Diffusion tensor image analysis along the perivascular space may serve as a potential biomarker for differentiating glioma recurrence from pseudoprogression: a preliminary study

Introduction

Glioma is the most common primary brain tumor in adults (1). Currently, the standardized clinical treatment regimen includes maximal surgical resection, concurrent chemoradiotherapy, and chemotherapy with temozolomide (TMZ) (2). The emergence of new or persisting enhancing lesions on magnetic resonance imaging (MRI) within the first 6 months after the completion of chemoradiotherapy presents a diagnostic challenge, as these lesions may indicate either tumor recurrence (TR) or pseudoprogression (PsP). Clinically, accurate differentiation between TR and PsP through a non-invasive approach is of vital importance.

Diffusion tensor imaging (DTI) captures microstructural alterations and provides useful information on the proliferation activity and genetic characteristics of gliomas (3-6). Moreover, previous studies have demonstrated that DTI metrics are promising in differentiating TR from PsP (7-9). However, controversial findings across studies have indicated that DTI metrics alone may be insufficient for accurate discrimination (7,10).

The glymphatic system (GS) plays a key role in maintaining brain extracellular homeostasis and clearing metabolic waste products via the perivascular pathway (11). In this pathway, cerebrospinal fluid (CSF) flows along the perivascular space (PVS), enters the brain parenchyma through aquaporin-4 (AQP4) water channels to exchange with interstitial fluid (ISF), and ultimately eliminates harmful solutes via perivenous pathways (12,13). Diffusion tensor image analysis along the PVS (DTI-ALPS) reflects the diffusivity along the PVSs around the medullary veins at the level of the lateral ventricle body (14). However, a study indicated that the ALPS index is an incomplete representation of GS function, as it is not exclusively reflective of PVS diffusion and is substantially confounded by underlying axonal geometry (15). Thus, the ALPS index may be useful for investigating fluid dynamics along axonal pathways but should be interpreted with caution in the context of glymphatic function (15). DTI-ALPS has shown potential as an imaging biomarker for many disorders such as Parkinson’s disease and Alzheimer’s disease (14,16-18). Previous studies have reported impaired CSF drainage through the GS in glioma (19). Several studies have used the ALPS index to investigate GS function in patients with treatment-naive glioma and revealed the potential of the ALPS index in glioma grading, genotyping, and prognosis prediction (20-23). The underlying mechanisms of GS dysfunction in glioma may involve tumor mass effect, blood-brain barrier (BBB) disruption, and abnormal AQP4 expression (23-25). Some studies implied that reduced glymphatic flow and clearance may cause toxic solutes to accumulate, hastening the growth of glioblastoma and providing a favourable environment for its progression (26). These studies suggest a potential connection between GS dysfunction and glioma recurrence, highlighting the necessity for further clinical investigations. Moreover, to the best of our knowledge, no previous studies have focused on the diagnostic value of the ALPS index in differentiating between TR and PsP. TR and PsP differ in microstructural alterations and pathological features, with TR often leading to tumor mass effect and BBB disruption (8). We hypothesized that the ALPS index in both hemispheres would be lower in TR than in PsP, and that the ALPS index could aid in distinguishing TR from PsP.

This preliminary study aimed to investigate the potential role of bilateral ALPS index in differentiating TR from PsP in glioma patients. Additionally, the added value of ALPS index to DTI metrics was further analyzed. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-2092/rc).

Methods

Study participants

Between October 2021 and December 2024, a total of 104 patients with pathologically confirmed unihemispheric glioma at initial diagnosis were retrospectively enrolled in this study. All patients had undergone gross total resection (GTR) followed by chemoradiotherapy and exhibited new or persisting enhancing lesions on follow-up MRI. The inclusion criteria were as follows: (I) patients had undergone GTR; (II) presence of new or persisting enhancing lesions after the completion of chemoradiotherapy; (III) conventional MRI and DTI performed at 2- to 3-month intervals; (IV) age range between 18 and 75 years. The exclusion criteria were as follows: (I) lack of DTI (n=5); (II) inadequate clinicoradiological follow-up (n=5); (III) poor image quality due to motion or susceptibility artifacts (n=1); (IV) severe brain deformation and completely damaged projection/association fibers in bilateral hemispheres (n=5); (V) enhancing lesions extending across the midline (n=16). In addition, age- and sex-matched healthy controls (HCs) were prospectively recruited, and these participants also underwent conventional MRI and DTI. The screening process for participants is shown in Figure 1.

Figure 1 The screening process for participants. CCRT, concurrent chemoradiation therapy; DTI, diffusion tensor imaging; GTR, gross total resection; MRI, magnetic resonance imaging; RT, radiotherapy.

This retrospective study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments, and was approved by the Institutional Review Board of the Nanfang Hospital of Southern Medical University (No. NFEC-202212-K3). Written informed consent from patients was waived due to its retrospective nature, and written informed consent was obtained from the HC volunteers.

Pathological diagnosis and reference criteria for diagnosis

All included patients had histologically confirmed glioma at initial diagnosis. Histologic diagnosis was established according to the 2016 World Health Organization (WHO) classification of tumors of the central nervous system (27). The isocitrate dehydrogenase (IDH) 1 and 2 alterations of the mutational hotspot codons R132 and R172 were investigated by Sanger sequencing (28). The O-6-methylguanine DNA methyltransferase (MGMT) promoter status was assessed by methylation-specific polymerase chain reaction through a two-step approach (29).

The final diagnosis of glioma recurrence or PsP was based on pathological confirmation after a second surgical operation or serial follow-up structural MRI. According to the Response Assessment in Neuro-Oncology (RANO) criteria, glioma recurrence was defined as the progressive enlargement of contrast-enhancing lesions, with at least a 25% increase in the sum of the products of perpendicular diameters of measurable lesions on more than two sequential follow-up MRI studies. PsP was diagnosed when enhancing lesions decreased or disappeared, accompanied by improvement in clinical symptoms during follow-up (30).

