Preoperative computed tomography perfusion and angiography predict the need for shunting in carotid endarterectomy: a multicenter study

Introduction

About 20% of strokes result from narrowing of the carotid artery (1), the main vessel supplying blood to the brain. Carotid endarterectomy (CEA) has been shown in large, well-designed randomized controlled trials (RCTs) to significantly reduce the relative risk of stroke in patients who have experienced a recent transient ischemic attack or minor stroke associated with severe symptomatic carotid artery stenosis (2-4). Despite its efficacy, temporary occlusion of the carotid artery during surgery can lead to cerebral ischemia, particularly in patients with compromised cerebral hemodynamics (5). To mitigate this risk, carotid shunting is often performed to maintain cerebral perfusion during carotid clamping.

Despite its potential benefits, routine shunting use introduces significant risks, including thrombotic and embolic complications, neural injury, and procedural difficulties (6). These risks are particularly pronounced when the procedure is performed by surgeons without specialized training in shunting placement techniques. During shunt insertion and manipulation, atherosclerotic debris from the carotid artery or thrombi within the shunt can dislodge and travel to the cerebral circulation, resulting in embolic events (7). Paradoxically, such events can precipitate focal or diffuse cerebral ischemia, directly counteracting the shunt’s intended protective effect. Additionally, the placement of a shunt can limit visualization of and access to the distal carotid plaque—a significant concern in urgent procedures—and has been associated with an elevated risk of carotid dissection (8). When preoperative evaluation indicates the need for shunting, comprehensive carotid exposure combined with technically skilled insertion by an experienced operator may substantially reduce procedure-related complications. Therefore, identifying patients who would benefit most from selective shunting is crucial (7).

Acetazolamide single-photon emission computed tomography (CT), which demonstrates reduced or absent vasodilatory capacity, has proven valuable for identifying patients most suitable for CEA or surgical revascularization (9). However, these nuclear imaging techniques require the administration of radioactive tracers, limiting their widespread clinical application. In contrast, CT perfusion (CTP) imaging has emerged as a robust alternative, offering comprehensive evaluation of cerebral vascular physiology and hemodynamic reserve. This imaging approach provides a quantitative assessment of key hemodynamic parameters—cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT), and time to peak (TTP)—which are established indicators of cerebral perfusion status. Specifically, reduced CBF reflects compromised perfusion, thereby identifying regions at elevated ischemic risk, whereas prolonged MTT indicates impaired cerebrovascular reactivity and diminished tolerance to perfusion pressure fluctuations (10). Multimodal CT imaging acquires multiple datasets from a single contrast injection, which can be processed to provide high-resolution vascular anatomy. This includes the degree and level of carotid stenosis, plaque morphology, and intracranial vasculature integrity—particularly of the circle of Willis (CoW)—thereby directly informing CEA surgical planning. Simultaneously, the protocol generates quantitative perfusion maps, enabling a comprehensive evaluation of cerebral hemodynamic alterations caused by carotid stenosis (11,12).

Previous studies have largely focused on modalities such as single-photon emission CT (SPECT) and magnetic resonance angiography (MRA) for cerebral hemodynamic evaluation (13). However, no study to date has specifically investigated the predictive value of CTP and CT angiography (CTA) for determining shunting requirement during CEA. Therefore, this study aimed to explore whether preoperative CTP combined with CTA can predict the need for selective carotid shunting during surgery. By analyzing the correlation between CTP- and CTA-derived parameters and the intraoperative occurrence of cerebral ischemia, we sought to determine whether CTP and CTA can serve as reliable and noninvasive tools for guiding surgical decision-making in CEA. The findings of this study may help to optimize the use of carotid shunting, enhance surgical outcomes, and improve patient safety. We present this article in accordance with the TRIPOD reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2319/rc).

Methods

Study design and patient selection

A total of 373 patients undergoing CEA were initially identified from two tertiary centers between January 2020 and December 2024. After applying the inclusion and exclusion criteria, 283 patients were included in the final analysis. Hospital I (Tianjin First Central Hospital, n=126) employed intraoperative monitoring to guide selective shunting, whereas hospital II (Tianjin Huanhu Hospital, n=157) utilized a non-shunting protocol. Predictive modeling for shunting requirement was developed using multivariate logistic regression in the hospital I cohort, followed by external validation of the derived model in the hospital II cohort. External validation was performed by applying the derived predictors from Hospital I to Hospital II to assess their ability to predict postoperative cerebral infarction—a clinically relevant endpoint reflecting the ischemic risk that shunting aims to prevent. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committees of Tianjin First Central Hospital (approval No. KYAP2025-109) and Tianjin Huanhu Hospital (approval No. 2025-026), and the requirement for informed consent was waived due to the retrospective nature of the study.

