Pabinger I, et al. - This study was conducted to create and externally validate a clinical prediction model for cancer-associated venous thromboembolism. Researchers proposed a clinical prediction model incorporating only one clinical factor (tumor-site category) and one biomarker (D-dimer). They succeeded at externally validating the model and noted that it successfully predicted the risk of venous thromboembolism in ambulatory patients with solid cancers. Compared with previous models used to predict cancer-associated venous thromboembolism, this simple model represents a considerably improved version and could help physicians in selection of patients who will likely benefit from thromboprophylaxis.

Methods

• Using data from the prospective Vienna Cancer and Thrombosis Study (CATS) cohort (n=1,423), researchers selected prognostic variables for inclusion in the model.
• The prospective Multinational Cohort Study to Identify Cancer Patients at High Risk of Venous Thromboembolism (MICA) cohort (n=832) was used to validate the model.
• They calculated c-indices to show how the predicted prevalence of objectively confirmed venous thromboembolism at 6 months compared with the cumulative 6-month incidences observed in both cohorts.

Results

• They selected two variables for inclusion in the final clinical prediction model: tumor-site risk category (low or intermediate vs high vs very high) and continuous D-dimer concentrations.
• They also found that the multivariable subdistribution hazard ratios were 1.96 (95% CI 1.41–2.72; p=0.0001) for high or very high vs low or intermediate and 1.32 (95% CI 1.12–1.56; p=0.001) per doubling of D-dimer concentration.
• In CATS and in MICA, the observed cross-validated c-indices of the final model were 0.66 (95% CI 0.63–0.67) and 0.68 (0.62–0.74), respectively.
• In both cohorts, adequate calibration of the clinical prediction model was reported.