Imaging acquisition

MRI data were acquired on a 3.0 T scanner (Ingenia, Philips Healthcare) with a 16-channel head coil. The conventional imaging protocol included T1-weighted imaging, T2-weighted imaging, fluid-attenuated inversion recovery, and three-dimensional contrast-enhanced T1-weighted imaging (CE-T1WI). The details of the sequences are presented in Table S1. DTI was performed using a single-shot echo planar imaging sequence with the following parameters: TR 8657 ms; echo time (TE) 89 ms; field of view (FOV) 224×224 mm²; matrix size 112×110; slice thickness 2 mm; acquisition time 5 minutes 20 seconds; b =0, 800 s/mm²; and diffusion gradient encoding in 15 directions.

Image analysis

DTI data were preprocessed with MRtrix3 (https://www.mrtrix.org), including denoising, Gibbs ringing artifact removal, and corrections for eddy currents, participant motion, and bias field. Subsequently, fractional anisotropy (FA) maps, mean diffusivity (MD) maps, radial diffusivity (RD) maps, axial diffusivity (AD) maps, color-coded FA maps, and diffusivity maps in the direction of the x-axis (right-left, Dx), y-axis (anterior-posterior, Dy), and z-axis (inferior-superior, Dz) were generated using the FMRIB Software Library version 6.0 (https://www.fmrib.ox.ac.uk/fsl), Analysis of Functional NeuroImages software (https://afni.nimh.nih.gov) and MRtrix3 (23,31).

DTI-derived metrics and DTI-ALPS index calculation

Two neuroradiologists (with 16 and 4 years of experience, respectively) who were blinded to the clinicopathological results manually defined regions of interest (ROI) with ITK-SNAP software (https://www.itksnap.org) at enhancing lesions on CE-T1WI, avoiding necrosis, cystic changes, and hemorrhage. Subsequently, FA, MD, AD, and RD maps of all patients were nonlinearly registered to the corresponding CE-T1WI using the Symmetric Normalization (SyN) algorithm in the Advanced Normalization Tools (http://stnava.github.io/ANTs). The mean values of FA, MD, RD, and AD were calculated based on these ROIs using the FMRIB Software Library.

At the level of the lateral ventricle body, the medullary veins are approximately parallel to the x-axis and perpendicular to the lateral ventricle walls. This structure of medullary vessels within the brain parenchyma supports mass transport that is the movement of substances from and to the outside of the brain parenchyma (32). The PVS around the medullary veins (x-axis) is perpendicular to both the projection (z-axis) and the association (y-axis) fibers, which restrict diffusion across fibers (Dy, proj and Dz, assoc). The DTI-ALPS method was developed to evaluate water diffusivity along the PVS while minimizing the strong influence of white matter fibers on diffusion (32,33). The diffusivity along the x-axis in both the projection and association fibers (Dx, proj and Dx, assoc) measures the impact of diffusion of water molecules along the PVS (14,23). In order to compare values of coefficients across different cases, it is desirable to normalize the coefficients in ratio to some internal reference (32). Therefore, the DTI-ALPS index was determined using the following formula: DTI-ALPS index = mean (Dx, proj, Dx, assoc)/mean (Dy, proj, Dz, assoc). A low ALPS index suggests reduced water diffusivity along the x-axis. However, a lower ALPS index may reflect neurodegeneration rather than solely indicating reduced diffusivity along the PVS (34). Based on the color-coded FA maps, two squared ROIs, each consisting of 2×2 voxels over one 2-mm-thick slice (corresponding to an area of 4×4 mm²) were manually drawn on the association fibers and projection fibers in the unilateral hemisphere using ITK-SNAP software by the two neuroradiologists. This ROI configuration was adopted based on a recent study recommending squared ROIs of 4 voxels for ALPS, due to their alignment with the voxel grid, reduced partial volume effects, and high inter-method reliability (35). An example of ROI placement is shown in Figure 2. For glioma patients without ipsilateral lateral ventricle body deformation, ROIs were placed in both hemispheres (Figure 2A); for those with such deformation, ROIs were delineated only in the contralateral hemisphere (Figure 2F). When brain symmetry was compromised by tumors, ROIs were placed in adjacent slices or asymmetric locations as necessary.

Figure 2 Illustration of ROI placement and representative cases of glioma recurrence and pseudoprogression. ROIs were placed on projection fibers (left, red area; right, blue area) and association fibers (left, yellow area; right, purple area) along the lateral ventricle body in bilateral hemispheres in patient without obvious deformation of the ipsilateral lateral ventricle body (A), and in only contralateral hemisphere in patient with obvious deformation of ipsilateral lateral ventricle body (F). (B-E) Representative case of pseudoprogression. The lesion in the left frontal lobe shows hyperintensity on MD map (B), hypointensity on FA map (C), displaced fiber tracts on color-coded map (D) and enhancement on CE-T1WI (E). (G-J) Representative case of glioma recurrence. The lesion in the left insular lobe shows hypointensity on MD map (G), hyperintensity on FA map (H), disrupted fiber tracts on color-coded map (I) and heterogeneous enhancement on CE-T1WI (J). CE-T1WI, contrast-enhanced T1-weighted imaging; FA, fractional anisotropy; MD, mean diffusivity; ROI, region of interest.

Statistical analyses

All statistical analyses were performed using SPSS version 26 (IBM Corp., Armonk, NY, USA) and R version 4.4.1 (The R Foundation for Statistical Computing). The Shapiro-Wilk test was applied to assess the normality of continuous variables. Clinical characteristics were compared between the TR group and PsP group using the chi-square test for categorical variables, and either the Student’s t-test or Mann-Whitney U test for continuous variables as appropriate. Interobserver variability was assessed using the intraclass correlation coefficient (ICC) based on the two-way random model, absolute agreement type (36). The ICC values were interpreted as follows: <0.50, poor agreement; 0.50–0.75, moderate agreement; 0.75–0.90, good agreement; >0.90, excellent agreement (36).