Inclusion criteria:

• Completion of preoperative evaluations, including CTA, CTP, magnetic resonance imaging (MRI), and carotid Doppler ultrasonography.
• Availability of postoperative MRI within 72 hours after surgery.
• Performance of a microscope-assisted standard CEA.
• For patients in hospital I: intraoperative monitoring with transcranial Doppler (TCD) and somatosensory evoked potentials.
• For patients in hospital II: no intraoperative monitoring during carotid clamping, with mean arterial pressure maintained at 120–130% of baseline levels.
• All procedures performed by a single designated chief neurosurgeon at each respective center.
• Administration of general anesthesia in all cases.

Exclusion criteria:

• Incomplete preoperative or postoperative imaging data.
• Use of carotid patch angioplasty or eversion endarterectomy.
• Combined cardiac and cerebrovascular procedures.
• Absence of intraoperative monitoring in hospital I.
• Procedures not performed by the designated operating surgeon.
• Surgery performed under local or regional anesthesia.

Figure 1 displays the patient distribution and overall study design.

Figure 1 Patient distribution and study design overview. Incomplete preoperative and postoperative imaging, preoperative computed tomography angiography, computed tomography perfusion, and magnetic resonance imaging, performed within 72 hours after surgery. Hospital I: Tianjin First Central Hospital; Hospital II: Tianjin Huanhu Hospital. CEA, carotid endarterectomy.

Data collection

Demographic and clinical characteristics: data on demographics and clinical characteristics were collected from medical records. Variables included sex, age, history of cerebral infarction, hypertension, diabetes mellitus, and preoperative symptomatic status. Patients were categorized as asymptomatic (stenosis detected incidentally) or symptomatic (presenting with a transient ischemic attack or established cerebral infarction).

Laboratory variables: assessed laboratory parameters included hyperlipidemia, serum homocysteine, and uric acid levels.

Imaging data: preoperative CTA was used to evaluate the degree of contralateral carotid artery stenosis and the integrity of the CoW. Significant contralateral stenosis was defined as ≥50% lumen narrowing. CoW integrity was defined as the anatomical presence of both the anterior communicating artery and the ipsilateral posterior communicating artery. Postoperative MRI was performed to assess for new cerebral infarctions.

CTP and CTA scanning

The imaging studies were performed using a 256-multidetector CT system (Revolution, GE Healthcare, Chicago, IL, USA). For the CTP scan, 40 mL of contrast material (370 mg iodine/mL; Ultravist 370, Bayer Schering Pharma, Berlin, Germany) was injected into the cubital vein using an 18 G needle at a rate of 5 mL/s followed by a 30 mL saline flush at a rate of 5 mL/s using a high-pressure syringe (MissouriXD2001, Ulrich, Ulm, Germany). A toggle-table technique (Jog mode) was employed with the following parameters: 80 kVp, 180 mAs, 128×0.625 mm collimation, and a 512×512 matrix. Scanning commenced 4 seconds after contrast injection, covered a 160 mm range from the skull base to the vertex (28 slices at 5 mm thickness; 672 total images), and comprised 24 cycles at 4-second intervals over 56 seconds.

For CTA, a separate 40 mL bolus of contrast was injected at 5 mL/s, followed by a 30 mL saline flush. The scanning parameters were 100 kVp, 150 mAs, 128×0.625 mm collimation, 512×512 matrix, and 0.9 mm slice thickness, covering from the aortic arch to the vertex. Using bolus tracking with a trigger threshold of 120 Hounsfield units (HU) in the aortic arch, data acquisition began 4 seconds after triggering.

CTP data were automatically processed on an offline workstation (AW 4.7, GE Healthcare). Following this, on multiplanar reconstruction images at the level of the centrum semiovale, a semi-circular region of interest (ROI) was manually delineated along the cerebral cortex edge. The mirror function was then applied to generate a symmetrical contralateral ROI. CBF values were recorded for both hemispheres, and the relative CBF (rCBF) ratio was calculated. All ROIs were delineated and CBF values were independently determined by two experienced neuroradiologists, with consensus reached in cases of disagreement.

Figure 2 illustrates the preoperative CTP-CTA findings from an individual case in hospital I, demonstrating the application of the defined CTP and CTA parameters.

Figure 2 Imaging findings from a 75-year-old male patient in Tianjin First Central Hospital. (A) CTA source image demonstrates severe stenosis of the right internal carotid artery (white arrow). (B) Post-processed CTA image reveals significant calcification within the carotid plaque (white arrow). (C) CTA shows an incomplete Circle of Willis configuration, characterized by absence of the ipsilateral posterior communicating artery (white arrow). (D) CTP map with manually delineated ROI placed in the bilateral centrum semiovale, displaying quantitative CBF values. The rCBF ratio was calculated as 38.07/59.85 =63.6%. CBF, cerebral blood flow; CTA, computed tomography angiography; CTP, computed tomography perfusion; MPR, multiplanar reconstruction; rCBF, relative cerebral blood flow; ROI, region of interest.

