Diagnostic performance of radiomics for predicting arterial plaque vulnerability: a systematic review and meta-analysis

Radiomics, in which medical images are analysed by extracting a large number of quantitative features, and is being increasingly applied in clinical research regarding medical imaging analysis. The aim of this study was to explore models based on radiomics for identifying the vulnerability of arterial plaques. Databases from PubMed, Web of Science, and EMBASE were used for this literature search. Evaluation of the diagnostic precision of radiomics in determining the susceptibility of various arterial plaque forms is one of the inclusion criteria. The Quality Assessment of Diagnostic Accuracy Study (QUADAS-2) and Radiomics Quality Score (RQS) were used to evaluate the risk of bias. The included radiomic studies were analysed by meta-analysis, using subgroup analysis to investigate heterogeneity. The precision of the radiomic models was assessed using the summary receiver operating characteristic (SROC) curve. Nine investigations were included, including four coronary artery studies, three carotid artery studies, and two intracranial artery studies of diagnostic accuracy. The radiological identification of vulnerable arterial plaques had an overall sensitivity of 81% (95% confidence intervals: 72–87%) and a specificity of 78% (95% confidence intervals: 69–82%). The area under the SROC curve was 0.84 (95% confidence intervals 0.81–0.87). The maximum and mean RQS percentages were 77.8% and 32.7%, respectively. The interrater agreement was good (ICC 0.98, 95% confidence intervals 0.95–0.95). Radiomics has demonstrated high diagnostic performance in identifying vulnerable arterial plaques. With more in-depth research, radiomics may alter the clinical diagnostic procedure for arterial plaques.