Abstract
Background
Fibroblast growth factor receptor (FGFR) alterations are potential oncogenic drivers that occur in a subset of patients with gliomas, but the natural history of these tumors is not clearly defined. An understanding of outcomes of FGFR-driven glioma has implications for targeted drug development for FGFR inhibitors, which are approved for other cancers.
Methods
We performed a retrospective cohort study of patients with gliomas who harbored FGFR alterations, including fusions, single nucleotide variant (SNV), and copy number variant (CNV) alterations seen at Massachusetts General Hospital between 2003 and 2023. The electronic medical record was searched to identify additional molecular data, treatment, and MRI scans. Kaplan–Meier analysis was used to assess progression-free and overall survival (OS).
Results
Thirty-one patients with glioblastoma (GBM), diffuse astrocytoma, and glioneuronal tumors with FGFR alterations were identified: 17 with FGFR fusions, 10 with SNV, and 4 with CNV. FGFR3-TACC3 was the most common fusion in patients with GBM or diffuse astrocytoma. Median OS in patients with GBM was 2.75 years, despite 55% of tumors having an unmethylated MGMT promoter. There were no clearly co-occurring mutations with an FGFR alteration. Of 7 patients who underwent subsequent surgery, 6 lost the original FGFR alteration. No patient received an FGFR inhibitor.
Conclusions
While FGFR alterations are rare in glioma, patients with FGFR-altered GBM may have prolonged survival, which has implications for clinical trial design. We found loss of FGFR alteration at the time of subsequent surgery, raising concern for the therapeutic potential of FGFR-targeting agents in recurrent gliomas.
Keywords: FGFR alterations, glioma, natural history, targeted therapy
Key Points.
mOS in GBM was 2.75 years, despite 55% being MGMT unmethylated.
An accurate survival benchmark is important in determining the efficacy of fibroblast growth factor receptor (FGFR) inhibitors in glioma.
Loss of FGFR alterations may be a resistance mechanism to FGFR inhibitors in recurrent glioma.
Importance of the Study.
Understanding the natural history of fibroblast growth factor receptor (FGFR) altered gliomas is vital to inform the development of FGFR inhibitors by providing an appropriate benchmark for comparison. Unfortunately, prior studies with FGFR-targeting drugs in gliomas have shown limited success. Our study suggests 1 possible explanation for this is that some patients have lost the FGFR alteration at recurrence, causing the inhibitor to be ineffective. Repeat tissue sampling may be needed to confirm the persistence of an FGFR alteration in future targeted studies.
Molecular profiling of brain tumors in recent years has identified important subgroups of gliomas that have distinct behavior and may benefit from targeted therapies.1 Fibroblast growth factor receptor (FGFR) genomic alterations occur in approximately 8% of gliomas and have been associated with different tumor grades.2–4 A subset of these (approximately 3%) harbors fusion rearrangements involving FGFR3 and TACC3 or FGFR1 and TACC1.5,6 FGFR-TACC fusions appear to be oncogenic drivers, but the exact mechanisms are not clear.7 Based on the oncogenic potential of the FGFR pathway, clinical trials are underway using the FGFR-targeting agents in gliomas and other cancer types, and 3 FGFR inhibitors (pemigatinib, erdafitinib, and futibatinib) have been approved by the FDA for cancers other than brain tumors.
There have been two clinical trials of FGFR inhibitors (infigratinib, fexagratinib) in recurrent high-grade gliomas with FGFR alterations showed modest benefit with 6-month progression-free survival (PFS) of only 16%–25% but a subset of patients had more durable disease control.8,9 Gaining a better understanding of the natural history and landscape of FGFR alterations in gliomas could provide insight into how best to target the FGFR pathway and influence drug development. Therefore, we investigated demographics, clinical course, radiologic characteristics, and genetic mutations in patients with gliomas harboring an FGFR alteration.
