Evaluation of Brain Tumors Using Amide Proton Transfer Imaging: A Comparison of Normal Amide Proton Transfer Signal With Abnormal Amide Proton Transfer Signal Value : Journal of Computer Assisted Tomography

Secondary Logo

Journal Logo

Neuroimaging: Brain

Evaluation of Brain Tumors Using Amide Proton Transfer Imaging: A Comparison of Normal Amide Proton Transfer Signal With Abnormal Amide Proton Transfer Signal Value

Sugawara, Kazuaki MS∗,†,‡; Miyati, Tosiaki PhD, DMSc; Wakabayashi, Hikaru MD§; Yoshimaru, Daisuke PhD; Komatsu, Shuhei MD; Hagiwara, Kazuchika MD, PhD; Saigusa, Kuniyasu MD, PhD#; Ohno, Naoki PhD

Author Information
Journal of Computer Assisted Tomography 47(1):p 121-128, 1/2 2023. | DOI: 10.1097/RCT.0000000000001378
  • Free

Abstract

Amide proton transfer (APT) imaging, a method of chemical exchange saturation transfer (CEST) imaging, is a novel molecular magnetic resonance imaging (MRI) technique that detects endogenous mobile proteins and peptides present in low concentrations in tissues. Amide proton transfer imaging was developed in 2003 by Zhou et al.1

In APT imaging, endogenous cellular protein information is obtained indirectly through a large amount of water signal used in MRI, showing a high signal intensity in lesions (eg, tumors) containing a large amount of tissue mobile proteins/peptides.1,2 This technique has been used as an imaging biomarker to characterize brain tumors and distinguish recurrent tumors from necrosis.3–5 In addition, APT imaging has been useful in many early studies, including studies focused on a differentiation between ischemic and hemorrhagic stroke, Alzheimer disease, Parkinson disease, and multiple sclerosis.6–11 These previous studies demonstrated the relationship between APT signal and disease, becoming useful for patient management, although the choice of APT imaging parameters used in clinical practice varied among centers.5,12–16 Therefore, the use of APT imaging conditions, especially the duration of saturation pulse and B1 intensity, and image analysis methods selected at each institution showed different results. For example, in a study comparing postradiotherapy radionecrosis with postradiotherapy brain tumors, there were inconsistent results in distinguishing the 2 lesions when APT images were used.5,13 Studies using APT imaging for glioblastoma and solitary brain metastases identified different indices to distinguish the 2 tumor types via APT signal characteristics.13,14 Studies analyzing the relationship between APT signal and age dependence showed opposite conclusions between age and APT signal.15,16

As described previously, there is no consensus on the parameters of APT imaging used in clinical practice.17 Therefore, a facility performing APT imaging may be making a facility-specific diagnosis based on the image contrast of a particular APT obtained from the pulse sequence used. This limits the reproducibility and comparability of studies across institutions and may cause problems in the APT imaging evaluation. Thus, standardization of the imaging parameters used is important to solve this problem. In addition, the normal APT signal value can be used as a reference value to compare the APT signals of disease subtypes, possibly being a tool to characterize the disease from the APT signal value. We examined the changes in APT signals in each normal brain region and reported that they were independent of age and sex differences under the parameter conditions we implemented.16 If APT imaging is performed under the same conditions as this imaging method, it may be possible to easily compare reproducibility and results between institutions.

This study aimed to compare the APT signal characteristics of brain tumors and normal brain tissue, as well as the relationships of these to one another. We also sought to show the potential to develop a standard for APT imaging.

MATERIALS AND METHODS

Study Population

From August 2018 to August 2021, 24 patients with glioblastoma or brain metastases underwent APT imaging in addition to a usual protocol, including Gd contrast. The effect of motion artifacts on images was carefully observed in all captured images. Four patients' data were excluded because they were affected by motion artifacts. Therefore, 20 patients were included in this study (Table 1). We classified 13 patients with glioblastoma (10 males: 59.6 ± 15.3 years, 3 females: 48.7 ± 10.1 years; age range, 26–76 years) and 7 patients with brain metastases (6 males: 68.8 ± 9.35 years, 1 female: 73 years; age range, 51–78 years). Primary cancer included lung in 5, stomach in 1, and colon in 1 (all adenocarcinoma) patient. One patient underwent pathological biopsy surgery before undergoing MRI. The pathological diagnosis was determined with specimens obtained at surgical resection according to the World Health Organization criteria18 by established neuropathologists. Our institutional review board approved this retrospective study, and the requirement for informed consent was waived (approval no. 680).

