Abstract
Objective
Glioblastoma (GB) is the deadliest primary brain tumor, largely due to inevitable recurrence of the disease after partial surgical resection or resistance to drug treatments. The study aimed to evaluate the predictive values of apparent diffusion coefficient (ADC) from diffusion-weighted imaging (DWI), plasma levels of placental growth factor (PGF) and glial fibrillary acidic protein (GFAP), and their combination in GB recurrence.
Methods
The study included 200 patients diagnosed with GB consisting of 136 with tumor recurrence and 64 without. DWI examinations and PGF and GFAP measurements in the plasma were performed preoperatively.
Results
The patients with recurrent disease exhibited a lower value of ADC with higher plasma PGF and GFAP levels than those with non-recurrence. The results of Pearson correlation analysis suggested the ADC shared negative correlations with plasma levels of PGF and GFAP in patients with recurrent GB. The plasma level of PGF was also found to be positively correlated with the plasma level of GFAP in patients with recurrent GB. The ADC, plasma levels of PGF and GFAP in predicting recurrent GB presented AUC: 0.75, 0.85, and 0.82, respectively. Combined analysis of two of three to evaluate the prediction showed AUCs of 0.98 (ADC and PGF), 0.96 (ADC and GFAP), and 0.84 (PGF and GFAP). Combined analysis of ADC and plasma PGF, ADC and plasma GFAP, PGF and GFAP to evaluate the prediction showed AUCs of 0.98, 0.96, and 0.84, respectively. Combined analysis of three of them to evaluate the prediction showed AUC as 0.98.
Conclusion
The study demonstrates that combined ADC and plasma PGF showed a similar predictive value in recurrent GB after standard treatment as combined ADC, plasma PGF and GFAP together. Considering limited sample size and lack of more sample types in this study, additional value of plasma GFAP in predicting recurrent GB should be further investigated in larger-scale population or different sample sources.
Keywords: Glioblastoma, Diffusion-weighted imaging, Placental growth factor, Glial fibrillary acidic protein, Prediction
Introduction
Glioblastoma (GB), classified as a grade IV glioma in the World Health Organization (WHO) classification of brain tumors, accounts for approximately 49% of primary malignant brain tumors, leading to substantial morbidity and mortality worldwide [1, 2]. The standard treatment for newly diagnosed GB is maximal safe surgical resection followed by concurrent radiotherapy with temozolomide (TMZ) and further adjuvant TMZ [3, 4]. Although the gold-standard first-line treatment for GB is available, this disease has poor prognosis with the median overall survival of about 15–18 months and the five-year survival rate below 10% [5, 6]. For most cases of GB, recurrence is an inevitable event after 6–9 months of standard treatment, and antiangiogenic therapy with an anti-vascular endothelial growth factor (VEGF) agent has been used for recurrent GB [7]. Therefore, accurately and earlier predicting GB recurrence hold the premise of tailored treatment enabling changes in therapy to prevent treatment ineffectiveness or adverse events [8]. Although earlier work has studied advanced imaging modalities, such as diffusion-weighted imaging (DWI) and perfusion-weighted imaging (PWI), for patient follow-up to differentiate true tumor progression/recurrence, there are still significant challenges due to limitations in acquisition variability, costs, and reproducibility [9, 10].
The initial lesion characteristics in the brain by magnetic resonance imaging (MRI) were found to be a prognostic marker of tumor recurrence and overall survival in GB [11, 12]. Recently, with the rapid development of molecular pathology, the combination of imaging examinations and molecular biomarkers has provided a promising noninvasive approach for predicting GB recurrence [13]. Placental growth factor (PGF) is a member of VEGF family that have been implicated in resistance to anti-angiogenic therapies in brain tumors [14]. Glial fibrillary acidic protein (GFAP), as a principal component of intermediate filament, is predominantly expressed in mature astrocytes of the central nervous system (CNS) [15]. Release of GFAP from tissue into the bloodstream resulting from astrocyte disintegration is found in neurological disorders including GB, suggesting GFAP can serve as blood biomarkers in the diagnosis and prognosis of GB [16, 17]. Plasma levels of PGF and GFAP exhibited promising potential in differentiating between unifocal GB and unifocal supratentorial intracerebral metastasis [18]. Herein, we propose a hypothesis that integrating plasma levels of PGF and GFAP with neuroimaging modalities offers the potential to improve disease characterization and monitor disease progression at a molecular level in GB. In this study, we evaluated the predictive values of DWI, plasma levels of PGF and GFAP, and their combination in GB recurrence.
