PB0591 - Coagulation parameters predict COVID-19-related thrombosis in a neural network with a positive predictive value of 98%

Monday, July 11, 2022

6:30 PM – 7:30 PM

Location: Exhibition

Presentation Format: In-Person

Presenting Author(s)

Romy De Laat-Kremers, PhD

Post-doctoral Research Project Leader
Department of Data Analysis and Artificial Intelligence, Synapse Research Institute, Maastricht, the Netherlands
Maastricht, Limburg, Netherlands

Co-Author(s)

Rd

Raf de Jongh

.
Department of Anesthesiology, Ziekenhuis Oost Limburg, Genk, Belgium
Genk, Limburg, Belgium
Marisa Ninivaggi, PhD

Research Project Leader
Department of Functional Coagulation, Synapse Research Institute, Maastricht, the Netherlands
Maastricht, Limburg, Netherlands
AF

Aernoud Fiolet

.
Department of Internal Medicine, Meander Medical Center, Amersfoort, the Netherlands
Amersfoort, Utrecht, Netherlands
RF

Rob Fijnheer

.
Department of Internal Medicine, Meander Medical Center, Amersfoort, the Netherlands
Amersfoort, Utrecht, Netherlands
JR

Jasper Remijn

.
Department of Clinical Chemistry, Meander Medical Center, Amersfoort, the Netherlands
Amersfoort, Utrecht, Netherlands
Bd

Bas de Laat, MD, PhD

General Director
Department of Functional Coagulation, Synapse Research Institute, Maastricht, the Netherlands
Maastricht, Limburg, Netherlands

Background: Thrombosis is a major complication of a SARS-CoV-2 infection. COVID-19 patients show changes in coagulation factor levels and functional coagulation tests that indicate an important role for the coagulation system in the pathogenesis of COVID-19. However, the multifactorial nature of thrombosis complicates the prediction of thrombotic events based on a single hemostatic variable.

Aims: We used a neural network to predict future COVID-19-related thrombosis.

Methods: We developed neural networks for the prediction of thrombosis in COVID-19 patients based on several dedicated coagulation parameters and general laboratory variables measured in plasma samples of 133 COVID-19 patients collected at the time of hospital admission (cohort 1). The neural network was validated in a second cohort of 16 COVID-19 patients admitted to the intensive care unit (cohort 2). In cohort 1 and 2, 19 and 7 patients respectively suffered from thrombosis during their hospital stay.

Results: The neural network predicts COVID-19-related thrombosis based on C-reactive protein (relative importance 14%), sex (10%), thrombin generation (TG) time-to-tail (10%), α2-macroglobulin (9%), TG curve width (9%), thrombin-α2-macroglobulin complexes (9%), plasmin generation lag time (8%), anti-SARS-CoV-2 serum IgM (8%), TG lag time (7%), TG time-to-peak (7%), thrombin-antithrombin complexes (5%), and age (5%). In developmental cohort 1, the neural network identified future thrombosis in COVID-19 patients with a positive predictive value of respectively 98%. The neural network accurately ruled out thrombosis in COVID-19 patients as the negative predictive value of the neural network was 86%. In validation cohort 2, the positive predictive value of the neural network was 100%, and a negative predictive value of 66%.

Conclusion(s): We developed a neural network that can accurately predict the occurrence of COVID-19-related thrombosis and is a promising algorithm to apply to other COVID-19 patient cohorts. The prediction of COVID-19 related thrombosis potentially can give clinicians the opportunity to increase anticoagulant therapy in high risk patients.