Biomarkers predictive of venous thromboembolism in patients with newly diagnosed high-grade gliomas

Abstract

Background: High-grade gliomas (HGGs) are among the most prothrombotic of malignancies.

Methods: We performed a prospective study to investigate 11 potential biomarkers for prediction of venous thromboembolism (VTE) in newly diagnosed HGG patients who had undergone a neurosurgical intervention. In addition, we tested 2 VTE risk assessment models (RAMs). The strongest predictors of VTE, which were identified by statistical forward selection, were used for the first RAM. The parameters used for the second RAM were both predictive of VTE and available in routine clinical practice.

Results: One hundred forty-one HGG patients were included in this study, and 24 (17%) of them developed VTE during follow-up. An association with the risk of future VTE was found for the following parameters: leukocyte count, platelet count, sP-selectin, prothrombin-fragment 1 + 2, FVIII activity, and D-dimer. The first RAM included low platelet count (<25th percentile of the study population) and elevated sP-selectin (≥75th percentile). The cumulative VTE probability after 12 months was 9.7% for score 0 (n = 76), 18.9% for score 1 (n = 59), and 83.3% for score 2 (n = 6). The second RAM included low platelet count (<25th percentile), elevated leukocyte count, and elevated D-dimer (≥75th percentile). The probability of VTE was 3.3% for score 0 (n = 63), 23.0% for score 1 (n = 53), and 37.7% for score 2 (n = 22) or score 3 (n = 3).

Conclusions: We identified biomarkers suitable for assessing the VTE risk in newly diagnosed HGG patients. The application of 2 RAMs allowed identification of patients at high risk of developing VTE. We could also define patients at low risk of VTE, who would most probably not benefit from extended primary thromboprophylaxis.

Keywords: biomarkers; high-grade gliomas; prediction; risk assessment models; venous thromboembolism.

Fig. 1.

Cumulative incidence of venous thromboembolism, accounting for competing risk (death from any cause other than fatal VTE) according to our experimental risk assessment model (including platelet count and soluble P-selectin) in HGG patients with scores 0, 1, 2.

Fig. 2.

Cumulative incidence of venous thromboembolism, accounting for competing risk according to our risk assessment model for routine clinical practice (including platelet count, leukocyte count, and D-dimer) in HGG patients with scores 0, 1, 2/3.