Methods
Study population
The study included 200 patients who had histological confirmation of GB (WHO grade IV) based on the 2021 WHO Classification of Tumors of the Central Nervous System [19] at Ganzhou People’s Hospital between January 2023 and December 2024. All patients underwent surgical resection followed by chemoradiotherapy and no more than one adjuvant TMZ-containing regimen. All patients routinely visited the outpatient clinic and underwent follow-up brain MRI with a brain tumor evaluation protocol at our institution. The first follow-up occurred 6 weeks after therapy was completed, and subsequent follow-ups were scheduled every 3 months during the follow-up period of 1 year. All the recurrence patients were followed up until first relapse. According to the Response Assessment in Neuro-Oncology (RANO) criteria [20], the patients were classified into disease recurrence and non-recurrence groups. The progression free survival (PFS) was defined as the interval between the initial diagnosis by MRI examination and the assessment of disease progression. All patients with disease progression were followed up until first relapse. Additional inclusion criteria were Karnofsky performance status score ≥ 70 [21], life expectancy more than 12 weeks [22], and aged 18 or older. Patients were excluded if they had non-standard treatment after surgery, a history of intracerebral or intratumoral hemorrhage, a history of significant cardiovascular disease, significant intercurrent illness, concurrent malignancy, pregnancy/breast feeding, incomplete imaging data or images of insufficient quality, or refusal to participate in this study.
DWI-MRI and imaging processing
The MRI examination was performed on a 3.0T scanner (Discovery MR750, GE Healthcare, Milwaukee, Wisconsin, USA) and a 32-channel torso coil. The DWI was performed in the axial plane with echo planner imaging sequence using the following parameters: repetition time (TR) = 3300 msec; echo time (TE) = 65.8 msec; field of view (FOV) = 240 × 240 mm2; matrix = 160 × 160; slice thickness = 4 mm; scan time = 26 s; pair of b-values: 0 and 1000 (sec/mm2). The DWI images obtained with b values of 1000 and 0 s/mm2 were transferred from the scanner to an off-line workstation (GE Advanced Work-station 4.6) in DICOM format and the apparent diffusion coefficient (ADC) was calculated by offline fitting the signal with these two b values. The region of interest (ROI) was delineated on T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), and T2WI-fluid attenuated inversion recovery (T2WI-FLAIR) sequences by selecting the brain glioma parenchyma, peritumoral edema area, relatively normal white matter, contralateral normal white matter, and avoiding necrosis, cystic degeneration, sulci, and ventricles. A minimum of three measurements were taken and subsequently averaged for accuracy. All measurements were taken by two radiologists who both had 5 years of experience in CNS diagnosis and were blinded to clinical outcomes. Any disagreement between the two neuroradiologists was resolved by consensus.
Measurements of PGF and GFAP
The blood samples were collected from each patient prior to the neurosurgical intervention and placed into a 2 ml vacutainer tube containing ethylenediaminetetraacetic acid (EDTA). After centrifugation at 3,000 rounds per minute for 10 min at 4 °C, the plasma was separated and stored at −80 °C. The plasma concentrations of PGF and GFAP were determined by using commercially available human ELISA kits (Product # EHPGF and Product # EEL079) according to the manufacture’s protocols.
Statistical analysis
The normality of the distributions of continuous variables was evaluated using the Shapiro-Wilk test. Descriptive statistics were performed to present a summary of continuous variables. When normal distribution was confirmed, continuous variables were summarized in the form of mean ± standard deviation (s.d.) and group differences were analyzed by the independent t‑test. When being out of normal distribution, continuous variables were summarized in the form of median with interquartile range (IQR) and group differences were analyzed by the Mann-Whitney U test. Categorical variables were reported as numbers with percentages and group differences were analyzed by Chi-square and Fisher’s exact tests. Pearson correlation analysis was performed to examine correlation between plasma PGF and GFAP, between plasma PGF and ADC, between plasma GFAP and ADC. The predictive values of three variables, individually and in combination, for recurrent disease were assessed by the receiver operating characteristic (ROC) curves and its summary statistics [area under the curve (AUC)] with 95% confidence intervals (CIs). The binary logistic regression model was performed to combine single variables and create the predictive models. The Proportion of Variance Explained (PRE) value in the model was used to yield the predictive values of combined variables. All statistical analyses were conducted by using IBM SPSS Statistics 27.0 for Windows (IBM, Armonk, NY, USA) and significant difference was indicated by a p-value of < 0.05 (two tailed).
