Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Feb 24.
Published in final edited form as: Neuro Oncol. 2025 Nov 1;27(11):3016–3026. doi: 10.1093/neuonc/noaf136

Impact of corticosteroid administration on contrast-enhancing volume and diffusion MRI in treatment naïve glioblastoma

Francesco Sanvito 1,2, Asher Kim 3,4, Catalina Raymond 5,6, Ashley Teraishi 7,8, Richard G Everson 9, Phioanh L Nghiemphu 10, Albert Lai 11, Robert A Chong 12, David A Nathanson 13, Noriko Salamon 14, Timothy F Cloughesy 15,16, Jingwen Yao 17,18,19,†, Benjamin M Ellingson 20,21,22,23,24,†
PMCID: PMC12908483  NIHMSID: NIHMS2135573  PMID: 40443040

Abstract

Background.

Corticosteroids impact the radiographic interpretation of glioblastoma, including artificial reduction in the contrast-enhancing tumor volume and intensity (i.e., a “pseudoresponse”) and in the apparent diffusion coefficient (ADC). This study aimed to estimate the influence of corticosteroids on these measurements in treatment naïve glioblastoma before surgery.

Methods.

Fifty-seven pairs of MRI scans from 54 patients with presurgical treatment-naïve glioblastoma were retrospectively grouped as increased (n = 29, all corticosteroid-free at baseline), stable (n = 25), or decreased (n = 3) corticosteroid dose between scans (median interval: 15 days). Tumor size and ADC changes between timepoints were compared between lesions with increased and stable corticosteroids. Volumetric changes ascribable to increased corticosteroid dose was modeled, adjusting for the time between scans.

Results.

Increased corticosteroid dose showed an observed volumetric shrinkage of the contrast-enhancing tumor (median: −23.7%) and reduction in estimated growth rates (median: −2.48% per day), significantly different (P < .0001) from the control group receiving a stable dose (median: +36.0% volume; +2.08% growth rate). When adjusting for the time interval between scans, the estimated corticosteroid-induced volumetric shrinkage was 44.0% (P < .0001, 95% CI, 25.7%–62.2%). Increased corticosteroid dose also decreased ADC in the contrast-enhancing tumor (median reduction: 180, IQR = 39 – 281 × 10−6 mm2/s, P = .0005).

Conclusion.

Corticosteroid administration can induce a significant “pseudoresponse” in glioblastoma, with an observed reduction in contrast-enhancing tumor volume of 23.7% and a time interval adjusted reduction of 44.0% (25.7%–62.2%), and an ADC drop of 180 × 10−6 mm2/s (14.2%). These data confirm that radiographic measurements are impacted by corticosteroids and provide benchmarks for the development of adjusted response criteria accounting for corticosteroid use.

Keywords: glioblastoma, corticosteroids, glucocorticoids, pseudoresponse, magnetic resonance imaging

The measurement of the contrast-enhancing tumor component is a crucial aspect of monitoring glioblastoma in both clinical trials1,2 and in clinical practice.3–5 Indeed, the contrast-enhancing component is currently considered the most reliable radiographic surrogate of disease burden in glioblastoma.6–9

Glucocorticoids, commonly referred to as “corticosteroids” for simplicity, are often used for the clinical management of glioblastoma. While corticosteroids do not warrant a survival benefit10, their use is still recommended for the management of neurological symptoms due to their ability to reduce vasogenic edema and local inflammation.11 Historical evidence12–15 and empirical observations in clinical practice indicate that corticosteroid administration can mimic treatment response, as it can result in a fainter contrast-enhancement12–15 and, in some anecdotal cases, even its complete disappearance (i.e., “vanishing” glioblastoma).16–18 This can be considered as a “corticosteroid-induced pseudoresponse,” ascribable to reduction in blood-brain barrier (BBB) permeability rather than an actual cytotoxic effect,19 and therefore similar to bevacizumab-induced pseudoresponse.20

While the effects of corticosteroids on contrast-enhancement intensity and extent are overall acknowledged, quantitative data about the extent of corticosteroid-induced pseudoresponse are scarce. Changes of glioblastoma size over serial imaging are influenced by the interplay of tumor growth and response to concomitant therapies; therefore, it is often difficult to isolate and quantify the unique contribution from alterations in corticosteroid dose. Due to the lack of benchmark data, current RANO 2.0 criteria still require an increase in the contrast-enhancing tumor volume of ≥40% (or ≥25% in bidimensional products) to categorize progressive disease, even in case of corticosteroid dose increase.1,2 However, this rule may lead to a substantial delay in identifying progression in such cases, since the interplay between corticosteroid-induced pseudoresponse and underlying tumor growth may result in an overall <40% volumetric increase. This scenario may occur rather frequently, as a growing tumor may cause worsened neurological symptoms that require an increase in corticosteroid dose for clinical management. Additionally, quantitative data about corticosteroid-induced pseudoresponse would expand the clinical trial eligibility of recurrent glioblastoma, for which the demonstration of active growth is required by RANO 2.0 criteria.1,2

In addition to contrast enhancing tumor size, the apparent diffusion coefficient (ADC) estimated from diffusion MRI has emerged as a relevant biomarker for glioblastoma, in particular as a tool for molecular profiling21–24 and treatment response prediction and monitoring.25–29 Similar to contrast-enhancement, diffusion MRI can also be impacted by corticosteroid administration through reduction in vasogenic edema30 and the precise impact of corticosteroids on ADC has not been well documented.

In this study, we retrospectively evaluated a cohort of treatment-naïve glioblastoma with at least 2 MRI scans available prior to surgical resection, in order to quantify the impact of increasing corticosteroid dose on contrast-enhancing tumor volume and ADC within contrast-enhancing tumor. We hypothesize that an increase in corticosteroid dose will significantly decrease contrast-enhancing tumor volume, tumor growth rates, and ADC values.

