Correlation between fractional flow and magnetic resonance perfusion in patients with intracranial artery stenosis in the anterior circulation

Introduction

Intracranial atherosclerotic stenosis (ICAS) is one of the most common causes of stroke worldwide (1). Hemodynamics is involved in ischemic risk through large vessel stenosis, collateral circulation, and perfusion in anterior ICAS patients (1-4). Moreover, perfusion imaging analysis based on computational tomography perfusion (CTP), magnetic resonance perfusion (MRP), or arterial spin labeling (ASL) is a commonly used method for evaluating cerebral vascular function. The time to maximum tissue residue function (Tmax) was reported as a potential predictive factor for stroke recurrence risk in symptomatic, anterior circulation, and moderate-to-severe ICAS patients (4). However, owing to the lack of an evaluation method, research on the relationship between large vessel function and cerebral perfusion is not yet sufficient.

Fractional flow reserve (FFR) is widely recognized as the functional standard for evaluating coronary artery stenosis (5). FFR is obtained by calculating the pressure ratio at the proximal and distal ends of coronary stenosis. To explore the value of functional evaluation in ICAS, various teams have utilized pressure wires to gauge pressure-derived parameters to assess the severity of ICAS (2,6). For example, fractional flow (FF) is obtained by calculating the pressure ratio at the proximal and distal ends of intracranial arterial stenosis (2). However, the high costs, operational complexity, and associated risks of pressure wires have constrained their clinical use. Consequently, nonwire techniques have been developed. By combining computational fluid dynamics (CFD) and medical imaging, such as computed tomography angiography (CTA), magnetic resonance angiography (MRA), or digital subtraction angiography (DSA), it is feasible to noninvasively evaluate the hemodynamics of ICAS (7-10).

Although some studies have proposed the value of functional evaluation in the diagnosis and treatment of ICAS, FF is not yet a standard (9,11). There is still a lack of sufficient data to discuss its association with the results of cerebral perfusion assessment and diameter stenosis (DS). Therefore, it is necessary to use commonly used methods to verify the clinical significance of FF. In this study, we computed FF via a CFD method based on DSA images and explored its correlation with MRP in anterior circulation intracranial stenosis patients. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-779/rc).

Methods

Patients

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Medical Ethics Committee of Zhejiang Hospital (approval No. 2022-80K) and individual consent for this retrospective analysis was waived.

Eighty-seven consecutive patients with suspected ICAS in the anterior circulation at our center from May 2021 to May 2024 were included in this retrospective study. Eligible patients had clinical symptoms consistent with ischemic events in the relevant vascular territory, including transient ischemic attack (TIA) or ischemic stroke within the preceding 90 days. Patients with suspected ICAS were first screened using non-invasive imaging modalities (CTA and/or MRA). DSA was performed if (I) non-invasive imaging showed possible DS ≥50%; (II) the lesion involved clinically relevant vascular territories; and (III) there were no contraindications for the procedure. ICAS was identified on DSA when DS ≥50% according to the Warfarin-Aspirin Symptomatic Intracranial Disease (WASID) criteria: DS = [(1 − (D_stenosis/D_normal)] ×100%, where D_stenosis is the diameter of the artery at the site of the most severe degree of stenosis and D_normal is the diameter of the proximal normal artery.

The inclusion criteria were as follows: (I) DS of the intracranial artery between 30% and 90%; and (II) ICAS lesions located between internal carotid artery (ICA) C2 segment and the beginning of middle cerebral artery (MCA) M2 segment. The exclusion criteria were as follows: (I) poor image quality; (II) severe overlap or distortion of target vessels; (III) acute cerebral infarction within 2 weeks; (IV) preoperative baseline modified Rankin Scale (mRS) ≥3; (V) proximal or distal aneurysms present in the stenotic segment; (VI) concomitant intracranial or extracranial stenosis of tandem vascular lesions; (VII) patients with bilateral lesions.

All enrolled patients underwent reasonable drug therapy and a preoperative examination. Combined with the patient’s symptoms, the anatomical severity of the stenosis and perfusion stages, balloon dilatation or intracranial artery stenting were performed in patients with surgical indications. These patients received standard dual-antiplatelet therapy (100 mg/d aspirin and 75 mg/d clopidogrel, loading dose 300 mg once each) 5 days prior to the procedure and continued thereafter.

Data acquisition

All magnetic resonance examinations were performed on a SIGNA Explorer 1.5T scanner (GE Medical System, Boston, USA) with a 32-channel head coil. All patients underwent diffusion-perfusion magnetic resonance imaging (MRI). Echoplanar imaging (EPI) was used to acquire cross-sectional images of the brain, and multitemporal scanning was performed. MRP was acquired and included following sequences: axial diffusion-weighted imaging [repetition time (TR)/echo time (TE) 5,000/87 ms; b values 0 and 1,000 s/mm², echo time 40 ms, flip angle 60°, matrix 128×128, slice thickness 5 mm, slice gap 5 mm, field of view (FOV) 240 mm], axial perfusion-weighted imaging (TR/TE 2,000/40 ms; echo time 40 ms, flip angle 60°, matrix 128×128, slice thickness 5 mm, slice gap 5 mm, FOV 240 mm), and other sequences not used are not stated. A total of 40 dynamics were acquired with a temporal resolution of 2.0 s.

