Dual Energy CT in Gastrointestinal Stromal Tumors: Unifying Advanced Treatment with Advanced Imaging

It's a pleasure to be here and talk with you about some background from our multicenter study and the purpose where we do believe in dual energy CT in GIST tumors. So kiss tumors, actually relatively rare tumors, but they can occur in the whole GI tract, and almost although they are rare tumors, they present 80% of all sarcoma tumors in the GI tract, and I think the main interesting thing. About Gist is that almost 85% of August tumors are associated with a certain pathologic pathway. The so called C kid pathway that makes it interesting for oncology as well as imaging, and that's it. Kid pathway, actually in space on certain mutations in the secret receptor, mainly those exons, exon 9/11, 13 and 17. And if this is mutated, we have uncontrolled activation of the secret receptor leading to. Cell survival and an controlled as cell proliferation, and that was the success story of the TK I therapies that led to that blockage. Blocking of the secret receptor an when that was identified in the 90s on the Cal cells increase tumors that led to the introduction of first line, and that's special about customers. First line targeted therapy, even in patients with very advanced stage in gifts. And it was a success story actually. For and a big business for Big Pharma, so relatively the milestones of distal therapies is relatively young. So first the covering discovering of the effectiveness of 1st land. Imagine it therapy in 2001 followed by FDA approval. Then Sunity Nip followed and then in the middle in 2017, people realized that we need new response criteria. If we have a certain tumor like this. And then also now novel treatment and Ativan therapy with Raffey Nip introduced in 2013. So why is it so important to be accurate in targeted therapy patients in GIST tumors? Because what you want is no false rating of these patients because the most effective treatment that people can get with customers is first line therapy with imatinib almost all patients develop resistant on first line imaging and therapy, and this is then followed by sunitinib therapy. But when you. Look at the median progression free survival you see you want to keep that time for all patients that you are certain they still benefit from the therapy because the median progression free survival with all the second and third line therapies is much less when compared to the first line treatment therapy and it's typical appearance of the tumors, so they look cystic. It's a more or less MCSO ET generation that we see in this tumors, that pathologic correlation we have not used that are. Highly vascularised with central cavitation and that makes it degeneration. So actually relatively often difficult situation to accurately predict response to that treatment. And that then led to a whole new introduction. Certainly developed for Kiss tumors by the group from Houston. The joy criteria which described the effect that we often do not see at decrease in size. But we see changed in the density of the lesion and the argument was then that we have to rate this as therapeutic response. Even size is not decreasing. And why is that important? Because when you also look at the median progression free survival, if you apply only size measurements you do not have a real good separation over the whole study population between good response responders and poor responders an which is much better when you strictly apply the joy criteria. So that led to a second publication in Chasey. Oh stating we should desist using resist, at least in gist. What are the criteria now? It's actually not that clear as it stands in here, because we either see a decrease in size or a decrease in Hounsfield unit by 15% and based on our experience that works well if you scan the patient exactly on the same scanner with exactly the same protocol all the time, but even Q current modulation or variations in cavi settings need to high variations that can. Be more than 50% in Hounsfield measurements and that was the purpose why we thought we should integrate tool Energy in this tumors. Another poor thing and this is something I think we have to learn is when you summarize all the registered trials that are currently ongoing. In this patients, you see that most of the trials still use resist for follow up and that points out how poor radiology's involved really in the setup of clinical trials. But we need the trials if we ever want to have like a technique like to Energy City, we have one trial. I'm talking now about to have a certain follow-up in ECMO, our school guidelines. So we have to be more active in this area to register our trials while we doing dual energy instead of dynamic perfusion imaging. When we talk about vascularization, that's a typical case because we see often a very diffuse spread of disease. With when you look at this case diffuse liver metastasis but also diffuse para toneel sarcoma ptosis and the main effect, while we desided doing dual energy CT instead of doing dynamic imaging is that we have one technique that covers the whole burden of the Seas relatively easily. In those patients, and the patients often go many follow-up studies. That was the initial idea from colleague of mine Doctor Abfalter, who looked at joy criteria and the so called iodine related attenuation based on dual energy CT. And that was the first study that found a very good correlation by measuring iodine