The importance of D-dimer in the diagnosis and treatment of cancer

Hello, everybody. I'm Doctor John Mutios, Medical Sciences Partner at Siemens Health and Ayers and I would like to welcome you to the next session of the Hemostasis Learning Institute 2025. Today's speaker is Doctor Chihan AI. Doctor AI currently serves as a Professor of Hematology in the Division of Hematology and Hemostasiology of the Medical University of Vienna. His research and clinical work focus on venous thromboembolism, anticoagulation and bleeding disorders. He has contributed to over 250 peer reviewed publications and serves as editor of several scientific journals. Professor AI is also an active member of both the GTH and ISTH in different capacities. Increased thrombotic risk is one of the most common but also underestimated symptoms of cancer. Identification and proper treatment of cancer patients with an increased risk of thrombosis can prevent unnecessary thrombotic events in patients already burdened by cancer therapy. This lecture addresses the central role of D dimer as a predictor of thrombotic risk in cancer patients. Professor I, the floor is yours. Dear colleagues, ladies and gentlemen, it's my great pleasure to give this presentation on the importance of D Dimer in the diagnosis, prevention and treatment of venous thromboembolism in patients with cancer. Here are my conflicts of interest. So during this presentation, I would like first to address why it is important to talk about this topic of venous thromboembolism in patients with cancer and what challenges we have in clinical practice in this patient population. I will also review some of the risk factors and risk assessment models for venous thromboembolism in patients with cancer. Some of them include biomarkers such as the dimer. And then I will review the evidence for the use of clinical prediction rules in cancer patients with suspected suspected deep vein thrombosis or pulmonary embolism. And finally, I will conclude by highlighting the role and limitations of D dimer testing to rule out venous thromboembolism in the cancer patient population. One out of five Vt events is related to cancer. This has been demonstrated in large registries such as the Reiter Registries Register, which is a multinational international registry where of around 100,000 patients, 22.7% had also a diagnosis of cancer, so where cancer was the trigger for venous thromboembolism. So cancer associate thrombosis is frequent but also from the oncologist perspective it's an important problem. During the course of cancer disease, up to 20% of patients with cancer will develop venous thromboembolism. Over the last decade we have observed an increasing number of so-called unsuspected incidental venous thromboembolic events. So these were pulmonary embolism that was incidentally detected on routine staging or restaging CTS. And here we also have a have a have a problem that signs and symptoms of venous thromboembolism of deep vein thrombosis and pulmonary embolism are very unspecific and are in some cases interpreted as signs of the underlying cancer. Another problem clinical problem is that the recurrence rate of venous thromboembolism is very high in cancer patients. So compared to non cancer patients, cancer patients with venous thromboembolism have a two to five fold higher risk of Vt recurrence. So we need, on the one hand, good tools for predicting and diagnosing an incident event of venous thromboembolism, but also good tools for predicting and diagnosing recurrent events of cancer associated thrombosis. The rates of venous thromboembolism are very heterogeneous and depend on the underlying type of cancer. So we have cancer types with a very high risk of venous thrombo of embolism. These are patients with stomach, primary brain tumors and pancreatic cancer. We have a large group of patients with a relatively high risk of venous thromboembolism, including patients with kidney cancer, hematological malignancies, colorectal cancer. And we have also group of patients with a low absolute risk of venous thromboembolism, so around 1 to 2%. This is the case in patients with prostate and and then breast cancer. However, these are the most prevalent cancer types in the population and their contribution to the overall prevalence of cancer associate thrombosis is still very huge. So what is the goal in clinical practice? So we want to prevent VTE and Vt related consequences in patients with cancer and to do this, so we