Methods

Patient selection

We screened patients in 2 retrospective cohorts of patients with newly diagnosed glioblastoma: (1) glioma patients with corresponding human-derived murine models of glioblastoma who were enrolled in our institutional glioma modeling program (subcohort A, not previously published) and (2) a previously characterized subcohort reported by Ellingson et al.31 (subcohort B). Patients were imaged between March 2003 and September 2021. All patients gave written informed consent to participate in research studies, and the use of their dataset for research was approved by the institutional review board (IRB-10–0655). Inclusion criteria for the present study were as follows: newly diagnosed glioblastoma with at least 1 cc of contrast-enhancing tumor component, availability of postcontrast T1-weighted images at 2 (or more) separate presurgical timepoints (<45 days apart), available documentation reporting corticosteroid doses at both MRI timepoints, histopathological diagnosis of glioblastoma from the surgical samples collected during the subsequent surgical resection. All patients meeting these inclusion criteria were included in the analysis of contrast-enhancement changes. In addition, a subset of patients for whom diffusion MRI images were available at both timepoints was also included in the diffusion MRI analysis.

Patient categorization based on corticosteroid administration

Patients were categorized in 3 groups based on changes in corticosteroid dose between the 2 MRI timepoints: patients with increased, decreased, and stable corticosteroid dose. As per modified RANO and RANO 2.0 guidelines1,2: an “increased” corticosteroid dose was defined as an increase in dexamethasone equivalent dose of ≥2 mg/day between timepoints, a “decreased” dose was defined as a decrease in dexamethasone equivalent dose of ≥2 mg/day between timepoints, while changes <2 mg/day were labeled as “stable.” In rare cases for which 3 presurgical MRI timepoints were available, the 3 scans were analyzed as 2 pairs of timepoints (i.e., by comparing timepoint 1 with 2, and timepoint 2 with 3), and each pair was categorized according to corticosteroid dose changes.

Image acquisition and processing

Image acquisition parameters were heterogeneous due to the retrospective nature of this study. Imaging datasets were acquired on different scanners at 1.5 or 3 Tesla field strength. The acquisition protocols for pre- and postcontrast T1-weighted images (T1-pre and T1-post) included either spin-echo and gradient-echo techniques, and either 2D and 3D imaging (Supplementary Table 1). Similarly, T2-weighted images (T2) and/or T2-weighted FLAIR (T2-FLAIR) images were acquired with heterogeneous acquisition parameters and voxel geometry. When available, diffusion-weighted images with b-values 0 and 1000 s/mm2 were used to calculate apparent diffusion coefficient (ADC) maps (Supplementary Table 2). Apparent diffusion coefficient images were registered to T1-post using flirt (FSL; University of Oxford; https://fsl.fmrib.ox.ac.uk/fsl/), after skull stripping (Supplementary Figure 1). Segmentations of the contrast-enhancing lesion component were obtained by contouring the brain tumor region on ITK-snap (http://www.itksnap.org/),32 applying a T1-post intensity threshold with the FSL function fslmaths, and performing manual adjustments with ITK-snap when needed. This approach allowed to include in the segmentations only contrast-enhancing tumor voxels, therefore excluding cysts, nonenhancing components representing necrotic tissue, and nonenhancing T2-hyperintense components. These steps were conducted by a neuroradiologist (F.S.) with 8 years of experience in neuroimaging, blinded to the clinical information. Segmentations were used to estimate tumor volume and to extract median ADC values within the contrast-enhancing lesion (Supplementary Figure 1).

Radiographic reads

A neuroradiologist (F.S.), blinded to corticosteroid doses and categorization, evaluated each pair of MRI scans with a side-by-side presentation, and patients were presented in randomized order. First, the size of the contrast-enhancing lesion was measured with bidimensional products calculated from bidimensional diameters, as per RANO 2.01,2. Of note, while RANO 2.0 are designed for treated glioblastoma and suggest avoiding cystic/necrotic areas in the measurements, in this cohort of newly diagnosed glioblastoma it was inevitable to include cystic/necrotic components in the measurements. This was conducted as an alternative measurement of size, in addition to the volumetric segmentation. Furthermore, the neuroradiologist annotated whether a qualitative reduction in contrast-enhancement intensity and extent, and/or a reduction of peri-enhancing edema on T2-weighted images was visible on timepoint 2 compared with timepoint 1. These qualitative changes are expected to be associated with corticosteroid exposure.12,33

Statistical Analyses

Statistical analyses focused on patients with “increased” and “stable” corticosteroid dose, while cases with “decreased” dose of corticosteroids were excluded due to their small sample size. Due to the rare occurrence of 2 pairs of timepoints available from a single patient (n = 3 patients), in such cases, the 2 paired sets of observations were simply treated as independent in statistical analyses. Volumes, median ADC, and bidimensional products were compared between timepoints with Wilcoxon signed-rank tests. The size percentage change, the growth rates (percentage size change per day), and the absolute ADC change were compared between groups with a Mann–Whitney U test. The dose dependency of volumetric changes and ADC changes depending on the change in corticosteroid dose was assessed with a linear regression. To determine the difference in percentual volumetric change between groups, while adjusting for the time interval between timepoints, a multivariate regression was performed. In the multivariate regression, the percentual volumetric change was set as the dependent variable, both the group (1 = “increased” and 0 = “stable,” categorical variable) and the time interval (continuous variable) were used as independent variables, and the intercept was forced to 0%, as we expect 0% volumetric change with a time interval of 0 days for the group with stable corticosteroid dose. In this model, the regression coefficient β associated with the group variable can be read as the percent volume change ascribable to corticosteroid dose increase, adjusted for the time interval. All linear regression analyses (univariate and multivariate) were conducted after excluding outliers with a ROUT test (Q threshold set to 0.1, as by default). Qualitative radiographic changes were compared between groups with a Fisher’s exact test. All analyses were performed with GraphPad Prism version 8.4.3. Statistical significance was set to P-value < .05.

