Risk assessment models for prediction of venous thromboembolism in ambulatory cancer patients – A literature review

Abstract

Background: Venous thromboembolism (VTE) represents a significant burden of health for those affected, and complications after a VTE are especially pronounced in cancer patients. One fifth of all cancer patients will develop a VTE during their disease. Routine primary thrombophylaxis is not recommend in ambulatory cancer patients due to lack of tools to identify those at especially high risk, as anticoagulants comes with a substantial risk of major bleeding. Aim: The aim was to perform a systematic literature search to gather existing information, summarize and evaluate the evidence on prediction models for VTE in ambulatory cancer patients. Methods: Published scientific articles were obtained by systematically searching the Medline and Embase databases, and in addition, a free-text search in PubMed was applied to ensure optimal coverage of the search. Results: A total of 22 articles on risk assessment models (RAMs) in ambulatory cancer patients was identified, of these 9 papers presented novel models and the rest validated one or more RAM. The identified scores were the Khorana risk score (2008), Vienna CATS score (2010), Protect score (2012), CONKO score (2013), ThroLy score (2016), ONKOTEV score (2017), COMPASS-CAT score (2017), Tic-onco index (2018) and Vienna nomogram (2018). The Khorana risk score has been validated the vast majority of the identified studies, with main critique is that >50% of patients fall into an intermediate risk category. Further, The Vienna CATS, Protecht, CONKO, COMPASS-CAT and Vienna nomogram has been externally validated and showed somewhat varying results. Conclusion: Several RAMs on cancer-related thromboembolism in ambulatory cancer patients are developed. Although promising performance is found for many of the scores, most validation studies are of suboptimal quality, and the clinical feasibility of the models are yet to be established. Additionally, continuously effort is needed to address the bleeding risk in this patient group.