Simultaneous PET-MRI imaging of cerebral blood flow and glucose metabolism in the symptomatic unilateral internal carotid artery/middle cerebral artery steno-occlusive disease

Ischaemic cerebrovascular disease is a chronic, occlusive cerebrovascular disease of unknown aetiology that is characterized by progressive steno-occlusive changes at the terminal portion of the internal carotid artery (ICA) or middle cerebral artery (MCA). Symptomatic patients are treated by various methods, including aggressive medical management, intracranial angioplasty, stenting, and cerebrovascular bypass surgery [1‐3]. The quantification of cerebral blood flow (CBF) is important and meaningful in ischaemic stroke [4]. Previous studies reported that while the hypoperfusion area (< 25 mL/100 g/min) was not primarily damaged by a reduction in blood supply, affected patients should undergo bypass surgery to improve CBF [5, 6]. Positron emission tomography (PET) imaging with ¹⁵O-H₂O is considered the gold standard for measuring CBF [7, 8]. However, the standard protocol is not widely used in the clinic due to its invasiveness. An alternative option, arterial spin labelling (ASL), is a promising non-invasive method that can be used to measure CBF that is also convenient to apply [9, 10]. Previous researchers who obtained CBF values using both ¹⁵O-H₂O PET and ASL-MRI found that there was a significant correlation between the two methods when used in healthy subjects under identical physiological conditions [11, 12]. In addition, decreased CBF (< 35.0 mL/100 g/min) was observed on the affected side, indicating chronic hypoperfusion [13, 14].

However, CBF alone is not sufficient to predict tissue fate. Normally, because glucose is the primary source of energy for brain cells, glucose metabolism also plays an important role in characterizing functional metabolism [15, 16]. ¹⁸Fluoro-2-deoxy-D-glucose (¹⁸F-FDG) is widely used in different diseases as a common tracer and in the measurement of brain glucose metabolism. Recently, there has been tremendous interest in evaluating the potential of ¹⁸F-FDG in cerebral ischaemia [17‐20]. Deng et al. also observed that ¹⁸F-FDG uptake was higher in the cerebral infarction area among drug treatment groups in a rat model of cerebral ischaemia reperfusion injury evaluated after 7 days [18]. Many previous studies have investigated perfusion and glucose metabolism using separate PET and MR procedures or integrated PET/MR in patients with cancer, Alzheimer’s disease, epilepsy, and other disorders [21‐23]. MR images and PET images can be obtained at different time points and fused for clinical and research purposes. However, various physiological processes may change between these imaging sessions. The optimal approach is to simultaneously acquire data using different modalities. Hybrid PET/MR solves this problem by offering complementary information obtained via conventional MR imaging, which has high spatial resolution and tissue contrast, and PET, which provides physiological information.

Superficial temporal artery MCA (STA-MCA) bypass is an effective method of surgical revascularization used in the treatment of ischaemic cerebral artery disease. The most important goal of surgical revascularization is to improve CBF as assessed by imaging [24]. Effective revascularization strategies improve haemodynamic parameters and reduce stroke occurrence [13, 25]. Cerebral FDG metabolism is also one of the factors used to evaluate cerebral metabolism [26]. ¹⁸F-FDG PET imaging shows promise as an indicator of the effectiveness of therapies in cerebral ischaemia [19]. Yoshida, K. et al. combined measurements of CBF and cerebral glucose metabolism to evaluate cognitive changes after carotid endarterectomy [27], and Yu Z et al. used ¹⁸F-FDG PET/CT to evaluated 7 patients who underwent STA-MCA bypass surgery [28]. However, few studies have focused on combining measurements of CBF and ¹⁸F-FDG uptake, especially using an integrated simultaneous PET/MR scanner, to evaluate patients after STA-MCA bypass surgery.

In this study, the effects of surgical treatment on symptomatic unilateral ICA/MCA steno-occlusive disease were evaluated using quantitative hybrid PET/MR. We first investigated the potential role of a simultaneous PET/MR system in patients with chronic and symptomatically severe steno-occlusive disease of the ICA or MCA in the same functional and physiological states. Second, we analysed changes in cerebral haemodynamic patterns and metabolism that occurred in patients who were offered superficial temporal artery (STA)–MCA bypass surgery as well as changes in imaging findings on follow-up scans obtained at 7 days after surgery. Third, we assessed the correlation between CBF measured by ASL and metabolism assessed by ¹⁸F-FDG in patients with symptomatic unilateral ischaemic cerebrovascular disease.