Results
Patient’s characteristics
A total of 200 patients with GB were included in the study, including 136 patients with evidence of unequivocal progression and 64 without progression (Fig. 1). The median PFS of the study patients were 10.0 months. Table 1 provides a summary of patient’s characteristics. No significant difference was noted regarding age, sex, Karnofsky performance status score, Isocitrate Dehydrogenase 1 (IDH1) status, and surgical type (p > 0.05).
Fig. 1.
Flowchart illustrating patient selection and inclusion/exclusion criteria. GB, glioblastoma
Table 1.
Patient’s characteristics
| Characteristic | Recurrence (n = 136) | Non-recurrence (n = 64) | p |
|---|---|---|---|
| Age, year (mean ± s.d.) | 54.7 ± 9.3 | 53.8 ± 9.9 | 0.532 |
| Sex, n (%) | 0.641 | ||
| Male | 86 (63.2%) | 47 (75.0%) | |
| Female | 50 (36.8%) | 17 (25.0%) | |
| Karnofsky performance status | 0.257 | ||
| Median | 80 | 80 | |
| IQR | 70–100 | 70–100 | |
| IDH1 status, n (%) | 0.078 | ||
| Wild-type IDH1 | 132 (97.1%) | 58 (90.6%) | |
| Mutated IDH1 | 4 (2.9%) | 6 (9.4%) | |
| Surgical type, n (%) | 0.349 | ||
| Partial resection | 52 (38.2%) | 20 (31.3%) | |
| Complete resection | 84 (61.8%) | 44 (68.8%) |
s.d., standard deviation; IQR, interquartile range; IDH1, Isocitrate Dehydrogenase 1; Data summarized as median with IQR were analyzed by Mann-Whitney U test and those summarized mean ± s.d. were analyzed by independent t test. Categorical variables are analyzed by using the chi-squared tests
The ADC, plasma PGF and GFAP in patients with recurrent GB
The ADC, plasma levels of PGF and GFAP in patients with GB prior to the neurosurgical intervention were summarized as 1.01 × 103 mm2/s (IQR: 0.94–1.12 × 103 mm2/s), 8.42 pg/mL (IQR: 0.50–14.07 pg/mL), and 0.48 ng/mL (IQR: 0.30–1.65 ng/mL). The value of ADC in the patients with recurrent disease were 0.99 × 103 mm2/s (IQR: 0.92–1.04 × 103 mm2/s), which was decreased compared to 1.14 × 103 mm2/s (IQR: 0.99–1.16 × 103 mm2/s) in those with non-recurrence (p < 0.0001; Fig. 2A). The values of plasma PGF and GFAP levels were 12.03 pg/mL (IQR: 6.46–14.26 pg/mL) and 0.96 ng/mL (IQR: 0.45–3.01 ng/mL), which were higher than 2.03 pg/mL (IQR: 0.00–4.81 pg/mL) and 0.32 ng/mL (IQR: 0.23–0.40 ng/mL) those with non-recurrence (p < 0.0001; Fig. 2B, C).
Fig. 2.
The ADC, plasma PGF and GFAP between patients with recurrent GB and those without. A, ADC, B, Plasma PGF level. C, Plasma GFAP level. Group difference was analyzed by Mann-Whitney U test. **** p < 0.0001. ADC, apparent diffusion coefficient; GFAP, glial fibrillary acidic protein; PGF, placental growth factor;
Correlation between ADC, plasma PGF and GFAP in patients with recurrent GB
The results of Pearson correlation analysis (Table 2) suggested the ADC shared negative correlations with plasma levels of PGF and GFAP in patients with recurrent GB (p < 0.0001). The plasma level of PGF was also found to be positively correlated with the plasma level of GFAP in patients with recurrent GB (p < 0.0001).
Table 2.
