Choroid Plexus as the Best Reference Region for Standardized Uptake Value Analysis on C11-Acetate PET/CT for Grading and Predicting Prognosis in Patients with Cerebral Gliomas

Abstract

Purpose

We aimed to compare different reference regions and select one with the most clinical relevance on C11-acetate (ACE) positron emission tomography/computed tomography (PET/CT) in patients with cerebral glioma.

Methods

We retrospectively reviewed 51 patients with cerebral glioma who underwent baseline ACE PET/CT at diagnosis. Other than the standardized uptake value (SUV) of the primary tumor, SUVs of the reference regions including the normal gray matter, white matter, choroid plexus, and cerebellum were measured. Then, the SUV ratio (SUV_R = tumor SUV_max/reference region SUV_mean) was calculated. The effect of patient age on the SUV_mean of each reference was examined and the SUV_Rs of each reference region were compared between grades. age, sex, tumor size, histological grades, SUV_R, and the presence of isocitrate dehydrogenase (IDH) mutation were included for survival analyses.

Results

Except for the cerebellum showing a mild negative correlation, we found no correlations between age and SUV_mean using the gray matter, white matter, and choroid plexus (r = − 0.280, P = 0.047). Only the SUV_R-choroid plexus was able to differentiate between the WHO grades (Grade II vs. III, P = 0.035; grade III vs. IV, P < 0.001; grade II vs. IV, P < 0.001). Multivariate Cox proportional hazards models found that the SUVR-choroid plexus and IDH mutation were statistically significant for predicting OS.

Conclusion

Of the different reference regions used for grading cerebral gliomas, the choroid plexus was found to be the most optimal. In addition, the SUV ratio is useful to predict the overall survival in the model with the choroid plexus as a reference region.

Introduction

The standardized uptake value (SUV) on positron emission tomography/computed tomography (PET/CT) is the most important parameter which measures the activity concentrations of regions of interest. It is typically adjusted for the total amount of activity injected and normalized using the physical characteristics of the patients, such as the total body weight or lean body mass, to allow comparison between patients. It is used as the ratio of the activity concentration measured in the region of interest (ROI) (kBq/mL) and the injected activity dose (MBq) divided by the body mass (kg). Although the SUVs are successfully used in many clinical studies, many biological and technical sources of variability should be taken into consideration. Since some of these factors cause common variances in different regions, the SUV ratio (SUVR) between the SUVs of a ROI and a reference region is better than SUVs alone in comparing different technical settings and patients with differing physical properties [1]. In fact, the liver is one of the most used reference regions in many cancers. The tumor-to-liver SUVR is better than the tumor SUV in the tumor grading, response evaluation, and prognostication [2].

C11-acetate (ACE), previously associated with fatty acid synthesis in prostate cancers, was developed to evaluate cardiac oxidative metabolism [3]. It is valuable in detecting those tumors with a low glycolytic phenotype and well-differentiated pathologic tumor grade. In hepatocellular carcinomas, ACE plays an important role in detecting non-F-18 fluorodeoxyglucose (FDG) avid low-grade tumors. Unlike tumors in the body, a recent study showed a high ACE uptake in high-grade cerebral gliomas and predicted survival better than the WHO grade [4]. However, the exact mechanism of ACE uptake in cerebral gliomas is unknown and requires further investigation. Compared to FDG with a high cortical background uptake in cerebral glioma imaging, low uptake in the cerebral cortex is one of the advantages of ACE. Despite the potential values of ACE, only limited studies have used ACE PET/CT in patients with cerebral gliomas and there was no established reference region for the SUVR.

In this study, we measured the ACE uptake in different reference regions including the gray matter, white matter, choroid plexus, and cerebellum, and compared each SUVR model in predicting the histological grades and overall survival (OS) in patients with cerebral glioma. Other than the known references, the choroid plexus was added based on a hypothesis that malignant tumor should show more C-11 ACE uptake than physiologic uptake in the choroid plexus.

Methods and Materials

Patients

We recruited 51 patients (29 men and 22 women; median age, 54 years; range, 23–74 years) with glioma suspicion on MRI. They underwent ACE PET/CT before surgery between January 2012 and July 2013 at our hospital. All patients with histologically confirmed cerebral gliomas (10 WHO grade II, 16 grade III, and 25 grade IV) were included in the analysis. The median duration of clinical follow-up was 29.0 months (range, 3.3–98.4 months). This was a retrospective analysis of a prospective study approved by the institutional review board (No. 4–2011-0697) to decide the best reference region for SUVR. Thanks to the pathologist added gene mutation information after glioma WHO 2016 classification released, patients before 2016 were also classified according to the WHO 2016 classification.