TABLE 1 - Study Population Characteristics
Histology Average Age, Age Range, y Sex (Male:Female) Location Tumor Size Range, Diameter (Min–Max), mm Primary Site
Glioblastoma multiforme 56.8 ± 15.6 6:0 Temporal 23.7–52.2
Grade IV (n = 13) 26–76 2:2 Frontal 20.6–56.5
1:1 Thalamus 25.1–31.6
1:0 Parietal 30.9–33.5
Brain metastases 70.4 ± 9 2:1 Frontal 19.5–43.1 Lung (2), colon (1)
Adenocarcinoma (n = 7) 51–78 4:0 Parietal 19.1–21.5 Lung (3), stomach (1)

Magnetic Resonance Imaging Scanning Protocol

The APT imaging conditions used in this study were as per a previous report that validated Dixon-based 3D-APT imaging.19 All patients underwent MRI using a 3-T unit (Ingenia; Philips Medical Systems, the Netherlands) equipped with a 32-channel head coil and a 2-channel parallel transmission coil. The APT imaging parameters were as follows: mDIXON 3D-APT for fast spin echo method; acquisition echo train length, 174; SENSE factor, 1.6; repetition time, 5864 milliseconds; echo time, 7.8 milliseconds; matrix, 128 × 100 (reconstructed to 256 × 256); field of view, 230× 180 mm2; thickness, 8 mm; saturation power, 2 μT; saturation duration, 2 seconds; and scan time, 5 minutes 32 seconds. The saturation power (2 μT) with saturation duration (2 seconds) was the maximum continuous pulse time allowed for the equipment used. After the second-order pencil-beam volume B0 shimming, 3D-APT imaging with intrinsic B0 correction using the Dixon method was performed using the fast spin echo sequence with driven equilibrium refocusing and the following parameters: 40 sinc-Gaussian pulses (50 milliseconds each) and 9 image volumes at multiple frequency offsets (S0, ±2.7, +3.5 [3], −3.5, ±4.3; value in parentheses is the number of acquisitions, which was considered as 1 if unspecified) were acquired. B0 map derived from 3 acquisitions at 3.5 ppm with slightly different echo shifts using an mDIXON algorithm was used for a voxel-by-voxel B0 correction. A Dixon B0 map was generated using an iterative least-squares method that decomposes water and fat images from source images acquired using 3 ES values (−0.4, 0, and +0.4 milliseconds). We performed B1 shimming with each scan to facilitate B1 inhomogeneity correction.18 Each voxel was corrected for image intensity based on the nominal saturation frequency offset by Lagrange interpolation among the neighboring Z-spectral images. This procedure corresponds to a frequency shift along the saturation frequency offset axis according to the measured B0 shift. The APT imaging signal value was calculated for each pixel using equation 1 as ±3.5 ppm of magnetic transfer ratio asymmetric or MTRasym, which is an index of the strength of the CEST effect.

APTsignal=MTRasym3.5ppm=Ssat3.5ppmSsat+3.5ppm/S0

where Ssat and S0 are the signal intensities obtained with and without selective saturation, respectively. The map obtained after performing this calculation for each pixel is called an APT-CEST image. For reference, several standard MR images, including T1-weighted image, T2-weighted image (T2WI), fluid attenuation inversion recovery (FLAIR), and contrast-enhanced T1-weighted images (Gd-T1W) were acquired. The APT images were performed before administration of gadolinium contrast agent in all patients because T1 shortening by gadolinium may alter APT-weighted signal. All imaging times were approximately 30 minutes.

Region of Interest Analyses

Two radiologists with 11 and 10 years of experience in diagnostic radiology performed image analysis independently. The region of interest (ROI) was set for all patients by selecting the slice section in which the largest tumor area was depicted. The tumor size in diameter ranged from a minimum of 19.1 mm to a maximum of 56.5 mm.