Results

Patient selection and characteristics

Fifty-seven pairs of MRI scans across 54 patients were selected for analysis (36 pairs from subcohort A, 21 pairs from subcohort B). The median time interval between paired MRI scans was 15 days (IQR 9–19). The majority of patients had 1 pair (2 timepoints) of presurgical MRI scans available, and only 3 patients had 2 pairs (3 timepoints) available. The corticosteroid dose was “increased” between scans for 29 pairs of scans, “stable” for 25 pairs, and “decreased” for 3 cases. All cases with an increased dose of corticosteroids were corticosteroid-free at timepoint 1 and were started on corticosteroids between the 2 timepoints. Most of the cases with stable dose of corticosteroids (n = 21) were corticosteroid-free at both timepoints 1 and 2, whereas only 4 pairs of MRI scans (2 from unique patients, 2 from follow-up imaging of patients included in the “increased” group) were acquired from patients on corticosteroids, initiated before timepoint 1. Among cases with increased corticosteroid dose, 2 cases received IV corticosteroids a few hours before timepoint 2. These cases were included in all analyses except the dose-dependency analysis, which was only conducted on cases with a daily oral corticosteroid dose reported in mg/day.

In the 54 patients retained for analyses, a total of 61% of patients were male (33 males vs 21 females), and the median age was 62 years (IQR 54–69). Thirty-three patients (subcohort A) received a histopathological and molecular diagnosis of IDH wild-type glioblastoma according to WHO 2021,34 while 21 (subcohort B) received only a histopathological diagnosis of glioblastoma without IDH testing, as they underwent surgery prior to the updated classification criteria and tissue was no longer available for testing. 35 pairs of MRI scans had diffusion MRI datasets available at both timepoints (18 labeled as “increased” corticosteroid dose, and 17 “stable”) and were eligible for diffusion MRI analysis. Corticosteroid doses, baseline tumor volumes, and symptoms at presentation are summarized in Supplementary Table 3.

Qualitative corticosteroid-induced radiographic changes

As expected, in treatment naïve glioblastoma prior to initial surgical resection, corticosteroid administration typically resulted in a size reduction of the contrast-enhancing tumor size, with fainter contrast-enhancement, consistent with a “corticosteroid-induced pseudoresponse,” associated with a decrease in ADC within the contrast-enhancing tumor. In blinded radiographic reads, a clear visual “decrease” in contrast-enhancement intensity and extent was more prevalent in the group with corticosteroid dose increase, compared with the control group (51.7% of cases vs 13.6% of cases, P = .003, Supplementary Figure 2A). Similarly, also a clear visual “reduction” in peri-enhancing edema tended to be more prevalent in the group with corticosteroid dose increase, compared with the control group (55.5% of cases vs 14.3% of cases, P = .064, Supplementary Figure 2B).

Figure 1A and B illustrates cases of a 69-year-old woman and a 50-year-old man exhibiting this generalized pattern, with overall reduction in the size of the contrast-enhancing lesion after administration of corticosteroids along with a thinner and fainter apparent rim of contrast enhancement (arrows in Figure 1A and B) surrounding cystic/necrotic areas. This was further accompanied by clear areas of reduced ADC within contrast-enhancing tumor regions (curved arrows in Figure 1A and B). Similarly, Figure 1C illustrates the case of a 54-year-old man exhibiting comparable thinning of the contrast-enhancement rim (arrows in Figure 1C) and ADC decrease in the primary contrast-enhancing lesion (curved arrows in Figure 1C). Of note, new areas of nodular contrast-enhancement appeared on the second scan (arrowheads in Figure 1C), consistent with continuous tumor growth even in the presence of corticosteroids, with an increase in volume around 16%, far below the threshold for disease progression via RANO 2.01,2. Arguably, this example suggests that the administration of corticosteroids results in the underestimation of underlying tumor growth, obscuring, and perhaps preventing the identification of true tumor progression.

Figure 1.

Figure 1.

Open in a new tab

Representative cases exhibiting typical radiographic changes associated with increased dose of corticosteroids. Corticosteroid dose increase determines a thinning of the contrast-enhancing rim and nodular components (arrows), which is visible both in cases with overall tumor shrinkage (A, B) and in cases with underlying tumor growth and progression (C, where tumor progression is also demonstrated by the appearance of a new contrast-enhancing nodule, indicated by the arrowheads). On ADC maps, corticosteroid dose increase determines a reduction of the ADC values within the contrast-enhancing tissue (curved arrows).

Figure 2A shows a complex case of a 55-year-old man with 3 MRI scans prior to surgery with varying doses of corticosteroids. This patient was started on 16 mg/day of dexamethasone after the initial MRI and then received a reduced dose of 8 mg/day after the second MRI and prior to the third exam. Interestingly, tumor pseudoresponse to alterations in steroid dose was heterogeneous. Some contrast-enhancing components within the rim and in the genu of the corpus callosum (arrows in Figure 2A) showed a pronounced size reduction, which was maintained even after dexamethasone dose reduction, while components in the centrum semiovale were stable after dexamethasone start and then increased remarkably after corticosteroid dose was reduced (arrowheads in Figure 2A). Figure 2B shows a patient imaged on 2 consecutive days, with IV corticosteroid injection during the night between the scans, which allows to visualize corticosteroid effects with negligible underlying tumor growth. Figure 2C shows a case of a 63-year-old man with a “vanishing glioblastoma” (Figure 2C), as described in other articles.16–18 In this case, the residual lesion is hypointense in postcontrast T1-weighted images, with only some partial and thin residuals of rim contrast-enhancement (curved arrows in Figure 2C).

Figure 2.

Figure 2.

Open in a new tab

Representative cases with peculiar scenarios. In case of corticosteroid dose increase (second scan in A) followed by a dose decrease (third scan in A), some contrast-enhancing tumor regions maintain the corticosteroid-induced pseudoresponse appearance with thinner rim and thinner nodular areas (arrows in A) even in presence of a reduced corticosteroid dose, while other areas show contrast-enhancement enlargement (arrowheads in A) as a result of both underlying tumor growth and underlying attenuated pseudoresponse. The evaluation of 2 scans with only one day of time interval offers a rare occasion to observe pure IV corticosteroid effects, arguably in the absence of underlying tumor growth (B). The findings are consistent with a typical corticosteroid-induced pseudoresponse, exhibiting thinner contrast-enhancing rims as well as resolution of contrast-enhancement in some nodular areas (arrows in B). A rare example of “vanishing glioblastoma” pattern was also seen in this cohort (C), where contrast-enhancement almost completely disappears after corticosteroids, and only an incomplete residual rim remains (curved arrows in C).

