Venous thromboembolism (VTE) is a major public health problem in cancer patients and there is a clear need to precisely define the risk of VTE and to administer thromboprophylaxis specifically to the group of high-risk patients who have a favourable risk-benefit ratio. Current risk assessment models (RAM) for identifying high-risk cancer patients, who could benefit from primary thromboprophylaxis, are based solely on clinical and laboratory parameters. Khorana score (KS) is the best validated RAM and recommended by the main clinical guidelines. All current RAM, including KS, have low positive predictive value and limited clinical applicability.

Our group was the first to determine that VTE, a multifactorial and complex disease, has a large genetic component with 60% heritability. We showed that the combination of genetic and clinical factors of thrombotic risk improves the predictive capacity of RAM of VTE in the general population. In this way, we have been pioneers in the development of a predictive tool to determine thromboembolic risk in general population, which is already being used in clinical practice. From these results and aware of the importance of the genetic base in the risk of VTE in general population, we wonder what role this genetic base already known could have in the increased risk of VTE in cancer patients. In cancer patients none of the current RAM to estimate the risk of VTE has taken into account the genetic factors of thrombotic risk, although some prothrombotic mutations such as Factor V Leiden and the G20210A mutation in the prothrombin gene have been associated with an increased risk of VTE in cancer patients (Vienna CATS study and the Tromsø cohort). The first data of our project indicate that a risk score including clinical and genetic data of the cancer patients at 6 months of clinical follow-up significantly improves KS in the identification of cancer patients undergoing out-of-hospital chemotherapy at high risk of suffering thromboembolic events. At this point in the study, the KS applied to our patients had a very low predictive capacity, in terms of discrimination, (AUC) of 0.58 (95% CI 0.51–0.65). However, the score developed from our study significantly increased the predictive capacity of thrombosis risk with the incorporation of genetic and clinical risk factors for thrombosis (AUC 0.73, 95% CI 0.67–0.79, P<0.0001). In addition to these results, it is important to mention that 56 patients who were classified as low and intermediate risk by the KS had an episode of VTE. Our score would have identified 42 of them (75%) as high risk of having a thromboembolic event, so they could have benefited from thromboprophylaxis.

With these results, we can conclude that our score has a better predictive capacity than the KS to identify cancer patients treated with chemotherapy at high risk of thrombosis, potentially, reducing the morbidity and mortality associated with this pathology. Therefore, our results have a direct applicability that may change the clinical practice.