density and compare that with the joy criteria and that was followed up then by a study in which we looked at the overall survival in those patients when we use resist joy criteria or the dual energy related criteria. And that shows you that you have poor overlap in here. When you use resist. But even in our Department, when we apply the joy criteria and when you scan them on various scanners, joy didn't really work to differentiate stable and progressive disease in those patients. And I talked about the effects that you mainly maybe put patients on the wrong therapy based on that decision. That's a case in which you see why is pseudo progression important. There's a leash and and to follow operation is progressive in size, but you now with the dual energy CT you have more than one Hounsfield measurement. We have to virtual noncontrast's information to contrast information and the mixed ratio and we see that here there's. Tremendous decrease in the iodine uptake of the nation, so that is a good example for pseudo progression and also in the other lesions that you that you see in here and another effect is why we look at the whole vital to more iodine now is that effect from cystic lesions. There's no change in the maximum diameter of this lesions. Everybody sees it. But when you have a patient in a study in which you only are allowed to measure the maximum diameter, you see that not you'll insist. That is often or curing and kiss tumors. Pure indicator for progressive disease, so that's also one of the main limitations that we see with resist. So. Another thing is what happens when we only measure a Hounsfield units, like using the joy criteria, and that's the main benefit that we see in dual energy imaging. So we have here diffuse spread of the Seas with metastatic lesion around the rectum here, and you see that the lesion is increased in density on this scan, so this could be either vascularization or tumor hemorrhage. So we now have the main effect that we can calculate the virtual noncontrast images from. All images and you see that this was mainly based that increased density by intratumoral hemorrhage, which often occurs during teeki therapy. So that's the main benefit differentiating tool Energy City from the joy criteria. Another example in here in which dual energy CT shows you based on the maximum diameter of this lesion. Here is stable disease, but it can also occur the other way around. When you measured the Hounsfield units, no difference, but the overall vital IO time content in that lesion is increasing. Here an relatively strong, so that's another round example for progressive disease, in which you make the right decision. Based on image Ng. There are some side effects that come along with that quantification of tumor lesions, so you also can calculate your image. This is even only first generation wanna energetic imaging, but you can calculate those. Locate vemana energetic images to also more accurately detect lesions, which is of particular importance for liver image Ng when you have small lesions. And now coming to our trial an that's the distribution of therapists of patients that are included. You can see the purpose of the trial and the dead registration number, so most patients are under first line therapy. Lot of them now, without metastatic disease, and then also alternative therapies that are now evaluated in clinical trust beyond hookup. Agathon IP an. Also, a lot of patients with second line therapy. So decide that participating centers so the Royal Marsden Clinic in London is participating. Milan and the University Medical Center, Tohoku, and we now follow up the patients. Based on all these criteria, we have 1/4 one which is the total vital iodine which we are currently looking at and it seems that this is the most accurate parameter. But even if we only do the iodine overlay, it seems to be much more accurate. Some of the examples, so there's. One of the scenarios so you would not be sure what to do with the toy criteria, because here you see on the density at decrease by 15%. But joy never mentioned what happens if there are. Is there a threshold? What do we accept in increase in size is only measured decrease or decrease in density, but it's really unclear what happens when there is a definite increase in size but also decrease in House field units. And that's an example. For maximum progressive disease even there is based on Hounsfield units at decrease in density or the other way around here, there's a small really almost too small to correct arise lesion in the liver that is increasing in size and based on density no decrease. So based on that second scan you should rate this patient as a progressive disease and put him on second line therapy. But follow up showed and dual energy also showed at this time of in point. That there's a respond less iodine object. It was caused by intratumoral hemorrhage. So these are only the 1st results on progression free survival. We're still now follow up the patients for overall survival, which is think much more important but progression free survival from this cohort I currently showed you these the recist criteria joy criteria that soul good results and when you compare that to the vital iodine progression free survival, there's the largest differentiation between responding patients and non responders in case so. With this I want to conclude and I'm very happy when we soon can show you the overall survival data. Thank you very much.