At the beginning of the collection of the first few phases, the high-pressure syringe was started to inject the contrast agent. Gadolinium contrast was injected with a flow rate set at 4–5 mL/s and a contrast dosage of 0.1–0.2 mmol/kg based on the concentration and relaxation rate of different contrast agents, individual weight differences, age, and vascular conditions, as recommended by American Society of Functional Neuroradiology (ASFNR) guidelines (12). No preload strategy was used, but correction was performed using T1 correction on the GE MRI post-processing workstation AW4.7.

DSA images were acquired via the Allura Xper FD20 angiographic system (Philips, Amsterdam, The Netherlands). A 5-F catheter was superselected via the radial or femoral artery to the proximal end of the target vessel. The responsible lesion was developed by the injection of contrast agent to obtain 2D and 3D images.

MRP parameters and FF computation

MRP image processing was performed via AccuMRP software (ArteryFlow Technology, Hangzhou, China). The automated analysis process includes image preprocessing, determination of arterial input function (AIF) (13) and venous output function (VOF), calculation of perfusion parameter maps, including cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT), and time-to-peak (TTP). The AIF was determined on a region of interest (ROI) automatically applied to the normal contralateral M1 segment of the MCA, and the VOF was determined on the superior sagittal sinus. The core step is the deconvolution algorithm based on the selected AIF, such as various singular value decomposition (SVD)-based methods and Bayesian inference method. The optimal AIF is found by an automatic algorithm based on the gamma fitting principle and k-means clustering algorithm. After the deconvolution calculation, the parameter maps can be calculated by the residual function curve R(t). Tmax was obtained. Patients are divided into four groups on the basis of Tmax ≤4 s, 4< Tmax ≤6 s, 6< Tmax ≤8 s, and Tmax >8 s.

Four round ROIs with a diameter of 18 mm were manually placed on affected cortical flow territories of the MCA on MRP maps at the same basal ganglia level (14), with mirrored ROIs automatically generated on the contralateral sides. The absolute perfusion value in each hemisphere was calculated by averaging the mean values of the four ROIs in the respective hemisphere (15). Relative perfusion parameters [relative MTT (rMTT) and relative TTP (rTTP)] were calculated as the ratio of the absolute values in the symptomatic and contralateral hemispheres, respectively. This step was completed by two operators independently. The inter-rater reliability for these measurements was quantitatively assessed using the intraclass correlation coefficient (ICC) based on a two-way mixed-effects model for absolute agreement. An ICC value greater than 0.90 was considered to indicate excellent reliability. If the results of the two operators differ greatly, the third operator will check and adjust the unreasonable results. Finally, the average of the two results was used for subsequent analysis.

Cerebral perfusion was classified into three stages (normal perfusion, hypoperfusion stage I, and hypoperfusion stage II) according to a previously published grading system (16). Hypoperfusion stage I refers to hypoperfusion with decreased CBF, increased TTP/MTT, and normal or elevated CBV, whereas stage II refers to hypoperfusion with decreased CBV.

The FF was calculated via AccuFFicas, a hemodynamic calculation software for ICAS based on DSA images and verified in a previous study by Wang et al. (9). A 3D DSA Digital Imaging and Communications in Medicine (DICOM) image series was imported to reconstruct the target vessel from the C2 segment at the ICA to more than 5 mm at the distal end of the lesion. According to the results of mesh independence test (Table S1 in Appendix 1), each case had more than 1 million volume elements containing hexahedra and split-hexahedra elements and the mesh size is set to 0.16 mm. Blood was considered a Newtonian fluid without regard to energy or gravity equations. The Reynolds number in the intracranial circulation, even at the stenosis throat, was estimated to be below the turbulent transition threshold (<2,000), justifying the laminar assumption. While pulsatile flow is physiologically more accurate, steady-state simulations have been widely used and validated in evaluating pressure gradients across focal stenoses, as they capture the dominant hemodynamic effect and significantly reduce computational cost. Thus, the basic equation governing the flow was the incompressible Navier-Stokes equation, and the laminar simulation was carried out in the steady state via the Semi-Implicit Method for Pressure-Linked Equations (SIMPLE) algorithm and an in-house solver. The viscosity coefficient of the blood was constant at 0.0035 Pa·s, and the density was 1,056 kg/m³. The patient-specific mean flow rates of the boundaries were calculated by combining DSA images and a 3D model, referring to the thrombolysis in myocardial infarction (TIMI) frame count method commonly used in coronary stenosis research (17). Two operators calculated FF independently, and the third operator checks and corrects the abnormal results. The inter-rater reliability for FF was quantitatively assessed using the ICC. The average of two results was used for analysis. The operators responsible for calculating FF did not know the perfusion results beforehand. A detailed description of the FF calculation process is provided in the Appendix 2. A representative example of AccuFFicas and AccuMRP analysis is illustrated in Figure 1.