PET/CT Imaging

For ACE PET/CT scan, all patients fasted for 4 to 6 h before undergoing ACE injection. Approximately 555 MBq (20 mCi) of ACE was administered to the patients intravenously and a 20-min emission scan was performed 10 min after injection. Head of the patient was positioned and fixed with the detector parallel on the orbitomeatal line. Patients underwent PET/CT using a PET/CT scanner (Discovery 600, General Electric Medical Systems, Milwaukee, WI, USA) with a scout view at 10 mA and 120 kVp, followed by a spiral CT scan for attenuation correction with a 0.8-s rotation time, 60 mA, 120 kVp, 3.75-mm-section thickness, 0.625-mm collimation, and 9.375-mm table feed per rotation. PET images were reconstructed using an iterative algorithm (iteration 2, subset 32).

Imaging Analysis

All PET/CT and MR images were reviewed and analyzed by two nuclear medicine physicians who reached consensus on a dedicated workstation. The PET/CT and MR images were registered using a fusion module in the imaging software MIM (MIM-6.5; MIM Software Inc., Cleveland, OH, USA). The maximum standardized uptake (SUV_max) of ACE by the tumor was measured using a volume of interest (VOI) within the tumor volume. The normal gray matter, white matter, choroid plexus, and cerebellum were used as the reference regions. The ipsilateral region of the cerebellum was used while the contralateral regions were used in the other organs. VOIs were manually drawn over the reference region to obtain the mean SUV (SUV_mean) of the reference regions (Fig. 1). Then, the tumor-to-reference region ratio (SUVR) was determined as the SUV_max of tumors to the SUV_mean of the reference region.

Fig. 1.

The normal gray matter, white matter, choroid plexus, and cerebellum were used as the reference regions. The ipsilateral region of the cerebellum was used while the contralateral regions were used in the others. The VOIs were manually drawn over the reference region to obtain the mean SUV (SUVmean) of the reference regions

Statistical Analysis

Correlation analysis was performed between the patients’ age and SUV_mean values of each reference region to see whether there would be any effect of age on SUVs of the reference region. The WHO grade was classified into low (grade II) or high (grades III and IV). The relationships between the histological WHO grades and SUV_R on PET/CT were assessed using the Bonferroni post hoc analysis. The OS time was calculated from the time from surgical resection to death or the last follow-up visit. Optimal cut-off values of the continuous parameters (age, tumor size, SUV_R) to predict OS were determined using the maximally selected log-rank statistics method using the Maxstat package in R software. Log-rank test and Cox proportional hazards regression analysis were used to evaluate the efficacy of clinico-pathological and imaging prognostic factors in univariate and multivariate analysis, respectively. Multiple multivariate Cox proportional hazards models were used due to the collinearity between 4 different SUVRs. All analyses were performed using IBM SPSS Statistics 25.0 (IBM Corp., Armonk, NY, USA), Medcalc version 19.3.1 (MedCalc Software Ltd., Acacialaan 22, 8400 Ostend, Belgium), and R 4.0.3 software (http://www.r-project.org; The R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was defined by a P value < 0.05 for all statistical analyses.

Results

Clinical Characteristics of Patients

Patient demographics are summarized in Table 1. Among the 51 patients (29 men and 22 women) with histologically confirmed cerebral gliomas, 10 (19.6%) had WHO grade II, 16 (31.4%) had grade III, and 25 (49.0%) had grade IV tumors. There were 5 patients with grade II oligodendroglioma (OD), and 5 patients with grade III anaplastic OD. The median SUV_R was 4.83 (interquartile range (IQR), 3.69 to 7.02) in the white matter, 2.48 (IQR, 1.63 to 3.32) in the gray matter, 1.65 (IQR, 1.23 to 2.41) in the choroid plexus, and 2.29 (IQR, 1.73 to 3.43) in the cerebellum.

Table 1.

Patient characteristics

Characteristic	Values
Age (year)	Median, 54 (range, 23–74)
Sex, n (%)
Male	29 (56.9%)
Female	22 (43.1%)
WHO 2016 grade, n (%)
Grade II	10 (19.6%)
Grade III	16 (31.4%)
Grade IV	25 (49.0%)
Histology, n (IDH-mutant/wildtype)	(19/32)
Astrocytoma	41
Grade II	5 (5/0)
Grade III	11 (3/8)
Grade IV	25 (1/24)
Oligodendroglioma	10
Grade II	5
Grade III	5
1p/19q co-deletion (non/co-deletion)	31/20
SUV ratio (IQR 25%–75%)
White matter	4.83 (3.69–7.02)
Gray matter	2.48 (1.63–3.32)
Choroid plexus	1.65 (1.23–2.41)
Cerebellum	2.29 (1.73–3.43)