The Gd-T1W and FLAIR and T2W were used to place 3 to 4 ROIs (15.4–21.6 pixels depending on the lesion size) within the tumor core, peritumoral edema (PE), and ipsilateral normal-appearing white matter (INAWM) and 1 ROI in the contralateral normal-appearing white matter (CNAWM; 21.6 pixels). Their mean ± standard deviation was calculated.

Figure 1 shows an example of ROI placement. Four ROIs were placed within the Gd-enhanced tumor area (black) and in the surrounding edema (blue), and 3 ROIs were placed in the ipsilateral normal white matter (gray) using T2WI (A), FLAIR (B), and Gd-T1W (C). One ROI was placed in the CNAWM (red). The ROIs placed in each image were copied to the APT image. The ROI size and number were taken into consideration for each patient so that the copy position of the ROI would not deviate significantly from the target. The ROI included (a) tumor core: a substantial mass in the Gd-enhancing region, (b) peritumor edema: 0.5 to 1 cm away from the tumor nucleus (Gd-enhancing region),20 (c) INAWM: 0.5 to 1 cm away from the edema margin,20 and (d) CNAWM: white matter that appeared normal on the opposite side.

F1
FIGURE 1:
Example of ROIs placement. Four ROIs were placed within the Gd-enhanced tumor area (black) and in the surrounding edema (blue), and 3 ROIs were placed in the ipsilateral normal white matter (gray) using T2WI (A), FLAIR (B), and Gd-T1W (C). One ROI was placed in the contralateral normal white matter (red). The ROIs placed in each image were copied to the APT image. Cystic, necrotic, or hemorrhagic components were avoided by referring to conventional MRI. The ROIs placed in each image were copied to the APT image. The ROI included (a) tumor core: a substantial mass in the Gd-enhancing region, (b) peritumor edema: 0.5 to 1 cm away from the tumor nucleus (Gd-enhancing region),20 (c) INAWM: 0.5 to 1 cm away from the edema margin,20 and (d) CNAWM: white matter that appeared normal on the opposite side. Among all patients, 2 glioblastomas did not show Gd enhancement; thus, tumor nuclei were estimated from FLAIR and T2WI. Certain areas, such as those of cystic masses, hemorrhagic components, or large blood vessels, were not included in the ROI. Figure 1 can be viewed online in color at www.jcat.org.

Among all patients, 2 glioblastomas did not show Gd enhancement; thus, tumor nuclei were estimated from FLAIR and T2WI. Certain areas, such as those of cystic masses, hemorrhagic components, or large blood vessels, were not included in the ROI.

Statistical Analyses

Kruskal-Wallis was used to compare the APT signal values of the tumor core, PE, INAWM, and CNAWM between glioma and brain metastases. We evaluated the receiver operating characteristic (ROC) curve and the area under the curve to determine the optimal cutoff value and distinguish the 2 tumor regions from the APT signal values. Wilcoxon signed-order sum test was used to analyze the difference between tumor core and PE APT signal values and contralateral normal white matter APT signal values (APTcore − CNAWM, APT PE − CNAWM). Then, the APT change rate calculated as APTcoreCNAWMCNAWM,APTPECNAWMCNAWM was analyzed. The interobserver agreement regarding the measurements by the 2 radiologists was analyzed by calculating the intraclass correlation coefficient (ICC). All statistical analyses were performed using R version 4.1.1 for Epi package. A P value less than 0.05 was considered statistically significant.

RESULTS

Interobserver Agreement

Interrater reliability, as determined by ICC, was acceptable for both glioblastoma and brain metastases. There was also a good consistency between raters with the ICC tumor core, PE, INAWM, and CNAWM values of 0.97, 0.96, 0.86, and 0.82, respectively.

Comparison of APT Signal Values of Each Tumor With Tumor Core, PE, INAWM, and CNAWM

Figure 2 shows the comparison of APT signal values between glioblastoma and brain metastases. The glioblastoma APT signal value was the highest in tumor core (3.41% ± 0.49%), followed by PE (2.24% ± 029%), which was significantly higher than in INAWM (1.35% ± 0.15%) and CNAWM (1.26% ± 0.12%, P < 0.001). There was no significant difference between INAWM and CNAWM (P > 0.05).

