Skip to main content
NCBI home page
JNSPG Special Collection logoLink to JNSPG Special Collection
. 2026 Jun 5:2026.1.JNS241395. doi: 10.3171/2026.1.JNS241395

Imaging, histological, and molecular characterization of a preclinical, orthotopic model of recurrent glioblastoma following image-guided laser ablation of the primary tumor

Tavarekere N Nagaraja 1,,7,✉, Seamus Bartlett 1, Glauber Cabral 2, Prabhu C Acharya 6,,2, Natalia Morosini 1, Indrani Datta 1, Laura Hasselbach 1, Jaspreet Singh 2, Parveen Parasar 2, Bhargav Ayloo 1,,4, Faith Avritt 1, Katelynn G Spanick 2, Pranavi Gowda 1, Ana C deCarvalho 1, Houtan Noushmehr 1, Robert A Knight 2,,4,,6,,7, Stephen L Brown 3,,5,,7, James R Ewing 1,,2,,7,,9, Ian Y Lee 1,,8
PMCID: PMC13640755  PMID: 42247692

In Brief

Researchers present a new preclinical model of MRI-guided tumor ablation that replicates the clinical laser cytoreduction approach to glioblastoma. The imaging biomarkers used to measure glioma vascular function, ablation efficacy, and recurrence are described, and the histological and molecular features of primary and postablation recurrent tumors were compared. This animal model replicated key brain tissue signatures of human glioma ablation and may provide an optimal platform to test new adjuvant drugs in efforts to delay or prevent glioblastoma recurrence.

Keywords: diffusion-weighted imaging, high-grade glioma, Ktrans, laser interstitial thermal therapy, LITT, magnetic resonance imaging, U251N glioma, oncology, tumor

ABBREVIATIONS : DCE = dynamic contrast-enhanced, DEG = differentially expressed gene, DW = diffusion-weighted, GBM = glioblastoma, H&E = hematoxylin and eosin, Ktrans = volumetric transfer constant, LITT = laser interstitial thermal therapy, MHC = major histocompatibility complex, RNA-Seq = RNA sequencing, VD = interstitial distribution volume, ve = extracellular distribution volume

Abstract

OBJECTIVE

Image-guided laser interstitial thermal therapy (LITT) is a minimally invasive cytoreductive treatment for recurrent gliomas and tumors in eloquent regions. This technique was adapted to develop an image-guided glioblastoma (GBM) ablation model of recurrence. The efficacy of imaging biomarkers for evaluating tumor ablation and recurrence was evaluated by comparing the cytopathology and molecular signatures of primary and recurrent tumors.

METHODS

Immune-compromised female rats were implanted with U251N tumor cells in 1 brain hemisphere (n = 20). Tumor growth was monitored using MRI and dynamic contrast-enhanced (DCE)-MRI. When tumors reached about 3.5–4.5 mm in diameter at their largest extent, they were ablated using a clinical LITT system (Visualase), guided by diffusion-weighted MRI. Five rats implanted with U251N tumors were used as unablated controls. Postablation DCE-MRI data were acquired at 24 hours and 2 and 4 weeks. Rats were euthanized at 2 and 4 weeks, and brain sections were stained for hematoxylin and eosin (H&E), human major histocompatibility complex (MHC), Ki-67, SOX2, OLIG2, and mesenchymal marker CD44. Additional rats with primary (n = 4) and postablation recurrent (n = 4) U251N tumors were used to compare molecular compositions using RNA sequencing (RNA-Seq).

RESULTS

All rats survived the LITT procedure. Sham controls showed increased tumor burden by 2 weeks and were euthanized. In the ablated groups, MRI revealed little tumor tissue at 24 hours after ablation, signs of recurrence at 2 weeks, and significant tumor tissue at 4 weeks. Tumor DCE-MRI parameters showed elevated intratumoral vascular permeability values (i.e., blood-to-tissue forward volumetric transfer constant [Ktrans]) at pre-LITT imaging that shifted to the ablation site periphery at 24 hours after ablation. A progressive decrease in Ktrans was seen until 1 week after LITT. Increasing Ktrans values at 2 weeks and after 4 weeks coincided with histological evidence of tumor recurrence. RNA-Seq showed that cell cycle, cellular movement, and inflammatory disease genes were the most differentially expressed genes (DEGs) in recurrent tumors compared to primary tumors.

CONCLUSIONS

These results suggest that DCE-MRI can serve as an efficient imaging biomarker for evaluating glioma cytoablation and for tracking post-LITT recurrence. Functions that influence cell cycle, cell motility, and inflammatory processes were upregulated in the recurrent tumor. This represents a new preclinical, image-guided tumor ablation model of recurrent GBM that can be used to test potential therapies.

Glioblastoma (GBM), a WHO class 4 high-grade glioma, is the most aggressive and prevalent primary brain tumor, presenting with a dismal prognosis and median survival of only 12 to 15 months.1 Recommended standard therapy includes a combination of safe maximal tumor resection followed by radiotherapy and temozolomide chemotherapy.2 However, aggressive, treatment-refractory tumor recurrence is almost certain.3 With safe maximal tumor resection being an important factor in delaying recurrence,4,5 intraoperative tumor visualization techniques are increasingly employed for glioma resection in the clinic.6 Several murine models are also used to test promising new therapies and adjuvant treatments against for this fatal disease. However, preclinical research reports with image-guided tumor cytoreduction are few and restricted to tumors implanted into the cortical surface in nearly all cases.7–9 Better animal models are recommended to enhance the relevance of preclinical data for potential clinical applications.10,11 Using a rat U251N glioma model, we previously reported an image-guided, laser interstitial thermal therapy (LITT)-based cytoreduction technique that can be performed even on embedded subcortical tumors.12 This technique also produces acute postablation changes in the surrounding tissue that are remarkably similar to the acute tissue changes reported after clinical LITT treatment of primary GBM.13

LITT is increasingly used for its cytoreductive capabilities to treat recurrent GBM.14 It is especially used for deep-seated lesions that would either be inaccessible by craniotomy or are close to eloquent structures.15–20 With consideration of these advantages in glioma management, the objective of this study was to extend our previous observations and develop an experimental recurrent GBM model using LITT. In addition, data on the possible effects of ablation on aggressive transformation of recurrent glioma are scarce. Being the preferred clinical imaging modalities for gliomas, MRI and dynamic contrast-enhanced (DCE)-MRI were used to evaluate tumor progression following LITT ablation. The rat orthotopic U251N glioma model was used for these studies, and a Visualase (Medtronic, Inc.) clinical laser system with diffusion-weighted (DW)-MRI guidance was employed for tumor ablation.12 The rats were longitudinally imaged following LITT ablation to ensure tumor absence or document recurrence, and they were euthanized at intervals to correlate imaging biomarkers with brain tissue biology.