CEA and shunting

The surgical indication for CEA strictly adhered to the criteria established by the North American Symptomatic Carotid Endarterectomy Trial (NASCET) (2). All procedures were performed under general anesthesia. Patients were positioned supine with the neck extended and rotated contralaterally. A longitudinal incision was made along the anterior border of the sternocleidomastoid muscle to expose the carotid bifurcation. Following systemic heparinization (typically a 5,000 IU intravenous bolus) (14), the common carotid artery (CCA), internal carotid artery (ICA), and external carotid artery were sequentially clamped.

At hospital I, somatosensory evoked potentials and TCD monitoring of bilateral middle cerebral artery flow velocities were established after anesthesia induction. Baseline readings were obtained, and continuous monitoring was maintained throughout the procedure by a dedicated electrophysiologist. An intraluminal shunt was inserted if either the mean blood flow velocity or somatosensory evoked potentials amplitudes decreased by more than 50% from baseline (5,13,15). Based on previous experience, a reduction in flow velocity typically preceded changes in potentials. The shunt was inserted with one end in the CCA and the other in the ICA, using gentle manipulation to minimize endothelial injury. Following complete plaque excision, the arteriotomy was closed with a continuous 6-0 polypropylene suture. The closure was initiated from both the distal and proximal ends, with the sutures meeting at the midpoint to ensure uniform tension distribution.

At hospital II, no intraoperative monitoring was employed. During carotid clamping, the mean arterial pressure was pharmacologically maintained at 120–130% of the baseline level. Plaque excision followed identical technical principles, but without shunt placement.

Statistical analysis

Univariate analysis was initially performed to identify potential factors associated with shunt placement during CEA. Continuous variables, which were non-normally distributed, are presented as medians with interquartile ranges (IQRs) and were compared using the Mann-Whitney U test. Categorical variables are expressed as numbers (percentages) and were compared using the chi-square test or Fisher’s exact test, as appropriate. Variables showing a significant association (P<0.05) in the univariate analysis were subsequently included in a multivariate logistic regression model to assess their independent effects. The results of the regression analysis are reported as odds ratios (ORs) with corresponding 95% confidence intervals (CIs). The discriminative ability of significant predictors was evaluated using receiver operating characteristic (ROC) curve analysis, and the area under the curve (AUC) was calculated. For the key continuous predictor, regional rCBF, the optimal cutoff value was determined by maximizing Youden’s index.

To evaluate the generalizability of the prediction model, external validation was conducted using an independent cohort from hospital II. The rCBF cutoff value derived from the ROC analysis in the development cohort (hospital I) was applied directly to the validation cohort. The predictive performance of this cutoff for postoperative cerebral infarction was assessed by calculating its sensitivity and specificity.

To evaluate the robustness of the predictive factors, we performed subgroup analyses based on key clinical characteristics-preoperative symptomatic status (asymptomatic or symptomatic) and contralateral carotid artery stenosis severity (significant stenosis or non-significant stenosis). Within each stratum, a binary logistic regression model incorporating interaction was fitted, and stratified ROC curves were generated to assess the model's discriminative performance.

All statistical analyses were performed using the software SPSS 30.0 (IBM Corp., Armonk, NY, USA) and GraphPad Prism (Version 10.0; GraphPad Software, San Diego, CA, USA). A two-sided P value <0.05 was considered statistically significant.

Results

Baseline characteristics of patients

A total of 283 patients were included from two centers (hospital I: n=126; hospital II: n=157). The cohorts were well-matched in baseline demographics and clinical profiles. However, a significant difference was observed in postoperative infarction rates (χ²=10.38, P=0.001) (Table 1). The cohort from hospital I served as the development set to identify risk factors for shunting requirement. These factors subsequently underwent external validation in the cohort from hospital II to assess their generalizability.

Analysis of predictive factors

Univariate analysis in hospital I revealed that contralateral carotid artery significant stenosis (P=0.03), reduced rCBF (P<0.001), and an incomplete CoW (P<0.001) were associated with the necessity for intraoperative shunting. In the subsequent multivariate model, both reduced rCBF (OR =0.35, 95% CI: 0.12–0.89, P<0.001) and an incomplete CoW (OR =5.47, 95% CI: 1.70–17.59, P=0.04) remained independent predictors for shunting (Table 2). Similarly, in hospital II, both univariate and multivariate analyses identified the same two factors as independent risk factors significantly associated with postoperative cerebral infarction: reduced rCBF (OR =0.21, 95% CI: 0.10–0.44, P<0.001) and an incomplete CoW (OR =7.22, 95% CI: 1.99–26.24, P=0.003) (Table 3).

