MELBOURNE – A venous thromboembolism risk score that combines clinical risk factors, such as lymphoma type and stage, along with genetic variables, could offer a better way to predict venous thromboembolism in patients with lymphoma, according to new findings presented at the International Society on Thrombosis and Haemostasis congress.
Cristina Pascual, MD, of the Hospital Universitario Gregorio Marañon in Madrid presented data from a development and validation study of a clinical-genetic risk model for thrombosis in lymphoma in 208 patients with lymphoma, 31 of whom experienced a venous thromboembolic event.
While the relationship between cancer and increased thrombosis risk is well recognized, lymphoma patients are at particularly high risk, with an estimated thrombosis incidence of 5%-10%, Dr. Pascual said.
Currently, the Khorana score is the most validated risk score for thrombosis in patients with solid tumors, using factors such as tumor site, platelet and leukocyte count, hemoglobin levels, and body mass index. However, Dr. Pascual pointed out that just 10% of the validation cohort for the Khorana score were lymphoma patients, and it had previously been found to be not as useful for that population.
More recently, researchers had developed the ThroLy score for predicting thromboembolic events specifically in patients with lymphoma, incorporating clinical variables such as mediastinal involvement and extranodal localization.
Another group took a different approach by incorporating genetic risk factors for thrombosis to create Thrombo inCode-Oncology (TiC-Onco) for solid tumors. This assessment included four genetic variants known to increase the risk of thromboembolic events in cancer patients, as well as the clinical risk factors of body mass index, family history of thrombosis, primary tumor site, and tumor stage.
Dr. Pascual and colleagues developed a unique risk factor model that combined both the ThroLy and TiC-Onco elements.
In 208 patients with lymphoma who were not receiving anticoagulant treatment, researchers identified five clinical factors that were most predictive of venous thrombosis: a history of thrombosis, immobilization for more than 3 days, lymphoma type, Ann Arbor score for lymphoma stage, and mediastinal extension.
They combined these clinical risk factors with the genetic risk factors from the TiC-Onco score to develop the TiC-Onco–associated lymphoma score (TiC-Lympho).
When validated in the same group of patients, the TiC-Lympho score had a sensitivity of 93.55%, a specificity of 54.49%, positive predictive value of 26.36%, and negative predictive value of 97.94%.
The researchers also compared TiC-Lympho’s performance with that of the ThroLy and TiC-Onco models, and found it performed better on sensitivity and negative predictive value. The area under the curve for TiC-Lympho (0.783) was significantly higher than that seen with the other two risk models.
Session chair, of the Peter MacCallum Cancer Centre in Melbourne, raised the question of how the score – and particularly the genetic risk factor assessment – might be applied in the real-world clinical setting.
In an interview, Dr. Pascual said the findings represented preliminary data only, so the model was not ready to be applied to clinical practice yet. She also stressed that this was based on retrospective data, and needed to be further validated in other cohorts of lymphoma patients.
No conflicts of interest were reported.
SOURCE: Pascual C et al. 2019 ISTH Congress, .