need to understand to identify patients at risk of venous thromboembolism. For identifying them, we have to know risk factors for venous thromboembolism and we also need improved risk assessment tools for venous thromboembolism. There have been many risk factors in predictors of venous thromboembolism described and reported in patients with cancer. Just to give you an overview, you can categorize them into patient related risk factors. These are general factors that are also associated with risk of venous thromboembolism in the general population like increasing age, presence of comorbidities, a prior history of venous thromboembolism, presence of thrombophilia. However, the most important risk factors are the so-called tumor related risk factors. So this type of tumor, so we have seen there are tumor types of a very high risk, others with a relatively high risk and again others with a low risk, the stage of the tumor. So the more advanced the cancer diseases, especially when the disease is metastasized, the risk is increasing further and also the time after. Are occurring in the first six months after the diagnosis of cancer and similarly important are the treatment related risk factors. So here nearly all types of therapies that are given for the treatment of cancer and this has been very well documented for specific chemotherapies like platinum based therapies or other chemotherapies for anti angiogenic agents. For anti hormone therapies, the risk of venous thrombone embolism is very high after major cancer surgery and also supportive treatments like frequent blood transfusions or the need of hospitalization and immobilization is further adding to the risk of VTE. And I think you all know that also adding central venous catheters which are needed to administer therapies in cancer patients are also associated with risk of venous thromboembolism. And there is a fourth group of risk factors that we can categorize as biomarkers which are helpful to predict risk of venous thromboembolism in cancer patients. So among them, so the different blood count parameters like like high platelet count, high leukocyte count have been reported to be linked to increased risk of cancer associate thrombosis. Hypercoagulability or in general activation of the hemostatic and fibrolytic system. Activation of platelets such as D dimer, P select improve hormone fragment one and two Trauma generation potential have been reported to be associated with risk of occurrence of venous thromboembolism in patients with cancer. And here I would like just to show you work that we have done in our group where we have shown that parameters that globally reflect coagulation activation and fibrolases predict cancer associated venous thromboembolism. So an increased peak trombone generation in one study was linked to a two to three fold higher risk of venous thromboembolism. And we have also documented that elevated D dimer levels are significantly increasing the risk of future development of venous thromboembolism in patients with cancer. These D dimer levels have been measured prior to a start of chemotherapy in most of the patients. So you have seen that there are multiple risk factors that contribute to the risk of venous thromboembolism in patients with cancer. And the question now is how can risk assessment in cancer be improved? And one approach was suggested already 15 years ago by Doctor Quran and colleagues who who developed a risk prediction models for venous thromboembolism. And here he suggested a risk scoring model that incorporates 5 readily available variables prior to the start of chemotherapy. This score is known as the Corona score. And these variables are the site of cancer. So patients with a very high risk cancer type get 2 points, those with a high risk type of cancer one point than patients with a high platelet count, low hemoglobin or the use of red cell growth factors, EPO, a high leukocyte count and high BMI get each one point and he could show that you. With the simple risk scoring model, you can categorize patients into three risk categories, into the high risk category with a score of three or higher in the intermediate risk category score of of one to two, and the low risk category with a score of 0. The Koran score has then been investigated in many core studies involving 35,000 patients. And interestingly, 90% of patients in this systematic riviant meta analysis that I'm showing here had a Koran score of 0. That means they had a low