely imaged on the scanner. The acquisition matrix is somewhat similar, a little lower on 1.5 tesla and a little higher on 3 tesla. Size thickness is typically 5 mm without any interface gap. And an acceleration factor of no more than 2 can be used for perfusion imaging by means of parallel imaging technique. Once perfusion acquisition is made, analysis can be done on one of the vendor provided software on the console or extended workstation. Or it can also be done on open source software like UMM perfusion on horrors or osirics on a Mac. For perfusion analysis, open the perfusion sequence in 4D viewer in osirics or horrors. Select the open source UMM perfusion plugin. It will warn you about its open source nature and that it is not certified for clinical use. Select fast deconvolution. Then select signal enhancement e2 star DSC. Change the window settings as required. Go to the level of skull base. Go to the time section where maximum susceptibility is seen arising from the middle cerebral artery. Draw a close polygon ROI, which will serve as the ROI for defining arterial input function. I am selecting here the baseline as two because the contrast injection was done too early in the perfusion sequence. Then you hit generate. This will create the various perfusion maps that is the plasma flow, which is nothing but cerebral blood flow. The volume distribution, which is nothing but cerebral blood volume and the empty to your main transit time for tumor imaging predominantly cerebral blood volume is localized. So I'll turn off the other two generated maps. You can open contrast channel sequence for comparison or defining ROIs in the region of tumor. Change the windowing so that the cerebral blood volume map is properly visualized. You can select color lookup table with the rainbow color view and adjust this map so that you get a proper perfusion color map. Adjust the contrast or window settings on the native perfusion image against sync all the images. One can then draw an ROI on the areas which are showing high CBV levels, which are appearing here in green colored or green and red hues. This can again be done using an open polygon or an overload circular ROI. Copy this ROI command C to the native perfusion image command V. Similarly, draw an ROI in control lateral unaffected white matter. Copy this ROI command C from the cerebral blood volume map to the native perfusion image command V. And then you can get various measurements. For example, here you can look at the mean measurement, which is in the region of tumor, which is showing a value of 25. And if you contrast that with the control lateral ROI, which is showing a measurement of approximately 10, thus we have an RCBV of the tumor compared to normal appearing control lateral white matter, which is giving us a value of approximately 2.5. For looking at the various signal intensity time curves of the ROI is just drawn, you have to go to all ROIs, select the newer drawn ROIs, check the boxes there and that will display the selected ROI signal intensity versus time curves. So the curve in blue is the signal intensity time curve from the tumor. The curve in yellow is the signal intensity time curve from the control lateral normal appearing white matter. And this red curve is the arterial input function. Similarly, ROIs can also be drawn on the enhancing portion of the tumor in the T1 post-contrast sequence. And this ROI can be copied on corresponding CBV and native perfusion images. Selecting this ROI will give you the signal intensity time curve for that ROI, which is seen here in dark blue. This is how one can perform perfusion analysis using open source software called UMM perfusion, which is freely available. The link will be shared in the comment section below. So perfusion imaging is a technique which utilizes the bolus trigger that is passage of contrast bolus through the blood vessels in the brain parenchyma. Typically, this has a baseline, a dip in signal intensity once the contrast arrives in the region of interest followed by baseline recovery and this last portion which is known as tail. So the time between administration of contrast and the minimum concentration of the contrast or minimum signal is known as time to peak. The area under the curve of this negative enhancement or this drop in signal intensity is known as negative enhancement integral and the cerebral blood volume is calculated from this negative enhancement integral. MTT is the time between appearance of the bolus and recovery to baseline. So since DSC employs T2 or T2 star weighted imaging, there is a drop in signal intensity in contrast to even base techniques where one would see enhancement or increase in signal intensity as would be seen with dynamic contrast enhancement techniques. DSC technique is the most common technique utilized for perfusion imaging of brain tumors. So once perfusion study is done, this technique generates maps which are known as cerebral blood volume map, blood flow or CBF map and MTT map. These maps are generated using