Figure 1 A 52-year-old female with MCA stenosis. (A) DSA image showing stenosis at the left-sided MCA. (B) Four round ROIs (red) on the affected cortical flow territories of the MCA on MRP maps at the same basal ganglia level with mirrored ROIs (green) on the contralateral sides. (C) The MRP map shows prolonged TTP and MTT and a slightly increased CBV on the ipsilateral side. (D) The FF result indicates that the distal FF of the stenosis is 0.39. CBV, cerebral blood volume; DSA, digital subtraction angiography; FF, fractional flow; MCA, middle cerebral artery; MRP, magnetic resonance perfusion; MTT, mean transit time; ROI, region of interest; TTP, time to peak.

Statistical analysis

The primary objective of this study was to explore the association between FF and hypoperfusion (defined as the combined stages I and II) in patients with anterior circulation ICAS. To eliminate the influence of examination intervals, sensitivity analysis was conducted on patients with MRP and DSA examination intervals ≤4 days. We hypothesized that FF would have a discriminative ability superior to chance [area under the curve (AUC) >0.5].

Cases with incomplete data were not included in the statistical analysis. Continuous variables were assessed via the Shapiro-Wilk test, and normally distributed data are presented as the means ± standard deviations (SDs), whereas nonnormally distributed data are presented as medians [interquartile ranges (IQRs)]. Categorical variables are expressed as counts (percentages). Spearman correlation analysis was applied to evaluate the associations between FF, DS, and MRP parameters. Group differences were compared via the Kruskal-Wallis test. Univariate and multivariate logistic regression analyses were performed to screen the independent risk factors associated with perfusion stage using enter logistic regression analysis.

The ability of DS and FF to differentiate hypoperfusion from normal perfusion was evaluated using receiver operating characteristic (ROC) curve analysis. The optimal cutoff value for FF was determined by maximizing the Youden index (J = sensitivity + specificity − 1). To account for overoptimism inherent in deriving and testing a cutoff from the same dataset, internal validation was performed using bootstrap resampling with 1,000 repetitions. The optimism-corrected AUC and performance metrics (sensitivity, specificity) are reported. Calibration of the FF-based logistic regression model was assessed by plotting the observed event rates against the predicted probabilities and by calculating the calibration slope and intercept. A slope of 1 and an intercept of 0 indicate perfect calibration. Furthermore, to evaluate the potential clinical utility of the FF model, decision curve analysis (DCA) was conducted to quantify the net benefit across a range of clinically reasonable probability thresholds. A confidence interval (CI) of 95% was used. A P value <0.05 was considered to indicate statistical significance. Statistical analysis was performed via R 4.3.0 software.

Results

Patient characteristics

Figure 2 shows the flow chart of this study. Among the 87 potentially eligible patients who underwent diffusion-perfusion MRI and had anterior circulation ICAS, 21 patients were excluded because of missing perfusion image series or poor-quality perfusion images, and 8 patients were excluded because of missing DSA image series, poor-quality DSA images, or occluded target vessels. The analysis demonstrated excellent consistency between the two independent operators, with an ICC of 0.97 (95% CI: 0.97–0.99) for rMTT, 0.97 (95% CI: 0.96–0.98) for rTTP, and 0.94 (95% CI: 0.91–0.97) for FF. Patient characteristics are summarized in Table 1. Fourteen (24%) patients had stenosis at the ICA, whereas 44 (76%) had stenosis at the MCA. Twenty-six (45%) patients were in the normal perfusion stage, 14 (24%) were in the hypoperfusion stage I, and 18 (31%) were in the hypoperfusion stage II. No adverse events occurred from performing MRI or DSA.

Figure 2 Flowchart of study patient enrollment. DSA, digital subtraction angiography; FF, fractional flow; ICAS, intracranial atherosclerotic stenosis; MRI, magnetic resonance imaging.

Associations between FF and MRP parameters

Correlation analysis between the FF and MRP parameters is shown in Figure 3A. FF was negatively correlated with rTTP (r=−0.41, P=0.002) and rMTT (r=−0.27, P=0.037). Figure 3B shows the distribution of FF in groups with different Tmax values. In groups with a larger Tmax, FF has a lower distribution interval.

Figure 3 Statistical analysis charts of FF and cerebral perfusion parameters. (A) Correlations between FF and MRP parameters. (B) FF distribution in groups with different Tmax values (left) and differences in FF among the three perfusion stages groups judged by MRP (right). Perfusion stage: 0 = normal perfusion, 1 = hypoperfusion stage I, 2 = hypoperfusion stage II. FF, fractional flow; MRP, magnetic resonance perfusion; rMTT, relative mean transit time; rTTP, relative time to peak; Tmax, time to maximum tissue residue function.

The FF was significantly different between normal perfusion patients and hypoperfusion patients (0.71±0.17 vs. 0.52±0.18, P<0.05), as shown in Table 1. The hypoperfusion patients had a lower DS, but the difference was not significant (0.66±0.17 vs. 0.54±0.13, P=0.087). Figure 3B shows the mean FF values in the normal perfusion (n=26), hypoperfusion stage I (n=14), and stage II (n=18) groups were 0.71, 0.52, and 0.52, respectively. However, there is a significant difference in the inspection time interval between the two groups [1 (IQR, 0, 2) vs. 3 (IQR, 1, 6), P<0.05]. Table S2 shows the parameter status of patients with a time interval of ≤4 days. In this subgroup, lesion located at MCA [16 (70%) vs. 20 (95%)], DS (0.53±0.14 vs. 0.64±0.18), and FF (0.73±0.16 vs. 0.49±0.16) showed significant difference between two groups.