risk of venous thromboembolism, 64% a score of one or two. That means they were in the intermediate risk category and 17% had a score of three or higher, falling into the high risk category. So the problem here is that patients in the high risk category, so only 20 to 25% of the patients who developed venous thromboembolism were in the high risk category. So the majority of patients who developed venous thromboembolism within six months were in the intermediate or in the low risk category. So that means that the Corona score is not perfect. Even if you change the cut off from three to two points or higher to define high risk category, you'd only identify approximately 50% of the patients that will eventually develop venous thromboembolism. So therefore so many other initiatives have been started to improve risk prediction by developing new clinical prediction models. And one of these clinical prediction models was developed by our group in collaboration with our colleagues from the Netherlands and from other parts in the in the world and here by developing this new clinical prediction rule. So we have followed the formal approach of developing clinical prediction model. So basically you put all candidate risk factors into a model and the model at the end tells you which of the risk factors or candidate risk factors are the most relevant and independent. And here we have identified that the most important risk factor was the tumor site category and the second one which independently predicted risk of venous thromboembolism was D dimer. And these two parameters were then put together into a clinical prediction model. D dimer was here entered as a continuous variable and we have then shown that with this model it was possible to better classify patients according to the risk of venous thromboembolum. So the net reclassification improvement compared to the original Corona score was zero 31. That means that 31% of the patients could be classified more appropriately, more accurately in, in, in, in, in, in the risk prediction into developing venous thromboembolism. And here is a just an example. So we have also a risk calculator that can be accessed online and also a nomogram that is available in the publication. And this new and simplified risk prediction models including just three, two variables allows also individual Vt risk assessment as shown here for instance. So we have a patient with pancreatic cancer falls into the very high risk category gets 100 points. You measure D dimer. D dimer is around 3.5, so you get 40 points that up to 140 points in total corresponding to A six month risk of 15%, which is quite high. And this risk assessment was also integrated in randomized controlled tries of dough works versus placebo. And this randomized controlled tries and two of them have been conducted and published have used the Korana score and they have set the cutoff of defining high risk at a score of two or higher based on previous publications also from our group. And patients of first patients were screened if they had a score of two or higher then they were randomized to a Pixaban 2.5 BID twice daily versus placebo in the AVERT study for six months or in the Cassini study to Rivaroxaman 10 milligrams once daily versus placebo for six months. And in both studies, you can see that the intervention with the dog has reduced the occurrence of venous thromboembolism, the risk of venous thromboembolism. Interestingly, if we look in the ADVERT study and there are different type of statistical analysis, but let's focus here for instance, on the primary analysis here, as you can see that around 10% in the patient in the in the placebo group developed venous thromboembolism. This was also in the expected range and with apixaban the risk could be reduced to 4.2%. So a relative risk reduction in the primary endpoint analysis with of 60%. There's also an other type of analysis, the on treatment analysis. So these were patients who stayed from the beginning until the end of month 6 on on therapy within the study. So and here the risk of Vt in the placebo group was 7.3%, it could be reduced to only 1% with a relative risk reduction of 86% and similar trends in the Cassini study. And what we so can conclude. So it is possible to identify with this corona score of two or higher patients with a baseline risk of Vt of 10% and but is this enough to convince oncologists and hemato oncologists to offer primary thromboprophylaxis to their patients? This