Methods
We queried the Massachusetts General Hospital (MGH) Pathology Department’s molecular database for tumors with FGFR alterations, including FGFR fusions, single nucleotide variant (SNV), and copy number variant (CNV) alterations. The molecular database includes all patients at MGH who underwent next-generation sequencing of their tumor as part of their clinical evaluation.10 Thirty-one with primary brain tumors with FGFR alterations at initial tissue diagnosis were identified. Specific tumor type, demographic data, treatment, and additional molecular information were collected from the electronic medical record. PFS was defined as the time from the date of diagnosis to progression determined by the treating investigator or death by any cause. Overall survival (OS) as defined as the time from the date of diagnosis to death. PFS1 was time to first progression, and PFS2 was time from first progression to second progression. Survival curves were calculated using the method of Kaplan–Meier. Initial diagnostic brain MRI scans were reviewed using the VASARI (Visually AcceSAble Rembrandt Images) criteria to standardize the image analysis.11 This study was approved by the Dana Farber/Harvard Cancer Center IRB, and a waiver of consent was granted given the retrospective nature of the research.
Results
Patient characteristics
Between 2003 and 2023, we identified 31 primary brain tumor patients with FGFR alterations: 17 with FGFR fusions, 10 with SNV, and 4 with CNV encompassing a variety of histologies, including glioblastoma, diffuse astrocytoma, and glioneuronal tumors. No patients were treated with an FGFR inhibitor. Table 1 lists the 17 patients with FGFR fusions—all the patients with GBM had an FGFR3-TACC3 fusion, except 1 with an FGFR3-PLEC fusion. Patients with glioneuronal tumors had more variability in the fusions detected. In the 11 patients with GBM, there were a variety of additional molecular variants, but none that consistently co-occurred with an FGFR fusion. Table 2 lists the 10 patients with SNV alterations in FGFR. There was no dominant alteration identified, with 5/10 having an FGFR2 mutation, 3/10 having an FGFR3 mutation, and 4/10 having an FGFR1 mutation. The specific SNV variants detected are shown in Supplementary Table 1. Table 3 lists the 4 patients with FGFR CNV alterations—all of which were loss of function, and thus not likely oncogenic drivers.
Table 1.
Characteristics of Patients with FGFR Fusions
| Characteristic | GBM (N = 11) | Diffuse Astrocytoma (N = 2) | Glioneuronal/ Glioma (N = 4) |
|---|---|---|---|
| Age range (yr) | 25–76 | 45–46 | 6–61 |
| Gender | Female (6) Male (5) |
Female (1) Male (1) |
Female (3) Male (1) |
| KPS range | 60–100 | 70–80 | 90 (2), Not available (2) |
| Histological grade | Grade 4 (11) | Low grade (1)* Not elsewhere classified (1)# |
Grade 1 (1) Grade 2 (1) Grade 3 (1) Not classified (1)* |
| MGMT | Unmethylated (6) Methylated (4) Not tested (1) |
Unmethylated (1) Insufficient tissue (1) |
Methylated (2) Not tested (2) |
| FGFR fusion | FGFR3-TACC3 (10) FGFR3-PLEC (1) |
FGFR3-TACC3 (2) | FGFR1-TACC1 (1) FGFR3-TACC3 (1) Intragenic FGFR1 exon 9 and exon 18 fusion (1) FGFR1 TDK (exon 19 and exon 11 fusion) (1) |
| Additional FGFR alteration | FGFR2 loss (1)+ | None | None |
| IDH status | Wildtype (11) | Wildtype (1) Not tested (1) |
Wildtype (4) |
| Other alterations present | TERT promotor (8), PTEN (1), CDKN2A loss (5), ATRX (2), ARID1A (1), RNF43 (1), PTPN11 (1), CDH1 (1), P1K3R1 (1), ATM (1), FBXW7 (1), CCND1 (1) | TERT promoter (1)# PTEN (1) Insufficient tissue (1) |
TERT promoter (1) Not tested (2) No significant variants (1) |
| Initial treatment | Radiation and temozolomide (11) | Radiation and temozolomide (1) Temozolomide (1) |
Radiation and temozolomide (1) Vincristine and carboplatin (1) No treatment (2) |
*No histological grade provided; # Diffuse glioma, IDH-wt, not elsewhere classified; + patient also included in Table 3. Where possible, tumors were classified based on WHO 2021 criteria.
Table 2.
