Utility of the Khorana and the new-Vienna CATS prediction scores in cancer patients of the HYPERCAN cohort.

Background: Risk assessment models (RAMs) are relevant approaches to identify cancer outpatients at high risk of venous thromboembolism (VTE). Among the proposed RAMs, the Khorana (KRS) and the new-Vienna CATS risk scores have been externally validated in ambulatory patients with cancer.

Objectives: To test KRS and new-Vienna CATS scores in 6-month VTE prediction and mortality in a large prospective cohort of metastatic cancer outpatients during chemotherapy.

Patients/methods: Newly diagnosed patients with metastatic non-small cell lung, colorectal, gastric, or breast cancers were analyzed (n = 1286). The cumulative incidence of objectively confirmed VTE was estimated with death as a competing risk and multivariate Fine and Gray regression.

Results: Within 6 months, 120 VTE events (9.7%) occurred. The KRS and the new-Vienna CATS scores showed comparable c-stat. Stratification by KRS provided VTE cumulative incidences of 6.2%, 11.4%, and 11.5% in the low-, intermediate-, and high-risk categories, respectively (p = ns), and of 8.5% vs. 11.8% (p = ns) in the low- vs. high-risk group by the single 2-point cut-off value stratification. Using a pre-defined 60-point cut-off by the new-Vienna CATS score, 6.6% and 12.2% cumulative incidences were obtained in the low- and high-risk groups, respectively (p < 0.001). Furthermore, having a KRS ≥2 = or a new-Vienna CATS score >60 points was also an independent risk factor for mortality.

Conclusion: In our cohort, the 2 RAMs showed a comparable discriminating potential; however, after the application of cut-off values, the new-Vienna CATS score provided statistically significant stratification for VTE. Both RAMs proved to be effective in identifying patients at increased risk of mortality.