Correlation between ADC, plasma PGF and GFAP in patients with recurrent GB
| ADC (103 mm2/s) | PGF (pg/mL) | GFAP (ng/mL) | |
|---|---|---|---|
| ADC (103 mm2/s) | – | r = 0.836 (95%CI: 0.777–0.881) | r = 0.706 (95%CI: 0.610–0.781) |
| PGF (pg/mL) | r = 0.836 (95%CI: 0.777–0.881) | – | r = 0.669 (0.564–0.753) |
| GFAP (ng/mL) | r = 0.706 (95%CI: 0.610–0.781) | r = 0.669 (0.564–0.753) | – |
ADC, apparent diffusion coefficient; PGF, placental growth factor; GFAP, glial fibrillary acidic protein
Prediction of recurrent GB using ADC, plasma PGF and GFAP
The ADC value, plasma levels of PGF and GFAP applied to predict recurrent GB showed AUCs of 0.75 (95%CI: 0.67–0.83), 0.85 (95%CI: 0.79–0.90), and 0.82 (95%CI: 0.76–0.87), respectively (Fig. 3A-C). Combined analysis of ADC and plasma PGF to evaluate the prediction showed an AUC of 0.98 (95%CI: 0.96–0.99) (Fig. 3D). Combined ADC and plasma GFAP showed an AUC of 0.96 (95%CI: 0.94–0.99) (Fig. 3E). Combined plasma levels of PGF and GFAP showed an AUC of 0.84 (95%CI: 0.78–0.89) (Fig. 3F). Combined three of them showed an AUC of 0.98 (95%CI: 0.97–1.00) (Fig. 3G). The sensitivity, specificity and cutoff values are listed in Table 3.
Fig. 3.
ROC curves of ADC, plasma PGF and GFAP, individually and in combination, in predicting recurrent GB. A, ADC, B, Plasma PGF level. C, Plasma GFAP level. D, Combined analysis of ADC and plasma PGF. E, Combined analysis of ADC and plasma GFAP. F, Combined analysis of plasma levels of PGF and GFAP. G, Combined analysis of three variables
Table 3.
The AUC with 95%CI, sensitivity, specificity, and cutoff value of plasma GFAP, plasma PGF, and their combination used to predict drug response in epileptic patients
| Item | Sensitivity | Specificity | Cutoff value |
|---|---|---|---|
| ADC, 103 mm2/s | 76.5% | 67.2% | 1.04 × 103 mm2/s |
| Plasma PGF, pg/ml | 80.1% | 85.9% | 5.47 pg/mL |
| Plasma GFAP, ng/ml | 77.2% | 84.4% | 0.43 ng/mL |
| ADC + PGF | 91.9% | 95.3% | PRE value: 0.70 |
| ADC+GFAP | 94.9% | 92.2% | PRE value: 0.65 |
| PGF+GFAP | 79.4% | 84.4% | PRE value: 0.63 |
| ADC + PGF+GFAP | 92.6% | 95.3% | PRE value: 0.68 |
AUC, area under the curve; 95%CI, 95% confidence interval; GFAP, glial fibrillary acidic protein; PGF, placental growth factor; PRE, Proportion of Variance Explained
Discussion
The study compared the ADC value, plasma levels of PGF and GFAP between patients with recurrent GB and those without, and analyzed the predictive values of the ADC value, plasma levels of PGF and GFAP, as well as their combination in recurrent GB. The findings obtained from the study showed good predictive values by combining DWI with plasma levels of PGF or GFAP for GB patients with tumor recurrence after standard treatment.
Conventional MRI examinations sometimes are inadequate for characterize glioma due to the enhancement of gliomas possibly caused by disruption of the blood-brain barrier rather than neovascularization [23]. DWI is an advanced functional imaging method that has been widely applied to detect and characterize GB, which provide additional information about tumor physiology that is not accessible with conventional sequences [24]. DWI is sensitive to the molecular movement of water with areas of reduced diffusion (low ADC) in relation to increased tumor cellularity [25]. DWI obtained 24 h after MRI-guided laser interstitial thermal therapy can help predict the location of GB recurrence months before the development of abnormal enhancement [26]. GB patients showing a diffusion MRI phenotype after chemoradiation has been found to predict a favorable response to anti-VEGF, suggesting the role of DWI in the prognostic evaluation of GB [27]. In the pretreatment setting, DWI has been used as tool to not only help narrow the differential diagnosis and pathological classification of gliomas, but also evaluate prognosis and treatment response [28]. Although DWI was performed at radiation planning for patients with GB in their study, Moore-Palhares et al. [29] believed temporal ADC changes are promising imaging biomarkers for treatment response monitoring and spatial recurrence prediction, which was partially similar as our results.