Predictive value in hospital I and external validation in hospital II

ROC curve analysis in hospital I demonstrated that rCBF possessed excellent discriminative ability for predicting shunting requirement, with an AUC of 0.82 (95% CI: 0.70–0.94). The optimal preoperative rCBF cutoff value was determined to be 53.5%, yielding a sensitivity of 92.5% and a specificity of 70.0%.

External validation was performed by applying this predefined rCBF cutoff (≤53.5%) to the cohort from hospital II. The ROC curve analysis for predicting postoperative infarction in this cohort yielded an AUC of 0.90 (95% CI: 0.84–0.96). The predefined cutoff of 53.5% maintained strong predictive performance, with a sensitivity of 79.9% and a specificity of 83.3%, thereby confirming its robust generalizability across different clinical settings (Figure 3).

Figure 3 Predictive value in hospital I and external validation in Hospital II. Hospital I: Tianjin First Central Hospital; Hospital II: Tianjin Huanhu Hospital. AUC, area under the curve.

Subgroup predictive value in hospital I

To assess the stability of the predefined rCBF cutoff (≤53.5%), subgroup analyses were conducted on clinically important variables in hospital I (preoperative symptomatic status and the severity of stenosis in the contralateral carotid artery). In hospital I, the predictive performance remained strong across all subgroups. For symptomatic patients, the AUC was 0.85 (95% CI: 0.70–0.99) with a sensitivity of 93.2% and a specificity of 75.0%. In asymptomatic patients, the AUC was 0.76 (95% CI: 0.56–0.97) with a sensitivity of 90.6% and a specificity of 62.5%. When stratified by contralateral carotid artery stenosis severity, the AUC was 0.83 (95% CI: 0.67–1.00) in patients with significant stenosis (sensitivity 94.0%, specificity 65.0%) and 0.82 (95% CI: 0.66–0.98) in those with non-significant stenosis (sensitivity 90.5%, specificity 77.8%) (Figure 4).

Figure 4 Subgroup predictive value in hospital I. (A) The severity of stenosis in the contralateral carotid artery; (B) preoperative symptomatic status. Hospital I: Tianjin First Central Hospital. AUC, area under the curve.

Subgroup consistency and interaction analysis

Results from tests for interaction confirmed the robustness of the predefined rCBF cutoff (≤53.5%) as a predictor for shunting requirement. Specifically, there was no significant effect modification by preoperative symptomatic status (P=0.70) (Table 4) or by the presence of contralateral carotid artery stenosis severity (P=0.84) (Table 5). These findings indicate that the strong association between the rCBF criterion and the need for shunting remains consistent and is not materially altered across these key clinical patient subgroups.

Discussion

This two-center study establishes that preoperative quantitative CTP and CTA effectively predict the need for shunt placement during CEA. We identified rCBF and CoW integrity as key independent predictors. The generalizability of the rCBF threshold was confirmed through external validation, and its robustness was affirmed by comprehensive subgroup analyses. These findings position CTP-CTA as a reliable tool for preoperative risk stratification.

At hospital I, the shunting usage rate was 15.9%, based on intraoperative TCD monitoring of middle cerebral artery blood flow velocity combined with somatosensory evoked potentials monitoring. This value falls within the previously reported range of 4–20% (16-18). The broad variability in shunting utilization across studies may be attributed to differences in anesthesia techniques and intraoperative monitoring protocols. In contrast, hospital II adopted a non-shunting protocol, relying solely on pharmacologic elevation of blood pressure to 120–130% of baseline during carotid clamping. However, this approach resulted in a postoperative cerebral infarction rate of 11.5%, markedly higher than the 1.6% observed at hospital I. Recent studies have also highlighted the role of CEA in modulating systemic blood pressure and cardiovascular outcomes, underscoring the importance of surgical technique in perioperative hemodynamic management (19,20).

A meta-analysis of 1,270 European patients reported that routine shunting use during CEA reduced postoperative infarction rates by approximately 4% compared with non-shunting procedures (21). Nevertheless, the risks associated with routine shunting primarily arise from potential complications during shunting insertion, such as air embolism, arterial injury, and technical errors related to operator experience. Previous studies have noted that most shunting-related complications occur in emergency operations (22). Previous strategies have focused on intraoperative monitoring (e.g., EEG, TCD, stump pressure) to guide shunting (15,23-25), whereas our protocol incorporates preoperative CTP and CTA to evaluate perfusion on the surgical side and assess collateral circulation adequacy. This strategy allows anticipatory identification of clamp-induced ischemia and facilitates judicious shunting placement—through extended carotid exposure by skilled surgeons—to minimize shunting-related complications.

