Finding a four-leaf clover—identifying long-term survivors in IDH-wildtype glioblastoma

See the article by Burgenske et al. in this issue, pp. 1458–1469.

Glioblastoma (GBM) is the most commonly diagnosed primary brain tumor and is associated with the most dismal clinical outcomes of all brain tumors. Patients face a median survival of 9 months with standard of care surgery and adjuvant chemo-radiotherapy increasing survival to 15 months, a modest improvement.¹

Standard of care for GBM has remained stagnant for the past two decades, involving a combination of surgery, radiation, and a small selection of chemotherapeutic agents despite significant discoveries in the genomic, biological, and clinical understanding of the disease.

One major advancement was the discovery of recurrent heterozygous isocitrate dehydrogenase 1 and 2 (IDH1/2) mutations in infiltrating gliomas and the significance on prognosis and impact on realignment of glass-based diagnoses.² These mutations occur in a vast proportion of World Health Organization (WHO) grades II and III infiltrating gliomas regardless of morphology and are considered to be “neomorphic” in nature. It results in the subversion of cellular metabolic pathways in the aberrant overproduction of 2-hydroxyglutarate, a potent oncometabolite, which leads to epigenome-wide dysfunction.³ Despite its central role in gliomagenesis, IDH-mutant gliomas exhibit favorable prognoses that supersede traditional histologic grading.⁴ Malignant transformation of IDH-mutant grades II/III gliomas to secondary GBM is virtually inevitable and these exhibit better survival compared with IDH-wildtype de novo or primary GBM.^2,5 Survival for IDH-wildtype GBM remains dismal. The search for therapeutic targets within IDH-wildtype GBMs has remained challenging due to the extraordinary amount of intra- and intertumoral heterogeneity, reflected in their molecular and clinical profiles.⁶ Prior research has shed light on the genomic drivers of IDH-wildtype GBM such as amplification of EGFR/PDGFRa/PTEN loss, and TP53 mutation.⁷ While integrated genomic analyses have identified 4 clinical subtypes based on molecular signatures,⁸ single cell sequencing has revealed that even within one GBM, several populations driven by different genomic events can exist in cohort,⁹ suggesting that no one single event drives each GBM.

Although the survival rates of GBM remain unfavorable, a small percentage of patients have exhibited extraordinary response to treatment and disease-free survival, some surviving over 10 years.¹ Therefore, there is very strong motivation to collate and profile GBM patients who have survived beyond the usual course of the disease. Numerous studies, beyond the scope of this editorial, have catalogued and profiled the clinical features and genomic landscapes of these GBM long-term survivors (LTS). However, beyond the usual metrics of age, performance status, O⁶-methylguanine-DNA methyltransferase promoter methylation, and IDH1/2 mutation status which predicts superior clinical status in GBM, there is little known about the molecular features of GBM LTS. This is even more remarkable in an IDH-wildtype population, which is normally associated with universally poor outcome.

In a report titled “Molecular Profiling of Long-Term IDH-Wildtype Glioblastoma Survivors,” Burgenske et al¹⁰ aimed to identify the biological differences between patients who survive over 5 years, or LTS, versus patients who succumb to their disease within 2 years, or short-term survivors (STS). The authors assembled an impressive cohort of clinical and molecular data of 12 pretreated IDH-wildtype GBM LTS utilizing targeted next-generation sequencing, copy number profiling, and transcriptomic and global methylation microarray analysis.

Using a gene panel of 50 commonly mutated glioma genes and copy number microarray, the authors found no significant enrichment of any genomic events between STS and LTS patients, supporting the notion that GBM is heterogeneous both intra- and intertumorally. No discernible differences in methylation profiles were also detected. However, at the transcriptomic level, LTS patients showed a propensity for increased sphingomyelin and ceramide-related metabolic pathways, while STS patients exhibited increased nucleotide excision repair and cell cycling pathways.

Clinically, the authors found that LTS patients were more likely to be younger at diagnosis, female, and to not have been given the current gold standard Stupp regimen of concurrent radiotherapy with temozolomide followed by adjuvant temozolomide. Altogether, the lack of any defining molecular feature useful for predicting LTS remains a disappointment. However, discovery of unique transcriptomic signatures associated with the STS and LTS cohorts may provide promising leads for future research.

While the authors have curated a remarkable cohort and generated a large amount of molecular data, further and more in-depth interrogation of this cohort may lead to more clinically informative biomarkers. A more exhaustive analysis of the genomic landscape of LTS using whole exome or even whole genome sequencing would provide a more comprehensive overview of genomic aberrations which may affect survival. Profiling of germline single nucleotide variants may further define the constitutional fabric of this subset of patients. Lastly, proteomic and metabolomic analyses may shed light on pathways that augment the effectiveness of current therapies leading to LTS. Hopefully, granular analyses of GBM LTS patients will enable us to pinpoint specific prognostic and possibly therapeutic clinical markers.