Compared to only depending on one biomarker category, integrating diverse biomarker types, such as circulating neuro-biomarkers and imaging biomarkers may likely improve diagnostic sensitivity and specificity and addressing several existing limitations with regard to liquid biomarkers for GB [30, 31]. Malignant gliomas are vascular tumors that produce VEGF, and PGF is one of members of VEGF family and binds exclusively to VEGF receptor 1 (VEGFR-1) [32]. Inhibition of PGF, either by genetic inhibition or by pharmacological blockade using distinct independently generated anti-PLGF antibodies, was found to reduce tumor growth and induce regression of pre-existing medulloblastoma [14]. Although remining controversial, anti-VEGF therapies achieved unprecedented improvements in radiographic response and longer survival for patients with recurrent GB, indicating a significant role of PGF in recurrent GB [33]. Pretreatment diffusion MRI could predict overall survival in patients with recurrent GB after anti-VEGF monotherapy at first or second relapse [34]. This may explain the ADC association with the plasma level of PGF in our study. Wang et al. [35] reported a humanized anti-PGF monoclonal antibody RO5323441 which could bind with high affinity to PGF and prevents the growth factor from binding to VEGFR-1 shared an apparent drug-drug interaction with bevacizumab, indicating the synergetic effect of anti-PGF antibody combined with bevacizumab in treating recurrent GB.
GFAP, as a cytoskeletal intermediate filament protein, is considered as a crucial indicator of astrocytic activation and neuroaxonal injury. An increased concentration of GFAP in cerebrospinal fluid (CSF) and blood-based samples (serum/plasma) has become a potential biomarker to differentiate multiple neurological diseases and monitor disease progression [36]. In recurrent GB, GFAP+VEGF bright+MMP2 + C5b-9 + small extracellular vesicles were detected and shared a positive correlation with O6 -methylguanine-DNA-methyltransferase (MGMT) gene promoter methylation levels, indicating the prognostic values of these markers in glioblastoma recurrence and treatment responses [37]. Serum GFAP levels were detectable in 40 out of the 50 GB patients, which were significantly increased compared to those of the non-GB tumor patients and healthy controls [38]. Consistent with our results, Kiviniemi et al. [39] found an elevated serum level of GFAP was associated with IDH1 mutation-negative high-grade gliomas and poor progression-free survival, indicating the potential of serum GFAP for detecting tumor recurrence in high-grade gliomas. Besides, Tichy et al. [40] found serum levels of GFAP were correlated with histopathological findings and MRI parameters and the levels were significantly higher in GBM patients than all other tumors. GB patients with high GFAP levels exhibited more necrosis as well as greater perilesional edema than those with low or non-detectable GFAP levels. These previous observations were similar as a positive association between plasma GFAP level and GB in patients with recurrent GB in our study.
Although radiomics-, genomics-, and proteometabolomics-based approaches have been increasingly applied to predict the prognosis, gene mutation, pathogenesis, and therapeutic responses of patients with GB [41, 42]. However, these approaches have limited availability in most resource-constrained settings due to their high costs and time-consuming steps. In our study, sampling and detection methods are easily obtained and the predictive performance was satisfactory.
The study has several limitations. One important limitation is that the plasma analyses for two neuro-biomarkers rely on commercially available ELISA tests which detect merely the levels above the level of detection given for each assay. Further studies should consider proteomic-based approaches that may enable the detection of a broader levels of these two neuro-biomarkers. A second limitation is lack of CSF sample detection comparing to plasma sample. Actually, CSF is a preferred sample type for liquid biopsy in brain tumor patients to reflect CNS pathology due to more abundance of ctDNA and protein in CSF than peripheral blood. Multiple sampling sourced from CSF and peripheral blood could validated our results. Thus, larger prospective trials with large sets of samples from multiple sources are warrant to support selection of a correct biological fluid for the differential diagnosis of GB recurrence.
In conclusion, our results demonstrate the combined ADC and plasma PGF showed a similar predictive value in recurrent GB after standard treatment as combined ADC, plasma PGF and GFAP together. The study suggests integrating DWI with candidate neuro-biomarkers may help narrow the differential diagnosis between true tumor recurrence and pseudoprogression, and alter future treatment planning. However, additional value of plasma GFAP in predicting recurrent GB should be further investigated in larger-scale population or different sample sources.
Acknowledgements
Not applicable.
Author contributions
YYJ conceived the study, did the methodology, and wrote the first draft of manuscript. DSM and ZZQ obtained the data and did data analysis. TWJ did the figure visualization and revised the manuscript. All authors reviewed and approved the manuscript.