F2
FIGURE 2:
Comparison of APT signal values of each tumor with tumor core, PE, INAWM, and CNAWM. The glioblastoma APT signal value was the highest in tumor core (3.41% ± 0.49%), followed by PE (2.24% ± 0.29%), INAWM (1.35% ± 0.15%), and CNAWM (1.26% ± 0.12%, **P < 0.001). There was no significant difference between INAWM and CNAWM (P > 0.05). The brain metastases APT signal value was the highest in tumor core (2.74% ± 0.34%), followed by PE (1.86% ± 0.35%), INAWM (1.17% ± 0.13%), and CNAWM (1.23% ± 0.09%, *P < 0.01). There was no significant difference between INAWM and CNAWM (P > 0.05). Figure 2 can be viewed online in color at www.jcat.org.

The brain metastases APT signal value was the highest in tumor core (2.74% ± 0.34%), followed by PE (1.86% ± 0.35%), which was significantly higher than in INAWM (1.17% ± 0.13%) and CNAWM (1.23% ± 0.09%, P < 0.01). There was no significant difference between INAWM and CNAWM (P > 0.05).

Parameters for Distinguishing Glioblastoma From Brain Metastases From APT Signal Values

Table 2 and Figure 3 show the results of ROC analysis of APT tumor core and PE signals of glioblastoma and brain metastases. The parameter that distinguished the 2 tumors by ROC analysis was the tumor nucleus. The optimal cutoff value was 3.07% with a sensitivity and specificity of 77% and 86%, respectively (P < 0.05).

TABLE 2 - Results of the ROC Analysis of Glioblastoma and Brain Metastases Using Tumor Core, PE, INAWM, and CNAWM as Parameters
Parameter APT Signal, Glioblastoma Metastasis, Mean ± SD AUC 95% CI Cutoff Sensitivity Specificity Accuracy P
Tumor core 3.41 ± 0.49 2.74 ± 0.34 0.885 0.73–1.00 3.07 77% 86% 80% 0.022
PE 2.24 ± 0.29 1.86 ± 0.35 0.769 0.52–1.00 2.0 92% 72% 85% 0.208
INAWM 1.35 ± 0.15 1.17 ± 0.13 0.835 0.64–1.00 1.23 93% 72% 85% 0.062
CNAWM 1.26 ± 0.12 1.23 ± 0.09 0.676 0.43–0.92 1.31 47% 100% 65% 0.818
AUC indicates Area Under Curve.

F3
FIGURE 3:
Results of the ROC analysis of tumor core and PE APT signals of glioblastoma and brain metastases. The parameter that distinguished the 2 tumors by ROC analysis was the tumor nucleus. The optimal cutoff value was 3.07% with a sensitivity and specificity of 77% and 86%, respectively (P < 0.05).

Table 3 and Figure 4 show the results of comparing APT difference and APT change rate for glioblastoma and brain metastases, respectively.

TABLE 3 - Comparison of APT Difference and APT Change Rate Between Tumor Core and Peritumor Edema
Parameter Tumor Core Peritumoral Edema
Glioblastoma Metastasis P Glioblastoma Metastasis P
APT difference, mean ± SD, % 2.15 ± 0.51 1.43 ± 0.3 0.011 0.99 ± 0.3 0.66 ± 0.32 0.036
APT difference, 95% CI 1.84%–2.46% 1.24%–1.84% 0.81%–1.17% 0.34%–0.98%
APT change rate, (%) 173 129 0.045 78 56 0.096

F4
FIGURE 4:
Results of comparing APT difference and APT change rate for glioblastoma and brain metastases. The APT difference for tumor core was 2.15% ± 0.51% and 1.43% ± 0.3% (P < 0.05) for glioblastoma and brain metastases, respectively. The APT change rate was 173% and 129% for glioblastoma and brain metastases, respectively (P < 0.05). The APT difference for peritumor edema was 0.99% ± 0.3% and 0.66% ± 0.32% (P < 0.05) for glioblastoma and brain metastases, respectively. The APT change rate was not significantly different at 78% and 56% for glioblastoma and brain metastases, respectively (P > 0.05). Figure 4 can be viewed online in color at www.jcat.org.