is the question because the guidelines already are recommending primary thromboprophylaxis. However, as I said, the corona score is probably imperfect and there is room for improvement and in a post hoc analysis of the AVERT study. So they have applied the CAT score with the nomogram that I have shown you to the population of the AVERT study. So they have calculated the score for each of the patients. And then they have divided patients into a predicted 6 months risk of below 8% and above 8% with the nomogram, the CAT score that I have presented just few slides ago. And here they could show that with this application of this CAT score. So the number needed to treat to prevent one Vt with apixaban could be reduced to six only in those with a predicted 6 month risk of above 8%. And I think this is really remarkable. So next I would like to talk about the dime of a prediction of venous tromboembolism recurrence in cancer patients. So the risk of Vt recurrence in cancer patients is high and different risk factors have been reported. So mainly presence of metastasis. So again this is very important. Specific type of cancers have a high risk of recurrence. We know this from also the incident risk of Vt. There are also some studies that have found that higher dedimal levels after stopping anticoagulator are predicting risk of Vt recurrence. So similarly to the general population to the non cancer population with venous thromboembols where D dimer is a established predicted for risk of recurrence. However. So this is just highlighting that persistent hypercoagulability is also associated with risk of Vt recurrence. However, I think the interaction of tumor cells and the hemostatic system. Actually lead to the activation of the hemostatic system for instance by release of procoagulant factors like tissue factors, cytokines by direct fibrolytic activity and this activated hemostatic system and this hypercoagulability also supports tumor cells to grow, supports tumors to invade for metastasis and also is supporting angiogenesis. And this hypercoagulability is also the basis for the increased risk of venous thrombosis, but also arterial risk. Risk of arterial thrombosis is relatively high in cancer patients and this is highlighted in a study that we have published more than 10 years ago, where we have shown that activation of the blood coagulation system he reflected by D dyma levels is associated with very poor overall survival in patients with cancer. So the higher the D dyma levels, the poorer the prognosis of patients was. So in patients who had a baseline before initiation of chemotherapy, the dymal levels in the highest quartile of the distribution in the overall cancer population of around 800 patients here had a one year survival of 54%, which decreased to 30% after two years as compared to those with lower D diamond levels in the first quartile of the distribution in the total population. So one year survival 88% and this went down to around 80% at two years. So D dimer is also a predictor of poor prognosis in patients with cancer even without the presence or of venous thromboembolism, even without development of venous thromboembolism. So we have summarized this in a recent review where we have I put together all the evidence which shows potential clinical application of hemostatic biomarkers in cancer patients for predicting overall survival, risk of cancer recurrence, risk of progression free survival or the association with the Disease Control rate. So we have shown that different hemostatic biomarkers and D dimer among them was the most promising and interesting one which was linked with prediction of poor prognosis, disease recurrence etcetera. So hemostatic biomarkers could have a potential for an individualized patient care. So they could contribute to risk stratified personalized oncological decision making in the future. So their ability to predict adverse outcomes in oncological patients is even better than most of the tumor specific biomarkers or the tumor markers. So let me summarize the first part of my presentation. So prediction based only on the tumor type will miss many cancer patients at risk of venous thromboembolism. Cancer types associated with a very high Vt risk are less prevalent, and the contribution to the overall prevalence of cancer associate thrombosis also is only minor. So there are multiple risk factors that contribute to the occurrence of venous thromboembolism and cancer patients there. And