Funding
The study was supported by the Science and Technology Program of the Health Commission of Jiangxi Province (Grant No. 2025110045) and the Ganzhou Science and Technology Planning Project (Grant No. 2025YLCE0069, GZ2024ZSF064).
Data availability
Data is provided within the manuscript.
Declarations
Ethics approval and consent to participate
The study was approved by the Ethics Committee of Ganzhou People’s Hospital (PJB2023-115-01) and performed in accordance with the Declaration of Helsinki. Each participant provided informed consent before use of their biological materials.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Singh S, Dey D, Barik D, Mohapatra I, Kim S, Sharma M, Prasad S, Wang P, Singh A, Singh G. Glioblastoma at the crossroads: current understanding and future therapeutic horizons. Signal Transduct Target Ther. 2025;10(1):213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Schaff LR, Mellinghoff IK. Glioblastoma and other primary brain malignancies in adults: a review. JAMA. 2023;329(7):574–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wen PY, Weller M, Lee EQ, Touat M, Khasraw M, Rahman R, Platten M, Lim M, Winkler F, Horbinski C, 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. 2025;27(11):2751–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Tan AC, Ashley DM, Lopez GY, Malinzak M, Friedman HS, Khasraw M. Management of glioblastoma: state of the art and future directions. CA Cancer J Clin. 2020;70(4):299–312. [DOI] [PubMed] [Google Scholar]
- 5.Krolikowska K, Blaszczak K, Lawicki S, Zajkowska M, Gudowska-Sawczuk M. Glioblastoma-a contemporary overview of epidemiology, classification, pathogenesis, diagnosis, and treatment: a review article. Int J Mol Sci. 2025;26(24):12162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sipos D, Raposa BL, Freihat O, Simon M, Mekis N, Cornacchione P, Kovacs A. Glioblastoma: clinical presentation, multidisciplinary management, and long-term outcomes. Cancers (Basel). 2025. 10.3390/cancers17010146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Drappatz J, Mantica M. Treatment strategies and innovation for recurrent high-grade glioma. J Neurooncol. 2025;176(1):118. [DOI] [PubMed] [Google Scholar]
- 8.Qi D, Li J, Quarles CC, Fonkem E, Wu E. Assessment and prediction of glioblastoma therapy response: challenges and opportunities. Brain. 2023;146(4):1281–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Tsakiris C, Siempis T, Alexiou GA, Zikou A, Sioka C, Voulgaris S, Argyropoulou MI. Differentiation between true tumor progression of glioblastoma and pseudoprogression using diffusion-weighted imaging and perfusion-weighted imaging: systematic review and meta-analysis. World Neurosurg. 2020;144:e100-9. [DOI] [PubMed] [Google Scholar]
- 10.Hilario A, Salvador E, Cardenas A, Romero J, Lechuga C, Chen Z, et al. Low rCBV values in glioblastoma tumor progression under chemoradiotherapy. Neuroradiology. 2024;66(3):317–23. [DOI] [PubMed] [Google Scholar]
- 11.Lundemann M, Munck Af Rosenschold P, Muhic A, Larsen VA, Poulsen HS, Engelholm SA, Andersen FL, Kjaer A, Larsson HBW, Law I, et al. Feasibility of multi-parametric PET and MRI for prediction of tumour recurrence in patients with glioblastoma. Eur J Nucl Med Mol Imag. 2019;46(3):603–13. [DOI] [PubMed] [Google Scholar]
- 12.Ismail M, Hill V, Statsevych V, Mason E, Correa R, Prasanna P, et al. Can tumor location on pre-treatment MRI predict likelihood of pseudo-progression vs. tumor recurrence in glioblastoma?-A feasibility study. Front Comput Neurosci. 2020;14:563439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Liu W, Li Z. Diagnostic performance of perfusion-weighted imaging combined with serum MMP-2 and − 9 levels in tumor recurrence after postoperative concomitant chemoradiotherapy of glioblastoma. J Clin Ultrasound. 2023;51(3):563–70. [DOI] [PubMed] [Google Scholar]
- 14.Dewerchin M, Carmeliet P. Placental growth factor in cancer. Expert Opin Ther Targets. 2014;18(11):1339–54. [DOI] [PubMed] [Google Scholar]