Methods

All materials used were analytical or pharmaceutical grade. The Visualase apparatus was a gift from Medtronic, Inc. It utilizes a saline-cooled diode laser fiber (0.64-mm diameter) with a near-infrared (980 nm) wavelength. Instantaneous cell death through coagulative necrosis occurs at 60°C. Between 43°C and 60°C, the combination of time and temperature results in cell death primarily through apoptosis. Compared to the relatively high-energy setting of 9–12 V and duration of several minutes for human tumor ablation, a low-energy setting of 1 V and duration of about 1 minute were used for rat glioma ablation.

U251N Rat Model of Orthotopic Glioma

All experimental procedures were approved by the Institutional Animal Care and Use Committee. Adult female, immune-compromised, RNU/RNU rats (Charles River Laboratories) underwent intracerebral implantation following published procedures.21–23 Briefly, animals were anesthetized with isoflurane (4% for induction, 1.5%–2.0% for maintenance, N2O/O2 ratio of 2:1). The scalp was cleansed with Betadine and alcohol, the eyes were coated with Lacri-lube, and the head was immobilized in a small animal stereotactic device (Kopf). After sterile draping, a 1-cm incision was made 2 mm right of the midline to expose the skull. A burr hole was drilled 3.50 mm to the right of bregma (anteroposterior 0.0), without penetrating the dura. A #2701 10-μl Hamilton syringe with a 26-gauge needle containing freshly harvested U251N tumor cells in the log growth phase (5 × 104 cells in 10 μl phosphate-buffered saline) was lowered to a depth of 3.0 mm and then raised back to a depth of 2.5 mm to create a pocket. Cells were injected into this pocket at a rate of 0.5 μl per 10 seconds until the entire volume was used. The needle was left in place for 5 minutes and then gently withdrawn, the burr hole was sealed with sterile bone wax, and the scalp incision was sutured.

Imaging

MRI and DCE-MRI were performed to measure tumor location and extent about 2 weeks after tumor implantation, following published techniques.21,22,24,25 Briefly, each animal was anesthetized with isoflurane (4% for induction, 1.5%–2.0% for maintenance, N2O/O2 ratio of 2:1) delivered through a nose cone and allowed to spontaneously respire. The tail vein was cannulated for administration of the magnetic resonance contrast agent (Magnevist, Bayer HealthCare Pharmaceuticals). Body temperature was maintained at 37°C with warm air and monitored via an intrarectal probe. Detailed methodologies of image acquisition and analyses have been previously published.12,21,22,25

DW-MRI–LITT Procedure

An initial MRI study was done to localize tumor extent and measure vascular kinetics. The LITT procedure was performed the next day under isoflurane anesthesia using the Visualase system with DW-MRI guidance following published methods (see Video 1 for an example).12

VIDEO 1. An example of DW-MRI–guided ablation of rat U251N orthotopic glioma. The ablation area encompasses the complete extent of the tumor mass with some sublethal heating effects noted in the ipsilateral hemisphere. After completion of ablation, the periablation regions looked normal, albeit with MR contrast enhancement in the ablation periphery. The effects of hyperthermia were limited to these regions within the ipsilateral hemisphere, and the contralateral hemisphere was unaffected. © Tavarekere N. Nagaraja, published with permission. Click here to view.

The ablated rats were allowed to recover for long-term studies, where longitudinal imaging sessions were conducted for the respective survival cohorts.

Histopathology

The rats were maintained in deep isoflurane anesthesia after the final postablation MRI session. For fluorescence microscopy, a cocktail of Evans Blue (Sigma-Aldrich) (1 ml of a 2% solution in normal saline) and 2000 kD fluorescent isothiocyanate–dextran (Sigma-Aldrich) (1 ml of 50 mg/ml in normal saline) was injected through the tail vein cannula and allowed to circulate for 10 minutes. The brain was quickly removed after decapitation and stored at 4°C in 10% paraformaldehyde solution. After 48 hours of immersion fixation, the brain was placed into a rat brain matrix and cut into three 2-mm-thick slabs encompassing the ablation/tumor area. One 2-mm-thick slab was cut into 100-μm slices in a vibratome (Vibratome Series 1000, Energy Beam Sciences). These slices were mounted in Glycergel medium (Dako, Denmark A/S) on glass slides to keep the slices hydrated and coverslipped. The slides were stored at 4°C until fluorescence microscopy.

The second or third 2-mm-thick slab of brain tissue was embedded in paraffin, de- and rehydrated in an ethanol series, and cut into 6-μm-thick sections. These sections were stained with hematoxylin and eosin (H&E) to evaluate tumor tissue status, major histocompatibility complex (MHC) to localize U251N cells,12,26 Ki-67 to identify proliferating cells, and SOX2 to identify stemlike cells. In addition, OLIG2 and CD44 were stained for proneural and mesenchymal markers, respectively, on separate sections (detailed methods given in Supplemental Data).

Fluorescence Microscopy

A Keyence All-in-One Fluorescence Microscope BZ-X800 (Keyence Corporation of America) was used for fluorescence microscopy. The vibratome sections were scanned with a ×2 objective lens at a numerical aperture of 0.1 and approximate depth of field of 55 μm. Subsequent microscopic data were acquired using a ×20 objective lens at a numerical aperture of 0.75 and approximate depth of field of 1.0 μm to examine the periablation tumor area in greater detail. For image acquisition, 20% excitation was used with a 3- to 3.5-second and 1/7.5- to 1/8.5-second brightness exposure for the ×2 and ×20 lenses, respectively. The red Evans blue and green fluorescent isothiocyanate fluorochromes within the brain tissue sections were excited with a laser beam at 560 nm and 470 nm, respectively, with corresponding absorption wavelengths of 630 nm and 525 nm. The red and green fluorophore images were merged using Keyence BZ-X800 analyzer software (Keyence) to visualize the periablation tumor microvasculature and to examine the relative distributions of the 2 plasma-borne tracers.