Angiographic studies demonstrated that ICA occlusion occurred in 10 patients, ICA stenosis in 1 patient, MCA stenosis in 1 patient, and MCA occlusion in 3 patients. A 3-dimensional region of interest (ROI) template was used to automatically place both ipsilateral and contralateral hemispheres (Fig. 1). Before the operation, ASL and PET were used to measure CBF and SUVR values, respectively, in 15 patients with ICA/MCA stenosis or occlusion. The mean value ± SD of the CBF-decreased regions on the affected side, not including the infarction zone, was 34.38 ± 3.76 mL·100 g⁻¹ min⁻¹. The mean value ± SD of the SUVR in the decreased regions on the affected side, not including the infarction area, was 0.88 ± 0.05. The synchronous region, defined as where both CBF and SUVR decreased (shown in Fig. 1d), showed that the CBF and SUVR values were 33.78 ± 4.00 (mL·100 g⁻¹ min⁻¹) and 0.90 ± 0.05, respectively. Figure 2 A and B show that CBF and SUVR were significantly lower in the side ipsilateral to the ICA/MCA stenosis or occlusion than in the contralateral side (P < 0.01). Figure 3 provides an example of data obtained in a patient before surgery.

A significant correlation was observed between the asymmetry index of preoperative CBF and preoperative SUVR values in 15 patients (r = 0.729, P < 0.01; Fig. 4), but there was no significant correlation between CBF and glucose uptake measurements obtained on the affected side (P > 0.05).

Figure 2 A and B also show the changes that occurred in CBF and SUVR values after the operation. The mean preoperative values ± SD of the regions with lower CBF on the zone on the affected side (without the infarction area) before the surgery was 41.59 ± 3.83 (mL·100 g⁻¹ min⁻¹). The mean preoperative values ± SD of the regions with lower SUVR on the affected side (without the infarction area) was 0.91 ± 0.05. The corresponding values in the synchronous region including the lower CBF and lower SUVR areas on the ipsilateral side were 40.84 ± 4.08 (mL·100 g⁻¹ min⁻¹) and 0.93 ± 0.04, respectively (Table 2).

CBF was significantly higher postprocedure than preprocedure for the same regions (P < 0.05 in both), but the postprocedural SUVR was not significantly higher than the preprocedural SUVR (P > 0.05). However, Fig. 5 shows that the postprocedural Al values obtained for of CBF and SUVR were significantly lower than the corresponding preprocedural values (P < 0.05). An example of the preoperative PET/MR map of obtained in an STA-MCA bypass patient is shown in Fig. 6.

This study highlights the potential of PET/MR as a noninvasive imaging tool for investigating haemodynamic and metabolic changes between before and after bypass surgery. Our results show that a combination of ASL and ¹⁸F-FDG PET can be used to optimally evaluate MR-negative regions based on structural images in unilateral ICA or MCA occlusive or stenosis disease.

Similar to ¹⁵O-H₂O PET, ASL sequences have shown great promise for the assessment of many brain disorders. CBF value plays an important role in diagnosis and treatment selection in patients with cerebrovascular disease. Amin-Hanjani et al. showed that regional hypoperfusion is an important predictive factor for stroke risk [13]. A previous study demonstrated that chronic hypoperfusion is indicated by decreased CBF (< 35.0 mL/100 g/min) in hemispheres with arterial disease [14]. Another study of 155 patients with chronic ischaemic cerebrovascular disease found that perfusion was 38.1 ± 10.9 (mL·100 g⁻¹ min⁻¹) on the affected side and that bypass surgery obviously improved CBF values [37]. In our study, the mean value ± SD of the CBF on the affected side was 34.38 ± 3.76 (mL·100 g⁻¹ min⁻¹) before the operation, comparable to previous studies [14, 37‐39]. Bypass surgery aims to increase blood flow to the affected hemisphere and reduce the risk of ensuing neuronal damage in patients. Most studies have documented the clinical benefits and hemodynamic improvements observed after revascularization surgery. These studies have shown that perfusion is greatly improved after treatment [9, 10, 25, 37]. Our results demonstrated that there was significant improvement between the preprocedural and postprocedural CBF values and that the CBF value for the contralateral side also increased after surgery. Soinne et al. investigated CBF changes in patients who underwent carotid endarterectomy and suggested that this procedure increases CBF values on both the side of the surgical procedure and the contralateral side [40]. Analogous findings have been presented in other studies [25]. This phenomenon may be due to the postoperative increase in blood flow on the affected side, which causes a compensatory increase in the contralateral blood flow.