The APT difference of tumor core was 2.15% ± 0.51% and 1.43% ± 0.3% (P < 0.05) for glioblastoma and brain metastases, respectively. The APT change rate was 173% and 129% for glioblastoma and brain metastases, respectively (P < 0.05). The APT difference for peritumor edema was 0.99% ± 0.3% and 0.66% ± 0.32% (P < 0.05) for glioblastoma and brain metastases, respectively. The APT change rate was not significantly different at 78% and 56% for glioblastoma and brain metastases, respectively (P > 0.05).

As representative examples, Figure 1 shows a patient with glioblastoma, and Figure 5 shows a patient with a brain metastases.

F5
FIGURE 5:
Representative example of brain metastases (single). Results of a 78-year-old man with postoperative histopathological evidence of a brain metastases (adenocarcinoma). The primary tumor was lung cancer. The tumor is located in the right frontal lobe with strong edema around the tumor. Inside the tumor, T2WI shows low to high nonuniform signal intensities (A), FLAIR shows medium to low signal intensities (B), and the tumor has a phyllodes structure with a ring of contrast enhancement on Gd-T1W (C). Amide proton transfer imaging shows nonuniform and high signal intensity from the margin to the outside of the tumor core (tumor core, APTmean = 3.02%, PE, APTmean = 2.46%, INAWM, APTmean = 1.31%, CNAWM, APTmean = 1.29%; D). Hematoxylin-eosin–stained section (×20) of the tumor core. Stained sections of tumor nuclei show increased vascularity and infiltration of moderately to highly differentiated adenocarcinoma (E). Figure 5 can be viewed online in color at www.jcat.org.

DISCUSSION

In this study, 3D-APT imaging was used in patients with brain tumors to evaluate the differences in APT signal characteristics between normal brain tissue and tumor. Using the same parameters as in the present method, the relationship between changes in amide proton/peptide levels and disease can be easily evaluated from changes in APT signal values. Using this approach, we believe that the reproducibility of results and the ability to compare results between institutions may be simplified.

Currently, APT imaging has many indications that are used in the field of brain tumors.2–5 The clinical added value of APT imaging demonstrated by these studies suggests that it may help guide treatment. The reason why the APT signal values of tumor nuclei and PE in glioblastoma and brain metastases were significantly higher than the contralateral normal APT signal values is described hereinafter.

First, glioblastoma and brain metastases may exhibit higher APT signal values than normal brain parenchyma because of rapid cell proliferation in the tumor core and an abundance of mobile proteins/peptides in surviving active tumor cells.2–5,21 In a study that had quantitatively evaluated tumor areas in the brain, tumor nuclei of glioblastoma had significantly higher T2 values than those of brain metastases.20 In addition, PE of glioblastoma includes both angiogenic edema and invasive tumor.22,23 Amide proton transfer imaging produces a signal from the chemical exchange of water protons with abundant amide protons/proteins in active tumor cells. Water content is related to the proton exchange rate and affects the signal intensity of APT.1 Therefore, the CEST effect on glioblastoma tumor nuclei and PE may be greater than that of brain metastases due to increased cellular protein content and abundant water content. Second, studies showed differences in perfusion for tumor core and PE. Compared with brain metastases, tumor core and PE of glioblastoma showed significantly higher microvascular density and cerebral blood flow.24,25 In addition, certain studies showed APT signal intensity in benign intratumoral angiogenesis-rich tumors to be comparable or higher than that in malignant brain tumors.26 Therefore, the higher APT signal values of tumor nuclei and periapical edema in glioblastoma compared with brain metastases may be related to the density of blood vessels grown by tumor cells and perfusion.