Quantitative impact of corticosteroids on contrast-enhancing volume

An increase in corticosteroid dose resulted in an overall shrinkage of the contrast-enhancing tumor volume (median −4.14 cc, P = .004, red in Figure 3A), while a stable corticosteroid dose (the “control” condition) was associated with a generalized growth in contrast-enhancing tumor volume (median +2.87 cc, P < .0001, blue in Figure 3A). Similarly, a median of 23.7% shrinkage (median −23.7%, IQR: −79.7% to +4.3%) in tumor volume was observed for patients with increasing corticosteroid dose and a 36.0% growth in tumor volume was observed in the control group with stable corticosteroid dose (Figure 3B). A direct group comparison revealed a significant difference both in univariate analysis (P < .0001, Figure 3B), and in multivariate analysis adjusting for the time interval between MRI scans (P < .0001, R2 = 0.34, Figure 3B). Importantly, this multivariate analysis revealed an estimated 44.0% shrinkage in contrast-enhancing volume ascribable to corticosteroid dose increase (95% CI, 62.2%–25.7%).

Figure 3.

Figure 3.

Open in a new tab

Impact of corticosteroids on contrast-enhancing volume and diffusion. The contrast-enhancing tumor shrank in the group with corticosteroid dose increase (red in A and F, all corticosteroid-free on timepoint 1), while it grew in the control group with stable dose (blue in A and F), with a significant difference in percentual volumetric change (B) and percentual volumetric growth rates (C) between groups. Volumetric changes (D) and volumetric growth rates (E) showed a corticosteroid dose dependency, after excluding the outliers (orange) from the analysis. ADC decreased between timepoints in the group with corticosteroid dose increase (red in G), but not in the control group (blue in G), with a significant difference between groups (H), and with dose dependency (I).

Similarly, patients with increased corticosteroids exhibited a “negative” growth rate (median −2.48% per day), while volumetric growth rates for the control group were “positive” (median +2.08% per day), with a significant difference between groups (P < .0001, Figure 3C). Importantly, this decrease in apparent growth rates during administration of increased steroid dose was in addition to any underlying tumor growth, as shown in the control group. The extent of both volumetric changes (P < .0001, R2 = 0.39, Figure 3D) and volumetric growth rates (P < .0001, R2 = 0.32, Figure 3E) showed a dependency on corticosteroid dose change.

Quantitative impact of corticosteroids on bidimensional tumor measurements

While not completely RANO compliant, as necrotic and/or cystic components were included in the measurement, bidimensional measurements were not significantly different between scans for patients with increasing steroid (median −0.02 cm2, P = .77, red in Supplementary Figure 3A), while the control group with stable corticosteroids showed a significant interval growth (median +2.24 cm2, P < .0001, blue in Supplementary Figure 3A). As a result, patients with increased corticosteroid dose exhibited a significantly smaller percent change in bidimensional size (median −0.2% vs +33.8%, P < .0001, Supplementary Figure 3B) and a smaller growth rate (median −0.01% vs +2.9% per day, P < .0001, Supplementary Figure 3C) compared with the control group of patients with no change in corticosteroid dose between scans.

Quantitative impact of corticosteroids on diffusion MRI

Patients with increase in corticosteroid dose showed a significant “reduction” in ADC values within the contrast-enhancing tumor between timepoints (median reduction 180 × 10−6 mm2/s with IQR = 39 – 281 × 10−6 mm2/s, P = .0005, red in Figure 3G), corresponding to a 14.2% median reduction (IQR 2.8%–21.1%). Conversely, a stable corticosteroid dose was not associated with significant ADC changes between timepoints (median + 30 × 10−6 mm2/s, P = .28, blue in Figure 3G). When directly comparing the 2 groups, increasing corticosteroid dose was associated with a significant ADC reduction compared with the control group with stable corticosteroid dose (difference between medians: −210 × 10−6 mm2/s, P = .0001, Figure 3H). Diffusion changes showed a significant dependency on corticosteroid dose changes (P = .0003, R2 = 0.37, Figure 3I).

Underlying differences between patients with and without corticosteroids initiation

Additional analyses were conducted to evaluate underlying differences between patients started on corticosteroids (n = 29, corresponding to the “increased corticosteroids” group, all off-steroids on timepoint 1) and patients never receiving corticosteroids (n = 21, representing 87.5% of the observations from the “stable corticosteroids” group) (Supplementary Table 3). Patients started on corticosteroids had larger contrast-enhancing tumors at diagnosis (volumetric P = .0002; bidirectional: P < .0001) than patients never receiving corticosteroids. As for the clinical presentation, patients started on corticosteroids had a higher prevalence of mass effect symptoms (P = .018), lower occurrence of seizures without focal deficits as presenting symptom (P = .005), and a trend towards a higher prevalence of neurological deficits at diagnosis (P = .25). These differences may partially explain the original decision to initiate corticosteroid administration in these patients.

Rerunning the multivariate analysis on these 2 groups while adjusting both for the time interval between scans and for the baseline contrast-enhancing volume, led to an estimated corticosteroid-induced shrinkage in contrast-enhancing volume of 41.0% (95% CI, 67.0%–15.0%, P = .003, R2 = 0.37), similar to the previous model.

Discussion

Our results confirm previous knowledge about MRI changes induced by corticosteroid use and provide a contemporary benchmark for the potential impact on response assessment. On postcontrast T1-weighted images, corticosteroid dose increase typically results in fainter and less extensive contrast-enhancement, that can be considered a “corticosteroid-induced pseudoresponse,” and that leads to a measurable shrinkage of the contrast-enhancing volume. Our current study estimates this volumetric change around 44% (using a comprehensive multivariate analysis), with a confidence interval between 26% and 62%. On diffusion imaging, corticosteroid dose increase typically results in an ADC reduction within the contrast-enhancing tumor component, with a median reduction of 180 × 10−6 mm2/s, or around 14% from the average estimates of ADC within the enhancing tumor.