Group comparisons of the ALPS index were conducted between the TR and HC groups, as well as between the PsP and HC groups, using the Student’s t-test or Mann-Whitney U test as appropriate. Effect sizes and corresponding 95% confidence intervals (CIs) were calculated to quantify the group separation. Since the ALPS index was also found to be negatively correlated with age (37), analysis of covariance (ANCOVA) with age adjustment was used to compare the ALPS index between the TR and PsP groups. Group comparisons of the DTI metrics were also conducted between the TR and PsP groups, using the Student’s t-test or Mann-Whitney U test as appropriate. Receiver operating characteristic (ROC) analysis was used to calculate the area under the curve (AUC), sensitivity, specificity and the cut-off value. Binary logistic regression analysis was used to construct integrated models combining the ALPS index and DTI metrics. The AUCs of single-parameter models and integrated models were compared using the DeLong test (38). Additionally, the integrated discrimination improvement (IDI) index and net reclassification index (NRI) were calculated to assess the added value of the ALPS index to DTI metrics (39). Additionally, correlations between DTI metrics and the ALPS index were analyzed using Spearman’s rank-order correlation because DTI metrics were not normally distributed. Statistical tests were two-sided, and P<0.05 was statistically significant.

Results

Baseline characteristics of the participants

A total of 72 glioma patients were eligible for inclusion, along with 34 age- and sex-matched HCs. The baseline demographic and clinical characteristics of the participants are summarized in Table 1. The baseline characteristics of the TR and PsP groups were not significantly different. Of the 72 patients, 35 (48.6%) had TR and 37 (51.4%) had PsP. Histopathological confirmation was available in 13 TR patients after repeat surgery, while the remaining patients were diagnosed based on follow-up MRI according to RANO criteria. In the TR group, 22 patients had lesions in the left hemisphere and 13 had lesions in the right hemisphere. In the PsP group, 17 patients had lesions in the left hemisphere and 20 had lesions in the right hemisphere. Representative MR images from patients with TR and PsP are shown in Figure 2, respectively.

Interobserver agreement for the ALPS index and DTI metrics measurements

The interobserver variability of measurements is reported in Table S2. The ICCs for the ALPS index, FA, MD, AD, and RD ranged from 0.698 to 0.966, showing moderate to excellent repeatability between the two observers. Thus, the final value of each metric was calculated as the average of the two independent measurements.

Differences in the ALPS index between glioma patients and HCs

The results of comparisons of the ALPS index in both cerebral hemispheres between glioma patient groups and matched HC groups are presented in Figure 3. Compared with the HC group, the TR group showed a significantly lower ALPS index in both the ipsilateral (1.22±0.16 vs. 1.55±0.17; P<0.001; Cohen’s d =1.96, 95% CI: 1.25–2.66) and contralateral hemispheres (1.38±0.19 vs. 1.53±0.22; P=0.003; Cohen’s d =0.74, 95% CI: 0.26–1.21). Compared with the HC group, the PsP group showed a significantly lower ALPS index in the ipsilateral hemisphere (1.39±0.19 vs. 1.60±0.23; P=0.001; Cohen’s d =0.96, 95% CI: 0.43–1.49), while a nonsignificant decrease was observed in the contralateral hemisphere (1.56±0.23 vs. 1.61±0.23; P=0.378; Cohen’s d =0.21, 95% CI: 0.24–0.66). There is no significant difference in the ALPS index between the right and left hemispheres in HC.

Figure 3 Comparison of the ALPS index in both cerebral hemispheres between glioma patient groups and matched HC groups. (A) ALPS index in patients with TR and matched HC; (B) ALPS index in patients with PsP and matched HC. ALPS, analysis along the perivascular space; HC, healthy control; PsP, pseudoprogression; TR, tumor recurrence.

Differences in the ALPS index and DTI metrics between TR and PsP

Compared with the PsP group, the TR group showed a significantly lower ALPS index in the ipsilateral hemisphere (P=0.001, partial ղ²=0.219) (Figure 4A). Furthermore, the contralateral ALPS index was significantly lower in the TR group than in the PsP group (P=0.001, partial ղ²=0.140) (Figure 4B). Comparisons of DTI metrics are presented in Table 2. Compared with the PsP group, the TR group showed significantly higher FA values (0.18 vs. 0.15, P=0.001) and significantly lower MD values (1.33 vs. 1.63, P=0.001), AD values (1.55 vs. 1.82, P=0.003), and RD values (1.21 vs. 1.49, P=0.001).

Figure 4 Comparison of the ipsilateral (A) and contralateral ALPS (B) between tumor recurrence and pseudoprogression. ALPS, analysis along the perivascular space; PsP, pseudoprogression; TR, tumor recurrence.

Diagnostic value of ALPS index and DTI metrics for differentiation between TR and PsP

The ROC curves for ALPS index and DTI metrics in differentiation between TR and PsP are shown in Figure 5, and the corresponding AUCs, sensitivities, and specificities are summarized in Tables 3,4. This is the first study to differentiate TR from PsP using the ALPS index. The AUC was 0.734 for the ipsilateral ALPS index and 0.706 for the contralateral ALPS index (Figure 5A). In the DTI univariate analysis, FA achieved the highest discriminatory performance between TR and PsP (AUC =0.728), followed by RD (AUC =0.725), MD (AUC =0.720), and AD (AUC =0.704) (Figure 5B).