RNA Sequencing Studies

A separate group of 8 rats bearing U251N tumors were used for RNA sequencing (RNA-seq) studies. The tumor was not ablated in 4 of these rats, while the tumor was ablated and monitored for recurrence in the other 4. All rats were euthanized under deep anesthesia. Fresh brains were cut rostrocaudally into 2-mm-thick slices using a rat brain matrix and instantly frozen by submerging them into 2-methylbutane chilled to −40°C with dry ice. Slices were stored at −80°C until further processing. Slices containing tumor tissue were placed on glass slides chilled with dry ice to prevent thawing, and any remaining tumor and contralateral brain tissues were dissected and placed in tared vials. Each sample was weighed and documented, along with the calculated protein concentration, and shipped to USC Molecular Genomics Core, University of Southern California, for analysis. This facility used an Agilent Bioanalyzer to prepare the RNA library of each sample. Each RNA library was assessed for quantity on a Qubit Fluorometer, while quantitative PCR was performed on a Thermo Fisher QuantStudio 7 and sequenced on an Illumina Nextseq500.

Data Analysis

MRI data were processed offline to produce quantitative parameter maps following published procedures.12,21,22,25 The values of the DCE-MRI parameters were estimated for the pre-LITT session and postablation session series. The final postablation MR parameter maps were used to verify the location and extent of the tumor by comparing them to the terminal histology slide images. The fluorescence images were examined at different magnifications to discern tracer leakage patterns in the periablation zone and/or recurrent tumor tissue. The images of the H&E- and immunostained specimens were examined using a Nikon E800M brightfield microscope equipped with a DXM1200C digital camera system and Image Pro-Plus software (Media Cybernetics) to identify tumor regions with the histological features of GBM and to assess positively staining cells.27 Cells stained positively for the selected antigens were counted in 3 random fields on each slide and averaged. Group means were computed and then compared across the primary and recurrent tumors.

RNA-Seq was used to profile 4 primary (tumor grown in rat brain) and 3 recurrent (postablation tumor grown in rat brain) tumors. The raw FASTQ data were aligned to STAR index version 39,28 and gene counts were quantified using feature count mapping to human reference genome h38 (version 39).29 Differential genes were selected based on the NOISeq method30 by comparing primary versus recurrent samples at a q-value threshold of 0.8. The biological roles of these genes were evaluated with Qiagen’s IPA knowledge base by further restricting these differential genes to those with greater than or equal to log2 fold change (QIAGEN, Inc., https://digitalinsights.qiagen.com/IPA).

Results

Pre-LITT MRI showed an embedded tumor located subcortically in many implanted rats. In some, the tumor extended into the neocortical region. Known radiological features of glioma, such as a space-occupying tumor with mass effect, contrast enhancement, and peritumoral edema, were also observed. Figure 1 exemplifies these features from an unablated rat along with corresponding H&E-, MHC-, and Ki-67–stained images. The tumor outline is identical in all these images, with relatively increased cell density toward the periphery and necrotic regions near the tumor core. The brain hemisphere ipsilateral to the tumor was swollen with a compressed ventricle and midline shift, which are signs of glioma mass effect (additional images of fluorescent tracer distributions in an intact glioma and a 24-hour postablation tumor are shown in Supplemental Data, Fig. S1).

FIG. 1.

FIG. 1.

Open in a new tab

Examples of an unablated tumor at 2 weeks (A–E) and approximately 3 weeks (F–J) after implantation. The parameter maps at these 2 time points shown from left to right are as follow: blood flow (A and F); intravascular distribution volume (vp) (B and G); blood-to-brain forward Ktrans (C and H); ve (D and I); and VD (E and J). The color scale bar on the right represents all maps, with the value range shown for each parameter on each map. K–N: Histological findings from this rat, from left to right, as follow: H&E (K), MHC (L), Ki-67 (M), and fluorescent tracer distributions (N). Note the increase in tumor size from 2 to 3 weeks that is visible on the MRI maps. This increased tumor burden typically necessitates euthanasia. Along with tumor enlargement, the development of necrotic regions toward the tumor core is evident (visible in panels G–N). Original magnification ×0.5 (K–M).

All rats survived the LITT procedure in all 3 study cohorts (24 hours and 2 and 4 weeks after ablation). The area of the T1-weighted contrast enhancement was confined to tumoral tissue before LITT and to the periablation area immediately after ablation. Figure 2 exemplifies the 2-week survival cohort along with histological and fluorescent images. The overall extent of the glioma tumor, as well as its appearance after ablation and recurrence, can be seen in these images. There appeared to be a latency period of a few days following ablation, where the surviving glioma cells remained quiescent. By the 2-week imaging session, however, the extent of glioma was greater than in the pre-LITT images and spanned a larger area. Nearly complete spatial agreement was observed between the final set of MR data and the histological and fluorescent images.

FIG. 2.

FIG. 2.

Open in a new tab

Images exemplifying the response to ablation and later recurrence at the 2-week endpoint. The displayed parameters from left to right are as follow: blood flow (cerebral blood flow [CBF]), T2-weighted images, and T1-weighted postcontrast minus precontrast images (T1-sub). A: Pre-LITT imaging. The far right image represents acute post-LITT T1-weighted contrast enhancement. B–D: Images obtained at different postoperative time points. E: Corresponding H&E, MHC, and Ki-67 histological and fluorescent images. Original magnification ×0.5 (E). Figure is available in color online only.