Our findings are in line with the well-established pathophysiological effects of corticosteroids on glioblastoma. Contrast-enhancement is a radiographic sign of BBB permeability to contrast agent molecules due to abnormalities in glioblastoma vasculature, such as fenestrations, defective tight junctions, and basal membrane anomalies.19 Corticosteroids are thought to reduce glioblastoma BBB permeability by inhibiting the release of mediators promoting vascular abnormalities and permeability, such as arachidonic acid,19 cytokines,19 and VEGF35 (formerly known as vascular permeability factor35,36), and also by directly upregulating the endothelial expression of occludin, a key component of tight junctions.19 This interpretation is consistent with MRI studies showing that corticosteroids cause a reduction of vascular permeability metrics,13 quantitative postcontrast T1 relaxation time,13 and quantitative contrast agent accumulation in tumor tissue.14 Such corticosteroid-induced reduction in BBB permeability also explains the observed ADC reduction following corticosteroid administration, which is probably linked to a decreased extravasation of plasmatic proteins and decreased vasogenic edema within the tumor tissue. This is also in line with our observation of corticosteroids determining a less pronounced peritumoral edema on T2-weighted images, and consistent with previous studies showing peritumoral ADC reduction.37,38

The overall extent of the contrast-enhancing tumor shrinkage found in the present study is consistent with previous findings from a seminal study on 11 patients with recurrent high-grade gliomas imaged with CT before and after corticosteroids,12 which reported an average shrinkage around 30%. Similarly, another historical article evaluating pre- and postcorticosteroid MRI of recurrent high-grade gliomas reported a measurable shrinkage of the contrast-enhancing component in 9 out or 10 of their cases, although the quantitative effect of the shrinkage seems to be less pronounced in this study (≥25% shrinkage in 3 cases out of 10).15 Notably, our study and these previous articles12,15 describe a heterogeneous extent of the shrinkage across different lesions. While technical factors may contribute to such heterogeneity, varying degrees of corticosteroid-induced shrinkage are probably driven by underlying biological factors. Indeed, tumors may differ in terms of intrinsic growth kinetics and responsiveness to corticosteroids, which influence the apparent growth rate observed after corticosteroid dose increase (as we speculate in Figure 4). Additionally, different sensitivity to corticosteroids may even be present among different habitats of the same lesion (as in Figure 2A). The interplay between tumor growth and corticosteroid effects may lead a progressing glioblastoma to exhibit a volumetric increase underneath the ≥40% threshold required to currently define disease progression according to RANO 2.0.1,2 If validated in further studies, our modeled volumetric shrinkage ascribable to corticosteroids could be used to calculate adjusted thresholds for progressive disease in the presence of corticosteroid dose increase. In more detail, since we estimated that corticosteroids can cause a 26% shrinkage at minimum (lower bound of the confidence interval), a theoretical adjusted threshold of ≥14% volumetric increase could potentially capture progressive disease in case of corticosteroid dose increase (i.e., a ≥40% ground truth growth minus a 26% corticosteroid-induced shrinkage). A prompt identification of these progressions can improve patient care by advocating for earlier retreatment. Additionally, it would also make clinical trial data interpretation more accurate and potentially improve the reliability of radiographic endpoints as surrogates of clinical outcomes.

Figure 4.

Figure 4.

Open in a new tab

Impact of corticosteroids on the apparent growth rate of the contrast-enhancing tumor component. In this diagram, we report a speculative interpretation describing how a corticosteroid-induced reduction of BBB permeability can result in an “apparent” growth rate (B) smaller than the ground truth growth rate in absence of corticosteroids (A) or in presence of stable corticosteroid dose (C). Additionally, depending on the combination of tumor-specific sensitivity to corticosteroids and underlying growth kinetics, even a “negative” apparent growth rate can be observed.

The median ADC reduction in the contrast-enhancing tumor seen in our cohort is in line with previous studies on smaller cohorts, which also described an ADC drop within brain tumor regions following corticosteroid administration.30,37,38 Overall, the magnitude of the corticosteroid-induced ADC reduction reported in these studies is slightly less pronounced than in our cohort: 7.6% in Sinha et al.,37 7.0% in Minamikawa et al.,30 and 5.8% in Bastin et al.38 This modest discrepancy can be explained with the inclusion of other tumor types in some of these studies (e.g., metastases and meningiomas) and with technical differences in the MRI protocol and measurements. An accurate quantification of corticosteroid-induced ADC decrease is crucial for a correct interpretation of ADC, which has been extensively proposed as a biomarker for both glioma molecular profiling at diagnosis,21–24 for treatment response prediction and monitoring.25–29 At diagnosis, the interpretation of ADC values for diagnostic and prognostic stratification should be informed by whether the MRI scan was acquired after corticosteroid initiation. Similarly, the comparison of longitudinal diffusion images should take into consideration that a decrease in ADC values can be driven by corticosteroid dose changes.

Limitations

First, this study focused on treatment-naïve glioblastoma, while pseudoresponse quantification is potentially more relevant in postchemoradiation lesions and lesions receiving active treatment. However, analyzing treatment-naïve glioblastoma can be considered a cleaner study design because treatment-naïve lesions are all expected to grow, while findings in lesions undergoing treatments are clouded by the underlying heterogeneous response to treatments. Another limitation is that it was not possible to confirm IDH wild-type status for a minority of patients (n = 21), who nevertheless had a histopathological diagnosis of glioblastoma. However, it should be noted that the incidence of IDH mutation in grade 4 gliomas is very low (reportedly around 3.8% among cases with known IDH status)39; therefore, it is likely that not more than 1 patient in our cohort may bear an IDH mutation, which arguably did not skew our results. Additionally, this cohort included patients imaged with heterogeneous MRI protocols, which arguably introduced some inconsistency in the volumetric and diffusion measurements. Another limitation is that, due to the observational design, underlying biological and clinical differences influenced the clinical decision to administer corticosteroids in this cohort; therefore, some mismatch between the corticosteroid-treated patients and the control group exists in terms of baseline tumor size and clinical presentations. These differences did not affect the observed shrinkage and diffusion changes in the increased corticosteroids group obviously, but possibly influenced the group comparisons and the estimated corticosteroid-induced shrinkage, which was based on the assumption that the only difference between groups was corticosteroid administration. As an attempt to correct for some of these confounding factors, an additional subanalysis adjusting for baseline tumor volume was conducted and showed similar results compared with the original multivariate model. Finally, our corticosteroid dose classification had some inherent limitations. Since we studied treatment-naïve glioblastomas at diagnosis, all the cases with increased dose were off-corticosteroids at baseline, and it was not possible to study the effect of further corticosteroid dose increase in patients already on corticosteroids. Therefore, further validation is needed before the proposed quantitative values are practically applied to glioblastomas previously exposed to corticosteroids. In addition, studying the effects of oral corticosteroids is inherently limited by a potential lack of patient compliance, as oral corticosteroid prescription does not guarantee actual administration. Last, for many cases, it was not possible to retrieve accurate data regarding the exact day of corticosteroid initiation; therefore, it was not possible to account for the time interval between corticosteroid start and second MRI scan. As for the scenario of decreased corticosteroid dose, the small sample size of this subgroup did not allow to provide quantitative estimates of the volumetric and diffusion changes following corticosteroid dose reduction.