Characteristics of Patients with FGFR SNV
| Characteristic | GBM (N = 4) | Diffuse Glioma (N = 1) | Glioneuronal (N = 5) |
|---|---|---|---|
| Age range (yr) | 46–74 | 46 | 1–44 |
| Gender | Female (4) | Female | Female (2), Male (3) |
| KPS range | 40-100 | 80-100 | 90-100, Not available (2) |
| Histological grade | Grade 4 (4) | Astrocytoma grade 3 | Low grade* (4) High-grade* (1) |
| MGMT | Unmethylated (1) Methylated (3) | Methylated | Unmethylated (1) Not tested (4) |
| FGFR fusion | None | None | None |
| SNV FGFR alteration | FGFR2 (1) + FGFR3 (4) + |
FGFR2 | FGFR1 (4)^ FGFR2 (2)^ |
| IDH status | Wildtype (4) | Mutant | Wildtype (4) Not tested (1) |
| Other alterations present | TERT promoter (4), PTEN (2), PIK3CA (1), PTCH1 (1), CDKN2A gain (1), CDKN2A loss (1), JAK2 (1), SMO (1), TP53 (1), RB1 (1), NF1 (2), APC (1), MDM4 (1) | CDKN2A, BRAC2, TP53, TP63, MSH6, ATRIX, ARID1A, SMARCA4, SDHB, FLT3 # | EGFR (1), PTPN11 (2), HNF1A (1), PIK3CA (2), PTCH1 (1), PTEN (1), BRAC2 (1), ERBB2 (1), ERBB3 (1), TP53 (1), TP63 (1), SMAD4 (1), TSC2 (1), PIK3R1 (1), ATM (1), CIC (1), PDGFRA (1), MEN1 (1), MSH6 (1), MAP2K1 (1), NRAS (1), RB1 (1), CDH1 (1), MYC (1), KDR (1), FBXW7 (1), DDX3X (1), MYCN (1), DAXX (1), RNF43 (1), RET (1), SUFU (1), ATRIX (1), AURKA (1), ALK (1), POLE (1), KEAP1 (1), ARID1A (1), TSC1 (1), SMARCA4 (1), NF1 (2), NF2 (1), APC (1), MDM4 (1)# |
| Initial treatment | Radiation and temozolomide (3) Radiation (1) |
Radiation and temozolomide (1) | Radiation (2) Vincristine (1) No treatment (2) |
*No histological grade provided; + One patient had both FGFR2 and FGFR3 alterations; ^ One patient had both FGFR1 and FGFR2 alterations; # Two patients had unusual profiles with multiple alterations.
Table 3.
Characteristics of Patients with FGFR CNV
| Characteristic | GBM (N = 3) | Diffuse Astrocytoma (N = 1) |
|---|---|---|
| Age range (yr) | 53–62 | 37 |
| Gender | Female (1), Male (2) | Male |
| KPS range | 70-90 | 100 |
| Histological grade | Grade 4 (3) | Grade 3 |
| MGMT | Unmethylated (1) Methylated (2) |
Methylated |
| FGFR fusion | FGFR3-TACC3 (1)* | None |
| CNV FGFR alteration | FGF19 loss (1)^ FGFR2 loss (3) |
FGFR3 loss |
| IDH status | Wildtype (3) | Mutant |
| Other alterations present | ATM loss (1), CCND1 loss (1), PTEN (1), CDKN2A loss (3), EGFR (2), PIK3CA (1) | TERT promoter, TP53, ATRX, GNA11 loss |
| Initial treatment | Radiation and temozolomide (3) | Radiation and temozolomide |
*One patient was also included in Table 1 because also had FGFR3-TACC3 fusion; ^ One patient had both FGF19 loss and FGFR2 loss.
Seven of the 31 patients had second surgeries at which time the FGFR fusion (2) and/or SNV (1) or CNV (1) alterations were not detected. One of these 4 had a third surgery, where the FGFR alteration was still not found. Only 1/7 retained the same FGFR fusion at the second surgery. One of the 7 was not tested for mutations at the second surgery but was retested at a third surgery, and the original FGFR alterations were not found. No patient developed a new FGFR alteration.