WHO, World Health Organization

Effects of Patient Age on the SUVs of the Reference Regions

The median SUV_mean was 0.50 (interquartile range (IQR), 0.41 to 0.57) in the white matter, 1.03 (IQR, 0.85 to 1.18) in the gray matter, 1.42 (IQR, 1.25 to 1.75) in the choroid plexus, and 1.05 (IQR, 0.87 to 1.18) in the cerebellum. We performed correlation analysis between patient age and SUV_mean values of each reference region to see the effect of age on the SUVs of the reference region. There was a mild, negative correlation between patient age and SUV_mean in the cerebellum (r = − 0.280, P = 0.047), while other reference regions showed no significant correlation with the patient age (r = − 0.047, P = 0.741 for the white matter, r = − 0.245, P = 0.0.083 for the gray matter, and r = − 0.224, P = 0.114 for the choroid plexus).

Relationship Between ACE Uptake on PET/CT and Tumor Histological Grades

We did ROC analyses using the SUVRs (white matter, gray matter, choroid plexus, and cerebellum) to divide gliomas into high grade (grade III and IV, n = 41) and low grade (grade II, n = 10). There were no statistically significant differences in the AUCs of the low- and high-grade gliomas (Fig. 2). However, only the SUVR-choroid plexus showed a significant difference in the three WHO grade groups. Other SUVRs showed significant differences between grade II and grade IV, and grade III and grade IV, but did not show significant differences between grade II and grade III (Table 2).

Fig. 2.

ROC analysis using SUVRs (gray matter, white matter, choroid plexus, and cerebellum) to differentiate high- and low-grade gliomas

Table 2.

Relationship between the histologic WHO grades and SUVR on PET/CT using the Bonferroni post hoc analysis

	Grade II vs, grade III	Grade III vs. grade IV	Grade II vs. grade IV
SUVR, white matter	P = 0.055	P < 0.001	P < 0.001
SUVR, gray matter	P = 0.135	P < 0.001	P < 0.001
SUVR, choroid plexus	P = 0.035	P < 0.001	P < 0.001
SUVR, cerebellum	P = 0.152	P < 0.001	P < 0.001

SUVR, standardized uptake value ratio

Cox Proportional Hazards Models for OS

The optimal cut-off values for age and tumor size to predict OS were > 39 years and > 5.9 cm, respectively. The optimal cut-off values for SUVRs of the white matter, gray matter, choroid plexus, and cerebellum to predict OS were > 2.17, > 3.72, > 1.52, and > 2.17, respectively. In univariate analysis, age, tumor size, WHO grade, IDH1 mutation, and all the SUV_Rs were significant prognostic variables for OS (Table 3). Because of the collinearity between the SUV_Rs, separate multivariate Cox proportional hazards models were used. Only the SUVR model using the choroid plexus as a reference had a statistically significant OS (P = 0.037). With a cutoff SUVR-choroid plexus of 1.52, the median OS was 15.63 months in patients with a SUVR-choroid plexus ≥ 1.52 and 84.9 months in patients with a SUVR-choroid plexus < 1.52 (P < 0.001). In the model, IDH mutation was found to be significant as well. In the other 3 models, the presence of IDH mutation was most significant with marginal statistical significances whereas SUVRs were found insignificant.

Table 3.

Univariate and multivariate Cox proportional hazards analyses for overall survival

Discussion

Although SUV is the nearly exclusive means for semi-quantitative evaluation of clinical PET/CT studies, there are problems with the standardization due to inter-changeability of the values with different scanners, protocols, patients, etc. One of the solutions to overcome some of the problems was the normalization of tumor SUV to SUV of a reference region. In evaluating cerebral gliomas on F-18 fluorodeoxyglucose or other amino acid-derived radiotracers, the contralateral gray matter is usually used as a normal reference. Regarding the ACE PET/CT, no study has been done to determine the best reference region for SUVR.

Of the limited number of studies using ACE PET/CT, all studies but one used the contralateral gray matter as a reference region [4–7]. When the contralateral gray matter was used as a reference region, two studies found that the SUVR of ACE in high-grade gliomas was significantly higher than those in low-grade gliomas [6, 7], whereas one study reported a limited value of SUVR in differentiating the histological grades [5]. In our study, we used 4 different reference regions (gray matter, white matter, choroid plexus, and cerebellum) to calculate the SUVR. All 4 SUVRs were excellent in differentiating high from low-grade gliomas and there were no statistically significant differences in their AUCs. The results were strongly supportive of the value of ACE PET/CT in glioma grading. In further analysis, the SUVR-choroid plexus was the only one which showed significant differences among the 3 WHO grade groups. In contrast, other SUVRs were able to differentiate between grade II and grade IV, and grade III and grade IV, but showed no significant differences between grade II and grade III. Based on our results, the choroid plexus seems to be the best reference region for the SUVR in glioma grading.