clinical prediction models with this regard seem to be promising because they can identify patients at risk of venous thromboembolism more accurately. We have seen some advances in risk assessment since the publication of the Khurana score and there were more newer and newer risk assessment models that are being published. And I think I have shown you today a novel and very easy clinical prediction model that was externally validated, which only includes 2 variables, the tumor site category and the dimer. That is very promising in predicting risk of venous thromboembolism. And I think this improvement of risk prediction might facilitate decision making on primary thromboprophylaxis. So and the risk of the role of the dimer as a risk factor for recurrence needs to be further confirmed. And we have good evidence that the dimer is a risk factor for an incident Vt event in cancer patients. Now in the next part of the presentation, I would like to talk a little bit about the diagnosis of venous thromboembolism in patients with cancer because this is a field of a potential application of T dimer. So we all know that in the general population in patients with suspected deep vein thrombosis and pulmonary embolism. So there are established diagnostic algorithms that start with clinical suspicion for for DVT or pulmonary embolism. And here you combine the pretest assessment of the clinical pretest probability with the dimer or without the dimer and directly going to imaging to rule out the vein thrombosis or pulmonary embolism. However, so and, and then maybe just to to mention, so the aim of this diagnostic algorithms is to reduce the number of unnecessary imaging, so unnecessary scans because the signs and symptoms of deep vein thrombosis and pulmonary embolism are very unspecific. But the question here is, are they accurate or useful in cancer patients? Such algorithms, such clinical prediction rules. So these algorithms, they include the clinic, the assessment of the pretest probability. The most prominent one is the wealth score for DVT and and PE. So they include variable symptoms suggestive of the brain thrombosis or pulmonary embolism. And interestingly here if you look, if you look at the clinical prediction rules, the well score for assessing the pretest probability of pulmonary embolism. We have also among this fact, this clinical features also cancer contributing to this clinical prediction tool with one point for assessing the pretest probability here and and this is as I said developed for the general population. And we have also other pretest probability tools of for assessing for instance, the clinical probability of pulmonary embolism like the Geneva score, some others. But the big question is what about their use in patients with cancer? And they may not be accurate or useful in cancer patients because of two reasons. So factors affecting the performance of clinical prediction rules in cancer patients. They include the high rate of non specific symptoms in cancer patients which may mimic symptoms of deep vein thrombosis and pulmonary embolism. So we have unspecific symptoms, they could be due to the cancer but they might mimic symptoms of the vein thrombosis, pulmonary embolism on the other hand. So cancer related risk factors like we have seen cancer type, metastatic disease etc that can affect the pretest probability of Vt are not included as variables in the clinical prediction rules like the Val score or the Geneva or the simplified Geneva score. And similarly the dimmer testing is an issue in cancer patients. So here is a summary of data that examined the accuracy and the accuracy of the dimer testing in cancer patients. And as you can see, if you look at the numbers very closely, so cancer compared to non cancer patients, the numbers of in cancer patients in each of the of these studies is is very, very low. So we have here only limited evidence, but in this evidence when we put this all together shows that the proportion of cancer patients with a negative D dimer result in these studies was very low. So it ranged from 8.5 to 30%. So cancer patients can have increased the dymal levels positive the dymal levels even in the absence of of thrombosis as I have shown you as a direct result of the tumor induced Coagulopathy and activation of coagulation. So what we have the following limitations here. So we have a lower specificity and higher false positive rates of D dimer tests in the population of cancer patients. If we