complex mathematical formula and the CBF map can be generated using what is known as the central volume principle which is given by this formula that is CBF is equal to CBB upon MTT. So what are the various applications of DSC contrast in us perfusion? These can be used to differentiate tumor from non-tumor pathologies, these can be used for tumor characterization that is whether a tumor is of glial origin, whether it is a lymphoma or is it a metastasis or is it a meningioma. Then this can also be used for grading of tumors that is differentiating between high and low grade tumors. Perfusion can be utilized for guiding biopsy of tumors and increasing yield and assigning a proper grade to the tumor. And last but not the least, perfusion imaging has an application in the response assessment that is it can be used to demonstrate response to treatment, progression, pseudo-progression or pseudo-response and differentiating between these entities. Moving on to diagnosing tumor versus non-tumor pathologies. So this is an example wherein on this actual T2 weighted image one can see a lesion which is showing central T2 hypo-intensity showing peripheral rim enhancement on post-contrast T1 weighted images and is hypo-perfusion perfusion imaging, implying that this is a non-tumor pathology by virtue of its hypo-perfusion and this was a tuberculoma. This is another example borrowed from literature wherein one can see in the top row there is a lesion which is showing irregular peripheral enhancement and is hypo-perfusion on perfusion weighted imaging, implying that this is either a non-tumorous pathology or a low grade pneumoplasm and this was a demyelinating lesion of baloconceptric type of multiple sclerosis. Whereas in the example below one can see that there is a heterogeneously enhancing lesion seen in the right frontal lobe which is showing areas of hyper-perfusion with RCBV values more than 2 that is 4.1 to be precise in this example and this was nothing but a glioblastoma. Moving on to tumor characterization, perfusion is a very good modality for characterization of tumors especially by looking at the baseline signal recovery. So this technique can be used to differentiate between various tumor types and this is one such example wherein in this first example we can see that there is a lesion which is seen in the right parietotemporal region which is showing areas of hyper-perfusion and it shows baseline overshoot that is the perfusion of the tumor region signal intensity is overshooting the baseline and this phenomenon is considered to be characteristic of lymphoma and this was a lymphoma on histopathology. This is an example wherein one can see a lesion in the right frontal lobe white matter which is also showing increased perfusion with decent baseline recovery however not complete that is there is a baseline recovery which is more than 50 percent of the signal which is recovered from the nadir of the signal intensity and this was a glioblastoma on histopathology. This is another example wherein a tumor or neoplasm is seen in the right frontal lobe which is showing hyper-perfusion and here one can see that there is not a good baseline recovery and the signal intensity continues to stay below the 50 percent of signal intensity from the nadir of the signal after administration of contrast and this was a metastasis at histopathology. So why do tumors show different kinds of baseline recovery so in cases of lymphoma there is something called as T1 weighted effects which predominate by virtue of the cellularity and vascular arrangement and lack of blood brain parenchyma in lymphomas. In cases of glioblastoma multi-form which is a destructive lesion there is there are leaky channels thus there is lack of blood brain barrier and there is extra position of contrast from these leaky vessels into the tumor interstitium as a result something called as the T2 star effects predominate and because of this predominant T2 star effect there is lack of complete baseline recovery. In cases of metastasis there is no blood brain barrier that is these tumors tend to have vessel characteristics which are similar to that of the primer in your plasm and because there is absence of blood brain barrier there is leakage of contrast into the tumor interstitium as a result giving rise to T2 star effects and lack of baseline recovery. This was another example wherein one can see that there is a lesion in the periventricular location which is showing restricted diffusion on diffusion weighted imaging which is intermediate to hyper intense on T2 weighted images and which is showing hyper perfusion and this was another case of lymphoma which was diagnosed at histopathology. However in this example we do not see the baseline recovery why so because this technique utilizes what is known as leakage correction algorithms and when such algorithms or techniques are utilized we would not have the typical baseline overshoot which can be seen or which is considered to be a characteristic of lymphoma. Remember that absence