Fig. 1: Study design. | Communications Medicine

Fig. 1: Study design.

From: Machine learning-assisted assessment of extracellular vesicles can monitor cellular rejection after heart transplant

Fig. 1

A total of 24 patients were enrolled and longitudinally evaluated for a median follow-up of 303 days, with a total number of visits ranging between 9 and 17 for each subject (total number of visits was 285). At each visit, patient underwent venous sampling and endomyocardial biopsy (EMB); a plasma sample was stored and used for EV profiling by flow cytometry (FC) according to two different protocols (see also methods): standardized immuno-capturing bead-based kit vs. customized EV profiling after capturing by membrane-sensing peptides (MSPs). From a predefined panel of 37 markers commonly expressed on EV membrane, we selected 14 EV antigens differentially expressed in rejecting, to be measured with our customized method (MSP-capturing). Levels of expression of EV antigens were combined in a biomolecular fingerprint by supervised learning, to build an AI random forest regressor (rRF) model and predict rejection episodes. Part of the figure was produced using the Servier Medical Art public domain (https://smart.servier.com).

Back to article page