Table 2 shows the results of univariate and multivariate logistic regression. Univariate analysis indicated that DS [odds ratio (OR) per 0.1 increment, 1.70; 95% CI: 1.17–2.46; P=0.006] was a significant risk factor, and FF (OR per 0.1 increment, 0.57; 95% CI: 0.41–0.80; P=0.001) was a protective significant factor. Multivariate analysis indicated FF (OR, 0.08; 95% CI: 0.01–0.80) was independent factor.

ROC curve analysis revealed that the AUC for differentiating hypoperfusion (stages I and II) from normal perfusion was 0.736 for DS (0.603–0.868) and 0.772 for FF (0.647–0.896), as shown in Figure 4A. The accuracy, sensitivity, and specificity were 75.9% (62.8–86.1%), 78.1% (60.0–90.7%) and 73.1% (52.2–88.4%), respectively, at the optimal cutoff value FF=0.63 (0.54–0.8). Youden Index J was 0.512 (0.212–0.666). After bootstrapping, the AUCs were 0.739 (0.597–0.850) for DS and 0.772 (0.632–0.876) for FF. The optimal cutoff value was FF=0.61, corresponding to an accuracy of 76.6% (58.8–92.1%), a sensitivity of 76.4% (56.5–93.3%), and a specificity of 78.7% (58.3–95.8%) Although FF has a higher AUC, the difference from DS is not significant (P=0.700 in DeLong test).

Figure 4 Statistical analysis charts of DS and FF on cerebral perfusion stages. (A) ROC curve for the ability of DS and FF to distinguish hypoperfusion from normal perfusion. (B) Calibration curve of DS and FF. (C) Decision curve analysis of DS and FF. AUC, area under the curve; CI, confidence interval; DS, diameter stenosis; FF, fractional flow; ROC, receiver operating characteristic.

The calibration plots for DS and FF were generated in Figure 4B. The slope and the intercept of DS was 0.954 and 0.026, and of FF was 1.049 and 0.027. The calibration performance of the two is similar. Figure 4C displayed DCA curves, indicating that DS was effective between 0.21–0.82, and FF was effective between 0.25–0.92.

Associations between treatment and parameters

Twelve patients had complete postoperative images, and MRP parameters and FF calculations were performed, as shown in Table 3. The FF values significantly increased after treatment (0.49 vs. 0.91, P<0.05). Eight patients showed improvement in perfusion, 2 patients maintained normal perfusion, and 2 patients maintained stage II hypoperfusion. Among the 8 patients whose perfusion improved, 6 patients had significantly smaller brain tissue volumes with Tmax >6 s and Tmax >4 s, whereas 2 patients had stable volumes. The FF values of these 8 patients were significantly increased (0.49 vs. 0.89, P<0.05).

Representative case results

Figure 5 shows 4 representative cases. Case 1 was a 76-year-old male with stenosis at right M1 segment and hypoperfusion stage 1. His DS was 58%, belonging to moderate stenosis, but FF was 0.62, indicating insufficient blood supply. Case 2 is a 71-year-old male with stenosis at right M1 segment and normal perfusion. His DS was 73%, but FF was 0.76, indicating acceptable blood supply. Case 3 is a 57-year-old male with stenosis at left M2 segment and hypoperfusion stage 2. His DS and FF were 83% and 0.46, both indicating insufficient blood supply. Case 4 is a 63-year-old male had with stenosis at left M1 segment and hypoperfusion stage 1. His DS was 67%, close to severe stenosis, but FF was 0.45, indicating insufficient blood supply.

Figure 5 Four representative cases with their perfusion analysis results, DSA images and FF results. CBF, cerebral blood flow; CBV, cerebral blood volume; DSA, digital subtraction angiography; FF, fractional flow; MTT, mean transit time; TTP, time to peak.

Discussion

This study explored the correlation between cerebral perfusion and functional ICAS assessment through quantitative analysis of MRP parameters and angio-based FF in anterior circulation ICAS patients. Negative correlations were found between FF and rTTP and rMTT on the hemisphere of the lesion, and a trend of lower FF distribution was observed in groups with higher Tmax values. There were significant differences in FF among patients with different hypoperfusion stages. The optimal FF cutoff for distinguishing hypoperfusion from normal perfusion was 0.63 in the original cohort, which remained robust at 0.61 after bootstrap validation, suggesting that patients with FF values below this threshold have a greater likelihood of hypoperfusion. Although the AUC of FF was not statistically superior to that of DS in differentiating hypoperfusion, DCA indicated that FF provided a broader range of clinical net benefit, supporting its potential utility in individualized decision-making. Multivariate logistic regression confirmed FF as an independent predictor of hypoperfusion, reinforcing its role beyond anatomical stenosis in assessing hemodynamic compromise. Furthermore, the significant post-treatment increase in FF, coupled with reductions in Tmax volumes in most patients, underscores the potential of FF as a real-time intraoperative indicator of hemodynamic improvement following revascularization. This study suggests that FF is associated with perfusion in patients with anterior circulation ICAS. Using DSA to quantify FF before and after stenting during surgery can predict improvements in cerebral perfusion in patients.

