The association between in vitro human thrombi composition, mechanics, and computed tomography imaging characteristics

Abstract

Background- A wide variation in composition is found within the thrombi retrieved from acute ischemic stroke patients. The composition of a thrombus and its mechanical properties have a significant influence on the procedural ease and success rate of mechanical retrieval procedures. If the composition and mechanical properties of a thrombus could be determined using admission computed tomography (CT) imaging, the treatment could be adjusted to ensure a good outcome in each case. This study aims to investigate and associate three thrombus characteristics related to this issue: composition, mechanics, and CT imaging characteristics.
Methods- Blood was drawn from three healthy volunteers. Analog thrombi of different compositions were prepared by combining red blood cells (RBCs) and platelet-rich plasma (PRP) in five different volumetric ratios: 0% RBCs and 100% PRP, 20% RBCs and 80% PRP, 40% RBCs and 60% PRP, 60% RBCs and 40% PRP, 80% RBCs and 20% PRP. To measure thrombus contraction, the weight of the resulting thrombus was expressed as a percentage of the weight of the original blood mixture. First, CT imaging was performed using micro CT and clinical CT imaging to obtain density and perviousness measurements of the thrombi. Secondly, the thrombi were subjected to a cyclic unconfined compression test to analyse stress-strain behaviour and to determine low and high strain secant moduli as a measure for thrombus stiffness. Lastly, thrombus composition was analysed using histological techniques to quantify RBCs, fibrin and platelet conglomerations, fibrin, and white blood cells.
Results- First, a positive linear relationship was found between the relative thrombus weight after
contraction and the volumetric RBC ratio (r=0.98, p<0.001). Secondly, with composition analysis, it was found that the thrombi with the 0% 20%, 40%, 60%, and 80% volumetric RBC ratios produced thrombi with non-equivalent compositions of 0.0%, 89.8%, 95.1%, 97.4%, and 99.4% RBCs. This showed that only 0% RBC and RBC-rich (>85% RBC) thrombi were created. Thirdly, on clinical CT imaging, the 0% RBC samples were found to be significantly lower in mean density compared to the RBC-rich thrombi (36 HU and 88 HU respectively, p<0.001). Fourthly, the perviousness was significantly higher for the 0% RBC thrombi compared to the RBC-rich thrombi (An increase of >81 HU and <14 HU respectively, p<0.005). Fifthly, for the mechanics, the secant modulus at high strain decreased with an increasing amount of RBCs in the original volumetric ratio. For donor one the high strain stiffness values were 263, 84, 72, 27, and 13 kPa for an increasing RBC content. Lastly, thrombus stiffness differed per donor. For example, the high strain stiffnesses for 40% RBC thrombi were 72, 432,
and 33 kPa between donors. Thrombus density and perviousness on clinical CT imaging were significantly different for the 0% RBC and RBC-rich group (both p<0.001). High strain stiffness was significantly correlated to the contracted thrombus weight (r=-0.68 p=0.006).
Conclusions- Both density and perviousness on clinical CT imaging were related to the RBC content of the analog thrombi. For the thrombus mechanics, the high strain stiffness of the thrombi was highly correlated to the relative thrombus weight after contraction and moderately correlated to the RBC content. This implicates that clinical imaging should strive to measure thrombus contraction as well as composition to estimate thrombus mechanics prior to mechanical retrieval.