Our findings established that an ipsilateral reduction in rCBF of ≥53.5% relative to the contralateral side served as the optimal shunting criterion, achieving a sensitivity of 92.5% and a specificity of 70.0%. The predictive performance of this rCBF cutoff was substantially superior to that of traditional parameters, such as contralateral carotid stenosis or preoperative symptomatic status, neither of which was an independent predictor of shunting requirement. Furthermore, subgroup analyses confirmed that these traditional factors did not significantly modify the predictive effect of the rCBF threshold. Our results align with previous studies that used impaired cerebrovascular reserve—assessed nonquantitative via acetazolamide single-photon emission computed tomography—as a shunting indicator (positive predictive value, 91%; negative predictive value, 94%) (26). A systematic review by Jaffer et al. [1950–2015] (17) on shunting prediction methods showed that both three-dimensional time-of-flight MRA and acetazolamide-challenged SPECT achieved high negative predictive values (96% and 94%, respectively) for determining shunting necessity during CEA. However, acetazolamide SPECT assessment of cerebrovascular reserve (CVR) relies on qualitative visual interpretation, making it prone to significant observer bias and limited reproducibility (26). In contrast, CTP—supported by advanced post-processing software—enables quantitative analysis of cerebral hemodynamic parameters within defined ROIs, substantially reducing subjectivity. Although combining SPECT with positron emission tomography (PET) allows for quantitative CBF assessment, its clinical adoption remains restricted by high cost, radiation exposure, and prolonged acquisition time. Multimodal CT imaging offers several practical advantages: a simplified workflow, shorter acquisition times that enhance patient comfort, and a comprehensive “one-stop-shop” evaluation that provides simultaneous vascular, perfusion, and parenchymal data. This integrated approach delivers a multifaceted pathophysiologic characterization, which significantly improves surgical planning and intraoperative decision-making. Consequently, CTP and CTA represent a noninvasive, robust, and quantitatively precise preoperative assessment toolkit that can stratify patients at higher risk of clamp-induced cerebral ischemia, thereby supporting well-prepared and judicious shunting.

CVR, which is typically assessed using acetazolamide SPECT, reflects the brain’s vasodilatory capacity in response to hemodynamic stress (27). Reduced CVR indicates impaired autoregulation and serves as an early, sensitive marker of hemodynamic failure, thereby identifying patients at high risk for ischemia during carotid clamping in CEA (28). However, measuring CVR requires a vasodilatory stimulus such as acetazolamide and involves radioactive tracers, both of which limit its clinical applicability and quantitative precision. The administration of acetazolamide is frequently associated with adverse reactions, including metabolic acidosis, hypokalemia, paresthesia, headache, tinnitus, gastrointestinal disorders, and, in rare cases, Stevens-Johnson syndrome. Clinical reports indicate that approximately 63% of patients experience side effects within 1–3 hours after acetazolamide administration, with symptoms persisting from 30 minutes to 72 hours. These reactions can substantially interfere with daily and occupational functioning (29). Following cerebrovascular stenosis, autoregulatory mechanisms may initially maintain CBF within normal ranges. Therefore, a significant reduction in CBF directly indicates compromised hemodynamic function and failure of compensatory vasodilation (30). Unlike provocative tests, CTP-derived CBF enables the direct and quantitative assessment of cerebral perfusion without pharmacologic stimulation or radioactive tracers. The centrum semiovale was selected as the ROI based on established hemodynamic principles. Although absolute CBF values in CTP can vary with ROI placement, the centrum semiovale is a strategic watershed zone located at the junction of terminal cortical branches (from the middle and anterior cerebral arteries) and deep perforating arteries (e.g., the lenticulostriate arteries). This boundary region is highly vulnerable to reductions in perfusion pressure and is among the first to manifest ischemic changes during systemic hemodynamic compromise. An et al. (31) demonstrated that chronic cerebral arterial occlusion produced more severe ischemia in deep white matter than in cortical gray matter, further validating the pathophysiological relevance of this region. Therefore, selecting the centrum semiovale as a standardized ROI provides a representative measure of perfusion impairment secondary to carotid stenosis. Furthermore, employing rCBF—calculated as the ratio of ipsilateral to contralateral values—as our primary analytical metric effectively minimizes confounding effects from individual physiological variability and technical factors, thereby enhancing the robustness and clinical applicability of the results.