Standardized Quantification JOURNAL OF CLINICAL JOURNAL OF rumor cell wmor cell Response Assessment of GIST DECT: M3GlST — Response assessment in patients undergoing JOURNAL OF CLINICAL ONCOLOGY JOURNAL OF CLINICAL ONTOLOGY GIST Multicenter M3 — Prospective international GIST Multicenter M3— Prospective international Why dual-energy instead of single source aquisition? GIST - GIST Kaplan Meier Curves for PFS Background False positive PD unter Imatinib False negative PR unter Sunitinib Progressive-free survival after adjuvant T Kl therapy Milestones in the therapy of GIST UMM We Should Desist Using RECIST, at Least in GIST Response evaluation of targeted therapies using Dual Energy CT Choi criteria 10 RECIST target therapy: Progressive disease based on DEC T target therapy: Stable disease based on RECIST Imaging therapy response of gastrointestinal stromal tumors 1.0 multicentertrail (WHO: UI 111-1171-6041) multicenter trail (WHO: UI 111-1171-6041) GIST - GIST W GIST (GIST) with PET, cr and MRI: a systematic review (GIST) with FOG PET, CT and MRI: a systematic review (GIST) with PET, CT and MRI: a systematic review (GIST) with FOG PET, cr and MRI: a systematic review Correlation of Computed Tomography Positron of Computed Tomography and Positron Multicenter study Tumors CT Treated a Single 04 07 06 27 — 0.7 O • • • • • • — stable/partial response App: CW Mixed 0.3 Institution With proposal of New Institution With Proposal Of New RECIST Response Evaluation Choi Criteria Response Evaluation 35 • • • Mean Criteria Stddev:l. Stddev: Area - 8. cm2 Area cm2 No Area: cm2 2007 University Medical Center Mannheim, Tumor response? Tumor progress? Tumor response? 0.2 Thomas Henzler, MD Thomas HenzLer (95% KIT positive) and the introduction of the KIT receptor blocker 20 15 25 30 JOURNAL OF CLINICAL ONCOLOGY ONCOLOGY 0.3 New vital tumor area New tumor hemorrhage Dietmar J. Dinter, MD, • peter Hohenberger, MD,t U. Joseph Schoepf: MD, to — progressive — prmressive — prmrossive — progressive progressive Mean 0.3 AB CT RECIST SIEMENS Thomas Henzler, MD SIEMENS .. Of 2S 25 20 University Medical Center Mannheim, University Medical Center 0.2 0.6 CT NO No 1.0 20 25 30 20 25 25 20 25 100 15 With 20 30 25 10 Prof. Of Radiology A in Of a in density (HU) Cr A in Of a in (HU) Cr Time (months) 06 NO Head of Computed Tomography GIST •4 NO Obvious Of 1.0 110 Healthineers Iodine Overlay Image University Medical Center Mannheim, University Medical Center university Medical Center Mannheim, University Medical Center Tohoku Iodine related attenuation meet the for CR. PR. PD the for CR. PR. PD DECT sum 15 Mannheim, Germany NO 0.5 0.6 0.2 0.9 0.9 RECIST 1.1 0.2 04 02 06 An in Of Of PR by p•00496 9-00496 9-0.0496 p•00496 0.2 0.3 0.9 Choi criteria 1.0 30 90 15 J App CW Mixed CW Mixed 03 . App CW Mixed 0.3 CM,' Mixed 03 (n • New 1.0 Mean' 29.0/20.9/49.5 29.0/'20.9,'49.5 29.0/20.9,'49.5 New in the of the existing _ 0.7 — 0.7 06 Iodine overlay Stddev: StddeV stddev Stddev 0.6 0.7 0.0 100 200 2007 15 18 18 15 20 200 1.0 700 30 80 M 0.3 0.0 0.3 0.9 25 70 30 Area, 2.8 Area 10 15 18 24 27 21 20 24 27 Response evaluation: Iodine Time (months) (months) 19.7 "u 19.7 Hu 04 • Complete response Thomas Henzler, MD 90 110 RECIST Area 39cmn2 Area 39 cm2 Area 39cm2 Mrea Mrean Area Mean Std Error std Error Mean Std Error std Error Group cr — Stable / Prof. Of Radiology Professor of Radiology —Stable respmse —Stable respmso —Stable respmsc 20 • 20 25 2007 0.3 Head of Computed Tomography 15 123,393 15, 7984 — progressive progressive 111,79 16,4225 12,0529 15,4165 Stabile disease 2. Lirw (Sunfiib) 2. Lino (Sunithib) 2. Line (Sunitinib) 3. Lino (Rogwahnm) 3. Line (Regorafinib) 2. Line (Sunitinib) 3. Lino (Rogorahnm) 3. Line (Regoafinib) 3. Lino (Rogorahnb) I. Line (Irmbnib) I. Lino (Irmbnib) 2. Line (Sunitinib) I. Line (Irmabnib) I. Lino (Irmabnib) Atornativo Thorapio Atornativo TKI Thorapio 2. 15,4165 15,0522 15,522 stable/partial response 3743,351 stable/partial response 3743,949 stable/partial response 334,54d stable/partial response 356,545 stable/partial response 3343,949 1,96131 1. Line (Imatinib) 3. Line (Regorafinib) 2. Line (Sunitinib) (Nilot"ib, Maitinib, Pazopanib) (Nilot"ib, Maitinib, Pazc»anib) 3. Line (Regorafinib) Alternative Therapie 1.0 1.0 20 25 15 1.0 25 40 20 7 15 70 30 20 00 100 30 18.09.2015 18.03.2016 18.09.2015 22.05.2016 27.05.2015 27.05.2015 18.03.2016 24.06.2015 16.09.2015 18.09.2015 10 20 30 40 50 60 70 80 90 100 (Nilotinib) (months) Time (months) (months) SUV imaging UMM IJMM UJMM FOG IRA SIEMENS .. Apfaltrer, ...Fink, Invest Radiology 2012 Sudarski S. T, Academic Radiology 2014 Meyer, Henzler et al. EJR 2013 Meyer, Henz/er et al. OR 2013 Meyer M et al. EJR 2013 4. Healthineers 40