However, differences between our results and those reported previously regarding APT signal characteristics of glioblastomas and of brain metastases may also be secondary to differences in study design and imaging parameters. Similar to the present study, 2 previous studies applied APT imaging to evaluate glioblastoma and brain metastases.13,14 Comparison of results between centers showed that while the PE APT signal value is a parameter that distinguishes the 2 tumors, the tumor core APT signal value may also be a parameter. The reasons for the difference between the results of this study and those of the present study are related to the following. First, the difference in saturation pulse duration significantly affects the APT image contrast. In earlier studies, a longer saturation pulse duration did not change the APT signal in low-grade tumors; however, it decreased the APT signal in normal white matter, improving the contrast between tumor and normal tissue.27 Long saturation pulse time could result from larger magnetization transfer and spillover effects and decreased magnetization transfer ratio asymmetry. Second, differences in methods of segmentation have been associated with differences in the results of earlier studies.28,29 In our study, 2 radiologists individually performed the ROI analysis. Although manual ROI analysis needs to account for interindividual and intraindividual variability, there was little difference in an analysis of results between observers. Third, the study imaged all patients using a 3D-APT sequence. When using 2D-multislice acquisition, the inherent contrast, which varies from slice to slice due to saturation loss caused by T1 relaxation between slices, can be problematic.30 The 3D-APT sequence has less interslice saturation loss and can provide uniform image contrast across the slice.30,31 Finally, APT studies need to consider the confounding issue of patient age.15 One of the results shown by Yu et al13 was that the brain metastases population was significantly older than the glioblastoma population. As a result, the contralateral normal APT signal value was lower in the metastases population than in the glioblastoma population.14 The present study used parameters that do not require consideration of age.16 Therefore, there was no change between the contralateral normal mean APT signal analyzed from the present case and the mean APT signal of each brain region derived from a previous study.16 Regarding reproducibility, it was shown that there is no change in the APT signal values when the parameter conditions and the MRI system used are the same.32

We believe that by comparing the mean tumor core APT signal value, and the differences, and change rate, between the tumor core and the normal APT signal value, glioblastoma can be distinguished from brain metastases with high accuracy. Specifically, if the mean tumor core APT signal value in glioblastoma is more than 3.0%, the difference between the tumor core APT signal value and the CNAWM APT signal value is 2.0%, and the rate of change is more than 170%, the tumor is likely to be glioblastoma. In addition, there was no significant difference in the ipsilateral normal white matter APT signal value, as related to the tumor, compared with the CNAWM APT signal value. This may help predict the spread of tumor cells from the APT signal value and plan the maximum tumor resection in surgery by referring to the Gd-T1–weighted images simultaneously.

In the 3 cases of recurrence among 13 glioblastoma cases we examined, we could evaluate the transition from postoperative to recurrence based on changes in APT signal values. Although statistical proof has not been performed, it suggests the possibility of an alternative follow-up to Gd-T1W in the future (Figs. 6,7).

F6
FIGURE 6:
Recurrent cases of glioblastoma. An MRI scan was performed on a 26-year-old man (after open head biopsy). Preoperative MRI showing ring-shaped contrast enhancement on Gd-T1W and enhanced structures at the margins of the cystic cavity (A). Amide proton transfer–weighted images show heterogeneous high signal tumors (tumor core: APTmean = 3.58%, PE: APTmean = 2.71%, INAWM: APTmean = 1.28%, CNAWM: APTmean = 1.33%; B). Hematoxylin-eosin–stained section (×20) of the tumor core. Postoperatively, glioblastoma was confirmed histopathologically (C): patient in Fig. 1. Figure 6 can be viewed online in color at www.jcat.org.
F7
FIGURE 7:
Recurrent cases of glioblastoma (patient in Fig. 1). In February 2021, FLAIR high signal was observed in the left parietal lobe (A), and the APT signal value in the same area was 1.27% (B). Gd-T1W showed no obvious contrast effect (C). In June 2021, the FLAIR high signal area in the left parietal lobe was increased (D), and the APT signal value in the same area was 1.90% (E). Gd-T1W showed no obvious contrast effect (F), but recurrence was strongly suspected based on the change in the APT signal value. Figure 7 can be viewed online in color at www.jcat.org.

