To construct a predictive model for neurological functional prognosis after endovascular thrombectomy in patients with acute anterior circulation large-vessel occlusion (AIS-LVO) based on multimodal cerebral blood flow imaging parameters, including transcranial Doppler, transcranial color-coded Doppler, and computed tomography perfusion, and to evaluate its clinical value.

A total of 174 patients with AIS-LVO who underwent mechanical thrombectomy were retrospectively enrolled. They were divided into a training set (121 cases) and an internal validation set (53 cases) at a ratio of 7∶3. According to the modified Rankin Scale (mRS) scores at 90-day follow-up after thrombectomy in the training set, patients were assigned to the favorable prognosis group and the poor prognosis group. Univariate difference analysis and multivariate Logistic regression analysis were performed to screen for independent risk factors, construct the predictive model, and validate its clinical application value.

Univariate and multivariate regression analyses identified age (OR=1.07, 95%CI: 1.02~1.11), the ratio of the mean velocity difference between the affected and healthy middle cerebral arteries (OR=1.34, 95%CI: 1.04~1.73), onset-to-recanalization time (OR=1.01, 95%CI: 1.01~1.04), and the volume of brain tissue with relative cerebral blood flow ＜30% (OR=1.18, 95%CI: 1.02~1.36) as independent risk factors for poor prognosis after thrombectomy (all P＜0.05). The model was established as follows: Logit (P, probability of poor prognosis) = −6.66 + 0.06 × age (years) + 0.01 × recanalization time (minutes) + 0.30 × ratio of the mean velocity difference between the affected and healthy MCAs + 0.16 × volume of brain tissue with rCBF ＜30% (mL). The area under the ROC curve (AUC) was 0.758 (95%CI: 0.67~0.85), with a theoretical predictive sensitivity of 63.2% and specificity of 83.7%. The Hosmer-Lemeshow goodness-of-fit test showed no significant lack of fit in the training set (P=0.301) or the internal validation set (P=0.126), indicating good calibration. Decision curve analysis suggested good clinical utility of the model.

Integrated models combining multimodal cerebral blood flow imaging and clinical data can improve the accuracy of predicting poor neurological outcomes after thrombectomy in AIS-LVO patients, thereby providing a basis for early warning of postoperative complications and personalized interventions.