Thrombosis in Diffuse Large B-cell Lymphoma:the PREDIC-TO (PREDICt ThrOmbosis) Study

DLBCL is the most frequent lymphoid malignancy in the adult population and the one with the highest risk of thrombotic events. The indications for thromboprophylaxis are, however, debated.

For such reason, risk scores (i.e., the Khorana, Throly and the recently published Model IX score), have been developed to identify patients bearing a high thrombotic risk at diagnosis. These models, however, were developed considering not only aggressive lymphomas, but also indolent lymphomas, characterized by a markedly lower risk of VTE when compared to DLBCL.

The performance of the Khorana, Throly and Model IX score has never been evaluated and compared in the sole DLBCL population. Such understanding in a cohort of "real world" patients could aid clinicians in deciding whether to initiate thromboprophylaxis and may direct randomised controlled trials on the use of prophylactic heparin in the future.

Moreover, the aforementioned scores assess independent variables at diagnosis, not taking into account the possible occurrence of risk factors that may arise over the course of DLBCL treatment or in case of placement of a central venous catheter (CVC).

Assessing the impact of thrombotic risk factors at diagnosis and during the disease, in a cohort of sole DLBCL patients, could allow the development of a new predictive model of thrombotic risk adapted to DLBCL.

1. OBJECTIVES AND ENDPOINTS OBJECTIVES Primary objective

• To assess the impact of thrombotic risk factors on the development of VTE in patients with Diffuse Large B-cell Lymphoma (DLBCL): real-life comparison of models predictive of thrombotic risk (Khorana, Throly and Model IX scores).

Secondary objectives

• To evaluate the statistical performance of the aforementioned predictive models in the subgroup of patients who did not have a thrombotic event at diagnosis
• To assess the impact of dynamic risk factors (i.e., not present at diagnosis) on thrombotic risk
• To evaluate the impact of risk factors at diagnosis which are not included in the published predictive models
• To evaluate overall survival and VTE-free survival
• Creation of a predictive model of thrombotic risk adapted to DLBCL

ENDPOINTS Primary endpoints

• Measurement of thrombotic risk factors present in the Khorana, Throly and Model IX scores: previous arterial or venous thrombotic event, ECOG, weight, extra nodal and/or mediastinal localization of disease, leukocytes, neutrophils, haemoglobin, platelet count, albumin, bulky mass
• Occurrence of arterial and venous thrombotic events, type of event, time to event after diagnosis and during disease

Secondary endpoints

• Data collection of static thrombotic risk factors at diagnosis and not included in the already published scores:
• de novo or secondary DLBCL status
• Neutrophil/lymphocyte ratio (ANC/ALC ratio)
• Cardio aspirin
• R-IPI
• Immunohistochemistry: Ki67, Myc, CD10, BCL-6, BCL-2
• Correlation between static risk factors at diagnosis and relative risk of thrombotic events
• Data collection of dynamic risk factors (as both categorical-dichotomous and temporal variable) present from diagnosis to expected event, death or lost at follow-up:
• Central venous catheter placement
• Therapy: radiotherapy, chemotherapy, anthracycline therapy, number of therapeutic lines, autologous haematopoietic stem cell transplantation
• LDH, haemoglobin < 10 g/dl if > than 10 at diagnosis, venous compression at CT scan, extra nodal localisation if not present at diagnosis, albumin < 3.5 g/dl if greater at diagnosis
• ANC/ALC ratio at the time of thrombotic event
• Correlation between acquisition of dynamic risk factors and relative risk of thrombotic event
• Date of thrombotic event, death or lost to follow-up to determine overall survival and VTE free survival
• Creation and validation of a predictive model adapted to DLBCL based on the Hazard Ratios of the independent variables