MRP imaging evaluation of ischemia

MRP can be used to evaluate ischemia effectively and compensate for brain tissue damage. The quantified hemodynamic parameters of the responsible blood vessel supply area, such as relative CBF (rCBF), relative CBV (rCBV), rMTT, and rTTP, provide reliable tools for precise clinical evaluation (18). The combination of these cerebral perfusion indicators can provide detailed perfusion information, such as hypoperfusion, collateral circulation establishment, and blood flow reperfusion or overperfusion, to guide the clinical decision of stent implantation for intracranial artery stenosis. In different areas and at different stages of cerebral ischemia, there are four main manifestations: (I) no perfusion or hypoperfusion: the rMTT is prolonged, the rCBV is reduced, and the rCBF is significantly reduced; (II) collateral circulation establishment: the rMTT is prolonged, the rCBV is increased or normal; (III) blood flow reperfusion: the rMTT is shortened or normal, the rCBV is increased, and the rCBF is slightly increased or normal; (IV) overperfusion: the rCBV and rCBF are significantly increased. The image processing software automatically completed and calculated the above perfusion parameters. According to the calculation results of AccuMRP, 14 (24%) patients had stage I hypoperfusion, and 18 (31%) patients had stage II hypoperfusion in this study.

Decreased cerebral perfusion in the cerebral hemisphere caused by ICAS can cause hemodynamic changes leading to TIA or ischemic stroke symptoms. The change in the intracranial perfusion level not only depends on the degree of cerebrovascular stenosis but is also is related to the intracranial collateral circulation and cerebral vascular reserve capacity. Perfusion weighted imaging (PWI) can be used to observe cerebral ischemia, but it does not always match the corresponding degree of vascular stenosis. Owing to good collateral circulation, some patients with moderate to severe cerebrovascular stenosis do not experience cerebral ischemia symptoms or obvious infarction, whereas some patients with mild stenosis exhibit obvious infarction (19). When anterior cerebrovascular stenosis or occlusion occurs, the Willis circle mostly does not participate in collateral compensation but relies mainly on the communication branch between cortical branches of cerebral surface blood vessels to compensate, and the endings of cerebral cortical branches form a diffuse vascular network in the pia meninges. However, anastomosis of the pia meninges had a limited compensatory effect on the distribution area of the MCA, so it was easy to cause cerebral ischemia symptoms due to insufficient collateral compensation.

Hemodynamic evaluation of ICAS

In the past, the DS was usually used as an indicator to evaluate the severity of stenosis. With the gradual deepening of research on ICAS and ischemia, functional evaluation has gradually attracted increasing attention. The SAMMPRIS and VISSIT trials revealed that the incidence of endpoint events within 30 days and 1 year of stent implantation combined with intensified drug therapy was significantly greater than that of drug therapy alone, which raises questions about the interventional treatment of ICAS (20,21). One potential reason is that the screening criteria for patients have not successfully identified patients who will truly benefit from interventional therapy. The degree of stenosis can reflect only the morphology of large blood vessels and cannot reflect the compensatory effect of collateral circulation blood supply, nor can it reflect individual differences in the demand of brain tissue for blood supply.

FF considers not only the obstructive effect caused by vascular stenosis but also the blood flow state. For patients with good collateral circulation or low blood supply demand, the blood flow of large blood vessels may be low, and they will not show high flow velocity at the stenosis site or be significantly affected by resistance, thus exhibiting a lower pressure drop and higher FF. In this study, patients with hypoperfusion stage I and patients with hypoperfusion stage II had significantly lower FF values than patients with normal perfusion. In contrast, there was no difference in the degree of stenosis between these groups of patients. The trend of correlation between FF and different perfusion parameters also provides evidence. FF is correlated mainly with time-dependent parameters such as rTTP, rMTT, and Tmax, while its correlation with rCBF and rCBV is poor, which is consistent with the research results of Suo et al. (22). A possible explanation is that time-dependent parameters are more related to blood flow, whereas CBVs and CBFs are related to brain tissue and may differ in long-term ischemic environments because of self-regulation in humans. There are other methods used to obtain cerebral vascular hemodynamic information. Transcranial Doppler (TCD) can quickly and noninvasively measure multiple intracranial blood vessels and obtain systolic and diastolic blood flow velocities. Xu et al. reported that the pulsatility index from TCD has a strong correlation with the FF from pressure wires (23). 4D flow MRI technology can obtain dynamic blood flow information and reflect complex flow states. Wu et al. successfully observed changes in the blood flow rate and peak velocity caused by ICAS via 4D flow MRI (24). However, neither of these detection methods can achieve real-time intraoperative evaluation. In contrast, FF based on DSA can be obtained during surgery, enabling real-time evaluation of changes in hemodynamics caused by stenting, and has unique application potential.