combine clinical prediction rules and D dimer testing for Vt diagnosis in cancer patients, it is not going to improve or going to be better. Only 6 to 9% of cancer patients have a low pretest probability and the negative D dimer. That means they then do not require further testing. You can rule out deep vein thrombosis or pulmonary embolism in these patients, but this is a very low number of patients. So the combination of clinic prediction rules and D dimer testing, so is less safe and less efficient in cancer patients compared to non cancer patients. So that means in most of the cases you need imaging to rule out the vein thrombosis pulmonary embolus. So what can we conclude on the one hand, so we have seen the many advances in the field of cancer associate thrombosis over the last two decades. So with regard to understanding risk factors, primary thromboprophylaxis, it, etc. But there have been no advances to improve diagnostic workup of a suspected deep vein thrombosal pulmonary embolism in patients with cancer. The best strategy to diagnose the vein thrombosal pulmonary embolism in cancer patients has not been formally established. So we need further dedicated studies to to develop cancer specific clinical prediction rules that we can use in conjunction with fixed or age adjusted or tumor adjusted D dimer testing or D dimer levels. With that, I would like to thank you for your attention.

2) 3) 4) 1) 20% 22.7 11.16 10 45.22 22.72 22.94 25 20 17.6 15 15.2 8.5 6. 5.6 5.1 4.5 1.5 7.6 0.20 0.15 0.10 0.05 100 200 300 400 500 600 700 217 180 157 136 115 107 94 691 589 370 330 272 409 8% 7% 7.1% 2% 1% 5% 15- 80 60 40 5- 5.0% 23.4% 100- 21.4% 10- 8.9% 30 50 90 0.1 16 32 120 160 12 29 24 140 8.8 6.0 6.4 4.2 2.6 1.0 10.2 7.3 5(6.0) 3(4.2) 7(4.5) 47 2014(37) 406 2018(56) 900 695 2019(17) 331 2020(47) 4573 2011(36) 166 3947 2017(42) 114 88% 80% 82% 66% 54% 50% 30% 0.6 0.0 0.4 Hemostasis Learning Institute presents The importance of D-dimer in the diagnosis and treatment of cancer Prof. Dr. Cihan Ay Powered by hemostasis experts John V. Mitsios PhD, Medical Sciences Partner Siemens Healthineers The importance of D-dimer in the diagnosis, prevention and treament of venous thromboembolism in cancer Cihan Ay Clinical Division of Haematology and Haemostaseology Department of Medicine I, Comprehensive Cancer Center Vienna Medical University of Vienna, Vienna, Austria MEDICAL UNIVERSITY OF VIENNA @Cihan_Ay_MD cihan.ay@meduniwien.ac.at SIEMENS Healthineers Disclosures · Speaker fees and advisory boards · Bayer, BMS Pfizer, Daiichi Sankyo, Sanofi Questions to be addressed in my presentation 1) Why is this topic important and what are the challenges? Risk factors and risk assessment models for VTE in patients with cancer Evidence for the use of clinical prediction rules (CPRs) in cancer patients with suspected VTE (DVT of the lower limb or PE) Role and limitations of D-dimer testing to rule out VTE in the cancer patient population Thrombosis and Cancer 1 of 5 VTE events is related to cancer Triggers of VTE (RIETE Registry) Risk factors Gender Unknown Any Immobility Age Cancer Surgery Hormonal use Travel All patients with Pregnancy 0.7 VTE Puerperium 0.56 10 15 20 25 30 35 40 45 50 Source: https://trombo.info/vte-risk-factors/?lang=en VTE, venous thromboembolism Grilz E (Ay C) et al. Eur Heart J. 2021 Mar 26:ehab171 Puerperium Cancer-associated VTE is frequent! Cancer and Thrombosis 1 in 5 patients with cancer will develop VTE* *increasing rates of unsuspected/incidental VTE/PE Ay C, Pabinger I & Cohen AT. Thromb Haemost. 2017;117(2):219-230. Signs and symptoms are unspecific! VTE recurrence rate is ~2-5-fold higher in patients with VTE and cancer compared with those with VTE and no cancer Rates of VTE in Patients With Cancer Vienna Cancer and Thrombosis Study (CATS)* Pooled analysis (*excluded patients with VTE within 3 months prior to study inclusion) Overall VTE risk of VTE VTE-incidence (%) during a median follow-up of 501 days [IQR, 255-731] in 825 patients with different types of cancer 13 per 1,000 person-years Highest VTE risk among patients with cancers of the pancreas, brain, and lung 68 to 200 per 1,000 person-years breast Annual incidence of VTE in cancer stomach pancreas colorectal kidney lymphoma prostate brain (glioma) between 1 to 20% multiple myeloma other solid tumors total study population Ay C. J Clin Oncol. 