Figure 5 Diagnostic performance for ALPS index and DTI metrics as assessed by ROC curves. (A) ROC curves for the ALPS index; (B) ROC curves for the DTI metrics; (C) ROC curves for combined ipsilateral ALPS index and DTI metrics; (D) ROC curves for combined contralateral ALPS index and DTI metrics. AD, axial diffusivity; ALPS, analysis along the perivascular space; DTI, diffusion tensor imaging; FA, fractional anisotropy; MD, mean diffusivity; RD, radial diffusivity; ROC, receiver operating characteristic.

The ALPS index and DTI metrics were incorporated into binary logistic regression models. Compared with corresponding DTI metric alone, the ipsilateral ALPS index combined with FA, MD, AD, or RD yielded improved AUC values of 0.852, 0.815, 0.789, and 0.821, respectively (Figure 5C). Although the DeLong test showed no significant AUC difference (P>0.05), all integrated models demonstrated significant improvements in discrimination (IDI: P<0.05). In addition, significant NRI values (NRI >0) were also observed in integrated models combining the ALPS index with MD, AD, or RD, while no significant NRI value (NRI >0) was found for the model integrating the ALPS index with FA (Table 3).

Compared with individual DTI metrics alone, the contralateral ALPS index combined with FA, MD, AD or RD also showed increased AUC values of 0.792, 0.810, 0.784 and 0.819, respectively (Figure 5D). The combination of RD and ALPS index significantly improved the diagnostic performance compared with RD alone (DeLong test, IDI and NRI; P<0.05). Other integrated models also demonstrated significant improvement in discrimination (IDI: P<0.05 and NRI: P<0.05), although the DeLong test showed no significant AUC differences (P>0.05) (Table 4).

Correlation between ALPS index and DTI metrics

None of the DTI metrics were correlated with the ipsilateral ALPS index (FA: rho =−0.130; MD: rho =0.177; AD: rho =0.156; RD: rho =0.167, all P>0.05) or the contralateral ALPS index (FA: rho =−0.228; MD: rho =0.100; AD: rho =0.093; RD: rho =0.101, all P>0.05). These results are shown in Table S3.

Discussion

Our results demonstrated that TR exhibited significantly lower bilateral ALPS indices compared with HC, whereas PsP showed a significant reduction only in the ipsilateral ALPS index. TR demonstrated lower ipsilateral and contralateral ALPS indices than PsP. Furthermore, bilateral ALPS indices showed good efficacy in distinguishing TR from PsP, and integrating ALPS indices with DTI metrics significantly improved discriminative ability.

GS dysfunction refers to impaired CSF influx from the PVS into the brain interstitium that may alter fluid exchange and perivenous ISF outflow (21). The DTI-ALPS index is a novel metric for evaluating diffusivity along the PVS (14). Recent studies reported significantly lower ipsilateral ALPS indices in patients with treatment-naive glioma compared with matched HCs (21-23). To the best of our knowledge, the present study is the first to investigate alterations of ALPS index in glioma patients with TR and PsP. The ipsilateral ALPS index was significantly reduced in both TR and PsP compared with HC, indicating decreased perivascular CSF flow in glioma patients. The underlying mechanisms may be twofold. Firstly, vascular hyperplasia in TR lesions and increased vascular permeability in PsP lesions both promote plasma extravasation, causing ISF retention and impaired perivascular fluid flow (40). Secondly, a rodent study indicated that surgery may exacerbate glymphatic impairment due to AQP4 depolarization (41). Tian et al. reported reduced postoperative ALPS indices in both hemispheres of glioma patients compared with preoperative levels, indicating possible damage to GS function following surgery (42). Additionally, sleep disorders are associated with a reduced ALPS index (43). As sleep disturbances are common in glioma patients (44), this may partly account for the lower ALPS index in glioma patients.

Of note, TR demonstrated a significantly lower ipsilateral ALPS index compared with PsP in this study. Several pathophysiological mechanisms may underlie this difference. Firstly, tumor mass, peritumoral edema, and infiltration in glioma can increase intracranial pressure (ICP), leading to elevated pressure within the PVS and brain interstitium (45). This hinders CSF influx along the PVS, CSF-ISF exchange and ISF outflux, decelerating perivascular CSF flow (23). Additionally, tumor cells infiltrate surrounding brain tissues, compressing and distorting the PVS, which may physically obstruct glymphatic pathways and impair CSF outflow (24). Recurrent gliomas exhibit higher neoplastic cellularity and aggressive biologic behaviors, compared with PsP (46). Consequently, the lower ALPS index in TR may be attributed to recurrent tumor growth. Secondly, tumor infiltration and tumor-associated inflammation may disrupt the BBB and impair GS function. Dynamic contrast-enhanced MRI studies have demonstrated higher BBB permeability in TR than in PsP (47,48). In recurrent gliomas, immature angiogenesis could disrupt the BBB and compromise fluid exchange. In contrast, PsP involves transient inflammation-related BBB hyperpermeability without pathological angiogenesis, leading to only temporary CSF-ISF exchange impairment (8). Thirdly, alterations in AQP4 expression or function may contribute to glymphatic dysfunction (19). AQP4, a key water channel protein expressed at astrocytic endfeet, plays a critical role in glymphatic transport by allowing CSF influx into the brain parenchyma, regulating CSF-ISF exchange, and facilitating perivenous clearance (19). Evidence from clinical and experimental studies suggests that AQP4 dysfunction in glioma is associated with impaired glymphatic transport (19,24). Recurrent glioma induces widespread inflammatory responses throughout the glioma-infiltrated cortex, releasing proinflammatory cytokines and chemokines that may disrupt AQP4 function (49). The lower ALPS index in TR may be partly attributed to lower AQP4 expression levels and impaired AQP4 function relative to PsP. However, histopathological evidence of AQP4 expression in TR and PsP lesions is lacking, and differences between the two remain unknown. Further investigations are warranted to elucidate the relationship between the ALPS index and AQP4 expression levels.

