To investigate the diagnostic value of real-time contrast-enhanced ultrasound (CEUS) quantitative parameters for assessment of liver graft‐versus‐host disease (aGVHD).

A total of 42 Wistar rats were included in the present study. Six rats were randomly selected as controls, and the remaining 36 rats were used as recipients. Ten Fischer 344 rats were included as donors. Bone marrow transplantation was performed to establish an aGVHD model. Six rats were randomly selected every week after surgery. Clinical scoring and ultrasonic examination were performed to obtain time-intensity curve (TIC) parameters [peak intensity (PI), time to peak (PPT), and mean transit time (MTT)]. Specimens were collected at the end of the examination and subjected to pathological diagnosis and scoring. The differences of clinical scores and TIC parameters among the control group, the group without aGVHD (nGVHD), and the aGVHD group were compared by the Kruskal-Wallis H test. According to the severity of pathology, rats in the aGVHD group were divided into mild (pathology level 2), moderate (pathology level 3), and severe (pathology level 4) groups. TIC parameters among different stage of aGVHD were analyzed. The correlation of clinical scores and TIC parameters with pathological scores was calculated by the Spearman test. Each TIC parameter and clinical scores were included in the binary Logistic regression analysis to obtain the joint prediction parameters, which were compared with the clinical indicators with regard to the diagnosis of liver aGVHD. Then, the receiver operating characteristic (ROC) curve was plotted. Clinical scores and contrast-enhanced ultrasound parameters in the diagnosis of liver aGVHD were compared by the Delong test.

Compared with the other two groups, PI was decreased, and TTP and MTT were increased in the aGVHD group (n=24; P<0.05). MTT had good ability to distinguish mild and severe aGVHD [(135.21±27.71) s vs (84.37±25.26) s]. There was a high correlation between PT and MTT and pathological scores (r=-0.790 and 0.733, respectively), and a moderate correlation between TTP and pathological scores (r=0.489). The area under the ROC curve values of clinical scores, TIC parameters, and their combination were 0.698, 0.885, and 0.910, respectively; the corresponding sensitivity and specificity were 79.2%, 87.5%, and 91.7%, and 75.0%, 83.3%, and 83.3%, respectively. The diagnostic efficacy of TIC parameters alone and TIC parameters combined with clinical scores was significantly better than that of the clinical scores (Z=2.290, P=0.022; Z=2.412, P=0.016).

The TIC changes of the liver parenchyma provide an imaging basis for the diagnosis of liver aGVHD and improve the diagnostic efficiency.