Waveform Analysis of Superficial Temporal Artery-Middle Cerebral Artery Bypass Graft in Revascularization Surgery for Moyamoya Disease

Superficial temporal artery (STA)-middle cerebral artery (MCA) bypass is an effective revascularization technique to prevent future strokes in moyamoya disease (MMD) [1‒5]. Postoperative hyperperfusion syndrome (PHS) is a well-known postoperative complication that can cause transient focal neurological deficits, vasogenic edema, and delayed intracranial hemorrhage [6‒9]. PHS remains a major clinical challenge in terms of the prescriptive factor of clinical outcomes after revascularization surgery. Postoperative radiological hyperperfusion (PRH), characterized by a focal increase in cerebral blood flow (CBF), is a useful predictor of PHS [6‒9]. Appropriate blood pressure control is crucial in reducing the risk of PHS after the detection of PRH by postoperative single-photon emission computed tomography (SPECT) [7]. However, PHS can occur immediately after or even during surgery [9], necessitating an earlier detection method. Intraoperative measurement of bypass blood flow using a flowmeter is a potential alternative for earlier detection of PHS [10‒12]. In addition to blood flow values, this device can provide waveforms containing information about peripheral vascular resistance (PVR) [13]. In MMD, chronic hypoperfusion induces vasodilation of the ischemic territories through the vascular autoregulatory system, leading to decreased PVR, one of the underlying mechanisms of PHS [9]. Therefore, we hypothesized that the measurement of PVR by intraoperative waveform analysis would be useful for the early detection of PRH.

We reviewed a consecutive series of 52 patients with MMD who underwent STA-MCA bypass and encephalo-duro-myo-synangiosis (EDMS) between January 2021 and March 2022. The inclusion criteria were as follows: (1) patients who underwent intraoperative blood flow measurement and (2) patients who were assessed by SPECT before and after surgery. The diagnosis of MMD followed the previous literature [14]. Surgical indication was decided on the basis of a comprehensive assessment, which considered the following: the presence of ischemic symptoms or a history of posterior hemorrhage [15], evident hemodynamic compromise on SPECT, independent activities of daily living, and the absence of cerebral infarction that exceeded the vascular territory of a major branch of the MCA.

All patients underwent STA-MCA anastomosis with EDMS under general anesthesia by a single surgeon (H.E.). Craniotomy was performed around the sylvian fissure end. The stump of the STA, either the frontal or parietal branch, was anastomosed to the M4 segment of the MCA, followed by EDMS. The patency of the STA-MCA anastomosis was confirmed by intraoperative indocyanine green video angiography and Doppler ultrasonography [7, 16].

Intraoperative measurement of bypass flow was conducted using a flexible perivascular flow probe (HQN-1.5 MB-1.5 mm, HQN-2 MB-2.0 mm; Transonic Systems Inc., Ithaca, NY, USA) and a transit time flowmeter (HT353, Transonic System, Ithaca, NY, USA). This device provided quantitative blood flow data, including maximum flow (mL/min), minimum flow (mL/min), mean flow (mL/min), and a waveform of the bypass graft. Using these quantitative blood flow data, the pulsatile index (PI), an indicator of PVR, was calculated using the following equation [17]:

$PI=maximumflow−minimumflow/meanflow$

In addition, we performed original waveform analysis to assess PVR based on the idea that waveform changes reflect PVR [13]. The averaged blood flow waveform was obtained from 10-s measurements (MATLAB 2019a, The MathWorks Inc., Natick, MA, USA). A peak of the averaged waveform was set as the reference point for the time course. The ratio of the time from peak to 50% decay and to 100% decay (RT50) was calculated using the following equation:

$RT50=timefrompeakto50%decay/timefrompeakto100%decay$

An example of the RT50 calculated from the averaged waveform is displayed in Figure 1.