Vasari characteristics
We reviewed the initial brain MRIs for all 31 patients (Supplementary Tables 2, 3, and 4). Across all glioma subtypes, there were no imaging features commonly shared across FGFR alterations (Supplementary Figures 1 and 2). Almost all tumors were unifocal, with the exception of one tumor having a gliomatosis cerebri pattern. There was no clear association of an FGFR alteration with the presence or absence of necrosis or edema, although 5/18 FGFR-associated GBMs had no necrosis and 3/18 FGFR-associated GBM had no or minimal edema (Supplementary Tables 2, 3, and 4).
Survival
The 11 FGFR fusion patients with GBM had a median OS of 2.75 years, despite 55% of the tumors having an unmethylated MGMT promoter. Median PFS1 was of 9.1 months, and median PFS2 of 5.3 months (Fig. 1). For the 2 patients with diffuse astrocytomas and FGFR fusions, the median OS was not reached and median PFS1 was 4.1 years. For the 4 patients with glioneuronal tumors and FGFR fusions, 3 have not progressed or died and 1 died after 28 months from tumor progression. These fusions are now known to be characteristic of more benign histologies, such as dysembryoplastic neuroepithelial tumor (DNET) and central neurocytoma. Two of the four patients diagnosed with FGFR SNV altered GBMs had a median OS of 1.6 years, with a median PFS1 of 1.0 years. Two of the five patients diagnosed with FGFR SNV altered glioneuronal tumors had a median OS that was undefined and a median PFS1 of 16.3 years. Of the 3 patients diagnosed with FGFR CNV altered GBMs, the median OS was 2.1 years, 3 had a median PFS1 of 9.75 months, 2/3 had a median PFS2 of 7.5 months. The one patient diagnosed with FGFR CNV altered diffuse astrocytoma progressed after 25.8 months and is still alive.
Figure 1.
Overall and progression-free survival in the cohort of patients with GBM and FGFR fusions.
Discussion
We identified 31 patients with primary brain tumors containing FGFR alterations at initial diagnosis, including fusions, SNV, and CNV alterations. The FGFR3-TACC3 fusion was the most common alteration in gliomas (particularly glioblastoma), but there was no dominant SNV alteration identified. Within our cohort, there were no clear mutations or imaging characteristics that co-occurred with specific FGFR alterations, as most of the co-mutations identified have commonly been reported in gliomas, making a clear association with an FGFR alteration challenging. Similar to prior reports, IDH mutation was exclusive of an FGFR fusion.6,9 Clinical trials enrolling patients with newly diagnosed GBM and both MGMT methylated and unmethylated tumors have estimated mOS to be 16–17 months, so the mOS of 2.75 years in our cohort of patients with GBM and an FGFR fusion suggests a more favorable prognosis despite 55% of patients in our cohort having an unmethylated MGMT promoter status.12–15 Prior profiling studies have also suggested improved survival with FGFR-TACC3 fusions, and awareness of this potential prolonged survival will be critical when designing clinical trials for FGFR-altered gliomas.4,16,17
Prior studies exploring multi-target tyrosine kinase inhibitors targeting FGFR in unselected populations of recurrent high-grade glioma have not shown efficacy.18,19 However, as our knowledge about the molecular landscape of gliomas grows, there is increased interest in developing targeted therapies for GBMs harboring specific molecular alterations. Two recent clinical trials with the FGFR inhibitor, infigratinib and fexagratinib, in selected patients with recurrent high-grade gliomas and FGFR alterations found mPFS to be 1.4–1.7 months and mOS to be 6.7–17.5 months, suggesting limited efficacy. However, a subset of patients at first progression or an activating FGFR1 or FGFR3 alteration or FGFR3-TACC3 fusion showed durable disease control for more than 1-year.8,9 The RAGNAR trial was a single arm, Phase 2 basket trial evaluating erdafitinib, a pan-FGFR tyrosine kinase inhibitor, in patients with solid tumors harboring FGFR1–4 alterations.20 Objective radiographic responses were seen in 2/7 (29%) patients with low-grade glioma and 3/30 (10%) patients with high-grade gliomas. In all of the trial patients, median PFS was 4.2 months and a median OS was 10.7 months so similar to other rGBM trials. As hinted at in the infigratinib trial, the type of FGFR alteration is important and likely needs to be a true activating mutation and not just a passenger event. Whether or not targeted FGFR inhibitors would be beneficial in first-line therapy, potentially before additional driver mutations are acquired, needs to be determined.
