MRI features and prognostic evaluation in patients with subventricular zone-contacting IDH-wild-type glioblastoma

Glioblastoma (GBM) is the most common malignant primary brain tumor in adults, accounting for approximately 49.1% of all malignant brain tumors.¹ It is characterized by high rates of disability, recurrence, and significant clinicopathological heterogeneity. The current standard-of-care for GBM involves surgical resection, followed by postoperative radiotherapy and adjuvant temozolomide chemotherapy.² Despite aggressive treatment, patient prognosis has not improved significantly, with a five-year survival rate of only 6.8%.³

In recent years, to improve the prognosis of patients with GBM, many researchers have paid attention to the cell origin of GBM. Although there is no consensus on the cellular origin of GBM, most scholars believe that neural stem cells (NSCs) in the subventricular zone (SVZ) may be the main cell of origin of GBM.^4,5 The SVZ is generally defined as the 3–5 mm lateral border of the lateral ventricles, which harbors the largest reservoir of NSCs in the adult brain. Due to its extensive contact with cerebrospinal fluid (CSF), the enrichment of various growth factors in the CSF may further enhance the malignant potential of tumor cells.^6,7 Previous studies have suggested that the CXCL12/CXCR4 signaling pathway, which is closely associated with the SVZ microenvironment, may attract glioma stem cells to the SVZ and promote their invasiveness.⁸ Furthermore, GBM with SVZ contact exhibits a distinct transcriptomic profile, with significant upregulation of Notch pathway-related genes, such as HES4 and DLL3, which are associated with poor survival.⁹ Additionally, basic studies have indicated that NSCs in the SVZ contribute to GBM development in distant brain regions and promote tumor proliferation, migration, and tumorigenicity, suggesting that SVZ is closely related to the origin and progression of glioblastoma.^10,11 In clinical research, GBM with SVZ contact at initial diagnosis has a poorer prognosis and tends to exhibit distinct MRI features.^12–15 In this study, we aim to investigate the association between MRI features and SVZ contact in IDH-wild-type GBM, as well as their prognostic significance to guide personalized diagnosis and treatment.

A total of 371 patients with IDH-wild-type glioblastoma were included in this study. Based on the analysis of the relationship between the tumor and the subventricular zone on preoperative T1C images, 239 patients (64.4%) were classified into the SVZ contact group, while 132 patients (35.6%) were classified into the SVZ non-contact group. Among the 239 patients in the SVZ contact group, 51 patients (21.3%) had frontal horn contact, 26 patients (10.9%) had body contact, 62 patients (26.0%) had temporal horn contact, and 100 patients (41.8%) had occipital horn contact. There were no significant differences between the two groups in terms of gender, age, peritumoral edema, enhancement degree, and multifocal lesions. However, significant differences were observed in OS, tumor size, tumor shape, enhancement type, cystic lesion, necrosis, and crossing the midline (Table 1).

Compared to the SVZ non-contact group, patients in the SVZ contact group had significantly shorter overall survival (11.0 vs. 17.5 months, p < 0.001), larger tumor size (5.07 vs. 3.31 cm, p < 0.001), a higher likelihood of crossing the midline (11.7% vs. 0%, p < 0.001), and were more prone to cystic lesion and necrosis. Univariate Cox regression analysis identified age, tumor size, enhancement type, multifocal lesions, crossing the midline, and SVZ contact as unfavorable prognostic factors for patients with IDH wild-type glioblastoma. Multivariate Cox regression analysis further revealed that SVZ contact (p = 0.027, HR = 1.364, 95% CI: 1.037–1.794), multifocal lesions (p = 0.010, HR = 1.548, 95% CI: 1.111–2.158), and age (p < 0.001, HR = 1.024, 95% CI: 1.014–1.034) remained statistically significant predictors of overall survival, indicating that these factors are independent prognostic risk factors for patients with IDH-wild-type glioblastoma (Table 2). Patients in the SVZ contact group had significantly worse overall survival compared to those in the SVZ non-contact group (p < 0.001, Figure 4A). Among patients in the SVZ contact group, the specific region of SVZ contact was associated with OS (p < 0.001, Figure 4B). Specifically, contact of the temporal horn was associated with longer OS, while contact of the body was associated with shorter OS.

FIGURE 4.

Kaplan-Meier curves for overall survival in glioblastoma: (A) Patients with and without SVZ contact; (B) Patients with tumors contacting specific SVZ regions.