Pre- and postoperative CBF was assessed using the autoradiographic method with ¹²³I-iodoamphetamine (IMP)-SPECT (Infinia Hawkeye 4; GE Healthcare, Tokyo, Japan) [18]. Preoperative CBF measurements were typically obtained within 2 months prior to surgery. We quantitatively measured the CBF by manual definition of a region of interest (ROI) with a diameter of 1 cm at the exact vascular territory supplied by the bypass and the ipsilateral cerebellum [18]. The exact vascular territory supplied by the bypass was determined by comparing the original axial slice on postoperative magnetic resonance arthrography (MRA) and ¹²³I-IMP SPECT. Both the pre- and postoperative CBF values were adjusted by the CBF value of the ipsilateral cerebellar hemisphere to calculate the postoperative/preoperative CBF ratio within the ROI [18]. PRH was defined as a greater than 120% increase in the postoperative/preoperative CBF ratio. Although the definition of RHP varies among previous studies [7‒9, 19], we defined the threshold of RHP as a 120% increase to detect mild hyperperfusion.

We performed a comparative analysis between the non-PRH and PRH groups using the RT50, PI, maximum flow, mean flow, minimum flow, preoperative CBF, temporary occlusion time, diameter of the recipient artery, and diameter of STA, which are reported potential risk factors for PHS [10‒12, 20‒22]. The Wilcoxon rank-sum test or Peason’s χ² test was applied for the comparisons between the two groups, and a linear regression analysis was conducted to examine the relationship between the postoperative/preoperative CBF ratio and the aforementioned parameters. Receiver operating characteristic (ROC) analysis was used to assess the accuracy of the RT50 and preoperative normalized CBF for the detection of PRH [9, 23]. Statistical analyses were performed using JMP Pro 12 (SAS Institute, North Carolina, USA), and significance was defined at a p value of 0.05.

Twenty-seven patients met the inclusion criteria. Fourteen (52%) of the 27 patients exhibited PRH; the median age at surgery was 42 (range: 17–66), 11 (79%) patients were female, 6 (43%) had left-side surgery, 2 (14%) had hemorrhagic onset, and 12 (86%) had ischemic onset. The clinical backgrounds of the 13 patients in the non-PRH group were as follows: the median age was 35 years (range: 24–65), 10 (77%) patients were female, 5 (38%) had left-side surgery, 4 (31%) had hemorrhage onset, and 8 (62%) had ischemic onset. There were no statistical differences in these background factors between the two groups. Four out of 27 patients (15%) developed PHS, and all of them were included in the PRH group (31%). In 4 patients with PHS, 1 patient had intracerebral hemorrhage, and the other three manifested transient neurological deficits. None of these events affected clinical outcomes.

Next, we compared the intraoperative and neuroradiological data between the PRH and non-PRH groups (Fig. 2a–i). Among the variables compared, only the RT50 was significantly different between the two groups; the PRH group demonstrated a higher RT50 than the non-PRH group (p = 0.047) (Fig. 2a). In the linear regression analysis investigating the correlation between the RT50 and other variables, a significant correlation was found only between PI and RT50 (y = 1.26–1.68x, R-square: 0.35, p = 0.001) (online suppl. Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000538548). In the ROC analysis to predict PRH, the area under the curve (AUC) of RT50 was 0.750, with a cutoff value of 0.255, a sensitivity of 0.928, and a specificity of 0.500, whereas the AUC of preoperative CBF was 0.565, with a cutoff value of 0.644, a sensitivity of 0.29, and a specificity of 0.00 (Fig. 3).