of this baseline overshoot pattern does not rule out a lymphoma. This is an example wherein an elderly gentleman on an MRI scan on this T1 weighted axial and T2 weighted axial images has a heterogeneous lesion in the right parietal lobe which is showing areas of blooming which is because of hemorrhage or hemocidal staining within the tumor and which is showing heterogeneous post-contrast enhancement and on perfusion studies one can see that there is a very high CBV about 3.4 compared to contralateral normal appearing white matter and there is not a complete baseline recovery that is the recovery is approximately 50% because of this high grade neoplasm. This is another example a case of meningioma wherein on this perfusion weighted imaging again one can see high RCB values compared to contralateral unaffected white matter and one can see that there is lack of complete baseline recovery because of absence of blood-brain barrier which can be seen with extra axial regions like meningioma. Moving on to grading of gliomas DHC perfusion technique has been well studied for a long time now and cutoff values have been established for differentiating between low grade and high grade glioma which is typically set at 1.75 that is a ratio of RCB between tumor and unaffected white matter which is higher than 1.75 then the tumor will be considered as a high grade tumor. This is an example wherein one can see that there is a well defined lesion which is hyper intense on T2 weighted images has more signal intensity on flare images doesn't show any post-contrast enhancement and this T2 and flare mismatch is considered to be characteristic of IDH mutant infiltrative astrocytoma which is a grade 2 or a low grade neoplasm and on perfusion weighted imaging we can see as expected with low grade neoplasms there is complete baseline recovery and if we look at the ratio between the tumor and the contralateral unaffected white matter the ratio is under 1.75. This is the same example which was discussed earlier is of glioblastoma which is showing a very RCB value of 3.5 implying that this is the high grade neoplasm which can be commented prospectively and this was also diagnosed to be a glioblastoma on astro pathology and as can be seen with glioblastoma there is a poor baseline recovery. Moving on perfusion can also be a good tool for guiding of biopsy also because if one does a biopsy from an area which is hyper perfused then the yield of such a biopsy is typically going to be high and a grade then so assigned to this tumor would be representative of the highly aggressive part of the tumor. Thus perfusion imaging is also a very good tool for guiding biopsies to improve the yield of the biopsy as well as assigning proper grading to the tumor which can further direct management of such lesions. Moving on perfusion imaging is also a very good technique for response assessment. Now this was an example of glioblastoma multi form which has been treated with radiotherapy and temozolomide and what we see on this contrast enhanced image is that there is heterogeneous post contrast enhancement which is identified in the region of left parietotemporal lobes. However these areas are showing high per perfusion compared to contralateral white matter implying that there is some residual or recurrent tumor. In addition one can also see on this contrast enhanced image is that there are this subependimal areas of enhancement which were showing high per perfusion on DSE perfusion images implying there was subependimal spread of disease and this is suggestive of tumor progression rather than pseudo aggression. This is another example of glioblastoma which has been treated with radiotherapy as well as temozolomide and what we can see here on the top row which is the baseline imaging after surgery that there was some enhancement which was identified in the medial aspect of the treated tumor bed which is showing mildly elevated RCBV of approximately 1.2 which was due to some residual tumor versus post treatment changes and on post treatment follow which was done approximately four months after administration of radiotherapy and temozolomide. One can see that there is extensive post contrast enhancement which is seen in the right cerebral hemisphere. In addition there is also increase in edema which is seen in the right cerebral hemisphere white matter. However on perfusion weighted imaging there is no significant increase in the perfusion which can be appreciated on this RCBV map and ROI in these affected areas is showing a RCBV value compared to contralateral side which is no more than one implying these are post treatment changes and no residual or recurrent tumor suggestive of pseudo progression. So pseudo progression is typically encountered in the sentence of radiotherapy or administration of temozolomide which peaks at approximately three to six months after administration of these therapies and a perfusion weighted imaging technique is a very good technique for differentiating between pseudo progression versus actual