For contralateral ALPS index analysis, a significant reduction was observed in TR compared with HC, but not in PsP. Furthermore, the contralateral ALPS index was significantly lower in TR than in PsP. These findings may be attributed to several aspects. Firstly, glioma has been identified as a systemic disease that can influence the functional connectivity and structural alterations even in the distant contralateral regions (21,50). Zeng et al. identified a lower contralateral ALPS in glioma patients compared with HC and indicated that increased ICP resulting from the tumor burden and associated edema may cause this reduction (23). Increased ICP may lead to widespread insufficient brain blood supply and further cause cerebral ischemia and edema, thereby impairing contralateral GS function (45). Moreover, Alibabaei et al. reported microstructural changes in the contralateral hemisphere of glioblastoma patients using different diffusion models (51). A rodent study demonstrated the immature vasculature in both the tumor-bearing and contralateral hemispheres (19). Recurrent glioma cells may infiltrate contralateral normal-appearing white matter, promoting immature angiogenesis and thereby impairing CSF outflow. Yet, to our knowledge, histological confirmation of contralateral hemisphere infiltration in glioma patients remains lacking, highlighting the need for further studies. Finally, some studies suggested that the compensation of the GS activity may increase the metabolic waste clearance in glioma and mild traumatic brain injury (21,25,52,53). Hu et al. confirmed extensive remodeling of dorsal meningeal lymphatic vessels, the downstream of the glymphatic pathway, in glioma-bearing mice (54). Hence, contralateral ALPS index was slightly reduced in PsP, which indicated that PsP may preserve contralateral GS function via remodeling of glymphatic pathways. In contrast, recurrent gliomas may limit such compensation by impairing lymphangiogenesis (23).

DTI provides a noninvasive method for multiparametric (FA/MD/AD/RD) evaluation of intratumoral microstructure, which has been extensively applied to distinguish TR from PsP (5,10). FA values reflect the integrity of fiber tracts and positively correlate with cell density, while MD values express the diffusion capacity of water molecules and negatively correlate with cell density (46). AD values reflect axonal injury, while RD values reflect demyelination and axonal disruption (22). In agreement with previous studies, our study showed that TR exhibited significantly lower MD values, RD values, and AD values, but higher FA values compared with PsP, and these DTI metrics demonstrated moderate diagnostic performance (7-9). This may be attributed to dense tumor cells and more severe white matter destruction in recurrent glioma, while almost no viable tumor cells are present in PsP (22). However, DTI metrics alone are insufficient to accurately differentiate TR from PsP because they are determined by multiple factors and their ability to detect microstructural deviations is limited (10,46).

In ROC analyses, the ALPS index demonstrated comparable efficacy to DTI metrics for differentiation between TR and PsP. Furthermore, compared with DTI metrics alone, combining the ipsilateral or contralateral ALPS index with DTI metrics improved the diagnostic performance. Previous studies demonstrated that the ALPS index added value to DTI metrics in glioma grading and genotyping (20,22). Their combination provides a more comprehensive understanding of the underlying pathophysiology, with the ALPS index indirectly reflecting glymphatic alterations and DTI metrics characterizing microstructural integrity. Therefore, their combination offers complementary value for the evaluation of treatment response. Additionally, no significant correlations were found between DTI metrics and the ALPS index in our study.

Recently, many studies have indicated that there are several pitfalls in the DTI-ALPS method (15,32,34,55). Firstly, the ALPS method may not represent the function of the GS as a whole because it evaluates glymphatic function in a limited region, which is the intrinsic weakness of the DTI-ALPS method. Secondly, several studies have pointed out that ALPS indices are influenced by white matter geometry (15,32,34,55). These anatomical and microstructural complexities may confound the interpretation of the ALPS index (15). Notably, radial asymmetry within white matter tracts declines with age and neurodegeneration, mirroring changes in the ALPS index (55). Therefore, reductions in the ALPS index should not be solely attributed to diffusivity alterations along the PVS, but also reflect age-related or neurodegeneration-associated white matter tract changes. Future studies are warranted to integrate advanced and complementary approaches to separate perivascular diffusion contributions from white matter structural influences and enable a more comprehensive assessment of glymphatic function.

Limitations

This preliminary study has a few limitations. Firstly, it was a retrospective study with a relatively small sample size that may have introduced inherent selection bias. Besides, only patients with ipsilateral recurrence were included, and ALPS index changes in cases of contralateral recurrences remain to be investigated. Secondly, the susceptibility-weighted images were not included in this study, which made it difficult to ensure that the periventricular vessels were perpendicular to the lateral ventricle. Additionally, the manual placement of ROIs remains an inherent limitation of this study, as it needs to accommodate peritumoral edema, postoperative changes, and hemispheric asymmetry. Thirdly, pathological confirmation was not available in all cases. Therefore, prospective studies with larger sample sizes and stronger evidence of TR are warranted to validate our results. Finally, secondary psychological symptoms and sleep disorders, which may potentially influence the ALPS index, were not systematically evaluated in the present study. Future studies should incorporate neuropsychological assessments to evaluate the direct impact of cognitive symptoms on the ALPS index in glioma patients.

Conclusions

This preliminary study found that bilateral ALPS indices have the potential to serve as useful imaging biomarkers for distinguishing glioma recurrence from PsP and that they provide added diagnostic value to DTI metrics. However, a comprehensive interpretation of alterations in the ALPS index is necessary to understand its complex nature. Further investigations integrating multiple approaches are warranted to elucidate the glymphatic dynamics in glioma and the mechanisms underlying these findings.

