Differentiation of acute coronary syndrome with radiomics of pericoronary adipose tissue.

Objective: To assess the potential values of radiomics signatures of pericoronary adipose tissue (PCAT) in identifying patients with acute coronary syndrome (ACS).

Methods: In total, 149, 227, and 244 patients were clinically diagnosed with ACS, chronic coronary syndrome (CCS), and without coronary artery disease (CAD), respectively, and were retrospectively analysed and randomly divided into training and testing cohorts at a 2:1 ratio. From the PCATs of the proximal left anterior descending branch, left circumflex branch, and right coronary artery (RCA), the pericoronary fat attenuation index (FAI) value and radiomics signatures were calculated, among which features closely related to ACS were screened out. The ACS differentiation models AC1, AC2, AC3, AN1, AN2, and AN3 were constructed based on the FAI value of RCA and the final screened out first-order and texture features, respectively.

Results: The FAI values were all higher in patients with ACS than in those with CCS and no CAD (all P < .05). For the identification of ACS and CCS, the area-under-the-curve (AUC) values of AC1, AC2, and AC3 were 0.92, 0.94, and 0.91 and 0.91, 0.86, and 0.88 in the training and testing cohorts, respectively. For the identification of ACS and no CAD, the AUC values of AN1, AN2, and AN3 were 0.95, 0.94, and 0.94 and 0.93, 0.87, and 0.89 in the training and testing cohorts, respectively.

Conclusions: Identification models constructed based on the radiomics signatures of PCAT are expected to be an effective tool for identifying patients with ACS.

Advances in knowledge: The radiomics signatures of PCAT and FAI values are expected to differentiate between patients with ACS, CCS and those without CAD on imaging.