Introducing an index on prediction of post-revascularization cerebral infarction using preoperative CT perfusion parameters in moyamoya disease

Abstract

Objective: To determine the value of preoperative CT perfusion (CTP) parameters for prediction of post-revascularization cerebral infarction (post-CI) in adults with moyamoya disease (MMD).

Methods: This retrospective study included 92 adults with MMD who underwent surgical revascularization. Preoperative quantitative CTP parameters, including cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT), time to drain (TTD), and transit time to maximum of the residue function (Tmax), along with clinical data, were compared between the groups with and without post-CI. Predictors of post-CI were identified and assessed using multivariable logistic regression and receiver-operating characteristic curve analyses.

Results: Post-CI occurred in 11 patients (12.0%). In univariate analysis, preoperative mean values for CBF, MTT, TTD, Tmax, initial presentation, infarction within the 2 months before surgery, surgical side, and modified Rankin Scale score on admission were associated with post-CI (all p < 0.05). Multivariable logistic regression revealed that the preoperative mean Tmax (OR 2.342, 95% CI: 1.267-4.330, p = 0.007) and infarction within the 2 months before surgery (OR 14.345, 95% CI: 2.108-97.638, p = 0.006) were independent predictors of post-CI. The preoperative mean Tmax produced the largest area under the curve (0.955, 95% CI: 0.914-0.997) with a cutoff of 3.590 s (sensitivity, 100%; specificity, 87.7%).

Conclusions: Adults with MMD are at risk of post-CI when the preoperative mean Tmax is > 3.590 s. Cerebral infarction during the 2 months before revascularization is also a risk factor for post-CI.

Critical relevance statement: Post-CI is a serious complication for adults with MMD following surgical revascularization. The risk of post-CI can be predicted using preoperative CTP parameters, which will assist neurosurgeons with surgical decisions and implementing individualized prophylactic strategies.

Key points: Predicting the risk of post-CI in MMD patients is beneficial to their prognosis. The preoperative mean Tmax was an excellent perfusion parameter for predicting post-CI. Preoperative CTP evaluation can help clinicians make cautious surgical decisions.

Keywords: Cerebral infarction; Cerebral revascularization; Moyamoya disease; Perfusion imaging; Tomography (X-ray computed).

Fig. 1

An illustration of the placement of 18 ROIs on four slices for quantitative CTP analysis. Nine ROIs were placed manually throughout the cortical flow territories on four standard levels, including the centrum semiovale (a), body of lateral ventricle (b), basal ganglia (c), and mesencephalon (d). Nine mirrored ROIs were automatically generated in the contralateral hemisphere. The quantitative values of CBF, CBV, MTT, TTD, and Tmax are automatically calculated for each ROI by the post-processing software (e)

Fig. 2

A 50-year-old male MMD patient with post-CI. Preoperative CTA showed MMD (a). The diffusion-weighted imaging performed two weeks before surgery demonstrated an acute cerebral infarction in the left parietal-occipital lobe (b). Preoperative CTP showed a significant decrease in CBF (c), a mild decrease in CBV (d), and a significant increase in MTT (e), TTD (f), and Tmax (g) in both hemispheres. Right hemiplegia appeared on the second day after surgery, followed by coma. Postoperative CT showed multiple newly developed cerebral infarctions in both hemispheres (h)

Fig. 3

A 50-year-old male MMD patient without post-CI. Preoperative CTA showed MMD (a). Preoperative diffusion-weighted imaging showed no acute cerebral infarction (b). Preoperative CTP showed no obvious changes in CBF (c), CBV (d), and a mild increase in MTT (e), TTD (f), and Tmax (g) in the left temporal-parietal lobe and right frontal lobe. Postoperative diffusion-weighted imaging showed no cerebral infarction (h)

Fig. 4

ROC curves prediction post-CI based on preoperative CTP parameters and clinical data. The mean Tmax produced the largest AUC with a cutoff of 3.590 s (sensitivity, 100%; specificity, 87.7%)