Figure 3, which is nearly identical to Fig. 2, illustrates tumor status changes in the 4-week survival cohort after ablation, along with corresponding terminal histological and fluorescent images. These images show an increasing tumor burden that begins 1 week after ablation. Temporal changes in volumetric transfer constant (Ktrans), extracellular distribution volume (ve), and interstitial distribution volume (VD) values decreased after a slight peak at 24 hours after ablation, then showed an increasing trend that coincided with glioma recurrence through 2 and 4 weeks. The quantified imaging biomarker values in the 2- and 4-week survival groups are shown in Fig. 4. Incidentally, there was 1 case in which the rat showed no signs of increasing tumor burden and was euthanized at 6 weeks. The MRI data from this rat showed all imaging biomarkers decreasing toward baseline after the initial expected postablation increase. Histology showed the complete absence of glioma in this rat, confirming the sensitivity of these MRI biomarkers to tumor ablation and subsequent absence of recurrence. These data are shown in Supplemental Data Fig. S5.

FIG. 3.

FIG. 3.

Open in a new tab

Images exemplifying the response to ablation and later recurrence at the 4-week endpoint. The displayed parameters from left to right are as follow: blood flow (cerebral blood flow [CBF]), T2-weighted images, and T1-weighted postcontrast minus precontrast images (T1-sub). A: Pre-LITT imaging. The far right image representing acute post-LITT T1-weighted contrast enhancement. B–D: Images obtained at different postoperative time points. E: Corresponding H&E, MHC, and Ki-67 histological and fluorescence images. Original magnification ×0.5 (E). Figure is available in color online only.

FIG. 4.

FIG. 4.

Open in a new tab

Temporal changes in MRI parameters from the 2- (A) and 4-week (B) survival cohorts. The abscissa indicates the imaging intervals, and the ordinate shows the mean parameter values. For the 2-week cohort, 1, 2, 3, and 4 on the abscissa represent the pre-LITT, 24-hour postablation, 1-week, and 2-week imaging sessions, respectively; for the 4-week cohort, they represent the pre-LITT, 24-hour postablation, 2-week, and 4-week imaging sessions, respectively. Note the increasing trend in the Ktrans estimates, coinciding with tumor recurrence. These data illustrate the temporal sensitivity and specificity of the DCE-MRI biomarkers to ablation efficacy, as well as glioma recurrence. Mean (dots) and standard deviation (error bars) are shown.

Examination of H&E-stained brain sections showed salient histopathological features that are employed in the clinic for GBM. These included a high cell density compared to normal brain regions, irregular vasculature with stasis, and pseudopalisading necrosis (Fig. 5 and Fig. S2). Mitotic bodies in H&E-stained specimens and the number of Ki-67–, SOX2-, and OLIG2-positive cells were counted in 3 random fields and averaged across the groups (Fig. 6). Except for OLIG2, the other indices did not show significant differences between primary and recurrent gliomas at the unablated and 2- and 4-week time points (Fig. S3). OLIG2-positive cells decreased from the unablated gliomas to the 2- and 4-week recurrent gliomas, suggesting decreased proneural features. This was complemented by increased CD44 staining in 2- and 4-week recurrent tumors, suggesting increased mesenchymal modification after recurrence. Regarding recurrent gliomas, infiltration into the peritumoral tissue was observed compared to the relatively well-defined tumor border in the primary glioma (Fig. 7A–F and Fig. S4). The tumor origin of the infiltrating cell clusters was confirmed with MHC, Ki-67, and CD44 staining (Figs. 6 and 7).

FIG. 5.

FIG. 5.

Open in a new tab

Representative high-magnification H&E histological images from unablated control (A–C), 2-week postablation recurrent tumor (D–F), and 4-week postablation recurrent tumor (G–I). These images demonstrate the presence of pseudopalisading necrotic regions, a typical feature of GBM, in these tumors. Such regions form due to intratumoral areas with severe hypoxia, from which the cells begin to recede and appear to form a ring around the margins. Bar = 100 μm. Arrows in panels A, B, and D point to irregular vascular segments with vascular stasis around glioma. The accompanying figure for these data (Fig. S2) shows the location and overall extent of these tumors, with examples of increased tumor cell density, enlarged images of tumor-associated irregular vasculature, and red cell stasis.

FIG. 6.

FIG. 6.

Open in a new tab

Magnified images (acquired at ×20 magnification) from H&E, Ki-67, SOX2, OLIG2, and CD44 staining. Increased cell density in the tumor is visible in the first 3 parameters. Ki-67– and SOX2-positive cells are distributed among the larger population of tumor cells, suggesting that some tumor cells are actively proliferating at any given time and only a limited number exhibit cancer stemlike cell properties. OLIG2 staining is relatively denser in primary tumors than in recurrent tumors, and CD44 exhibited the opposite pattern with increased invasion in recurrent tumors. The accompanying figure for these data (Fig. S3) depicts the quantification of mitotic bodies, Ki-67, SOX2, and OLIG2 counts from the 3 time points.

FIG. 7.

FIG. 7.

Open in a new tab

Increased infiltration in recurrent tumors compared to primary gliomas. Brain sections stained with H&E (A, C, and E) and Ki-67 (B, D, and F) in the study groups were used to assess this parameter. Bar = 300 μm. In unablated tumors, a relatively well-defined tumor periphery was seen (A and B). In 2-week recurrent gliomas, the periphery was somewhat discontinuous and isolated clusters of tumor cells invading the adjacent brain tissue were observed on both H&E and Ki-67 staining (arrows in C and D). Such infiltration increased further in the 4-week recurrent group, with tumor cell infiltration along perivascular Virchow-Robin spaces (arrows in E and F). Refer to Fig. S4 (Supplemental Data) for images of brain sections with the complete extent of tumor from which these magnified images were derived.