One potential explanation for the limited efficacy is loss of the original FGFR alteration. Of the 8 patients, 7 in our cohort lost the FGFR alteration at a subsequent surgery so, as opposed to IDH mutations, FGFR alterations may not be persistent driver mutations or may occur in a heterogeneous subset of tumor cells, raising concerns about the therapeutic potential of targeting FGFR alterations. This finding is in conflict with prior studies that have suggested persistence of the FGFR3-TACC3 fusions at time of recurrence, although these were small studies including 1–5 patients.21,22 Most prior trials, including the infigratinib trial, did not require confirmation of persistence of FGFR alterations prior to trial enrollment, as it is not always feasible to obtain new tumor tissue in glioma patients at time of recurrence. However, the loss of FGFR alterations in our cohort suggests that tissue should be obtained to document persistence of these alterations in recurrent glioma trials testing FGFR-targeting drugs.
Limitations to our study include the small sample size and the variety of histologies, which limits conclusions within specific subtypes of glioma. As clinical NGS profiling of brain tumors at our institution is driven by clinical indication and tissue availability, assessment for FGFR alterations is not universal for all patients with gliomas. In 3 patients with gliomas, FGFR3-TACC fusions co-occurred with FGFR3 gene amplification. While it is technically challenging to determine whether the fusion underwent subsequent amplification or if the FGFR3 gain involved the wildtype allele, the outcome of these alterations is a strong activation of the FGFR pathway.
In conclusion, FGFR alterations in patients with gliomas are rare; however, patients with FGFR-altered GBM may have prolonged survival, which has implications for clinical trial design. We found loss of the original FGFR alteration at the time of tumor progression in most patients, raising concerns for the therapeutic potential of FGFR-targeting agents in recurrent gliomas.
Supplementary Material
Contributor Information
Maya Viera, Massachusetts General Hospital Cancer Center, Boston, MA 02114, USA.
Christopher W Mount, Massachusetts General Hospital Department of Pathology. Boston, MA 02114, USA.
Dora Dias-Santagata, Massachusetts General Hospital Department of Pathology. Boston, MA 02114, USA.
Julie J Miller, Harvard Medical School, Boston, MA 02115; Massachusetts General Hospital Cancer Center, Boston, MA 02114, USA.
Elizabeth R Gerstner, Harvard Medical School, Boston, MA 02115; Massachusetts General Hospital Cancer Center, Boston, MA 02114, USA.
Acknowledgments
This study used funding from Miles for Mary. Dr. Julie J. Miller received consulting fees from Servier.
Conflict of interest statement
There are no conflict of interests for the authors MV, CM, DD, JJM, ERG.
Authorship statement
Designed the study and contributed to the data analysis and interpretation (ERM, MV, JJM). Contributed to the interpretation of pathology and molecular genetics (CM, DD).
Data availability
All the data are available upon reasonable request to the corresponding author and in compliance with the applicable regulations.