2009 Sep 1;27(25):4124-9. Horsted F. PLoS Med. 2012;9(7):e1001275. What is the Goal? To prevent VTE and VTE-related consequences in patients with cancer How to prevent VTE in patients with cancer? Identify patients at risk of VTE Know risk factors for VTE in cancer Improve risk assessment of VTE Risk Factors and Predictors for VTE in Cancer Patients Patient-related Tumor-Related Site of cancer . Very High: stomach, pancreas, brain . Presence of varicose veins . Prior VTE High: lung, hematologic, gynaecologic, . Hereditary risk factors (eg, factor V renal, bladder Leiden) . Histological grade of a tumour Stage of cancer/metastases Cancer-Associated VTE Risk Treatment-related Biomarkers . Platinum-based and other chemotherapy Hematologic biomarkers (eg, platelet, . Anti-angiogenesis agents . Hormonal therapy D-dimer . Surgery P-selectin Radiotherapy . Prothrombin fragment 1 + 2 Blood transfusion Central venous catheters . MV-tissue factor activity . Hospitalization and immobility C-reactive protein, VEGF, MPV, etc. Neutrophil extracellular traps (NETS) Adapted from: Ay C, Pabinger I & Cohen AT. Thromb Haemost. 2017;117(2):219-230. Parameters that Globally Reflect Coagulation Activation and Fibrinolysis Predict Cancer-related VTE Peak Thrombin generation elevated D-dimer (275th percentile) >75th percentile non-elevated D-dimer (<75th percentile) Probability of VTE Cumulative probability of VTE (%) No. of patients at risk Time (days) Observation time (days) Ay et al, J Clin Oncol. 2011 Ay C et al. J Clin Oncol. 2009. Pabinger et al, Blood 2013 Multiple factors contribute to risk of VTE in patients with cancer! How can risk assessment in cancer be improved? Risk Prediction Model of VTE in Patients With Cancer . Prediction of cancer-associated VTE during chemotherapy with the „Khorana-Score“ (follow-up 2.4 months) Development cohort Risk Patient characteristic score Very high risk (stomach, pancreas) High risk (lung, lymphoma, gynecologic, bladder, testicular) Rato of VTE (%) Prechemotherapy platelet count 350 × 109/L or more Hemoglobin level less than 100 g/L or use of red cell growth factors BMI 35 kg/m2 or more Intermediate [1-2) Risk category score Khorana et al, Blood 2008 to of VTE (%) High (>3) Khorana Score for Prediction of VTE in Cancer Patients: a Systematic Review and Meta-analysis 55 cohorts enrolling almost 35 000 cancer patients 19% of patients had a Khorana score of 0 points, 64% had a score 1 or 2, and 17% had a score ≥3 points Estimated incidence of VTE and proportion in the high-risk group over 6 months Traditional threshold (3 points or more considered high risk): Low (0) Intermediate (1-2) High (23) allocated to high risk group (%) Proportion of all VTEs that were VTE incidence first 6 months (%) Mulder Fl et al. Haematologica. 2019;104:1277-1287. ortion of all VTEs th f all VTEs that were Proportion of all VTEs t 11 0% ortion of all VTEs that were n of all VTEs that were Proportion of all V tion of all VT ortion of all VTEs t Proportio that were Lower threshold (2 points or more considered high risk): Low (s1) High risk group (>2) of all VTEs th Proporti Proportion of all VTEs th on of all VTEs that were A NEW PREDICTION MODEL A clinical prediction model for cancer-associated venous thromboembolism: a development and validation study in two independent prospective cohorts Ingrid Pabinger, Nick van Es, Georg Heinze, Florian Posch, Julia Riedl, Eva-Maria Reitter, Marcello Di Nisio, Gabriela Cesarman-Maus, Noémie Kraaijpoel, Christoph Carl Zielinski, Harry Roger Büller, Cihan Ay Clinical Prediction Rule · Tumour site category "Low/intermediate" Multivariable SHR 1.96 Breast, prostate "high" Lung, colorectal, lymphoma, genitourinary excluding prostate, gynecologic excluding breast, esophageal, others "Very high" Stomach, pancreas as continuous variable Pabinger I et al. Lancet Haematol. 2018 *This model compared to Khorana score: Population-weighted net reclassification improvement (NRI)=0.31 A new and simplified risk prediction model for assessing individual VTE risk Pabinger et al. Lancet Hematol 2018; 5: e289-98 Tumour-site risk category Low or intermediate Breast Prostate OtherS High Points Lung 0 10 60 70 80 Colorectal Oesophagus Kidney Tumour-site risk Lymphomat Low or Bladder or urothelial intermediate Uterus Total points Cervical Ovarian Cumulative 6-month Very high incidence (%) Pancreas Case: ~15% Stomach 0 10 20 70 80 20 24 20 24 29 MED 6 7 0 24 0 24 29 Em Primary thromboprophylaxis With DOAC based on Risk Assessme *Patients with a Khorana Score of 2 or higher were eligible for the study (estimated baseline VTE risk of 