A reliable threshold is crucial for FF to become the diagnostic criterion for ICAS. However, there is currently no recognized threshold for FF used in the diagnosis and treatment of ICAS. Wang et al. reported a threshold value of the distal pressure/proximal pressure (Pd/Pa) =0.67 with a Tmax >6 s in the MRP results of 18 symptomatic ICAS patients (9). Suo et al. obtained a threshold value of the quantitative flow ratio (QFR) =0.82 on the basis of the CTP hypoperfusion stage of 62 patients with symptomatic anterior circulation ICAS with previous ischemic stroke or TIA (22). Li et al. obtained Pa/Pd =0.81 or Pa − Pd =8 mmHg on the basis of the rCBF from ASL results of 25 TIA or minor stroke patients with contralateral >50% stenosis (25). In our study, a threshold of 0.63 was reported. There are multiple reasons for the difference in threshold values. (I) The calculation methods for FFs or similar parameters are different. (II) The inclusion and exclusion criteria for establishing cohorts are different. (III) The definitions of positive patients are different. Therefore, a widely recognized threshold requires more rigorous and large-scale research to establish higher-level clinical evidence.

Clinical contributions

The findings of this study suggest that FF provides hemodynamic information that is distinct from and complementary to the anatomical assessment of DS. While DS remains the primary criterion for patient selection in current guidelines, which caution against routine stenting for most symptomatic ICAS patients, FF holds promise as an investigational adjunct for hemodynamic assessment. The ability to compute FF intraoperatively could, in the future, provide real-time feedback on physiological improvement following revascularization. However, its role in guiding management decisions must be validated in prospective studies before it can be considered for integration into clinical practice guidelines.

Limitations

This study has several limitations. First, this study was conducted in a single center with a small retrospective sample size. Although statistical differences were observed in the FF of patients with different perfusion stages in this study, the correlation between FF and cerebral perfusion parameters was low. This may be related to factors such as the duration of the patient’s illness, disease progression, and medication use. However, due to the limitation of the number of cases, this study did not investigate these factors. A larger sample size is needed to establish more convincing results.

Second, due to the requirements for the integrity and quality of MRP and DSA images, we only collected images of some patients after endovascular treatment. This leads to a lack of analysis between cerebral perfusion parameters, FF and clinical manifestations of patients after endovascular treatment. The follow-up results of the patients were not included in the analysis. In the following research, we will collect more complete data based on the data collection standards established in this study to conduct more comprehensive analysis.

Third, owing to the scope of application of the software AccuFFicas, this study was limited to ICAS patients with moderate to severe stenosis of the intracranial anterior circulation. After demonstrating the clinical value of FF in the diagnosis and treatment of ICAS, the scope of software application will be expanded to explore the significance of FF in different types of ICAS patients.

Conclusions

There is a correlation between cerebral perfusion and FF in patients with anterior circulation ICAS. A low FF may indicate a high Tmax, high TTP and poor hypoperfusion. The FF has the potential to serve as a hemodynamic evaluation index for ICAS patients.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-779/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-779/dss

Funding: This work was supported by the Basic Public Welfare Project of the Natural Science Foundation of Zhejiang Province (No. GF22H096743); Zhejiang Provincial Key Research and Development Plan (No. 2024C03095); Hangzhou Leading Innovation and Entrepreneurship Team Project (No. TD2022007); Special Support Program for High Level Talents of Zhejiang Province (No. 2022R52038); Medical Health Science and Technology Project of Zhejiang Provincial Health Commission (No. WKJ-ZJ-2340); Basic Public Welfare Project of the Natural Science Foundation of Zhejiang Province (No. LGF22H090017); Zhejiang Provincial Key Research and Development Plan (No. 2025C02151); and the Key Project of Zhejiang Administration of Traditional Chinese Medicine (No. 2023009022).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-779/coif). J.W. reports that this study was funded by the Basic Public Welfare Project of the Natural Science Foundation of Zhejiang Province (No. LGF22H090017), Zhejiang Provincial Key Research and Development Plan (No. 2024C03095), Zhejiang Provincial Key Research and Development Plan (No. 2025C02151), and the Key Project of Zhejiang Administration of Traditional Chinese Medicine (No. 2023009022). R.Z. reports that this study was funded by Zhejiang Provincial Key Research and Development Plan (No. 2024C03095), Hangzhou Leading Innovation and Entrepreneurship Team Project (No. TD2022007), Medical Health Science and Technology Project of Zhejiang Provincial Health Commission (No. WKJ-ZJ-2340), and Zhejiang Provincial Key Research and Development Plan (No. 2025C02151). R.Z. is an employee of ArteryFlow Technology Co., Ltd. Y.H. reports that this study was funded by Zhejiang Provincial Key Research and Development Plan (No. 2024C03095) and Hangzhou Leading Innovation and Entrepreneurship Team Project (No. TD2022007). Y.H. is an employee of ArteryFlow Technology Co., Ltd. J.X. reports that this study was funded by the Basic Public Welfare Project of the Natural Science Foundation of Zhejiang Province (No. GF22H096743), Zhejiang Provincial Key Research and Development Plan (No. 2024C03095), Hangzhou Leading Innovation and Entrepreneurship Team Project (No. TD2022007), Medical Health Science and Technology Project of Zhejiang Provincial Health Commission (No. WKJ-ZJ-2340), Basic Public Welfare Project of the Natural Science Foundation of Zhejiang Province (No. LGF22H090017), and Zhejiang Provincial Key Research and Development Plan (No. 2025C02151). J.X. is an employee of ArteryFlow Technology Co., Ltd. S.W. reports that this study was funded by the Basic Public Welfare Project of the Natural Science Foundation of Zhejiang Province (No. GF22H096743), Zhejiang Provincial Key Research and Development Plan (No. 2024C03095), Special Support Program for High Level Talents of Zhejiang Province (No. 2022R52038), Medical Health Science and Technology Project of Zhejiang Provincial Health Commission (No. WKJ-ZJ-2340), Basic Public Welfare Project of the Natural Science Foundation of Zhejiang Province (No. LGF22H090017), Zhejiang Provincial Key Research and Development Plan (No. 2025C02151), and the Key Project of Zhejiang Administration of Traditional Chinese Medicine (No. 2023009022). The other author has 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 Medical Ethics Committee of Zhejiang Hospital (approval No. 2022-80K) and individual consent for this retrospective analysis was waived.