SVZ =subventricular zone

In our study, among the 371 glioblastoma patients, 239 cases (approximately 64%) exhibited subventricular zone contact on MRI, consistent with previous observations indicating that 60%–70% of newly diagnosed GBM patients show SVZ contact.¹³ We analyzed the clinical and imaging characteristics of IDH-wild-type GBM patients with SVZ contact and further confirmed that SVZ contact serves as an independent negative prognostic factor for GBM patients. In addition, age, tumor size, multifocal lesions, and crossing the midline were also identified as prognostic risk factors. Compared to the SVZ non-contact group, patients in the SVZ contact group exhibited shorter overall survival, larger tumor size, higher proportions of crossing the midline and multifocal lesions, as well as an increased tendency for cystic lesion and necrosis. These findings suggest that SVZ contact not only influences the biological behavior of the tumor but is also closely associated with clinical prognosis.

In the analysis of survival and prognosis, we identified SVZ contact as an independent prognostic factor in patients with IDH-wild-type glioblastoma. This finding is consistent with a previous meta-analysis¹⁸, and supported by clinical studies. For instance, Hallaert et al.¹³ analyzed 214 glioblastoma patients and found that those with SVZ contact had shorter overall survival and progressionfree survival. Additionally, a study by Najbauer et al.¹⁹ further supported SVZ contact as a significant indicator of poor prognosis and investigated the relationship between molecular markers in the SVZ microenvironment and patient outcomes through genomic sequencing. Our study demonstrated a significant correlation between the specific SVZ contact region and OS, with patients exhibiting temporal horn contact having a relatively longer OS, whereas those with body contact having a shorter OS. However, previous studies have reported inconsistent findings. Some studies have suggested that occipital horn contact is associated with poorer prognosis^20,21, while others found no significant differences in OS among different SVZ contact regions.^15,22 These discrepancies may be attributed to heterogeneity in study design, subjective criteria for SVZ subregion classification, and variations in the biological characteristics of the SVZ microenvironment. Previous studies have indicated that different SVZ regions exhibit distinct cellular compositions and molecular characteristics^23,24, which may lead to tumor contacting specific regions exhibiting distinct biological behaviors and invasion patterns, ultimately affecting patient prognosis. Moreover, the current classification of SVZ regions is primarily based on anatomical subdivisions, which may not fully reflect regional molecular and microenvironmental differences. Although our study provides new insights into the prognostic significance of SVZ contact, the underlying molecular mechanisms remain unclear. Future research should focus on large-scale prospective studies, single-cell sequencing, spatial transcriptomics, and radiomics to further elucidate the molecular characteristics and microenvironmental dynamics of different SVZ contact regions, thereby uncovering their potential associations with tumor invasion and prognosis. SVZ contact may increase the surgical complexity and postoperative complication risk in glioblastoma patients. Mistry et al.²⁵ demonstrated that SVZ contact was associated with an increased risk of post-treatment hydrocephalus and leptomeningeal dissemination. Additionally, another study reported that glioblastoma with SVZ contact was more likely to disseminate along the cerebrospinal fluid pathways, and patients with intraoperative ventricular entry had a higher risk of postoperative leptomeningeal dissemination than those without ventricular entry.²⁶ Therefore, to reduce postoperative complications, neurosurgeons are advised to minimize ventricular entry during surgery. However, the impact of ventricular opening on prognosis remains controversial. Saito et al.²⁷ found that a greater extent of ventricular opening was associated with longer survival, possibly due to the removal of a larger proportion of tumor stem cells residing within the SVZ. Nevertheless, surgical trauma may potentially affect neurological function, thereby reducing patients’ quality of life and overall survival. In summary, GBM with SVZ contact is often associated with poor prognosis and poses technical challenges for neurosurgical resection. Future research should focus on optimizing the balance between maximal tumor resection and minimizing postoperative complications to improve patient outcomes.

