OC 03.4 - The Vienna CATScore for predicting VTE risk in patients with cancer: an external validation across multiple time-points

Saturday, June 22, 2024

13:45 – 14:00 ICT

Room: Ballroom B3

Presenting Author(s)

Cornelia Englisch, MD

PhD Student
Division of Hematology and Hemostaseology, Department of Medicine I, Medical University of Vienna
Vienna, Wien, Austria

Co-Author(s)

Stephan Nopp

Physician
Division of Hematology and Hemostaseology, Department of Medicine I, Medical University of Vienna
Vienna, Wien, Austria
Florian Moik, MD PhD

Medical University of Graz, Austria
AS

Angelika M. Starzer

Physician
Division of Oncology, Department of Medicine I, Medical University of Vienna
Vienna, Wien, Austria
PQ

Peter Quehenberger

Professor
Department of Laboratory Medicine, Medical University of Vienna
Vienna, Wien, Austria
MP

Matthias Preusser

Professor
Division of Oncology, Department of Medicine I, Medical University of Vienna
Vienna, Wien, Austria
AB

Anna Berghoff

Professor
Division of Oncology, Department of Medicine I, Medical University of Vienna
Vienna, Wien, Austria
Cihan Ay, MD

Professor
Division of Hematology and Hemostaseology, Department of Medicine I, Medical University of Vienna
Vienna, Wien, Austria
Ingrid Pabinger

Professor
Division of Hematology and Hemostaseology, Department of Medicine I, Medical University of Vienna
Vienna, Wien, Austria

Background: Venous thromboembolism (VTE) is a well-recognized cancer complication. Risk assessment models were designed to aid in identifying high-risk patients. Novel anticancer treatments could impact the risk of VTE and the validity of established prediction models. Furthermore, an ideal model should also be applicable after start of anti-cancer treatment.

Aims: We aimed to externally validate the Vienna CATScore (Pabinger et al. Lancet Haematology 2018), which is a nomogram-based model including tumor type and D-dimer, in a prospective cohort of patients initiating systemic anti-cancer therapies, including immune-checkpoint inhibitors [ICIs].

Methods: Three time-points were assessed: at study inclusion, three weeks, and three months after start of therapy, measuring D-dimer with the Tina-quant® D-Dimer Gen2 Assay (Roche Diagnostics, Rotkreuz, Switzerland). C-statistics and calibration plots were assessed.

Results: This study encompassed 598 patients (49%women, median age: 62 years [interquartile range: 53-70]). Among solid tumor patients, 381 (63.7%) had metastatic disease at study inclusion (most common treatments: chemotherapy [45.2%], ICIs [16.1%], and both combined [17.2%]). Over a six-month observation period, VTE was diagnosed in 54 patients (6-month cumulative incidence: 9.2%, 95% confidence interval [CI]: 6.8-11.5). The Vienna CATScore evaluated at up to three time-points per patient, demonstrated good discriminatory ability (c-statistic 0.68, 95%CI: 0.63-0.73, figure 1A) with comparable c-statistics at all three time-points. The model showed excellent calibration for patients with predicted VTE probabilities below 10%, yet tended to underestimate risk in patients with very high predicted probabilities (figure 1B). Applying a 6-month predicted VTE risk threshold of 8% as recommended in current guidelines, the CATScore effectively distinguished between a low- and high-risk group at all three time-points (6.3% versus 13.6% observed risk, p< 0.001, figure 2).

Conclusion(s): This external validation of the Vienna CATScore confirms its effectiveness in predicting VTE risk in the initial three months of anti-cancer therapy and underscores its utility in selecting patients for primary thromboprophylaxis.