The brain is a highly metabolically active organ that relies on glucose as its energy source. ¹⁸F-FDG is a tracer of regional brain glucose metabolism. Although it is a widely used radiopharmaceutical agent in the metabolic evaluation of oncology and cardiology patients, it may also offer valuable insights into cerebral glucose metabolism during cerebral ischaemia. In fact, many studies have evaluated the potential of ¹⁸F-FDG in cerebral ischaemia [17‐19, 41]. Fukumoto et al. carried out a serial PET study in which ¹⁸F-FDG was used to investigate changes in glucose metabolism before and at 1, 3, 7, and 14 days after acute stroke in an animal model and found that glucose metabolism was comparable between the peri-ischaemic areas and normal brain regions on days 1 and 3 but significantly higher in the later on days 7 and 14. Furthermore, some studies have observed that ¹⁸F-FDG uptake is higher in peri-ischemic regions, similar to findings in animal studies [19]. However, all the studies we read explored acute stroke, and little research has focused on chronic stroke. Nagasawa H et al. found that there were significant decreases in cerebral glucose metabolic rate (CMRGlu) in areas in which no lesions had been detected based on using ¹⁸F-FDG PET in MRI or CT scan in 7 patients with chronic ischemic cerebrovascular disease [20]. Yu Z et al. previously reported seven patients who demonstrated improved glucose metabolism on ¹⁸F-FDG PET after bypass surgery [28]. In the present study, we also focused on the changes that occurred in glucose metabolism between before and after bypass surgery in chronic ischaemic cerebrovascular disease. Our results show there was no significant increase in SUVR between preprocedural and procedural measurements. However, SUVR AI values were significantly lower after surgery than before surgery. This change in SUVR and AI values might indicate that surgery plays a pivotal role in preventing irreversible ischaemic damage in steno-occlusive artery disease. However, a follow-up duration of 7 days after bypass surgery might be too short. CBF values increased significantly during these 7 days. Hence, the promotion of glucose metabolism seems to involve a slower process and requires a longer follow-up.

Cha et al. compared resting state cerebral perfusion evaluated using ASL and glucose metabolism evaluated using ¹⁸F-FDG PET in 20 normal volunteers and found a good overall correlation between perfusion and glucose metabolism [42]. Yu Z et al. found that bypass surgery changed CBF and FDG uptake of chronic ischaemic cerebrovascular disease [28]. However, they did not obtain these data simultaneously. The optimal way to avoid functional and physiological variations in a multimodal study is to simultaneously acquire data using different modalities. The advantage of our study is that we established a method of measuring cerebral perfusion and glucose metabolism while also assessing collateral flow patterns with a single modality in a single session. Our data showed that although there was an overall good correlation between the AI of CBF measured by ASL and the AI of SUVR measured by ¹⁸F-FDG PET, both were significantly lower after surgery. It is clear that the changes that occurred in CBF and SUVR were more dramatic in the synchronous region than in either the CBF-decreased areas alone or the SUVR-decreased areas alone. Ischaemic cerebrovascular disease is a fatal condition that arises due to insufficient blood flow and metabolic impairment to damages brain cells [43, 44]. Theoretically, vascular stenosis or occlusion immediately affected perfusion and causes metabolic changes in the area of the vascular distribution. However, our results show that the area with blood flow reduction and the area with metabolic reduction were not completely overlapping. Some researchers have suggested that this blood flow metabolism mismatch is due to a compensatory or inflammatory response [45, 46]. In the case of FDG PET measurements, test/retest studies have shown that there is variability in their ranges [47, 48]. However, we chose to explore the difference between the SUVR of the ipsilateral region and the SUVR of the contralateral region to directly evaluate the metabolism of glucose. Thus, rCMRglc can be regarded as a stable value in our studies and showed improvements were achieved in some areas. In the common area, both CBF and SUVR decreased and were negative in a structural image. Based on stroke mechanisms, I propose that the common area might have a higher risk of infarction. The effectiveness of treatment is important in patients with chronic ischaemic cerebrovascular disease. However, its exact impact needs further research

This study also has some limitations. First, the follow-up time was short: Since the natural temporal course of brain FDG metabolism and changes in CBF in carotid artery stenosis or occlusion are unknown, the optimal timing remains unclear. Hence, the study follow-up times should be extended in both number and duration. Second, we should enlarge the study sample size. Third, our study shows that the postlabelling delay (PLD) is the most important parameter when measuring CBF; hence, multiple PLDs should be chosen. While SUVR is a relative parameter when measuring glucose metabolism, absolute CMRGlu could be measured by dynamic ¹⁸F-FDG PET using image-derived input functions (IDIF) in an integrated PET/ MRI system. Finally, a cognitive score could be used to systematically investigate cognitive differences among these patients.

The present study demonstrates that combining CBF measured by ASL and metabolism assessed by ¹⁸F-FDG using hybrid PET/MR allows the simultaneous analyse of changes in cerebral haemodynamic patterns and metabolism both before and after STA–MCA bypass surgery and may therefore represent an essential tool for the evaluation of critical haemodynamic and metabolic status in patients with symptomatic unilateral ischaemic cerebrovascular disease.