Through the process of significant gene selection, 5109 genes were identified as differentially expressed between primary and recurrent tumors (Fig. 8). For the enrichment analysis, differential genes that had a log2 fold change were selected (662 upregulated and 695 downregulated) (see Supplemental Excel Spreadsheets in Supplemental Data). The biological functions of these genes were evaluated with an enrichment test (IPA knowledge base) via the Fisher exact method to derive p values. The enrichment test revealed several biological functions and gene-gene networks. DNA replication, recombination, and repair (1.59 × 10−5 to 3.84 × 10−2), cell cycle (1.17 × 10−3 to 4.56 × 10−2), and cellular movement (1.17 × 10−3 to 4.56 × 10−2) were among the most involved biofunctions. Cellular development, cellular growth and proliferation, hematological system development and function (score 41, focus molecule 22), cell-to-cell signaling and interaction, cellular assembly and organization, and cellular development (score 34, focus molecule 19) were the highest ranking gene-gene networks (Fig. S6B). Furthermore, the focus molecules in these gene-gene networks were among the significant differentially expressed genes (DEGs) mentioned earlier, which were connected with other genes from the IPA knowledge base, and used to construct this network in relation to a particular disease or biological function.31 The score provided here for each network is a numerical value that represents the rank of the respective network to its degree of relevance between differential genes and the network-eligible molecules in the IPA knowledge base.

FIG. 8.

FIG. 8.

Open in a new tab

RNA-Seq data for DEGs in recurrent tumors compared to those of primary gliomas. The volcano plot shows the distribution of DEGs in recurrent tumors expressed as the fold change in expression compared to that in primary tumors. Blue dots show no change, orange dots indicate upregulated genes, and green dots indicate downregulated genes (refer to Figs. S6A and B in Supplemental Data that show the top gene-gene networks from the enrichment analysis of the data presented here).

Discussion

Herein, we present evidence that a clinical LITT apparatus can be adapted to perform image-guided ablation of orthotopic brain tumors to generate a preclinical model of glioma cytoreduction and subsequent recurrence. All rats survived the procedure and showed few adverse effects. The ablation procedure was efficient in achieving significant cytoreduction. The model resulted in acute brain tissue changes, such as a nearly acellular tumor core, and periablation contrast enhancement surrounded by a transition zone with edematous neuropil and astrogliosis. These features were comparable to the acute postablation histopathological observations of a newly diagnosed GBM tumor.13 Ablation extended overall survival in this model. An unablated rat with a U251N tumor typically survives about 3–4 weeks after implantation. On the other hand, LITT extended the lifespan of implanted rats to as long as 6 weeks, with LITT intervention performed at 2 weeks after implantation. Recurrence was observed after a latent period of about a week after LITT ablation.

To assess biomarkers for this recurrence, the imaging parameters, histopathological results, and genetic signatures were compared between primary and recurrent tumors. Along with standard MRI indices, several novel imaging biomarkers derived from DCE-MRI were used to assess tumor status. These included the Patlak plot– and Logan plot–derived parameters of Ktrans, ve, and VD. The combination of these biomarkers was useful for assessing glioma microvascular kinetics before LITT and immediately after LITT in order to determine the efficacy of ablation, as well as to assess the subsequent quiescence and recurrence. Due to its sensitivity to tumor vascular function, Ktrans represented ablation efficacy by tracking recurrence better than the other biomarkers. With consideration of its prevalence in previous reports on tumor treatment evaluations, Ktrans may be a useful measure of glioma physiology, even if it is derived by different DCE-MRI data–processing methods. The other imaging indices reported here have been previously validated for evaluating human glioma and may have utility in human glioma staging.

Histologically, the primary and recurrent tumors exhibited some similarities and some differences. Large dysregulated blood vessel profiles with red cell stasis were observed in the tumor periphery of most samples. The numbers of such vascular segments appeared to increase in recurrent tumor samples. Both primary and recurrent tumors showed a central necrotic region surrounded by pseudopalisading necrosis. These represent hypoxic areas with receding tumor cells, forming a ringlike perimeter. The cell density around such regions in recurrent tumors trended higher than in primary tumors. Quantification of mitotic bodies and Ki-67–positive and SOX2-positive cells from randomly chosen regions in the primary and 2- and 4-week recurrent samples showed areas with relatively low and high numbers of these indices. However, the mean values were not statistically different between these groups. Compared to these markers, the expression levels of OLIG2 and CD44 were different between naïve and recurrent tumors, suggesting that these are better suited to differentiate these tumor types. The main difference in histology was the relatively well-demarcated tumor periphery in the primary tumors, as opposed to the tumor cell clusters that infiltrated normal brain tissue in the recurrent tumors, at both the 2- and 4-week time points. Notably, in the 4-week group, perivascular infiltration was evident (Fig. 7E and F). The differences in infiltration patterns between the control and 2- and 4-week recurrent tumors were similar to those observed in orthotopic preclinical GBM tumors grown using immortalized glioma cell lines such as U251N and 9L. U251N reportedly represents primary GBM in its presentation, and 9L represents gliosarcoma or recurrent GBM with mesenchymal modification. The U251N glioma presented with a defined tumor outline and very little invasion into surrounding brain tissue, whereas 9L showed a relatively indistinct tumor border with infiltration into host brain tissue around the tumor and into perivascular spaces.32 In comparison, the unablated tumors in this study showed a fairly well demarcated tumor border with little infiltration, the 2-week recurrent tumors exhibited clumps of tumor cells invading the host tissue, and increased invasion into host tissue was observed at 4-week recurrence. This invasion pattern that is similar in several respects to that of 9L suggests increased invasion after recurrence in the U251N tumor. The increased CD44 staining observed in infiltrating tumor cells suggests elevated stemness and/or mesenchymal modification, although it needs to be investigated further in this new recurrent GBM model. Perivascular infiltration is reportedly a primary mechanism of treatment resistance to antiangiogenic therapies.33 A detailed examination of this phenomenon is warranted so that strategies that attenuate such infiltration can be implemented to improve standard and novel antiglioma therapy responses.

The molecular signatures of primary and recurrent glioma tissues showed some differences. Genes involved in cell cycle and cellular movement were the top DEGs in recurrent tumors. These DEGs may underlie the increased cell division (Ki-67 staining) in some regions and the infiltration observed in recurrent tumors during examination of histological characteristics. Although further evidence is required, these results suggest that LITT can somewhat affect the genetic fingerprints of recurrent gliomas, resulting in increased infiltration and, thus, evasion of standard treatment.