References
- 1. Lim-Fat MJ, Song KW, Iorgulescu JB, et al. Clinical, radiological and genomic features and targeted therapy in BRAF V600E mutant adult glioblastoma. J Neurooncol. 2021;152(3):515–522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Métais A, Dangouloff-Ros V, Garcia J, et al. ; contributors of the Biopathology RENOCLIP‐LOC network. Phenotypic and epigenetic heterogeneity in FGFR2-fused glial and glioneuronal tumours. Neuropathol Appl Neurobiol. 2024;50(6):e13014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Dono A, El Achi H, Bundrant BE, et al. Infiltrating gliomas with FGFR alterations: Histologic features, genetic alterations, and potential clinical implications. Cancer Biomark. 2023;36(2):117–131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Bae H, Lee B, Hwang S, et al. Clinicopathological and molecular characteristics of IDH-wildtype glioblastoma with FGFR3::TACC3 fusion. Biomedicines. 2024;12(1):150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Helsten T, Elkin S, Arthur E, et al. The FGFR landscape in cancer: analysis of 4,853 tumors by next-generation sequencing. Clin Cancer Res. 2016;22(1):259–267. [DOI] [PubMed] [Google Scholar]
- 6. Singh D, Chan JM, Zoppoli P, et al. Transforming fusions of FGFR and TACC genes in human glioblastoma. Science (1979). 2012;337(6099):1231–1235. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Brennan C. FGFR-TACC approaches the first turn in the race for targetable GBM mutations. Neuro Oncol. 2017;19(4):461–462. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Picca A, Di Stefano AL, Savatovsky J, et al. TARGET: A phase I/II open-label multicenter study to assess safety and efficacy of fexagratinib in patients with relapsed/refractory FGFR fusion-positive glioma. Neurooncol. Adv. 2024;6(1):vdae068. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Lassman AB, Sepúlveda-Sánchez JM, Cloughesy TF, et al. Infigratinib in patients with recurrent gliomas and FGFR alterations: a multicenter phase II study. Clin Cancer Res. 2022. (11):2270-2277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Zheng Z, Liebers M, Zhelyazkova B, et al. Anchored multiplex PCR for targeted next-generation sequencing. Nat Med. 2014;20(12):1479– 84. [DOI] [PubMed] [Google Scholar]
- 11. Gevaert O, Mitchell LA, Achrol AS, et al. Glioblastoma multiforme: exploratory radiogenomic analysis by using quantitative image features. Radiology. 2014;273(1):168–174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Raizer JJ, Giglio P, Hu J, et al. ; Brain Tumor Trials Collaborative. A phase II study of bevacizumab and erlotinib after radiation and temozolomide in MGMT unmethylated GBM patients. J Neurooncol. 2016;126(1):185–192. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Gilbert MR, Dignam JJ, Armstrong TS, et al. A randomized trial of bevacizumab for newly diagnosed glioblastoma. N Engl J Med. 2014;370(8):699–708. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Chinot OL, Wick W, Mason W, et al. Bevacizumab plus radiotherapy–temozolomide for newly diagnosed glioblastoma. N Engl J Med. 2014;370(8):709–722. [DOI] [PubMed] [Google Scholar]
- 15. Gilbert MR, Wang M, Aldape KD, et al. Dose-dense temozolomide for newly diagnosed glioblastoma: a randomized phase III clinical trial. J Clin Oncol. 2013;31(32):4085–4091. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Di Stefano AL, Picca A, Saragoussi E, et al. ; TARGET study group. Clinical, molecular, and radiomic profile of gliomas with FGFR3-TACC3 fusions. Neuro Oncol. 2020;22(11):1614–1624. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Wu Z, Lopes Abath Neto O, Bale TA, et al. DNA methylation analysis of glioblastomas harboring FGFR3-TACC3 fusions identifies a methylation subclass with better patient survival. Acta Neuropathol. 2022;144(1):155–157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Sharma M, Schilero C, Peereboom DM, et al. Phase II study of Dovitinib in recurrent glioblastoma. J Neurooncol. 2019;144(2):359–368. [DOI] [PubMed] [Google Scholar]
- 19. Norden AD, Schiff D, Ahluwalia MS, et al. Phase II trial of triple tyrosine kinase receptor inhibitor nintedanib in recurrent high-grade gliomas. J Neurooncol. 2015;121(2):297– 302. [DOI] [PubMed] [Google Scholar]
- 20. Pant S, Schuler M, Iyer G, et al. Erdafitinib in patients with advanced solid tumours with FGFR alterations (RAGNAR): an international, single-arm, phase 2 study. Lancet Oncol. 2023;24(8):925–935. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Di Stefano AL, Fucci A, Frattini V, et al. Detection, characterization, and inhibition of FGFR-TACC fusions in IDH wild-type glioma. Clin Cancer Res. 2015;21(14):3307–3317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Wang J, Cazzato E, Ladewig E, et al. Clonal evolution of glioblastoma under therapy. Nat Genet. 2016;48(7):768–776. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All the data are available upon reasonable request to the corresponding author and in compliance with the applicable regulations.