- 15.Li D, Liu X, Liu T, Liu H, Tong L, Jia S, Wang YF. Neurochemical regulation of the expression and function of glial fibrillary acidic protein in astrocytes. Glia. 2020;68(5):878–97. [DOI] [PubMed] [Google Scholar]
- 16.Heimfarth L, Passos FRS, Monteiro BS, Araujo AAS, Quintans Junior LJ, Quintans JSS. Serum glial fibrillary acidic protein is a body fluid biomarker: a valuable prognostic for neurological disease - a systematic review. Int Immunopharmacol. 2022;107:108624. [DOI] [PubMed] [Google Scholar]
- 17.Ahmadipour Y, Gembruch O, Pierscianek D, Sure U, Jabbarli R. Does the expression of glial fibrillary acid protein (GFAP) stain in glioblastoma tissue have a prognostic impact on survival? Neurochirurgie. 2020;66(3):150–4. [DOI] [PubMed] [Google Scholar]
- 18.Ilhan-Mutlu A, Wagner L, Widhalm G, Wohrer A, Bartsch S, Czech T, Heinzl H, Leutmezer F, Prayer D, Marosi C, et al. Exploratory investigation of eight circulating plasma markers in brain tumor patients. Neurosurg Rev. 2013;36(1):45–55 discussion 55 – 46. [DOI] [PubMed] [Google Scholar]
- 19.Louis DN, Perry A, Wesseling P, Brat DJ, Cree IA, Figarella-Branger D, Hawkins C, Ng HK, Pfister SM, Reifenberger G, et al. The 2021 WHO classification of tumors of the central nervous system: a summary. Neuro Oncol. 2021;23(8):1231–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Wen PY, Macdonald DR, Reardon DA, Cloughesy TF, Sorensen AG, Galanis E, Degroot J, Wick W, Gilbert MR, Lassman AB, et al. Updated response assessment criteria for high-grade gliomas: response assessment in neuro-oncology working group. J Clin Oncol. 2010;28(11):1963–72. [DOI] [PubMed] [Google Scholar]
- 21.Chamberlain MC, Johnston SK, Glantz MJ. Neoplastic meningitis-related prognostic significance of the Karnofsky performance status. Arch Neurol. 2009;66(1):74–8. [DOI] [PubMed] [Google Scholar]
- 22.Scoccianti S, Perna M, Olmetto E, Delli Paoli C, Terziani F, Ciccone LP, Detti B, Greto D, Simontacchi G, Grassi R, et al. Local treatment for relapsing glioblastoma: a decision-making tree for choosing between reirradiation and second surgery. Crit Rev Oncol Hematol. 2021;157:103184. [DOI] [PubMed] [Google Scholar]
- 23.Eraky AM, Beck RT, Treffy RW, Aaronson DM, Hedayat H. Role of advanced MR imaging in diagnosis of neurological malignancies: current status and future perspective. J Integr Neurosci. 2023;22(3):73. [DOI] [PubMed] [Google Scholar]
- 24.Kang XW, Xi YB, Liu TT, Wang N, Zhu YQ, Wang XR, Guo F. Grading of Glioma: combined diagnostic value of amide proton transfer weighted, arterial spin labeling and diffusion weighted magnetic resonance imaging. BMC Med Imaging. 2020;20(1):50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Saglam SY, Kokcu F, Ceker ME. Importance of apparent diffusion coefficient in differentiation of cerebral tumors. Niger J Clin Pract. 2025;28(5):608–14. [DOI] [PubMed] [Google Scholar]
- 26.Mahammedi A, Bachir S, Escott EJ, Barnett GH, Mohammadi AM, Larvie M. Prediction of recurrent glioblastoma after laser interstitial thermal therapy: the role of diffusion imaging. Neuro-oncol Adv. 2019;1(1):vdz021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Sanvito F, Kryukov I, Yao J, Teraishi A, Raymond C, Gao J, et al. Advanced imaging characterization of post-chemoradiation glioblastoma stratified by diffusion MRI phenotypes known to predict favorable anti-VEGF response. J Neurooncol. 2025;173(3):583–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Wu J, Su R, Qiu D, Cheng X, Li L, Huang C, et al. Analysis of DWI in the classification of glioma pathology and its therapeutic application in clinical surgery: a case-control study. Transl Cancer Res. 2022;11(4):805–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Moore-Palhares D, Lawrence LSP, Myrehaug S, Stewart J, Detsky J, Tseng CL, Chen H, Dinakaran D, Maralani P, Ruschin M, et al. Temporal apparent diffusion coefficient changes during chemoradiation: an imaging biomarker for tumor response monitoring and spatial recurrence prediction in glioblastoma. Int J Radiat Oncol Biol Phys. 2025;122(3):592–604. [DOI] [PubMed] [Google Scholar]