During carotid cross-clamping in CEA, perfusion of the ipsilateral hemisphere is predominantly sustained by collateral pathways through the CoW (32). Hemodynamic models indicate that the anterior circulation contributes more substantially than the posterior circulation during acute unilateral carotid occlusion. However, the coexistence of ipsilateral and contralateral carotid stenosis can reduce ipsilateral middle cerebral artery blood flow by approximately 40% (33). Benjamin et al. reported that only 16.6% of patients with severe contralateral ICA stenosis developed significant cerebral ischemia during carotid clamping—a rate comparable to that in patients without contralateral stenosis (34). These findings support the conclusion that contralateral carotid stenosis is not an independent predictor of ischemia during clamping, which aligns with the results of the present study. In patients with coexisting contralateral stenosis, the posterior circulation serves as a critical compensatory pathway (32). Consequently, both the anterior and posterior circulations are indispensable for maintaining adequate cerebral perfusion during cross-clamping. The CoW exhibits considerable anatomical variability. Studies have reported a complete CoW in only 13–59% of individuals, whereas imaging-based analyses indicate a prevalence of 12–79% (35,36). Lengyel et al. (37) demonstrated that patients with an incomplete CoW undergoing CEA—particularly those with an isolated middle cerebral artery configuration—are at a significantly greater risk of ischemia during clamping. Similarly, other studies have confirmed that a nonfunctional CoW is strongly associated with an increased risk of intraoperative ischemia (38). Consistent with these findings, our study demonstrated that incomplete CoW is a key determinant of ischemic risk. In hospital I, incomplete CoW was significantly more prevalent in the shunting group than in the non-shunting group, and multivariate analysis confirmed it as an independent predictor of shunting requirement. Similarly, in hospital II, incomplete CoW correlated strongly with postoperative cerebral infarction. Collectively, these results from two independent centers underscore the predictive value of CoW configuration in assessing shunting necessity during carotid clamping and suggest that preoperative CoW evaluation may help reduce postoperative infarction rates.

This study has several limitations. First, its retrospective design—despite demonstrating strong predictive value for shunting necessity in hospital I and validating the rCBF threshold in hospital II—lacks the robustness of a large-sample, prospective, RCT. Second, although the rCBF cutoff showed stability in subgroup analyses, the wide CIs observed, likely due to the limited number of events after stratification, indicate a need for caution. Due to the limited sample size, our analysis employed a dichotomous cutoff (≥50%) for contralateral carotid stenosis and was unable to support a more granular stratification (50–69%, 70–99%, or occlusion). Future studies with larger samples are required to confirm these findings. Third, all procedures were performed under general anesthesia, which reduces cerebral metabolic demand; therefore, the results may not be generalizable to CEA performed under local anesthesia. Fourth, our predictive model was developed from a cohort that underwent a standardized surgical technique. Its applicability to patients undergoing eversion endarterectomy or patch angioplasty remains uncertain, as these approaches may differ in operative duration, clamp time, and hemodynamic stress on collateral circulation. Finally, although the combination of electrophysiological and TCD monitoring was used to enhance diagnostic accuracy, this approach may limit the generalizability of our findings to centers that base shunting decisions on other modalities, such as carotid stump pressure, electroencephalography, or cerebral oximetry.

Conclusions

Our two-center study establishes that preoperative rCBF and CoW integrity are independent predictors for shunt necessity during CEA. Preoperative CTP-CTA integration thus provides a reliable, quantitative basis for surgical decision-making, offering a strategy to optimize shunt utilization, improve outcomes, and enhance perioperative patient safety.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2319/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2319/dss