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. Gutierrez J, Turan TN, Hoh BL, Chimowitz MI. Intracranial atherosclerotic stenosis: risk factors, diagnosis, and treatment. Lancet Neurol 2022;21:355-68. [Crossref] [PubMed]
2. Miao Z, Liebeskind DS, Lo W, Liu L, Pu Y, Leng X, Song L, Xu X, Jia B, Gao F, Mo D, Sun X, Liu L, Ma N, Wang B, Wang Y, Wang Y. Fractional Flow Assessment for the Evaluation of Intracranial Atherosclerosis: A Feasibility Study. Interv Neurol 2016;5:65-75. [Crossref] [PubMed]
3. Wabnitz AM, Derdeyn CP, Fiorella DJ, Lynn MJ, Cotsonis GA, Liebeskind DS, Waters MF, Lutsep H, López-Cancio E, Turan TN, Montgomery J, Janis LS, Lane B, Chimowitz MI. SAMMPRIS Investigators. Hemodynamic Markers in the Anterior Circulation as Predictors of Recurrent Stroke in Patients With Intracranial Stenosis. Stroke 2019;50:143-7. [Crossref] [PubMed]
4. de Havenon A, Khatri P, Prabhakaran S, Yeatts SD, Peterson C, Sacchetti D, Alexander M, Cutting S, Mac Grory B, Furie K, Liebeskind DS, Yaghi S. Hypoperfusion Distal to Anterior Circulation Intracranial Atherosclerosis is Associated with Recurrent Stroke. J Neuroimaging 2020;30:468-70. [Crossref] [PubMed]
5. De Bruyne B, Pijls NH, Kalesan B, Barbato E, Tonino PA, Piroth Z, et al. Fractional flow reserve-guided PCI versus medical therapy in stable coronary disease. N Engl J Med 2012;367:991-1001. [Crossref] [PubMed]
6. Han YF, Liu WH, Chen XL, Xiong YY, Yin Q, Xu GL, Zhu WS, Zhang RL, Ma MM, Li M, Dai QL, Sun W, Liu DZ, Duan LH, Liu XF. Severity assessment of intracranial large artery stenosis by pressure gradient measurements: A feasibility study. Catheter Cardiovasc Interv 2016;88:255-61. [Crossref] [PubMed]
7. Liu J, Yan Z, Pu Y, Shiu WS, Wu J, Chen R, Leng X, Qin H, Liu X, Jia B, Song L, Wang Y, Miao Z, Wang Y, Liu L, Cai XC. Functional assessment of cerebral artery stenosis: A pilot study based on computational fluid dynamics. J Cereb Blood Flow Metab 2017;37:2567-76. [Crossref] [PubMed]
8. Leng X, Lan L, Ip HL, Abrigo J, Scalzo F, Liu H, et al. Hemodynamics and stroke risk in intracranial atherosclerotic disease. Ann Neurol 2019;85:752-64. [Crossref] [PubMed]
9. Wang M, Leng X, Mao B, Zou R, Lin D, Gao Y, Wang N, Lu Y, Fiehler J, Siddiqui AH, Wu J, Xiang J, Wan S. Functional evaluation of intracranial atherosclerotic stenosis by pressure ratio measurements. Heliyon 2023;9:e13527. [Crossref] [PubMed]
10. Yang P, Wan S, Wang J, Hu Y, Ma N, Wang X, Zhang Y, Zhang L, Zhu X, Shen F, Zheng Q, Wang M, Leng X, Fiehler J, Siddiqui AH, Miao Z, Xiang J, Liu J. Hemodynamic assessment for intracranial atherosclerosis from angiographic images: a clinical validation study. J Neurointerv Surg 2024;16:204-8. [Crossref] [PubMed]
11. Li J, Wang L, Zhang Y, Zhu X, Zhang X, Hua W, Chen R, Liu H, Yin W, Xiang J, Xing P, Li Z, Zhao R, Zhang Y, Liu J, Dai D, Zhang L, Yang P. Hemodynamic evaluation of symptomatic and asymptomatic intracranial atherosclerotic stenosis using cerebral angiographic images: an exploratory study. J Neurointerv Surg 2025;jnis-2024-022455. [Crossref] [PubMed]
12. Welker K, Boxerman J, Kalnin A, Kaufmann T, Shiroishi M, Wintermark M. American Society of Functional Neuroradiology MR Perfusion Standards and Practice Subcommittee of the ASFNR Clinical Practice Committee. ASFNR recommendations for clinical performance of MR dynamic susceptibility contrast perfusion imaging of the brain. AJNR Am J Neuroradiol 2015;36:E41-51. [Crossref] [PubMed]