Acknowledgments

We thank Yiming Wang (Bayer healthcare, Guangzhou, China) for insightful discussion and help with manuscript editing and language polishing.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-2092/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-2092/dss

Funding: This work was supported by the Natural Science Foundation of Guangdong Province, China (No. 2023A1515011453) and the Clinical Research Fund of Nanfang Hospital, Southern Medical University (No. 2023CR029).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-2092/coif). All authors report that this study received funding from the Natural Science Foundation of Guangdong Province, China (No. 2023A1515011453) and the Clinical Research Fund of Nanfang Hospital, Southern Medical University (No. 2023CR029). The authors have no other 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. This retrospective study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments, and was approved by the Institutional Review Board of the Nanfang Hospital of Southern Medical University (No. NFEC-202212-K3). Written informed consent from patients was waived due to its retrospective nature, and written informed consent was obtained from the HC volunteers.

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. Miller KD, Ostrom QT, Kruchko C, Patil N, Tihan T, Cioffi G, Fuchs HE, Waite KA, Jemal A, Siegel RL, Barnholtz-Sloan JS. Brain and other central nervous system tumor statistics, 2021. CA Cancer J Clin 2021;71:381-406. [Crossref] [PubMed]
2. Stupp R, Brada M, van den Bent MJ, Tonn JC, Pentheroudakis G. High-grade glioma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2014;25:iii93-101. [Crossref] [PubMed]
3. Beppu T, Inoue T, Shibata Y, Yamada N, Kurose A, Ogasawara K, Ogawa A, Kabasawa H. Fractional anisotropy value by diffusion tensor magnetic resonance imaging as a predictor of cell density and proliferation activity of glioblastomas. Surg Neurol 2005;63:56-61; discussion 61. [Crossref] [PubMed]
4. Kinoshita M, Hashimoto N, Goto T, Kagawa N, Kishima H, Izumoto S, Tanaka H, Fujita N, Yoshimine T. Fractional anisotropy and tumor cell density of the tumor core show positive correlation in diffusion tensor magnetic resonance imaging of malignant brain tumors. Neuroimage 2008;43:29-35. [Crossref] [PubMed]
5. Xiong J, Tan WL, Pan JW, Wang Y, Yin B, Zhang J, Geng DY. Detecting isocitrate dehydrogenase gene mutations in oligodendroglial tumors using diffusion tensor imaging metrics and their correlations with proliferation and microvascular density. J Magn Reson Imaging 2016;43:45-54. [Crossref] [PubMed]
6. Price SJ, Peña A, Burnet NG, Jena R, Green HA, Carpenter TA, Pickard JD, Gillard JH. Tissue signature characterisation of diffusion tensor abnormalities in cerebral gliomas. Eur Radiol 2004;14:1909-17. [Crossref] [PubMed]
7. Feng A, Yuan P, Huang T, Li L, Lyu J. Distinguishing Tumor Recurrence From Radiation Necrosis in Treated Glioblastoma Using Multiparametric MRI. Acad Radiol 2022;29:1320-31. [Crossref] [PubMed]
8. Wang S, Martinez-Lage M, Sakai Y, Chawla S, Kim SG, Alonso-Basanta M, Lustig RA, Brem S, Mohan S, Wolf RL, Desai A, Poptani H. Differentiating Tumor Progression from Pseudoprogression in Patients with Glioblastomas Using Diffusion Tensor Imaging and Dynamic Susceptibility Contrast MRI. AJNR Am J Neuroradiol 2016;37:28-36. [Crossref] [PubMed]
9. Razek AAKA, El-Serougy L, Abdelsalam M, Gaballa G, Talaat M. Differentiation of residual/recurrent gliomas from postradiation necrosis with arterial spin labeling and diffusion tensor magnetic resonance imaging-derived metrics. Neuroradiology 2018;60:169-77. [Crossref] [PubMed]
10. Wu XF, Liang X, Wang XC, Qin JB, Zhang L, Tan Y, Zhang H. Differentiating high-grade glioma recurrence from pseudoprogression: Comparing diffusion kurtosis imaging and diffusion tensor imaging. Eur J Radiol 2021;135:109445. [Crossref] [PubMed]
11. Iliff JJ, Wang M, Liao Y, Plogg BA, Peng W, Gundersen GA, Benveniste H, Vates GE, Deane R, Goldman SA, Nagelhus EA, Nedergaard M. A paravascular pathway facilitates CSF flow through the brain parenchyma and the clearance of interstitial solutes, including amyloid β. Sci Transl Med 2012;4:147ra111. [Crossref] [PubMed]
12. Rasmussen MK, Mestre H, Nedergaard M. The glymphatic pathway in neurological disorders. Lancet Neurol 2018;17:1016-24. [Crossref] [PubMed]
13. Plog BA, Nedergaard M. The Glymphatic System in Central Nervous System Health and Disease: Past, Present, and Future. Annu Rev Pathol 2018;13:379-94. [Crossref] [PubMed]
14. Taoka T, Masutani Y, Kawai H, Nakane T, Matsuoka K, Yasuno F, Kishimoto T, Naganawa S. Evaluation of glymphatic system activity with the diffusion MR technique: diffusion tensor image analysis along the perivascular space (DTI-ALPS) in Alzheimer's disease cases. Jpn J Radiol 2017;35:172-8. [Crossref] [PubMed]