Our study has the following limitations. First, the target sample was small and the diseases studied were limited. Only grade IV glioblastoma was included, and all brain metastases were adenocarcinoma. Thus, larger sample sizes and case comparisons are needed to confirm the relationship between normal APT signal values and disease in the future. Second, it should be mentioned that the results were obtained in a single institution. Because APT studies may show different results depending on the imaging conditions and equipment environment, caution should be exercised in making comparisons between sites. In the future, multicenter studies are needed to establish the correlation with clinical outcomes. Third, the correlation with APT imaging is unclear because histopathological data on PE were lacking. Fourth, artifacts due to CSF flow and pulsation and partial volume effects might have affected the APT signal measurement in the ROI. Finally, manual ROI settings may have had a small impact on the reproducibility of measurement results. In the future, it is necessary to use automatic segmentation to clear tissue exclusion.

In conclusion, performing APT imaging using the same parameters as used in this study may help in the identification of brain tumors. Using similar parameters to those used in our study could simplify the reproducibility and comparison of results across centers. We hope that our findings may aid in the standardization of parameters for APT signal assessment in just a few percent of water. We would also note that until APT imaging parameters are standardized, caution should be exercised when comparing APT imaging results across sites.

ACKNOWLEDGMENTS

The authors acknowledge the valuable guidance with Dr Saito Akira, Department of Pathology, Tokyo Bay Urayasu Ichikawa Medical Center, Chiba. The authors thank the staff of the Department of Radiology at Tokyo Bay Urayasu Ichikawa Medical Center for helping with the MR imaging data acquisition. The authors also thank the anonymous reviewers for their valuable comments.