Conclusions

Corticosteroid dose increase resulted in a “pseudoresponse” on postcontrast T1-weighted images, consisting in a reduction of intensity and extent of contrast-enhancement of glioblastoma, and reflected in an apparent shrinkage of the contrast-enhancing tumor estimated as 44.0%, with a confidence interval ranging 25.7%–62.2%. This was associated with an ADC reduction around 180 × 10−6 mm2/s, or 14.2%, within the contrast-enhancing tumor. The interpretation of contrast-enhancing tumor size changes and of ADC measurements should take into consideration the interplay of corticosteroid-induced effects.

Supplementary Material

Supplement

Supplementary material is available online at Neuro-Oncology (https://academic.oup.com/neuro-oncology).

Key Points.

  • The observed median shrinkage of the contrast-enhancing tumor after corticosteroid initiation was −23.7% (IQR: −79.7% to +4.3%).

  • The estimated corticosteroid-induced shrinkage of the contrast-enhancing tumor, after adjusting for time interval between scans and comparing with the control group, was around −44.0% (95% CI, −25.7 to −62.2%).

  • The corticosteroid-induced ADC reduction was around −180 (IQR: −39 to −281) × 10−6 mm2/s, or −14.2% (IQR: −2.8% to −21.2%).

Importance of the Study.

Corticosteroids are commonly administered for the clinical management of glioblastoma. However, their use influences tumor contrast-enhancement and diffusion MRI features. The contrast-enhancing tumor volume is currently crucial for treatment response assessment, both in the clinical practice and in clinical trials. Diffusion MRI is an emerging tool that provides valuable insights in glioma molecular profiling. Therefore, corticosteroid administration can confound MRI evaluations of glioblastoma, impacting diagnosis, prognosis stratification, and treatment monitoring. This study provides both a qualitative depiction and quantitative estimate of corticosteroids effects on the contrast-enhancing tumor volume and on diffusion MRI. Knowing the quantitative impact of corticosteroids on the contrast-enhancing tumor volume paves the way to the future development of adjusted response criteria (e.g., RANO) that account for corticosteroid dose variations. A quantitative estimate of diffusivity reduction induced by corticosteroids can inform the correct interpretation of diffusion MRI and potentially lead to corticosteroid-adjusted cutoffs for clinical applications.

Funding

UCLA SPORE in Brain Cancer (P50 CA211015) (Sanvito, Yao), National Institutes of Health / National Cancer Instituite (NIH/NCI) R01CA270027 (Ellingson) and R01CA279984 (Ellingson).

Footnotes

Conflict of interest statement

B.M.E. is on the advisory board and is a paid consultant for Alpheus Medical, Carthera, Chimerix, Ellipses Pharma, Erasca, Global Coalition for Adaptive Research (GCAR), Imaging Endpoints, Medicenna, Voiant, Medscape, Monteris, Neosoma, Nerviano Medical Sciences, Nuvation Bio, Orbus Therapeutics, Sagimet Biosciences, Sapience Therapeutics, Servier Pharmaceuticals, Siemens, SonALAsense, Sumitomo Dianippon Pharma Oncology, Telix, the Sontag Foundation, the National Brain Tumor Society, and Third Rock Ventures. T.F.C. is cofounder, major stock holder, consultant and board member of Katmai Pharmaceuticals, holds stock for Erasca, member of the board and paid consultant for the 501c3 Global Coalition for Adaptive Research, holds stock in Chimerix and receives milestone payments and possible future royalties, member of the scientific advisory board for Break Through Cancer, member of the scientific advisory board for Cure Brain Cancer Foundation, has provided paid consulting services to Blue Rock, Vida Ventures, Lista Therapeutics, Stemline, Novartis, Roche, Sonalasense, Sagimet, Clinical Care Options, Ideology Health, Servier, Jubilant, Immvira, Gan & Lee, BrainStorm, Katmai, Sapience, Inovio, Vigeo Therapeutics, DNATrix, Tyme, SDP, Kintara, Bayer, Merck, Boehinger Ingelheim, VBL, Amgen, Kiyatec, Odonate Therapeutics QED, Medefield, Pascal Biosciences, Bayer, Tocagen, Karyopharm, GW Pharma, Abbvie, VBI, Deciphera, VBL, Agios, Genocea, Celgene, Puma, Lilly, BMS, Cortice, Novocure, Novogen, Boston Biomedical, Sunovion, Insys, Pfizer, Notable labs, Medqia, Trizel, Medscape and has contracts with UCLA for the Brain Tumor Program with Roche, VBI, Merck, Novartis, BMS, AstraZeneca, Servier. The Regents of the University of California (T.F.C. employer) has licensed intellectual property co-invented by TFC to Katmai Pharmaceuticals.

Ethics statement

All patients provided written informed consent to have their data used for research. The institutional review board (IRB-10–0655) approved the usage of patient data for research purposes. This study was performed in line with the principles of the Declaration of Helsinki and in compliance with the Health Insurance Portability and Accountability Act.