Although the prediction of grading is important, histopathological information including the grade can be eventually obtained after surgery. Therefore, the prediction of patient outcomes using imaging modalities may have more clinical relevance than the grading. In this study, univariate analyses found that the patient age, tumor size, histological grade, and IDH mutation were all significant. In multivariate Cox proportional hazards analysis, the model with SUVR-choroid plexus was the only one in which SUVR on ACE PET/CT found to be significant for OS. In the model, the presence of IDH mutation was significant in predicting OS. However, other models with SUVR-gray matter, SUVR-white matter, or SUVR-cerebellum showed no statistical significance for predicting OS and IDH mutation showed marginal significances. It was likely that the selection of a reference region seemed important and affects the performances of ACE PET/CT in predicting patient outcomes.

In our study, the choroid plexus showed the highest SUV_mean value among the reference regions. As a tissue barrier, it is responsible for the synthesis of the cerebrospinal fluid (CSF) and its major proteins, metabolites, and a diversity of other molecules. It receives the highest local blood flow in the brain, despite a surface area similar to that of the blood brain barrier (BBB) [8]. Monocarboxylate transporter (MCT) facilitates the transport of ACE across the BBB. Part of the ACE uptake in the choroid plexus could be related to the expression of MCT in the choroid plexus [9]. Furthermore, the fenestrated capillaries of the choroid plexus are permeable to molecules of different sizes [10]. In physiological conditions, macromolecules with molecular weights up to ∼ 800 kDa or size ∼ 12 nm may diffuse into the choroid plexus [8]. Therefore, contrast agents used in magnetic resonance imaging (MRI) lead to a homogenous enhancement of the choroid plexus which could be used to normalize abnormal brain enhancement [11]. All the factors described above could have affected the physiologic uptake of ACE in the choroid plexus on PET/CT. Like contrast agents in MRI, the use of ACE uptake in the choroid plexus to normalize ACE uptake in the gliomas was useful for grading and also for predicting patient survival.

There are radiotracers showing physiologic uptake in the choroid plexus such as C-11 methionine. Studies have found the relationship between brain methionine uptake and age. O’Tuama et al. showed a significant age-dependent decline of 11C-MET uptake in maturing adults [12], and Uda et al. also reported a slightly negative linear regression, although no statistically significant difference was observed [13]. In this study, we found a mild, negative correlation between age and the SUV_mean of the cerebellum. However, ACE uptake in the other reference regions showed no significant correlation with the patient age. The independency of physiologic ACE uptake especially in the choroid plexus can be a favorable feature to use it as a reference tissue in patients with different ages.

This study had some limitations. First, although we utilized a prospectively obtained brain tumor database, the study was retrospective in nature which could be subject to selection bias. Second, volumetric parameters, such as metabolic tumor volume and total lesion glycolysis, were not evaluated. Third, in tumors without ACE uptake, there were difficulties to localize the tumor boundary and measure SUVs on PET/CT. Co-registration to MRI was used to exclude the normal cortex as much as possible. Fourth, there might be a potential problem of using choroid plexus as a reference region related to partial volume averaging effect. However, we found that the choroid plexus was the one with most physiologic uptake of all the reference regions and there were stable C-11 ACE uptake in the choroid plexus over patient’s age.

Conclusion

On ACE PET/CT, the choroid plexus was the best reference tissue able to differentiate between the grades II, III, and IV groups of gliomas. Of the survival models with different reference regions, the one with the choroid plexus was the only one in which SUV_R on ACE PET/CT predicted the overall survival in addition to IDH mutation.

Compliance with Ethical Standards

Conflict of Interest

Dongwoo Kim, Arthur Cho, Sang Hyun Hwang, KwanHyeong Jo, and Jong Hee Chang, Mijin Yun declare that they have no conflict of interest. This research was supported by the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (NRF-2016R1E1A1A01943303). The funders of the study were not involved in the study design, data collection, data interpretation, writing of the report, or the decision to submit the paper for publication. The authors report no other potential conflict of interest relevant to this article.

Ethical Approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the Helsinki declaration as revised in 2013 and its later amendments or comparable ethical standards.

Informed Consent

The institutional review board of our institute approved this retrospective study, and the requirement to obtain informed consent was waived.