13. Calamante F. Arterial input function in perfusion MRI: a comprehensive review. Prog Nucl Magn Reson Spectrosc 2013;74:1-32. [Crossref] [PubMed]
14. Damasio H. A computed tomographic guide to the identification of cerebral vascular territories. Arch Neurol 1983;40:138-42. [Crossref] [PubMed]
15. Lan L, Leng X, Abrigo J, Fang H, Ip VH, Soo YO, Leung TW, Yu SC, Wong LK. Diminished Signal Intensities Distal to Intracranial Arterial Stenosis on Time-of-Flight MR Angiography Might Indicate Delayed Cerebral Perfusion. Cerebrovasc Dis 2016;42:232-9. [Crossref] [PubMed]
16. Huang CC, Chen YH, Lin MS, Lin CH, Li HY, Chiu MJ, Chao CC, Wu YW, Chen YF, Lee JK, Wang MJ, Chen MF, Kao HL. Association of the recovery of objective abnormal cerebral perfusion with neurocognitive improvement after carotid revascularization. J Am Coll Cardiol 2013;61:2503-9. [Crossref] [PubMed]
17. Gibson CM, Cannon CP, Daley WL, Dodge JT Jr, Alexander B Jr, Marble SJ, McCabe CH, Raymond L, Fortin T, Poole WK, Braunwald E. TIMI frame count: a quantitative method of assessing coronary artery flow. Circulation 1996;93:879-88. [Crossref] [PubMed]
18. Vilela P, Rowley HA. Brain ischemia: CT and MRI techniques in acute ischemic stroke. Eur J Radiol 2017;96:162-72. [Crossref] [PubMed]
19. Lau AY, Wong EH, Wong A, Mok VC, Leung TW, Wong KS. Significance of good collateral compensation in symptomatic intracranial atherosclerosis. Cerebrovasc Dis 2012;33:517-24. [Crossref] [PubMed]
20. Chimowitz MI, Lynn MJ, Derdeyn CP, Turan TN, Fiorella D, Lane BF, et al. Stenting versus aggressive medical therapy for intracranial arterial stenosis. N Engl J Med 2011;365:993-1003. [Crossref] [PubMed]
21. Zaidat OO, Fitzsimmons BF, Woodward BK, Wang Z, Killer-Oberpfalzer M, Wakhloo A, Gupta R, Kirshner H, Megerian JT, Lesko J, Pitzer P, Ramos J, Castonguay AC, Barnwell S, Smith WS, Gress DR. Effect of a balloon-expandable intracranial stent vs medical therapy on risk of stroke in patients with symptomatic intracranial stenosis: the VISSIT randomized clinical trial. JAMA 2015;313:1240-8. [Crossref] [PubMed]
22. Suo S, Zhao Z, Zhao H, Zhang J, Zhao B, Xu J, Zhou Y, Tu S. Cerebral hemodynamics in symptomatic anterior circulation intracranial stenosis measured by angiography-based quantitative flow ratio: association with CT perfusion. Eur Radiol 2023;33:5687-97. [Crossref] [PubMed]
23. Xu X. Raynald, Li X, Li R, Yang H, Zhao X, Miao Z, Mo D. New evidence for fractional pressure ratio prediction by pulsatility index from transcranial Doppler in patients with symptomatic cerebrovascular stenosis disease. Quant Imaging Med Surg 2024;14:264-72. [Crossref] [PubMed]
24. Wu C, Schnell S, Vakil P, Honarmand AR, Ansari SA, Carr J, Markl M, Prabhakaran S. In Vivo Assessment of the Impact of Regional Intracranial Atherosclerotic Lesions on Brain Arterial 3D Hemodynamics. AJNR Am J Neuroradiol 2017;38:515-22. [Crossref] [PubMed]
25. Li L, Yang B, Dmytriw AA, Li Y, Gong H, Bai X, Zhang C, Chen J, Dong J, Wang Y, Gao P, Wang T, Luo J, Xu X, Feng Y, Zhang X, Yang R, Ma Y, Jiao L. Correlations between intravascular pressure gradients and cerebral blood flow in patients with symptomatic, medically refractory, anterior circulation artery stenosis: an exploratory study. J Neurointerv Surg 2024;16:608-14. [Crossref] [PubMed]