Contributor Information

Francesco Sanvito, UCLA Brain Tumor Imaging Laboratory (BTIL), Center for Computer Vision and Imaging Biomarkers, University of California at Los Angeles, Los Angeles, California; Department of Radiological Sciences, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California.

Asher Kim, UCLA Brain Tumor Imaging Laboratory (BTIL), Center for Computer Vision and Imaging Biomarkers, University of California at Los Angeles, Los Angeles, California; Department of Bioengineering, Henry Samueli School of Engineering and Applied Science, University of California at Los Angeles, Los Angeles, California.

Catalina Raymond, UCLA Brain Tumor Imaging Laboratory (BTIL), Center for Computer Vision and Imaging Biomarkers, University of California at Los Angeles, Los Angeles, California; Department of Radiological Sciences, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California.

Ashley Teraishi, UCLA Brain Tumor Imaging Laboratory (BTIL), Center for Computer Vision and Imaging Biomarkers, University of California at Los Angeles, Los Angeles, California; Department of Radiological Sciences, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California.

Richard G. Everson, Department of Neurosurgery, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California

Phioanh L. Nghiemphu, Department of Neurology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California

Albert Lai, Department of Neurology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California.

Robert A. Chong, Department of Neurology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California

David A. Nathanson, Department of Pharmacology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California

Noriko Salamon, Department of Radiological Sciences, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California.

Timothy F. Cloughesy, Department of Neurology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California UCLA Neuro-Oncology Program, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California.

Jingwen Yao, UCLA Brain Tumor Imaging Laboratory (BTIL), Center for Computer Vision and Imaging Biomarkers, University of California at Los Angeles, Los Angeles, California; Department of Radiological Sciences, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California; Department of Bioengineering, Henry Samueli School of Engineering and Applied Science, University of California at Los Angeles, Los Angeles, California.

Benjamin M. Ellingson, UCLA Brain Tumor Imaging Laboratory (BTIL), Center for Computer Vision and Imaging Biomarkers, University of California at Los Angeles, Los Angeles, California; Department of Radiological Sciences, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California; Department of Bioengineering, Henry Samueli School of Engineering and Applied Science, University of California at Los Angeles, Los Angeles, California; Department of Neurosurgery, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California; Department of Psychiatry, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California.

Data availability

Datasets are available from the authors upon request.