15. Schilling KG, Newton A, Tax C, Nilsson M, Chamberland M, Anderson A, Landman B, Descoteaux M. White Matter Geometry Confounds Diffusion Tensor Imaging Along Perivascular Space (DTI-ALPS) Measures. Hum Brain Mapp 2025;46:e70282. [Crossref] [PubMed]
16. Pang H, Wang J, Yu Z, Yu H, Li X, Bu S, Zhao M, Jiang Y, Liu Y, Fan G. Glymphatic function from diffusion-tensor MRI to predict conversion from mild cognitive impairment to dementia in Parkinson's disease. J Neurol 2024;271:5598-609. [Crossref] [PubMed]
17. Zhang J, Liu S, Wu Y, Tang Z, Wu Y, Qi Y, Dong F, Wang Y. Enlarged Perivascular Space and Index for Diffusivity Along the Perivascular Space as Emerging Neuroimaging Biomarkers of Neurological Diseases. Cell Mol Neurobiol 2023;44:14. [Crossref] [PubMed]
18. Zhang Q, Wei CG, Cui C, Huang T, Liu SW, Li M, Zhi YQ, Liang XY, Hu H, Jiang Z, Zhu JT, Liu R. Association of magnetic resonance imaging glymphatic function with gray matter volume loss and cognitive impairment in Alzheimer's disease: a diffusion tensor image analysis along the perivascular space (DTI-ALPS) study. Quant Imaging Med Surg 2025;15:6160-74. [Crossref] [PubMed]
19. Xu D, Zhou J, Mei H, Li H, Sun W, Xu H. Impediment of Cerebrospinal Fluid Drainage Through Glymphatic System in Glioma. Front Oncol 2021;11:790821. [Crossref] [PubMed]
20. Zhu H, Xie Y, Li L, Liu Y, Li S, Shen N, Zhang J, Yan S, Liu D, Li Y, Zhu W. Diffusion along the perivascular space as a potential biomarker for glioma grading and isocitrate dehydrogenase 1 mutation status prediction. Quant Imaging Med Surg 2023;13:8259-73. [Crossref] [PubMed]
21. Wang L, Liu L, Zhang H, Yue Q. Diffusion tensor imaging changes along the perivascular spaces may serve as a prognostic factor for high-grade glioma. Eur J Radiol 2025;184:111993. [Crossref] [PubMed]
22. Fang K, Wang C, Zhang Y, Wu S, Liu L, Wen X. Noninvasive assessment of glymphatic dysfunction and IDH mutation in glioma with DTI-ALPS and DTI metrics. Quant Imaging Med Surg 2025;15:5007-22. [Crossref] [PubMed]
23. Zeng S, Huang Z, Zhou W, Ma H, Wu J, Zhao C, Yang Z, Qiu H, Chu J. Noninvasive evaluation of the glymphatic system in diffuse gliomas using diffusion tensor image analysis along the perivascular space. J Neurosurg 2025;142:187-96. [Crossref] [PubMed]
24. Scalia G, Ferini G, Silven MP, Noto M, Costanzo R, Maugeri R, Iacopino DG, Zagardo V, Viola A, Nicoletti GF, Umana GE. Alterations of the Glymphatic System in Glioblastomas: A Systematic Review. Anticancer Res 2024;44:3223-30. [Crossref] [PubMed]
25. Gao M, Liu Z, Zang H, Wu X, Yan Y, Lin H, Yuan J, Liu T, Zhou Y, Liu J. A Histopathologic Correlation Study Evaluating Glymphatic Function in Brain Tumors by Multiparametric MRI. Clin Cancer Res 2024;30:4876-86. [Crossref] [PubMed]
26. Lan YL, Wang H, Chen A, Zhang J. Update on the current knowledge of lymphatic drainage system and its emerging roles in glioma management. Immunology 2023;168:233-47. [Crossref] [PubMed]
27. Louis DN, Perry A, Reifenberger G, von Deimling A, Figarella-Branger D, Cavenee WK, Ohgaki H, Wiestler OD, Kleihues P, Ellison DW. The 2016 World Health Organization Classification of Tumors of the Central Nervous System: a summary. Acta Neuropathol 2016;131:803-20. [Crossref] [PubMed]
28. Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N, Weller M, Kros JM, Hainfellner JA, Mason W, Mariani L, Bromberg JE, Hau P, Mirimanoff RO, Cairncross JG, Janzer RC, Stupp R. MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med 2005;352:997-1003. [Crossref] [PubMed]
29. Reed GH, Kent JO, Wittwer CT. High-resolution DNA melting analysis for simple and efficient molecular diagnostics. Pharmacogenomics 2007;8:597-608. [Crossref] [PubMed]
30. Park YW, Ahn SS, Kim EH, Kang SG, Chang JH, Kim SH, Zhou J, Lee SK. Differentiation of recurrent diffuse glioma from treatment-induced change using amide proton transfer imaging: incremental value to diffusion and perfusion parameters. Neuroradiology 2021;63:363-72. [Crossref] [PubMed]
31. Tatekawa H, Matsushita S, Ueda D, Takita H, Horiuchi D, Atsukawa N, Morishita Y, Tsukamoto T, Shimono T, Miki Y. Improved reproducibility of diffusion tensor image analysis along the perivascular space (DTI-ALPS) index: an analysis of reorientation technique of the OASIS-3 dataset. Jpn J Radiol 2023;41:393-400. [Crossref] [PubMed]
32. Taoka T, Ito R, Nakamichi R, Nakane T, Kawai H, Naganawa S. Diffusion Tensor Image Analysis ALong the Perivascular Space (DTI-ALPS): Revisiting the Meaning and Significance of the Method. Magn Reson Med Sci 2024;23:268-90. [Crossref] [PubMed]
33. Yu L, Hu X, Li H, Zhao Y. Perivascular Spaces, Glymphatic System and MR. Front Neurol 2022;13:844938. [Crossref] [PubMed]
34. Georgiopoulos C, Werlin A, Lasic S, Hall S, van Westen D, Spotorno N, Hansson O, Nilsson M. Diffusion tensor imaging along the perivascular space: the bias from crossing fibres. Brain Commun 2024;6:fcae421. [Crossref] [PubMed]