Funding: This work was supported by Tianjin Key Medical Discipline Construction Project (grant No. TJYXZDXK-3-002A-3), and Tianjin Municipal Health Commission (No. TJWJ2021QN017).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2319/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committees of Tianjin First Central Hospital (approval No. KYAP2025-109) and Tianjin Huanhu Hospital (approval No. 2025-026), and the requirement for informed consent was waived due to the retrospective nature of the study.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Grau AJ, Weimar C, Buggle F, Heinrich A, Goertler M, Neumaier S, Glahn J, Brandt T, Hacke W, Diener HC. Risk factors, outcome, and treatment in subtypes of ischemic stroke: the German stroke data bank. Stroke 2001;32:2559-66. [Crossref] [PubMed]
2. Barnett HJM, Taylor DW, Haynes RB, Sackett DL, Peerless SJ, Ferguson GG, Fox AJ, Rankin RN, Hachinski VC, Wiebers DO, Eliasziw M. Beneficial effect of carotid endarterectomy in symptomatic patients with high-grade carotid stenosis. N Engl J Med 1991;325:445-53. [Crossref] [PubMed]
3. MRC European Carotid Surgery Trial: interim results for symptomatic patients with severe (70-99%) or with mild (0-29%) carotid stenosis. European Carotid Surgery Trialists' Collaborative Group. Lancet 1991;337:1235-43. [Crossref] [PubMed]
4. Rothwell PM, Eliasziw M, Gutnikov SA, Fox AJ, Taylor DW, Mayberg MR, Warlow CP, Barnett HJCarotid Endarterectomy Trialists' Collaboration. Analysis of pooled data from the randomised controlled trials of endarterectomy for symptomatic carotid stenosis. Lancet 2003;361:107-16. [Crossref] [PubMed]
5. Uno M, Takai H, Yagi K, Matsubara S. Surgical Technique for Carotid Endarterectomy: Current Methods and Problems. Neurol Med Chir (Tokyo) 2020;60:419-28. [Crossref] [PubMed]
6. Squizzato F, Siracuse JJ, Shuja F, Colglazier J, Balachandran Wilkins P, Goodney PP, Sands Brooke B, DeMartino RR. Impact of Shunting Practice Patterns During Carotid Endarterectomy for Symptomatic Carotid Stenosis. Stroke 2022;53:2230-40. [Crossref] [PubMed]
7. Levin SR, Farber A, Goodney PP, Schermerhorn ML, Patel VI, Arinze N, Cheng TW, Jones DW, Rybin D, Siracuse JJ. Shunt intention during carotid endarterectomy in the early symptomatic period and perioperative stroke risk. J Vasc Surg 2020;72:1385-1394.e2. [Crossref] [PubMed]
8. Hommery-Boucher X, Fortin W, Beaudoin N, Blair JF, Stevens LM, Elkouri S. Editor's Choice - Safety of Shunting Strategies During Carotid Endarterectomy: A Vascular Quality Initiative Data Analysis. Eur J Vasc Endovasc Surg 2024;68:695-702. [Crossref] [PubMed]
9. Kaneta T. PET and SPECT imaging of the brain: a review on the current status of nuclear medicine in Japan. Jpn J Radiol 2020;38:343-57. [Crossref] [PubMed]
10. Xu H, Han H, Liu Y, Huo R, Lang N, Yuan H, Wang T, Zhao X. Perioperative cerebral blood flow measured by arterial spin labeling with different postlabeling delay in patients undergoing carotid endarterectomy: a comparison study with CT perfusion. Front Neurosci 2023;17:1200273. [Crossref] [PubMed]
11. Sha A, Yu F, Zhang M, Shan Y, Guo D, Shakya MR, Bai X, Ma Q, Jiao L, Lu J. Multimodal CT imaging characteristics may predict post-reperfusion infarct volume in wake-up stroke patients. Quant Imaging Med Surg 2023;13:878-88. [Crossref] [PubMed]
12. Sabarudin A, Subramaniam C, Sun Z. Cerebral CT angiography and CT perfusion in acute stroke detection: a systematic review of diagnostic value. Quant Imaging Med Surg 2014;4:282-90. [PubMed]
13. Udesh R, Natarajan P, Thiagarajan K, Wechsler LR, Crammond DJ, Balzer JR, Thirumala PD. Transcranial Doppler Monitoring in Carotid Endarterectomy: A Systematic Review and Meta-analysis. J Ultrasound Med 2017;36:621-30. [Crossref] [PubMed]
14. Přibáň V, Dostál J, Mork J, Mraček J. Carotid endarterectomy from the neurosurgeon's perspective. Rozhl Chir 2025;104:242-6. [Crossref] [PubMed]
15. Yang N, Wang Q, Qi H, Song Z, Zhou C, Zhang S, Zhang B. TCD-Guided management in carotid endarterectomy: a retrospective study. J Cardiothorac Surg 2024;19:588. [Crossref] [PubMed]
16. Myrcha P, Pinheiro F, Rocha-Neves J, Myrcha J, Gloviczki P. The effect of the collateral cerebrovascular circulation on tolerance to carotid artery cross-clamping and on early outcome after carotid endarterectomy. J Vasc Surg 2024;80:1603-1613.e6. [Crossref] [PubMed]
17. Jaffer U, Normahani P, Harrop-Griffiths W, Standfield NJ. Pre-operative methods to predict need for shunting during carotid endarterectomy. Int J Surg 2015;23:5-11. [Crossref] [PubMed]
18. Wain RA, Veith FJ, Berkowitz BA, Legatt AD, Schwartz M, Lipsitz EC, Haut SR, Bello JA. Angiographic criteria reliably predict when carotid endarterectomy can be safely performed without a shunt. J Am Coll Surg 1999;189:93-100; discussion 100-1. [Crossref] [PubMed]