REFERENCES

1. Zhou J, Lal B, Wilson DA, et al. Amide proton transfer (APT) contrast for imaging of brain tumors. Magn Reson Med. 2003;50:1120–1126.
2. Zhou J, Blakeley JO, Hua J, et al. Practical data acquisition method for human brain tumor amide proton transfer (APT) imaging. Magn Reson Med. 2008;60:842–849.
3. Togao O, Yoshiura T, Keupp J, et al. Amide proton transfer imaging of adult diffuse gliomas: correlation with histopathological grades. Neuro Oncol. 2014;16:441–448.
4. Zhou J, Zhu H, Lim M, et al. Three-dimensional amide proton transfer MR imaging of gliomas: initial experience and comparison with gadolinium enhancement. J Magn Reson Imaging. 2013;38:1119–1128.
5. Zhou J, Tryggestad E, Wen Z, et al. Differentiation between glioma and radiation necrosis using molecular magnetic resonance imaging of endogenous proteins and peptides. Nat Med. 2011;17:130–134.
6. Wang M, Hong X, Chang CF, et al. Simultaneous detection and separation of hyperacute intracerebral hemorrhage and cerebral ischemia using amide proton transfer MRI. Magn Reson Med. 2015;74:42–50.
7. Wang R, Li SY, Chen M, et al. Amide proton transfer magnetic resonance imaging of Alzheimer's disease at 3.0 Tesla: a preliminary study. Chin Med J (Engl). 2015;128:615–619.
8. Li C, Peng S, Wang R, et al. Chemical exchange saturation transfer MR imaging of Parkinson's disease at 3 Tesla. Eur Radiol. 2014;24:2631–2639.
9. Li C, Wang R, Chen H, et al. Chemical exchange saturation transfer MR imaging is superior to diffusion-tensor imaging in the diagnosis and severity evaluation of Parkinson's disease: a study on substantia nigra and striatum. Front Aging Neurosci. 2015;7:198.
10. Dula AN, Asche EM, Landman BA, et al. Development of chemical exchange saturation transfer at 7T. Magn Reson Med. 2011;66:831–838.
11. By S, Barry RL, Smith AK, et al. Amide proton transfer CEST of the cervical spinal cord in multiple sclerosis patients at 3T. Magn Reson Med. 2018;79:806–814.
12. Mehrabian H, Desmond KL, Soliman H, et al. Differentiation between radiation necrosis and tumor progression using chemical exchange saturation transfer. Clin Cancer Res. 2017;23:3667–3675.
13. Yu H, Lou H, Zou T, et al. Applying protein-based amide proton transfer MR imaging to distinguish solitary brain metastases from glioblastoma. Eur Radiol. 2017;27:4516–4524.
14. Kamimura K, Nakajo M, Yoneyama T, et al. Histogram analysis of amide proton transfer-weighted imaging: comparison of glioblastoma and solitary brain metastasis in enhancing tumors and peritumoral regions. Eur Radiol. 2019;29:4133–4140.
15. Zhang Z, Zhang C, Yao J, et al. Amide proton transfer-weighted magnetic resonance imaging of human brain aging at 3 Tesla. Quant Imaging Med Surg. 2020;10:727–742.
16. Sugawara K, Miyati T, Ueda R, et al. Quantitative analysis of mobile proteins in normal brain tissue by amide proton transfer imaging: age dependence and sex differences. J Comput Assist Tomogr. 2021;45:277–284.
17. Zhou J, Heo HY, Knutsson L, et al. APT-weighted MRI: techniques, current neuro applications, and challenging issues. J Magn Reson Imaging. 2019;50:347–364.
18. Louis DN, Ohgaki H, Wiestler OD, et al. The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol. 2007;114:97–109.
19. Togao O, Keupp J, Hiwatashi A, et al. Amide proton transfer imaging of brain tumors using a self-corrected 3D fast spin-echo Dixon method: comparison with separate B0 correction. Magn Reson Med. 2017;77:2272–2279.
20. Oh J, Cha S, Aiken AH, et al. Quantitative apparent diffusion coefficients and T2 relaxation times in characterizing contrast enhancing brain tumors and regions of peritumoral edema. J Magn Reson Imaging. 2005;21:701–708.
21. Tan Y, Wang XC, Zhang H, et al. Differentiation of high-grade-astrocytomas from solitary-brain-metastases: comparing diffusion kurtosis imaging and diffusion tensor imaging. Eur J Radiol. 2015;84:2618–2624.
22. Nicholas M, Prados M, et al. Malignant astrocytomas. In: Black P, Loeffler J, eds. Cancer of the Nervous System. Cambridge, United Kingdom: Blackwell Science; 1997:p464–p491.
23. Goplen D, Bougnaud S, Rajcevic U, et al. αB-crystallin is elevated in highly infiltrative apoptosis-resistant glioblastoma cells. Am J Pathol. 2010;177:1618–1628.
24. Weber MA, Zoubaa S, Schlieter M, et al. Diagnostic performance of spectroscopic and perfusion MRI for distinction of brain tumors. Neurology. 2006;66:1899–1906.
25. Sunwoo L, Yun TJ, You SH, et al. Differentiation of glioblastoma from brain metastasis: qualitative and quantitative analysis using arterial spin labeling mr imaging. PLoS One. 2016;11:e0166662.
26. Bohara M, Kamimura K, Nakajo M, et al. Amide proton transfer imaging of cavernous malformation in the cavernous sinus. Magn Reson Med Sci. 2019;18:109–110.
27. Togao O, Hiwatashi A, Keupp J, et al. Amide proton transfer imaging of diffuse gliomas: effect of saturation pulse length in parallel transmission-based technique. PLoS One. 2016;11:e0155925.
28. Cox SR, Ferguson KJ, Royle NA, et al. A systematic review of brain frontal lobe parcellation techniques in magnetic resonance imaging. Brain Struct Funct. 2014;219:1–22.
29. Tisserand DJ, Pruessner JC, Sanz Arigita EJ, et al. Regional frontal cortical volumes decrease differentially in aging: an MRI study to compare volumetric approaches and voxel-based morphometry. Neuroimage. 2002;17:657–669.
30. Zhao X, Wen Z, Zhang G, et al. Three-dimensional turbo-spin-echo amide proton transfer MR imaging at 3-tesla and its application to high-grade human brain tumors. Mol Imaging Biol. 2013;15:114–122.
31. Zhu H, Jones CK, van Zijl PC, et al. Fast 3D chemical exchange saturation transfer (CEST) imaging of the human brain. Magn Reson Med. 2010;64:638–644.
32. Lee JB, Park JE, Jung SC, et al. Repeatability of amide proton transfer-weighted signals in the brain according to clinical condition and anatomical location. Eur Radiol. 2020;30:346–356.
Keywords:

amide proton transfer; contralateral normal white matter; ipsilateral normal-appearing white matter; peritumoral edema

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.