35. Ulloa P, Rudolf JC, Kremer J, Schmidt A, Schramm P. Influence of orientation, size and shape of the region of interest in diffusion MRI along perivascular spaces index. MAGMA 2025;38:761-9. [Crossref] [PubMed]
36. Koo TK, Li MY. A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. J Chiropr Med 2016;15:155-63. [Crossref] [PubMed]
37. Zhang Y, Zhang R, Ye Y, Wang S, Jiaerken Y, Hong H, Li K, Zeng Q, Luo X, Xu X, Yu X, Wu X, Yu W, Zhang M, Huang P. The Influence of Demographics and Vascular Risk Factors on Glymphatic Function Measured by Diffusion Along Perivascular Space. Front Aging Neurosci 2021;13:693787. [Crossref] [PubMed]
38. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837-45. [Crossref] [PubMed]
39. Pencina MJ, D'Agostino RB Sr, Demler OV. Novel metrics for evaluating improvement in discrimination: net reclassification and integrated discrimination improvement for normal variables and nested models. Stat Med 2012;31:101-13. [Crossref] [PubMed]
40. Liang W, Sun W, Li C, Zhou J, Long C, Li H, Xu D, Xu H. Glymphatic system dysfunction and cerebrospinal fluid retention in gliomas: evidence from perivascular space diffusion and volumetric analysis. Cancer Imaging 2025;25:51. [Crossref] [PubMed]
41. Zhu X, Lin J, Yang P, Wu S, Lin H, He W, Lin D, Cao M. Surgery induces neurocognitive disorder via neuroinflammation and glymphatic dysfunction in middle-aged mice with brain lymphatic drainage impairment. Front Neurosci 2024;18:1426718. [Crossref] [PubMed]
42. Tian B, Jiang X, Luo X, Zhang W. Analysis of the glymphatic system function in high-grade glioma patients using diffusion tensor imaging along perivascular spaces. BMC Neurol 2025;25:181. [Crossref] [PubMed]
43. Saito Y, Hayakawa Y, Kamagata K, Kikuta J, Mita T, Andica C, et al. Glymphatic system impairment in sleep disruption: diffusion tensor image analysis along the perivascular space (DTI-ALPS). Jpn J Radiol 2023;41:1335-43. [Crossref] [PubMed]
44. Chen Y, Yu W, Huang Y, Jiang Z, Deng J, Qi Y. Causal associations between sleep traits, sleep disorders, and glioblastoma: a two-sample bidirectional Mendelian randomization study. J Neurophysiol 2025;133:513-21. [Crossref] [PubMed]
45. Madhok DY, Vitt JR, Nguyen AT. Overview of Neurovascular Physiology. Curr Neurol Neurosci Rep 2018;18:99. [Crossref] [PubMed]
46. Wu X, Zhang M, Jiang Q, Li M, Wu Y. Diagnostic accuracy of magnetic resonance diffusion tensor imaging in distinguishing pseudoprogression from glioma recurrence: a systematic review and meta-analysis. Expert Rev Anticancer Ther 2024;24:1177-85. [Crossref] [PubMed]
47. Cao H, Huang Z, Qiu R, Xiao X, Li Z, Pillai JJ, Hua J, Huang G, Xu Y, Liang W, Wu Y. Predicting progression of enhancing non-measurable disease in high-grade glioma by using dynamic contrast-enhanced MR imaging. Magn Reson Imaging 2025;120:110394. [Crossref] [PubMed]
48. Yun TJ, Park CK, Kim TM, Lee SH, Kim JH, Sohn CH, Park SH, Kim IH, Choi SH. Glioblastoma treated with concurrent radiation therapy and temozolomide chemotherapy: differentiation of true progression from pseudoprogression with quantitative dynamic contrast-enhanced MR imaging. Radiology 2015;274:830-40. [Crossref] [PubMed]
49. Dadario NB, Boyett DM, Teasley DE, Chabot PJ, Winans NJ, Argenziano MG, Sperring CP, Canoll P, Bruce JN. Unveiling the Inflammatory Landscape of Recurrent Glioblastoma through Histological-Based Assessments. Cancers (Basel) 2024;16:3283. [Crossref] [PubMed]
50. Kinno R, Muragaki Y, Maruyama T, Tamura M, Tanaka K, Ono K, Sakai KL. Differential Effects of a Left Frontal Glioma on the Cortical Thickness and Complexity of Both Hemispheres. Cereb Cortex Commun 2020;1:tgaa027. [Crossref] [PubMed]
51. Alibabaei S, Yousefipour M, Rahmani M, Raminfard S, Tahmasbi M. Evaluating machine learning models for post-surgery treatment response assessment in glioblastoma multiforme: a comparative study of gray level co-occurrence matrix (GLCM), curvelet, and combined radiomics features selected by multiple algorithms. BMC Med Imaging 2025;25:362. [Crossref] [PubMed]
52. Villacis G, Schmidt A, Rudolf JC, Schwenke H, Küchler J, Schramm P, Ulloa P. Evaluating the glymphatic system via magnetic resonance diffusion tensor imaging along the perivascular spaces in brain tumor patients. Jpn J Radiol 2024;42:1146-56. [Crossref] [PubMed]
53. Dai Z, Yang Z, Li Z, Li M, Sun H, Zhuang Z, Yang W, Hu Z, Chen X, Lin D, Wu X. Increased glymphatic system activity in patients with mild traumatic brain injury. Front Neurol 2023;14:1148878. [Crossref] [PubMed]
54. Hu X, Deng Q, Ma L, Li Q, Chen Y, Liao Y, et al. Meningeal lymphatic vessels regulate brain tumor drainage and immunity. Cell Res 2020;30:229-43. [Crossref] [PubMed]
55. Wright AM, Wu YC, Chen NK, Wen Q. Exploring Radial Asymmetry in MR Diffusion Tensor Imaging and Its Impact on the Interpretation of Glymphatic Mechanisms. J Magn Reson Imaging 2024;60:1432-41. [Crossref] [PubMed]