Testing a New Diagnostic Approach for Suspected Deep Vein Thrombosis Using Point-of-Care D-dimer, AI Prediction, and Ultrasound

What this Trial Is About

Researchers are evaluating a new diagnostic approach for patients suspected of having deep vein thrombosis DVT in the lower limbs. The study aims to compare a machine learning-based algorithm combined with point-of-care POC D-dimer testing to the standard use of laboratory D-dimer tests and compression ultrasound. The main goal is to determine if this new strategy can exclude DVT in more patients than the current clinical assessment methods alone. All participants will receive the usual diagnostic care including physician examination, laboratory D-dimer testing, and ultrasound by a radiologist if the D-dimer test is positive. In addition, POC D-dimer testing and POC ultrasound performed by emergency department physicians will be conducted, along with blood sampling for biobanking and photographs of the lower limbs. The machine learning model will be tested against these results to assess its accuracy and whether it could potentially reduce the need for ultrasound. Participants will be monitored from enrollment through a 90-day primary assessment period. Researchers will evaluate the safety of the new diagnostic approach, the agreement between POC and laboratory tests, and compare the timing and efficiency of the new strategy versus standard care. Ultrasound results performed by emergency physicians will also be compared to those by radiologists. The study involves no changes to patient treatment based on the additional tests, and participants will be followed to assess safety and outcomes over the study period.

Systematic Machine Learning Algorithm for Rapid Thrombosis Detection