Clinically, LITT is used in the cytoreduction of recurrent gliomas more frequently than primary tumors. However, one report of primary glioma ablation, followed by histopathological assessment, is available. In this regard, our model replicates several findings on post-LITT ablation histopathology in human GBM.13 These authors also reported that LITT caused concentric destruction of tumor tissue with viable tumor cells just beyond the central necrosis and granulation zones. GFAP-positive astrocytes were also found immediately beyond these zones.13 The histological features of glioma in this model tracked along with the imaging biomarkers. The effects of laser ablation on tumor tissue have been reported in a mouse brain tumor model.34 Cortical GL261 or MGG8 tumors were localized by bioluminescence and ablated on the bench using an Nd-YAG laser fiber set to 2 W, with laser energy delivered for as long as 3 minutes with an accompanying temperature probe. Pre- and post-LITT MRI showed ablation of tumor tissue with periablation increases in blood-brain barrier permeability and LITT-extended lifespans of the treated mice. The present data confirm some of these observations and, in addition, demonstrate that LITT under imaging guidance is adaptable to a preclinical GBM model and allows precise image-guided ablation.

In clinical cases, patients diagnosed with GBM first undergo debulking to decrease the tumor burden before initiating standard treatment. Many preclinical models used for testing novel GBM therapeutics employ naïve orthotopic tumors. Considering this, some studies have reported preclinical resection models of GBM, including laser ablation. With the goal of safe maximal tumor resection, intraoperative tumor visualization techniques are increasingly used for glioma resection in the clinic.6 However, preclinical models with image-guided glioma cytoreduction are scarce. In most resection models, tumor implantation has been limited to the cerebral cortex due to the ease of visual localization and tumor removal without inflicting much mechanical injury to rest of the brain.8,35,36 Most experimental gliomas are implanted subcortically and are not amenable to surgical debulking without inflicting significant injury to the surrounding brain tissue.35 Laser-based cytoreduction has been employed in mouse glioma models with imaging confirmation before and after ablation.34 Resection of deep-seated mouse glioma has also been reported.37 Compared to these models, the image-assisted ablation presented here offers several advantages, primarily the ability to observe ablation in real time and achieve maximal safe cytoreduction, as in the clinic. The ability to longitudinally follow up with any intervening treatments after ablation allows the additional benefit of minimally invasive evaluation for treatment efficacy, presenting a model of post-cytoreduction recurrence to test new therapies. In future studies, post-LITT recurrent tumors may need to be subjected again to ablation to replicate the clinical scenario of LITT being used more frequently on recurrent tumors, be it after surgical debulking/resection or LITT.

Current clinical practice is to initiate the Stupp treatment protocol 3–4 weeks after GBM debulking.38 A median duration of 26 days between postoperative MRI and pre–radiotherapy MRI was reported in one study.39 Over the course of GBM disease progression, however, the immediate post-debulking phase has the least tumor burden and, in the case of LITT, an open blood-brain barrier in the vicinity of tumor cells that survived ablation. These data suggest that the acute post-cytoreduction stage may be the optimal time to suppress tumor cell repair and proliferation with adjuvant treatments before beginning standard therapy, with the goal of delaying or preventing treatment-resistant recurrence. Reports of GBM progression between cytoreduction and the beginning of standard treatment protocols support consideration of earlier initiation of post-cytoreduction adjuvant therapies.39,40

Study Limitations

We note that this proof of principle study was performed using only female, immune-suppressed RNU/RNU rats at relatively low laser power settings and short durations. These factors impose some restrictions on wider generalizations of the data. Additionally, although it captures the radiological features and mass effect of embedded solid primary glioma, the U251N cell line does not reflect the genetic variations and invasive properties of human GBM. These limitations necessitate confirmation of the results in immune-competent rats of both sexes bearing orthotopic glioma and subjected to LITT with follow-up imaging. Although the physical responses to local hyperthermia are likely comparable in immune-competent and -suppressed rats, as well as in female and male rats, employing an immune-competent animal model will generate data on acute and chronic immune and inflammatory responses in periablation brain tissue. With the reported suggestion that LITT can favorably predispose glioma to immune therapy,41 the generation of such preclinical data will be crucial to testing promising new immune therapies that can follow LITT for better control of recurrence.

Conclusions

Overall, the present data have several important implications to developing preclinical testing for GBM treatment. It is known that most recurrences occur close to the resected primary tumor margin in a majority of GBM patients.42–44 Akin to these reports, the persistence of tumor cells in periablation regions after LITT therapy for primary GBM has been also reported,13 resulting in local recurrence. The approach presented in this study, showing post-LITT glioma cell survival within the ablation vicinity, can be a useful preclinical model to study drug delivery and to evaluate putative treatment efficacy on either delaying or preventing GBM recurrence.

Acknowledgments

We thank Drs. Meser Ali and Asm Iskander for providing the U251N cells and Jun Xu and Andrea Transou for excellent technical assistance. We also thank Katie H. Hurrish, PhD, at Henry Ford Health, for editorial assistance. This study was supported by the Henry Ford Health System (grant no. A20050) and NIH/National Cancer Institute (NCI) (grant no. CA218596).

Disclosures

Dr. Lee reported personal fees from Medtronic and Monteris during the conduct of the study.

Author Contributions

Conception and design: Nagaraja, Cabral, Knight, Lee. Acquisition of data: Nagaraja, Bartlett, Cabral, Acharya, Hasselbach, Singh, Parasar, Spanick, Gowda, Knight, Ewing. Analysis and interpretation of data: Nagaraja, Bartlett, Cabral, Acharya, Morosini, Datta, Ayloo, Spanick, Gowda, deCarvalho, Noushmehr, Knight, Brown, Lee. Drafting the article: Nagaraja, Datta, Ayloo, Brown. Critically revising the article: Nagaraja, Bartlett, Acharya, Morosini, deCarvalho, Knight, Brown, Lee. Reviewed submitted version of manuscript: Nagaraja, Morosini, Datta, Hasselbach, Ayloo, deCarvalho, Knight, Brown, Ewing, Lee. Approved the final version of the manuscript on behalf of all authors: Nagaraja. Statistical analysis: Avritt, Gowda. Administrative/technical/material support: Ayloo, Ewing. Study supervision: Lee.