19. Germanova O, Galati G, Vachev A, Germanov A, Biondi-Zoccai G. The operation of carotid endarterectomy: is it an effective method of treatment of arterial hypertension? Panminerva Med 2023;65:211-9. [Crossref] [PubMed]
20. Germanova O, Galati G, Vachev A, Germanov A, Biondi-Zoccai G. Carotid endarterectomy: the optimal surgical technique from the point of view of cardiologist. Chirurgia 2022;35:355-62. [Crossref]
21. Counsell C, Salinas R, Naylor R, Warlow C. Routine or selective carotid artery shunting for carotid endarterectomy (and different methods of monitoring in selective shunting). Cochrane Database Syst Rev 2000;CD000190. [PubMed]
22. Chuatrakoon B, Nantakool S, Rerkasem A, Orrapin S, Howard DP, Rerkasem K. Routine or selective carotid artery shunting for carotid endarterectomy (and different methods of monitoring in selective shunting). Cochrane Database Syst Rev 2022;6:CD000190. [PubMed]
23. Aburahma AF, Stone PA, Hass SM, Dean LS, Habib J, Keiffer T, Emmett M. Prospective randomized trial of routine versus selective shunting in carotid endarterectomy based on stump pressure. J Vasc Surg 2010;51:1133-8. [Crossref] [PubMed]
24. Simon MV, Malcharek M, Ulkatan S. Monitoring in carotid endarterectomy. Handb Clin Neurol 2022;186:355-74. [Crossref] [PubMed]
25. Jovanovic A, Jonsson M, Roy J, Eriksson J, Mutavdzic P, Trailovic R, Koncar I. Comparison of Methods for Monitoring Intra-operative Cerebral Perfusion in Patients Undergoing Carotid Endarterectomy with Selective Shunting: A Systematic Review and Network Meta-Analysis of Randomised Controlled Trials and Cohort Studies. Eur J Vasc Endovasc Surg 2023;65:233-43. [Crossref] [PubMed]
26. Kim JS, Moon DH, Kim GE, Cho YP, Kim JS, Ryu JS, Lee HK. Acetazolamide stress brain-perfusion SPECT predicts the need for carotid shunting during carotid endarterectomy. J Nucl Med 2000;41:1836-41. [PubMed]
27. Nemoto EM, Yonas H. Cerebrovascular Reserve (CVR) and Stages of Hemodynamic Compromise. Adv Exp Med Biol 2022;1395:363-6. [Crossref] [PubMed]
28. Stoll M, Hamann GF. Cerebrovascular reserve capacity. Nervenarzt 2002;73:711-8. [PubMed]
29. Saito H, Ogasawara K, Suzuki T, Kuroda H, Kobayashi M, Yoshida K, Kubo Y, Ogawa A. Adverse effects of intravenous acetazolamide administration for evaluation of cerebrovascular reactivity using brain perfusion single-photon emission computed tomography in patients with major cerebral artery steno-occlusive diseases. Neurol Med Chir (Tokyo) 2011;51:479-83. [Crossref] [PubMed]
30. Jiang M, Li G, He Q, Zhang Y, Li W, Gao Y, Yan J. Multimodal imaging evaluation of early neurological deterioration following acute ischemic stroke. Quant Imaging Med Surg 2024;14:4763-78. [Crossref] [PubMed]
31. An P, Shi H, Ren Q, Zhang G, Ren G, Guan S, et al. Blood CT pefusion analysis of cerebral gray matter and white matter in supply region of chronic cerebral artery occlusion. Chinese Journal of Radiolog 2019;53:877-81.
32. Lengyel B, Magyar-Stang R, Pál H, Debreczeni R, Sándor ÁD, Székely A, Gyürki D, Csippa B, István L, Kovács I, Sótonyi P Jr, Mihály Z. Non-Invasive Tools in Perioperative Stroke Risk Assessment for Asymptomatic Carotid Artery Stenosis with a Focus on the Circle of Willis. J Clin Med 2024;13:2487. [Crossref] [PubMed]
33. Alastruey J, Parker KH, Peiró J, Byrd SM, Sherwin SJ. Modelling the circle of Willis to assess the effects of anatomical variations and occlusions on cerebral flows. J Biomech 2007;40:1794-805. [Crossref] [PubMed]
34. Benjamin ME, Silva MB Jr, Watt C, McCaffrey MT, Burford-Foggs A, Flinn WR. Awake patient monitoring to determine the need for shunting during carotid endarterectomy. Surgery 1993;114:673-9; discussion 679-81. [PubMed]
35. Hindenes LB, Håberg AK, Johnsen LH, Mathiesen EB, Robben D, Vangberg TR. Variations in the Circle of Willis in a large population sample using 3D TOF angiography: The Tromsø Study. PLoS One 2020;15:e0241373. [Crossref] [PubMed]
36. Sibiya G, Omotoso BR, Harrichandparsad R, Lazarus L. Exploring the anatomical configurations of the cerebral arteries in a cohort of South African patients. Sci Rep 2024;14:6060. [Crossref] [PubMed]
37. Lengyel B, Vecsey-Nagy M, Peter C, Varga A, Kovats T, Bayerle P, Sotonyi P, Banga PV. The Circle of Willis Status Influences Neurological Complications of Carotid Endarterectomy with Contralateral Carotid Occlusion. Ann Vasc Surg 2024;108:410-8. [Crossref] [PubMed]
38. Gyöngyösi Z, Farkas O, Papp L, Bodnár F, Végh T, Fülesdi B. The value of transcranial Doppler monitoring of cerebral blood flow changes during carotid endarterectomy performed under regional anesthesia - A case series. Transl Neurosci 2022;13:476-82. [Crossref] [PubMed]