References

  • 1.Wen PY, Van Den Bent MJ, Youssef G, et al. RANO 2.0: update to the Response Assessment in Neuro-Oncology criteria for high- and low-grade gliomas in adults. J Clin Oncol. 2023;41(33):5187–5199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Ellingson BM, Sanvito F, Cloughesy TF, et al. A neuroradiologist’s guide to operationalizing the Response Assessment in Neuro-Oncology (RANO) criteria version 2.0 for gliomas in adults. Am J Neuroradiol. 2024;45(12):1846–1856. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Sanvito F, Castellano A, Cloughesy TF, Wen PY, Ellingson BM. RANO 2.0 criteria: concepts applicable to the neuroradiologist’s clinical practice. Curr Opin Oncol. 2024;36(6):536–544. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Weinberg BD, Gore A, Shu H-KG, et al. Management-based structured reporting of posttreatment glioma response with the brain tumor reporting and data system. J Am Coll Radiol. 2018;15(5):767–771. [DOI] [PubMed] [Google Scholar]
  • 5.Essien M, Cooper ME, Gore A, et al. Interrater agreement of BT-RADS for evaluation of follow-up MRI in treated primary brain tumor patients. Am J Neuroradiol. 2024;45(9):1308–1315. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Ellingson BM, Wen PY, Cloughesy TF. Evidence and context of use for contrast enhancement as a surrogate of disease burden and treatment response in malignant glioma. Neuro Oncol. 2018;20(4):457–471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Ramakrishna R, Barber J, Kennedy G, et al. Imaging features of invasion and preoperative and postoperative tumor burden in previously untreated glioblastoma: correlation with survival. Surg Neurol Int. 2010;1:40. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Ellingson BM, Abrey LE, Nelson SJ, et al. Validation of postoperative residual contrast-enhancing tumor volume as an independent prognostic factor for overall survival in newly diagnosed glioblastoma. Neuro Oncol. 2018;20(9):1240–1250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Sanvito F, Kaufmann TJ, Cloughesy TF, Wen PY, Ellingson BM, Standardized brain tumor imaging protocols for clinical trials: current recommendations and tips for integration. Front Radiol. 2023;3:1267615. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Pitter KL, Tamagno I, Alikhanyan K, et al. Corticosteroids compromise survival in glioblastoma. Brain. 2016;139(Pt 5):1458–1471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Wen PY, Weller M, Lee EQ, et al. Glioblastoma in adults: a Society for Neuro-Oncology (SNO) and European Society of Neuro-Oncology (EANO) consensus review on current management and future directions. Neuro Oncol. 2020;22(8):1073–1113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Cairncross JG, Macdonald DR, Pexman JH, Ives FJ. Steroid-induced CT changes in patients with recurrent malignant glioma. Neurology. 1988;38(5):724–726. [DOI] [PubMed] [Google Scholar]
  • 13.Andersen C, Jensen FT. Differences in blood-tumour-barrier leakage of human intracranial tumours: quantitative monitoring of vasogenic oedema and its response to glucocorticoid treatment. Acta Neurochir. 1998;140(9):919–924. [DOI] [PubMed] [Google Scholar]
  • 14.Armitage PA, Schwindack C, Bastin ME, Whittle IR. Pretreatment tumoural perfusion correlates with an imaging-based response to dexamethasone in patients with glioblastoma multiforme. J Neurol Neurosurg Psychiatry. 2010;81(4):446–448. [DOI] [PubMed] [Google Scholar]
  • 15.Watling CJ, Lee DH, Macdonald DR, Cairncross JG. Corticosteroid-induced magnetic resonance imaging changes in patients with recurrent malignant glioma. J Clin Oncol. 1994;12(9):1886–1889. [DOI] [PubMed] [Google Scholar]
  • 16.Zaki HS, Jenkinson MD, Du Plessis DG, Smith T, Rainov NG. Vanishing contrast enhancement in malignant glioma after corticosteroid treatment. Acta Neurochir. 2004;146(8):841–845. [DOI] [PubMed] [Google Scholar]
  • 17.Romano A, De Giorgi S, Romano A, et al. “Vanishing” glioblastoma: a case report and review of the literature. Radiol Case Rep. 2024;19(8):3276–3282. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Goh JJ, See SJ, Ang E, Ng WH. Vanishing glioblastoma after corticosteroid therapy. J Clin Neurosci. 2009;16(9):1226–1228. [DOI] [PubMed] [Google Scholar]
  • 19.Kotsarini C, Griffiths PD, Wilkinson ID, Hoggard N. A systematic review of the literature on the effects of dexamethasone on the brain from in vivo human-based studies: implications for physiological brain imaging of patients with intracranial tumors. Neurosurgery. 2010;67(6):1799–815; discussion 1815. [DOI] [PubMed] [Google Scholar]
  • 20.Arevalo OD, Soto C, Rabiei P, et al. Assessment of glioblastoma response in the era of bevacizumab: longstanding and emergent challenges in the imaging evaluation of pseudoresponse. Front Neurol. 2019;10:460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Maynard J, Okuchi S, Wastling S, et al. World health organization grade ii/iii glioma molecular status: prediction by MRI morphologic features and apparent diffusion coefficient. Radiology. 2020;296(1):111–121. [DOI] [PubMed] [Google Scholar]
  • 22.Thust SC, Hassanein S, Bisdas S, et al. Apparent diffusion coefficient for molecular subtyping of non-gadolinium-enhancing WHO grade II/III glioma: volumetric segmentation versus two-dimensional region of interest analysis. Eur Radiol. 2018;28(9):3779–3788. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Leu K, Ott GA, Lai A, et al. Perfusion and diffusion MRI signatures in histologic and genetic subtypes of WHO grade II–III diffuse gliomas. J Neurooncol. 2017;134(1):177–188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Cho NS, Sanvito F, Le VL, et al. Diffusion MRI is superior to quantitative T2-FLAIR mismatch in predicting molecular subtypes of human non-enhancing gliomas. Neuroradiology. 2024;66(12):2153–2162. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Pope WB, Lai A, Mehta R, et al. Apparent diffusion coefficient histogram analysis stratifies progression-free survival in newly diagnosed bevacizumab-treated glioblastoma. Am J Neuroradiol. 2011;32(5):882–889. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Ellingson BM, Sahebjam S, Kim HJ, et al. Pretreatment ADC histogram analysis is a predictive imaging biomarker for bevacizumab treatment but not chemotherapy in recurrent glioblastoma. Am J Neuroradiol. 2014;35(4):673–679. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Hagiwara A, Oughourlian TC, Cho NS, et al. Diffusion MRI is an early biomarker of overall survival benefit in IDH wild-type recurrent glioblastoma treated with immune checkpoint inhibitors. Neuro Oncol. 2022;24(6):1020–1028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Ellingson BM, Cloughesy TF, Zaw T, et al. Functional diffusion maps (fDMs) evaluated before and after radiochemotherapy predict progression-free and overall survival in newly diagnosed glioblastoma. Neuro Oncol. 2012;14(3):333–343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Ellingson BM, Cloughesy TF, Lai A, et al. Graded functional diffusion map-defined characteristics of apparent diffusion coefficients predict overall survival in recurrent glioblastoma treated with bevacizumab. Neuro Oncol. 2011;13(10):1151–1161. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Minamikawa S, Kono K, Nakayama K, et al. Glucocorticoid treatment of brain tumor patients: changes of apparent diffusion coefficient values measured by MR diffusion imaging. Neuroradiology. 2004;46(10):805–811. [DOI] [PubMed] [Google Scholar]
  • 31.Ellingson BM, Nguyen HN, Lai A, et al. Contrast-enhancing tumor growth dynamics of preoperative, treatment-naive human glioblastoma. Cancer. 2016;122(11):1718–1727. [DOI] [PubMed] [Google Scholar]
  • 32.Yushkevich PA, Piven J, Hazlett HC, et al. User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. Neuroimage. 2006;31(3):1116–1128. [DOI] [PubMed] [Google Scholar]
  • 33.Kural C, Atac GK, Tehli O, et al. The evaluation of the effects of steroid treatment on the tumor and peritumoral edema by DWI and MR spectroscopy in brain tumors. Neurol Neurochir Pol. 2018;52(4):495–504. [DOI] [PubMed] [Google Scholar]
  • 34.Louis DN, Perry A, Wesseling P, et al. The 2021 WHO classification of tumors of the central nervous system: a summary. Neuro Oncol. 2021;23(8):1231–1251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Criscuolo GR, Lelkes PI, Rotrosen D, Oldfield EH. Cytosolic calcium changes in endothelial cells induced by a protein product of human gliomas containing vascular permeability factor activity. J Neurosurg. 1989;71(6):884–891. [DOI] [PubMed] [Google Scholar]
  • 36.Klagsbrun M, Soker S. VEGF/VPF: the angiogenesis factor found? Curr Biol. 1993;3(10):699–702. [DOI] [PubMed] [Google Scholar]
  • 37.Sinha S, Bastin ME, Wardlaw JM, Armitage PA, Whittle IR. Effects of dexamethasone on peritumoural oedematous brain: a DT-MRI study. J Neurol Neurosurg Psychiatry. 2004;75(11):1632–1635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Bastin ME, Carpenter TK, Armitage PA, et al. Effects of dexamethasone on cerebral perfusion and water diffusion in patients with high-grade glioma. Am J Neuroradiol. 2006;27(2):402–408. [PMC free article] [PubMed] [Google Scholar]
  • 39.Ostrom QT, Price M, Neff C, et al. CBTRUS statistical report: primary brain and other central nervous system tumors diagnosed in the United States in 2016–2020. Neuro Oncol. 2023;25(12 Suppl 2):iv1–iv99. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement
NIHMS2135573-supplement-Supplement.docx (2.5MB, docx)

Data Availability Statement

Datasets are available from the authors upon request.

RESOURCES