Supplemental Information

Videos

Video 1. https://vimeo.com/1163214473.

Online-Only Content
Supplemental material is available with the online version of the article.

Previous Presentations

Portions of this work were presented at the 6th Quadrennial Meeting of the World Federation of Neuro-Oncology Societies (WFNOS), Seoul, Korea, March 24–27, 2022.

References

  • 1. Johnson DR, O’Neill BP. Glioblastoma survival in the United States before and during the temozolomide era. J Neurooncol. 2012;107(2):359–364. doi: 10.1007/s11060-011-0749-4. [DOI] [PubMed] [Google Scholar]
  • 2. Stupp R, Mason WP, van den Bent MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 2005;352(10):987–996. doi: 10.1056/NEJMoa043330. [DOI] [PubMed] [Google Scholar]
  • 3. Delgado-López PD, Corrales-García EM. Survival in glioblastoma: a review on the impact of treatment modalities. Clin Transl Oncol. 2016;18(11):1062–1071. doi: 10.1007/s12094-016-1497-x. [DOI] [PubMed] [Google Scholar]
  • 4. Youngblood MW, Stupp R, Sonabend AM. Role of resection in glioblastoma management. Neurosurg Clin N Am. 2021;32(1):9–22. doi: 10.1016/j.nec.2020.08.002. [DOI] [PubMed] [Google Scholar]
  • 5. Serventi J, Behr J. Surgery and evidence-based treatments in patients with newly diagnosed high-grade glioma. Semin Oncol Nurs. 2018;34(5):443–453. doi: 10.1016/j.soncn.2018.10.009. [DOI] [PubMed] [Google Scholar]
  • 6. Nagaraja TN, Lee IY. Cerebral microcirculation in glioblastoma: a major determinant of diagnosis, resection, and drug delivery. Microcirculation. 2021;28(3):e12679. doi: 10.1111/micc.12679. [DOI] [PubMed] [Google Scholar]
  • 7. Emerich DF, Winn SR, Snodgrass P, et al. Injectable chemotherapeutic microspheres and glioma II: enhanced survival following implantation into deep inoperable tumors. Pharm Res. 2000;17(7):776–781. doi: 10.1023/a:1007591721877. [DOI] [PubMed] [Google Scholar]
  • 8. Emerich DF, Winn SR, Bartus RT. Injection of chemotherapeutic microspheres and glioma. IV: eradicating tumors in rats. Cell Transplant. 2002;11(1):47–54. doi: 10.3727/096020198389771. [DOI] [PubMed] [Google Scholar]
  • 9. Zhu H, Leiss L, Yang N, et al. Surgical debulking promotes recruitment of macrophages and triggers glioblastoma phagocytosis in combination with CD47 blocking immunotherapy. Oncotarget. 2017;8(7):12145–12157. doi: 10.18632/oncotarget.14553. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Agnihotri S, Burrell KE, Wolf A, et al. Glioblastoma, a brief review of history, molecular genetics, animal models and novel therapeutic strategies. Arch Immunol Ther Exp (Warsz) 2013;61(1):25–41. doi: 10.1007/s00005-012-0203-0. [DOI] [PubMed] [Google Scholar]
  • 11. Aldape K, Brindle KM, Chesler L, et al. Challenges to curing primary brain tumours. Nat Rev Clin Oncol. 2019;16(8):509–520. doi: 10.1038/s41571-019-0177-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Nagaraja TN, Bartlett S, Farmer KG, et al. Adaptation of laser interstitial thermal therapy for tumor ablation under MRI monitoring in a rat orthotopic model of glioblastoma. Acta Neurochir (Wien) 2021;163(12):3455–3463. doi: 10.1007/s00701-021-05002-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Elder JB, Huntoon K, Otero J, et al. Histologic findings associated with laser interstitial thermotherapy for glioblastoma multiforme. Diagn Pathol. 2019;14(1):19. doi: 10.1186/s13000-019-0794-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Chu KF, Dupuy DE. Thermal ablation of tumours: biological mechanisms and advances in therapy. Nat Rev Cancer. 2014;14(3):199–208. doi: 10.1038/nrc3672. [DOI] [PubMed] [Google Scholar]
  • 15. Carpentier A, Chauvet D, Reina V, et al. MR-guided laser-induced thermal therapy (LITT) for recurrent glioblastomas. Lasers Surg Med. 2012;44(5):361–368. doi: 10.1002/lsm.22025. [DOI] [PubMed] [Google Scholar]
  • 16. Oertel MF, Stieglitz LH, Bozinov O. A stereotactic frame-based drill guide-aided setting for laser interstitial thermal therapy. Acta Neurochir (Wien) 2021;163(12):3447–3453. doi: 10.1007/s00701-021-04869-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Lee I, Kalkanis S, Hadjipanayis CG. Stereotactic laser interstitial thermal therapy for recurrent high-grade gliomas. Neurosurgery. 2016;79(Suppl 1):S24–S34. doi: 10.1227/NEU.0000000000001443. [DOI] [PubMed] [Google Scholar]
  • 18. Leuthardt EC, Duan C, Kim MJ, et al. Hyperthermic laser ablation of recurrent glioblastoma leads to temporary disruption of the peritumoral blood brain barrier. PLoS One. 2016;11(2):e0148613. doi: 10.1371/journal.pone.0148613. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Rammo R, Asmaro K, Schultz L, et al. The safety of magnetic resonance imaging-guided laser interstitial thermal therapy for cerebral radiation necrosis. J Neurooncol. 2018;138(3):609–617. doi: 10.1007/s11060-018-2828-2. [DOI] [PubMed] [Google Scholar]
  • 20. Rammo R, Scarpace L, Nagaraja T, Lee I. MR-guided laser interstitial thermal therapy in the treatment of recurrent intracranial meningiomas. Lasers Surg Med. 2019;51(3):245–250. doi: 10.1002/lsm.23045. [DOI] [PubMed] [Google Scholar]
