A prediction model of silent brain infarction constructed based on ultrasonic features of carotid plaques and resistance of intracranial blood flow.

OBJECTIVE

We aimed to construct a prediction model of silent brain infarction (SBI) based on ultrasonic features of carotid plaques and resistance of intracranial blood flow.

METHODS

A retrospective cohort study was designed to collect patients' clinical data from the database of our hospital. An SBI group was set up to include 168 SBI patients confirmed by cranial magnetic resonance imaging (MRI) between January 2022 and December 2024, and a non-SBI group was created to enroll 168 patients with carotid plaques. The influencing factors for SBI were identified through multivariate analysis. A prediction model of SBI was established and verified for its efficiency.

RESULTS

Regarding the parameters of intracranial blood flow resistance, the middle cerebral artery pulsatility index (MCA-PI) and middle cerebral artery resistance index (MCA-RI) were higher in the SBI group than in the non-SBI group (p < 0.05). The ultrasonic features of carotid plaques [plaque thickness and carotid intima-media thickness (CIMT)] as well as the parameters of intracranial blood flow resistance (MCA-PI and MCA-RI) were risk factors for SBI (odds ratio > 1, p < 0.05). The calibration curve resembled the straight line Y = X, the model was well-calibrated, and the concordance index was 0.889, suggesting a good discrimination of the model.

CONCLUSION

The SBI prediction model constructed on the basis of ultrasonic features of carotid plaques (plaque thickness and CIMT) plus parameters of intracranial blood flow resistance (MCA-PI and MCA-RI) has potential clinical value.