- 30.Seyhan AA. Circulating liquid biopsy biomarkers in glioblastoma: advances and challenges. Int J Mol Sci. 2024. 10.3390/ijms25147974. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ghaderian M, Shahbazi-Gahrouei D, Nikzad S, Didehban E, Hafezi H, Laher I, et al. Bridging neuro-biomarkers and MR imaging: the synergistic role of glial fibrillary acidic protein in early CNS disease diagnosis. IBRO Neurosci Rep. 2025;18:739–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Norden AD, Drappatz J, Wen PY. Novel anti-angiogenic therapies for malignant gliomas. Lancet Neurol. 2008;7(12):1152–60. [DOI] [PubMed] [Google Scholar]
- 33.Ellingson BM, Hagiwara A, Morris CJ, Cho NS, Oshima S, Sanvito F, Oughourlian TC, Telesca D, Raymond C, Abrey LE, et al. Depth of radiographic response and time to tumor regrowth predicts overall survival following anti-VEGF therapy in recurrent glioblastoma. Clin Cancer Res. 2023;29(20):4186–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Ellingson BM, Gerstner ER, Smits M, Huang RY, Colen R, Abrey LE, Aftab DT, Schwab GM, Hessel C, Harris RJ, et al. Diffusion MRI phenotypes predict overall survival benefit from anti-VEGF monotherapy in recurrent glioblastoma: converging evidence from phase II trials. Clin Cancer Res. 2017;23(19):5745–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wang K, Stark FS, Schlothauer T, Lahr A, Cosson V, Zhi J, Habben K, Tessier J, Schick E, Staack RF, et al. An apparent clinical pharmacokinetic drug-drug interaction between bevacizumab and the anti-placental growth factor monoclonal antibody RO5323441 via a target-trapping mechanism. Cancer Chemother Pharmacol. 2017;79(4):661–71. [DOI] [PubMed] [Google Scholar]
- 36.Park Y, Kc N, Paneque A, Cole PD. Tau, glial fibrillary acidic protein, and neurofilament light chain as brain protein biomarkers in cerebrospinal fluid and blood for diagnosis of neurobiological diseases. Int J Mol Sci. 2024. 10.3390/ijms25126295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Yunusova N, Tulendinov E, Svarovsky D, Ryabova A, Kondakova I, Ponomaryova A, et al. Levels of proangiogenic molecules and terminal complement complex C5b-9 in the crown of circulating sEVs in patients with recurrent glioblastomas: relationship with tumor molecular characteristics. Curr Issues Mol Biol. 2025. 10.3390/cimb47020132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Jung CS, Foerch C, Schanzer A, Heck A, Plate KH, Seifert V, Steinmetz H, Raabe A, Sitzer M. Serum GFAP is a diagnostic marker for glioblastoma multiforme. Brain. 2007;130(Pt 12):3336–41. [DOI] [PubMed] [Google Scholar]
- 39.Kiviniemi A, Gardberg M, Frantzen J, Parkkola R, Vuorinen V, Pesola M, et al. Serum levels of GFAP and EGFR in primary and recurrent high-grade gliomas: correlation to tumor volume, molecular markers, and progression-free survival. J Neurooncol. 2015;124(2):237–45. [DOI] [PubMed] [Google Scholar]
- 40.Tichy J, Spechtmeyer S, Mittelbronn M, Hattingen E, Rieger J, Senft C, et al. Prospective evaluation of serum glial fibrillary acidic protein (GFAP) as a diagnostic marker for glioblastoma. J Neurooncol. 2016;126(2):361–9. [DOI] [PubMed] [Google Scholar]
- 41.Zhang T, Zhu H, Sun H, Chen Y, Sun X, Wu Y, Wang B, Zhu Y, Zhang A, Wang K, et al. Radiomics in glioblastoma recurrence: advances in prediction, localization, and differentiation from treatment-related effects. J Transl Med. 2026;24(1):516. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Zhang J, Wang G, Yan B, Yang G, Yang Q, Hu Y, et al. Integrative analysis of transcriptome and proteome profiles in primary and recurrent glioblastoma. Proteomics Clin Appl. 2024;18(3):e2200085. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data is provided within the manuscript.