  • 21. Elmghirbi R, Nagaraja TN, Brown SL, et al. Toward a noninvasive estimate of interstitial fluid pressure by dynamic contrast-enhanced MRI in a rat model of cerebral tumor. Magn Reson Med. 2018;80(5):2040–2052. doi: 10.1002/mrm.27163. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Ewing JR, Nagaraja TN, Aryal MP, et al. Peritumoral tissue compression is predictive of exudate flux in a rat model of cerebral tumor: an MRI study in an embedded tumor. NMR Biomed. 2015;28(11):1557–1569. doi: 10.1002/nbm.3418. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Nagaraja TN, Aryal MP, Brown SL, et al. Cilengitide-induced temporal variations in transvascular transfer parameters of tumor vasculature in a rat glioma model: identifying potential MRI biomarkers of acute effects. PLoS One. 2013;8(12):e84493. doi: 10.1371/journal.pone.0084493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Brown SL, Nagaraja TN, Aryal MP, et al. MRI-tracked tumor vascular changes in the hours after single-fraction irradiation. Radiat Res. 2015;183(6):713–721. doi: 10.1667/RR13458.1. [DOI] [PubMed] [Google Scholar]
  • 25. Nagaraja TN, Elmghirbi R, Brown SL, et al. Imaging acute effects of bevacizumab on tumor vascular kinetics in a pre-clinical orthotopic model of U251 glioma. NMR Biomed. 2021;34(7):e4516. doi: 10.1002/nbm.4516. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Nagaraja TN, deCarvalho AC, Brown SL, et al. The impact of initial tumor microenvironment on imaging phenotype. Cancer Treat Res Commun. 2021;27:100315. doi: 10.1016/j.ctarc.2021.100315. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years of image analysis. Nat Methods. 2012;9(7):671–675. doi: 10.1038/nmeth.2089. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Dobin A, Davis CA, Schlesinger F, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29(1):15–21. doi: 10.1093/bioinformatics/bts635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics. 2014;30(7):923–930. doi: 10.1093/bioinformatics/btt656. [DOI] [PubMed] [Google Scholar]
  • 30. Danecek P, Bonfield JK, Liddle J, et al. Twelve years of SAMtools and BCFtools. Gigascience. 2021;10(2):giab008. doi: 10.1093/gigascience/giab008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Krämer A, Green J, Pollard J, Jr, Tugendreich S. Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics. 2014;30(4):523–530. doi: 10.1093/bioinformatics/btt703. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Nagaraja TN, Elmghirbi R, Brown SL, et al. Reproducibility and relative stability in magnetic resonance imaging indices of tumor vascular physiology over a period of 24h in a rat 9L gliosarcoma model. Magn Reson Imaging. 2017;44:131–139. doi: 10.1016/j.mri.2017.09.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Baker GJ, Yadav VN, Motsch S, et al. Mechanisms of glioma formation: iterative perivascular glioma growth and invasion leads to tumor progression, VEGF-independent vascularization, and resistance to antiangiogenic therapy. Neoplasia. 2014;16(7):543–561. doi: 10.1016/j.neo.2014.06.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Salehi A, Paturu MR, Patel B, et al. Therapeutic enhancement of blood-brain and blood-tumor barriers permeability by laser interstitial thermal therapy. Neurooncol Adv. 2020;2(1):vdaa071. doi: 10.1093/noajnl/vdaa071. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Emerich DF, Winn SR, Hu Y, et al. Injectable chemotherapeutic microspheres and glioma I: enhanced survival following implantation into the cavity wall of debulked tumors. Pharm Res. 2000;17(7):767–775. doi: 10.1023/a:1007576405039. [DOI] [PubMed] [Google Scholar]
  • 36. Ji X, Chen S, Guo Y, et al. Establishment and evaluation of four different types of patient-derived xenograft models. Cancer Cell Int. 2017;17:122. doi: 10.1186/s12935-017-0497-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Das A, Henderson F, Lowe S, et al. Single agent efficacy of the HDAC inhibitor DATS in preclinical models of glioblastoma. Cancer Chemother Pharmacol. 2018;82(6):945–952. doi: 10.1007/s00280-018-3684-7. [DOI] [PubMed] [Google Scholar]
  • 38. Mutter N, Stupp R. Temozolomide: a milestone in neuro-oncology and beyond? Expert Rev Anticancer Ther. 2006;6(8):1187–1204. doi: 10.1586/14737140.6.8.1187. [DOI] [PubMed] [Google Scholar]
  • 39. Kraus RD, Weil CR, Frances Su FC, Cannon DM, Burt LM, Mendez JS. Incidence and extent of disease progression on MRI between surgery and initiation of radiotherapy in glioblastoma patients. Neurooncol Pract. 2022;9(5):380–389. doi: 10.1093/nop/npac044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Villanueva-Meyer JE, Han SJ, Cha S, Butowski NA. Early tumor growth between initial resection and radiotherapy of glioblastoma: incidence and impact on clinical outcomes. J Neurooncol. 2017;134(1):213–219. doi: 10.1007/s11060-017-2511-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Chandar JS, Bhatia S, Ingle S, et al. Laser interstitial thermal therapy induces robust local immune response for newly diagnosed glioblastoma with long-term survival and disease control. J Immunother. 2023;46(9):351–354. doi: 10.1097/CJI.0000000000000485. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Hochberg FH, Pruitt A. Assumptions in the radiotherapy of glioblastoma. Neurology. 1980;30(9):907–911. doi: 10.1212/wnl.30.9.907. [DOI] [PubMed] [Google Scholar]
  • 43. Lemée JM, Clavreul A, Menei P. Intratumoral heterogeneity in glioblastoma: don’t forget the peritumoral brain zone. Neuro Oncol. 2015;17(10):1322–1332. doi: 10.1093/neuonc/nov119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Petrecca K, Guiot MC, Panet-Raymond V, Souhami L. Failure pattern following complete resection plus radiotherapy and temozolomide is at the resection margin in patients with glioblastoma. J Neurooncol. 2013;111(1):19–23. doi: 10.1007/s11060-012-0983-4. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Online-Only Content
Supplemental material is available with the online version of the article.
Supplemental_Data_and_Excel_Spreadsheets.zip (5.2MB, zip)

Articles from Journal of Neurosurgery are provided here courtesy of American Association of Neurological Surgeons

RESOURCES