9.6% during the first 6 months of chemotherapy according to results the Vienna CATS study, Ay C et al. Blood 2010) AVERT study (n=574) CASSINI study (n=841)# Placebo Apixaban 2.5 BID PlaceboRivaroxaban 10 mg OD High 10.2 60 70 Rate of VTE at 6 months Primary endpoint On treatment 6 Primary endpoint 20 HR 0.14 HR 0.66 HR 0.40 95% CI 0.26-0.65 95% CI 0.05-0.42 95% CI 0.40-1.09 95% CI 0.20-0.80 *Ultrasound screening at baseline was performed, and if thrombosis was found, patients were excluded Carrier M et al. N Engl J Med. 2019;380(8):711-719 .; Khorana A et al. N Engl J Med 2019;380:720-8. Assessment* (Randomized-controlled Trials) Placebo Apixaban 2.5 BID HR 0.41 Is baseline VTE risk of 10% (identified by Khorana Score >=2) high enough to convince oncologists and haematologists to offer primary thromboprophyalxis to their patients? Vienna catscore risk stratification* in AVERT study population ** (Avert study population, all Khorana score >2, mITT analysis) http://catscore.meduniwien.ac.at/ CATScore ≥ 8 % Outcome Apixaban Placebo HR n= 154 n = 157 (95 % CI) n = 83 n = 72 VTE, n (%) 6 (3.9) 7 (4.5) 0.89 (0.30-2.65) 7 (8.4) 19 (26.3) 0.33 (0.14-0.81) Major bleeding, n (%) 4 (2.6) 2 (1.3) 2.07 (0.38-11.3) 3 (4.2) 1.91 (0.44-8.19) *Pabinger et al, Lancet Haematol 2018, ** Kumar et al, Oncologist ., 2020 Apixaban 5 (6.0) Number needed to treat (NNT) in CATScore > 8 % n=6 D-dimer for prediction of VTE recurrence in patients Risk Factors for a recurrent VTE in cancer patients Study Study type Treatment Finding Risk estimate Chee CE et al. Cohort study Long term treatment: Metastasis HR 1.5 [95% CI: 1.03-2.17] 74% heparin followed HR 6.38 [95% CI: 2.69-15.13] by warfarin, 9% IVC Stage IV pancreatic cancer Brain cancer Myeloproliferative or myelodysplastic disorders Ovarian cancer Stage IV cancer (nonpancreatic) lung cancer neurological disease with leg paresis cancer stage progression Khorana AA et Posthoc Tinzaparin or warfarin 900 al. 2017(40) venous compression SHR 3.0 [95% CI: 1.8-4.9] analysis of for 6 months hepatobiliary cancer SHR 2.9 [95% CI: 1.2-7.0] RCT Young AM et al. 2018(41) Dalteparin or rivaroxaban for 6 stomach or pancreas versus other months malignancies HR 2.69 [95% CI: 1.11-6.53] lung, lymphoma, gynecologic, or bladder versus other symptomatic VTE versus incidental PE Englisch C, Moik F and Ay C. Thrombosis Update 2021 Long term treatment Cohort study Bauersachs R et al. Posthoc analysis of RCT Tinzaparin or warfarin for 6 renal impairment Sakamoto J et al. Retrospective cohort warfarin 83%, DOACs Mulder FI et al. Post hoc analysis of Edoxaban or LMWH for 6- ECOG 2 versus 0 for recurrent 12 months incidental VTE Louzada ML et al. Systematic review (6 Vitamin K antagonist or prospective studies) metastatic versus localized LMWH disease Oto J et al. 2020(50) Prospective study More than 3 months of male sex anticoagulation than stop ratio basal D-dimer to 21 days D- and blood draw, 21 days dimer >2 later blood draw again increasing hs-CRP (day 21) increasing p-selectin (day 21) Prospective registry Long-term treatment: 67% LMWH, 20% Vitamin K lung cancer versus breast cancer antagonists, 1.1% rivaroxaban, 1.6% fondaparinux At least 3 months of hs-CRP >4.5 mg/L LMWH, then D-dimer >600 ng/mL SHR 5.81 [95% CI: 1.06-31.72] anticoagulation stop and blood draw, 21 days later blood draw again Cells Procoagulant Activity Growth Cytokines Invasion Fibrinolytic Activity Angiogenesis Hemostatic System Hypercoagulability s breast can Arterial Venous thrombosis Tumor Growth Factors Activation of Blood Coagulation is Associated With Poor Overall Survival in Patients With Cancer + Censored D-dimer levels 1.quartile Survival probability 1 year 2 years Observation time Ay C et al, Haematologica 2012 0.6 - Potential clinical application of haemostatic biomarkers in patients with cancer Association between cancer and components of the Prediction of ... haemostatic system Overall survival Risk of cancer recurrence Progression free survival Disease control rate Increased levels of haemostatic biomarkers: Individualized Fibrinogen patient care Extracellular vesicle TF-activity Prothrombin fragment 1+2 sP-selectin Biomarkers of haemostasis might be used to contribute to risk-stratified, personalized oncologic decision making in the future. Moik F et al. Thromb Res. 2020;187:9-17. Moik F et al. Cancers. 2020;12(6):1619. Moik F et al. Arterioscler Thromb Vasc Biol. 2021;41(11):2837-2847. Moik F & Ay C. J Thromb Haemost. 2022; 20(12): 2733-2745.