progression of disease. Again follow up of these patients is very important since follow up has been a proven method of proving or displaying a diagnosis of pseudo progression versus tumor progression wherein in this patient on a follow up one would expect that these changes of extensive edema in the right shredded cerebral hemisphere and the enhancement to subside in the follow up. So these patients are kept on a short interval follow up which will be done approximately three months after the scan which is done for response assessment or may even be done earlier in case the patient clinically deteriorate. Next entity which can be differentiated using perfusion imaging is pseudo response versus actual response or progression of disease. So this entity is encountered in patients with glioblastoma receiving anti angiogenic therapy using monoclonal antibody called bevacizumab and in this example one can see at baseline that is before administration of bevacizumab there is a tumor which is seen in the left frontal lobe which is showing irregular thick peripheral enhancement suggestive of residual tumor along the tumor bed in this postoperative setting and at follow up study after administration of bevacizumab one can see that there is marked reduction in the enhancement which was seen on the earlier baseline study. However if one sees on perfusion weighted imaging the periphery continues to show hyper perfusion implying there is still some residual tumor which can be seen in this patient thus even though bevacizumab causes many positive vascular effects in cases of glioblastoma it may not always cause reduction in the tumor and bevacizumab has been shown to have a very good progression free survival however there has not been any significant difference in overall survival in patient receiving bevacizumab versus those who don't receive this therapy. Moving on to pitfalls of dynamic susceptibility contrast perfusion so this is a technique which involves meticulous administration of contrast agent hence if the administration rate of contrast agent is low or if there is any extravization of contrast one would not be able to acquire good images again there are certain effects which occur because of in case of contrast through the capillaries because of break in the blood brain barrier as a result there are something known as T1 leakage effects and T2 leakage effects and a dynamic susceptibility contrast perfusion weighted imaging is also highly susceptible to any blood product classification metal or air or air and bone interface which can be seen especially in the brain parenchyma adjacent to the sinuses because of utilization of gradient base sequences so this is one such example wherein because of a lower rate of contrast infusion one can see that there is a very tour signal to noise ratio which is seen on the CBV map again because of deposition of blood or some lesions which can be hemorrhagic there can be very poor signal coming from these lesions and as a result these lesions cannot be very well interrogated by DSC based perfusion technique. Another pitfall of perfusion imaging which is because of tumor characteristics is that a certain low-grade tumors like great to oligodendroglioma tend to show very high perfusion values which is because of the chicken wire capillaries present in the tumor and this is one pitfall which radiologist should be aware that perfusion is usually a technique which is an adjunct to conventional imaging wherein in this patient we can see that there is a tumor which is involving the left frontal lobe as well as the cortex associated with thickening of the cortex which is usually encountered with oligodendrogliomas. This is another example which I can never forget wherein we can see that there is a mass which is present in the frontal lobe which is showing elevated RCBV values with RCBV approximately 2 and on coronal T2 weighted images we can see that there is a heterogeneous lesion which is seen in the right frontal lobe causing midline shift is showing heterogeneous post-contrast enhancement and certain very fairly enhancing areas with central necrotic areas and which are showing areas of susceptibility which we considered was because of bleed in a possible high-grade tumor but no sooner did the surgeon perform a derotomy immediately called us and told us that look there is frank pus which is coming from this tumor so this is one such pitfall of dynamic susceptibility weighted contrast image wherein one might get fooled by certain lesions as in this case these were cerebral abscesses which lead to spuriously high perfusion values along the abscesses because of reactive angiogenesis in the brain parenchyma adjacent to the abscesses and if we look at these lesions retrospectively we can see some fluid debris level in some of these lesions implying that these were abscesses so sometimes the diagnosis can be quite difficult to be made prospectively even by using adjunctive techniques like perfusion weighted imaging we need to be born in mind thank you