A 45-year-old male presented with transient left upper limb weakness and was diagnosed with MMD through neuroradiological assessment, including digital subtraction angiography, which revealed Suzuki stage 3 MMD in the right hemisphere (Fig. 4a). MRI revealed scattered old infarctions, and SPECT showed a significant CBF decrease in the right MCA area (not shown). Because the patient was diagnosed with symptomatic MMD with impaired cerebral hemodynamics, we performed STA-MCA anastomosis in the right hemisphere (Fig. 4b). Intraoperative flowmeter analysis for the STA graft showed a maximum flow of 165.7 mL/s, a mean flow of 118.8 mL/s, a minimum flow of 85.0 mL/s, a PI of 0.68, and an RT50 of 0.27 (Fig. 1a). The patient was awake after surgery without any neurological deterioration. Postoperative MRA showed good patency of the bypass graft (Fig. 4c). SPECT the day after surgery revealed a 233.9% CBF increase at the site of anastomosis, indicating RHP (Fig. 4d, right). Based on the results of waveform analysis and SPECT, we managed this patient with strict blood pressure control to prevent the development of HPS. The patient remained symptom-free and was discharged 8 days after surgery. Figure 1a is the averaged waveform of this patient, and Figure 1b is the averaged waveform of a non-RHP patient (RT50: 0.14, 113.4% CBF increase).

This study is the first to demonstrate the utility of waveform analysis of the bypass graft for predicting PRH in MMD. The RT50, the original value calculated from the waveform analysis as the index for PVR, was found to be significantly higher in the PRH group. ROC analysis showed that the RT50 can be a more useful factor than preoperative CBF status for predicting PRH.

A blood flow wave is observed as the sum of a forward wave from the heart and a backward reflected wave. In general, the amplitude and speed of the backward wave are positively correlated with impedance, such as PVR, and the backward wave causes flow to change in the opposite direction [13]. Hence, the amplitude and speed of the backward wave become lower and slower under the condition of decreasing PVR. Consequently, the amplitude of the observed flow wave rises, and the downhill slope of the observed wave is gradual, increasing the RT50. The PI can be used to evaluate the PVR, and an increased PI can indicate an increased PVR in the distal vascular territories [17]. We confirmed a strong negative correlation between RT50 and PI, which suggests that a higher RT50 indicates decreased PVR. Therefore, the higher RT50 in the PRH group was thought to reflect decreased PVR in the PRH group compared with that in the non-PRH group. In contrast, there were no significant differences in PI between the two groups. It is important to note that factors such as age and sex can influence the PI [24], affecting the results of this study.

Contrary to previous reports, there were no significant differences in blood flow value, preoperative CBF, and the diameter of the bypass graft between the two groups in our study. Differences in sample size, timing of SPECT, definition of PRH, and surgical methodology are plausible reasons for this. Blood flow values are easily influenced by factors such as blood pressure, cardiac output, and the diameter of the bypass graft. In contrast, waveform analysis is less susceptible to these factors because both forward and backward waves are equally affected, canceling out the influence of these factors.

One of the limitations of this study is the small sample size collected from a single center. This may have affected the non-significant correlation between the RT50 and the postoperative/preoperative CBF ratio (online suppl. Figure), whereas the RT50 was significantly higher in the PRH group. The limited sample size also prevented reliable multivariate analysis, emphasizing the necessity for further investigations with larger cohorts. Moreover, we did not compare the waveform analysis and cerebrovascular reserve capacity (CVR), which are also associated with PVR [25]. The ROC analysis of the RT50, which indicated high sensitivity and low specificity, facilitates the creation of multimodal models, including waveform analysis and CVR, in the future.

Waveform analysis is a noninvasive and real-time tool that can be useful for the early detection of PRH in patients with MMD.

The present study conforms to the guidelines issued in the Declaration of Helsinki. This study was approved by the Ethics Committees of Kohnan Hospital (Approval Number: 2020-0520-03). Written informed consent was obtained from all the participants. We have also received written informed consent for the publication of identifying details and accompanying images for representative case.

The authors have no conflicts of interest to declare.

JSPS KAKENHI Grant No. 23K08491 supported English proofreading fee.

R.K. and H.E.: conception and design. R.K., H.E., R.T., A.K., and H.U.: acquisition of data. R.K., H.E., and R.T.: analysis and interpretation. R.K.: drafting. H.E.: critical revision of the article. H.E.: study supervision.

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants. Further inquiries can be directed to the corresponding author.