In this study, we observed that GBM with SVZ contact exhibited larger tumor size, consistent with previous studies.^13,19,28,29 This characteristic may be attributed to the potential malignant transformation, strong proliferative and migratory capacities of NSCs within the SVZ microenvironment. Additionally, the SVZ is in extensive contact with cerebrospinal fluid, which can enrich various growth factors and promote the proliferation of cancer cells. We also observed that multifocal lesions were more frequently present in the SVZ contact group, concordant with findings from previous studies.^15,30,31 Lim et al.³⁰ were the first to report that GBM with SVZ contact was more likely to exhibit multifocal lesions and distant recurrence. Gupta et al.³² revealed that GBM with SVZ contact displayed distinct tangential, radial, or multipolar migratory patterns, which may explain the higher incidence of multifocal lesions in these tumors. Additionally, these tumors more frequently exhibit a butterflylike growth pattern across the midline, suggesting corpus callosum invasion. This invasion disrupts the barrier function of the corpus callosum’s dense white matter tracts against interstitial edema, tumor spread, and inflammatory infiltration.³³ These imaging characteristics indicate that GBM with SVZ exhibits greater invasiveness, proliferation, and biological heterogeneity. These tumors are often located in anatomically complex deep brain regions and frequently exhibit multifocal characteristics, increasing surgical difficulty of complete tumor resection, often leading to poorer outcomes. Therefore, SVZ contact not only indicates a more aggressive biological behavior of the tumor but also poses challenges for clinical treatment, warranting further investigation and attention.

We also observed a higher incidence of cystic lesion in GBM with SVZ contact. However, Zhao et al.²⁸ reported a lower incidence of cystic lesion in these tumors. It has been reported that the formation of intratumoral cystic degeneration in gliomas is primarily associated with the disruption of the blood-brain barrier, leading to the exudation and accumulation of plasma proteins.³⁴ Additionally, factors such as tumor liquefactive necrosis and hemorrhagic liquefaction may also contribute to cyst formation. Regarding the prognostic significance of cystic degeneration in gliomas, Park et al.³⁵ conducted a retrospective analysis of preoperative MRI images from 158 patients with low-grade gliomas and found that tumor cystic degeneration might be a favorable prognostic factor. However, another study suggested that cystic degeneration are not an independent prognostic factor for lowgrade gliomas.³⁶ For GBM, Zhou et al.³⁷ analyzed imaging characteristics from 180 patients and observed that cystic degeneration may be associated with longer survival. Pathological studies have indicated that cystic GBM exhibits less peritumoral edema, clearer boundaries, and fewer infiltrating tumor cells, which may lead to reduced residual tumor burden postoperatively and potentially better prognosis.³⁸ However, a large cohort study did not confirm a significant correlation between cystic degeneration and GBM prognosis.³⁹ In this study, although there was a difference in the incidence of cystic lesion between the SVZ contact and non-contact groups, no significant association was identified between cystic lesion and overall survival. Therefore, the prognostic impact of cystic degeneration requires further validation through larger sample sizes and multicenter studies.

This study has several limitations. First, as a single-center retrospective study, it is inevitably subject to selection bias. Second, due to limited data availability, this study did not include other potential prognostic factors associated with glioblastoma, such as KPS scores, MGMT status, and extent of resection. Third, the tumor margins and SVZ contact were assessed solely using T1C images, which may not accurately reflect the true tumor boundaries of glioblastoma. Therefore, incorporating multiparametric imaging analysis is critical for a more comprehensive assessment. Fourth, the interpretation of SVZ subregions and tumor contact remains somewhat subjective. Future studies could integrate microenvironmental characteristics and utilize automated segmentation methods to more objectively identify SVZ contact, thereby reducing interpretation bias and minimizing experimental error. Future studies should include larger multicenter cohorts and combine multiparametric imaging analyses with clinical and molecular characteristics to further validate our findings.

Our study demonstrates that SVZ contact is an independent prognostic factor for overall survival in IDH-wild-type glioblastoma patients. GBM with SVZ contact is typically characterized by larger tumor size, a higher proportion of crossing the midline and multifocal lesions, suggesting that these tumors may exhibit increased aggressiveness, proliferative and migratory capabilities. However, further studies are needed to verify whether differences in cystic degeneration and necrosis exist between SVZ contact and SVZ non-contact groups, and to elucidate their prognostic significance. Early and systematic evaluation of tumor imaging characteristics can provide valuable insights for clinical neurosurgeons, enabling better prediction of patient outcomes and the development of individualized treatment strategies. Moreover, strategies targeting NSCs in the SVZ region and eradicating tumor-initiating cells through radiotherapy could help suppress tumor progression at its source, potentially prolonging patient survival. In the future, the SVZ may emerge as a critical target for radiotherapy and other targeted therapies in glioblastoma treatment.