Shorting vorasidenib for IDH mutant low-grade glioma

KEYWORDS:

• Glioma
• IDH mutation
• vorasidenib
• survival
• randomized clinical trial

Management of low-grade glioma is challenging. Despite decades of translational research that has deciphered some of the molecular alterations underpinning the tumor’s pathogenesis, neuro-oncologists still have an incomplete understanding on the drivers of progression [1,2]. This makes their job much more difficult. In fact, managing patients with low-grade gliomas not only requires therapies that work but also treatments that are not too toxic, in the long run, to the central nervous system due to the prolonged nature of patient survival. Therefore, it is important to put into context the recent data on vorasidenib’s benefit for isocitrate dehydrogenase (IDH) mutated low-grade gliomas [3].

The development of vorasidenib is based on the premise that mutation of IDH is associated with the pathogenesis of IDH-mutant low-grade gliomas [4,5]. This mutation is said to occur early on, and additional somatic mutations accumulate progressively over time leading to histological anaplasia and malignant phenotype [2]. However, defining these additional mutations requires tissue analysis, and repeated neurosurgical biopsies in these patients are not always feasible due to risk of catastrophic neurological complications. Therefore, the molecular alterations associated with evolution of low-grade gliomas cannot be precisely defined. More importantly, IDH-mutated low-grade gliomas do not appear to be addicted to the mutation for survival, and this is evident from the extremely low response rate of 1.2% in the randomized phase III vorasidenib trial INDIGO [3].

Vorasidenib is not a targeted therapy. Drugs targeting an activating mutation interfere with an essential signal that enables the tumor to survive and grow, and tumor regression ensues when the signal is blocked [6]. For example, both receptor tyrosine kinase inhibitors, erlotinib and osimertinib, can inhibit signal overdrive from deletion of exon 19 and L858R point mutation in exon 21 of epidermal growth factor receptor (EGFR) in non-small cell lung carcinoma, and the osimertinib can also suppress the acquired resistance T790M mutant cells [7,8]. Tumor regression is expected at a median of 8.0 weeks after initiation of erlotinib and 6.1 weeks after initiation of osimertinib [9,10]. This is because these activating mutations bias the EGFR toward an activated state, allowing survival and growth signals to transduce intracellularly. When this signal is curtailed or blocked, the addicted tumor cells die and the tumor regresses in size. However, unlike the activating EGFR mutations, the IDH-1 enzyme with R135H mutation converts α-ketoglutarate to 2-hydroxyglutarate, an oncometabolite that induces a hypermethylated phenotype by blocking Tet methylcytosine dioxygenase 2 and Jumonji histone demethylases [11]. After the genome is altered epigenetically, broad changes in gene expression occur and the effectors may remain within the viable cells causing persistent changes that are difficult to reverse. Therefore, the low-grade glioma may not be truly addicted to the IDH mutation or dependent on its downstream signal alterations for survival and growth, and therefore vorasidenib should be viewed more like an epigenetic disease modifier that stabilizes or delays tumor progression, malignant transformation, or both.

The phase III vorasidenib trial INDIGO is a double-blinded randomized study investigating the effect on patients with IDH-1 or IDH-2 mutated low-grade gliomas [3]. Typically, these patients are categorized as high risk for progression if they are older than 40 years of age or cannot undergo complete resection of the tumor. Treatment may be delayed at the time of diagnosis due to the slow pace of tumor progression, and patients then undergo serial monitoring by magnetic resonance imaging. The investigators took advantage of this observation period and introduced vorasidenib to control the tumor. Under blinded imaging-based independent assessment, they found the drug significantly improved progression-free survival (PFS) and delayed time to next intervention (TTNI) compared to the control population. However, these two benchmarks of may not be independent of each other, and TTNI is probably not a good metric for primary efficacy assessment. This is because a prolonged PFS would invariably lead to a delay in TTNI. Furthermore, patients in the control group can cross over to receive vorasidenib, and their psychological threshold for crossing over to another pill is probably lower compared to those in the experimental cohort considering higher risk interventions such as brain surgery, radiation, and/or cytotoxic chemotherapy. The impetus for next intervention may be inherently greater among patients taking the placebo. Therefore, the trial only demonstrated vorasidenib’s efficacy assessment in the PFS domain, and this small molecule did not produce the expected radiological response and its benefit in prolonging patient overall survival (OS) is still uncertain due to the relatively short median follow-up duration of 14.2 months [3].

Extent of resection is a major prognostic factor in low-grade gliomas. Despite a lack of randomized controlled clinical trial, multiple studies from single institutions [12,13], meta-analysis [14], post hoc analysis from the European Organization for Research and Treatment of Cancer (EORTC) prospective trials 22844 and 22845 [15], and comparative analysis of different practice patterns from two Norwegian hospitals [16] all indicated that extent of resection matters. Still, the investigators in INDIGO chose not to report the impact of this important prognostic factor even though data on biopsy, subtotal and gross total resection were captured as an inclusion criterion. Imbalance in the number of subjects who underwent gross total resection in the two cohorts could potentially bias the outcome favoring vorasidenib treatment.

Long duration of follow-up is needed for clinical trial of low-grade gliomas. This period can easily surpass a decade. For example, in RTOG 9802, a randomized trial of radiotherapy plus procarbazine, lomustine and vincristine for supratentorial low-grade gliomas, data analysis after only 7.5 years of follow up revealed only statistically significant prolongation of PFS but not OS [17]. However, a repeated analysis after more than a decade of follow-up eventually showed an OS benefit [18]. Because treatment is not expected to effect a response or shrinkage of these gliomas, PFS and OS are therefore the only independent measures of efficacy. Detecting a statistically significant difference in OS will take a much longer duration of follow-up compared to PFS. Furthermore, post-progression crossover to the vorasidenib treatment arm may attenuate or even negate any detectable difference in OS. A good example of this is the autologous tumor-lysate-loaded dendritic cell vaccination or DCVax trial, in which subjects were crossed over to receive the vaccine, and no significant OS was subsequently observed [19]. Therefore, a sound and rigorous trial in low-grade gliomas will probably require a prolonged follow-up period of at least a decade or longer so that enough events can accumulate in the comparative cohorts for PFS and OS to be statistically appraised.

Despite the inadequate assessments, vorasidenib still represents a significant advance for the treatment of IDH-mutated low-grade glioma patients. The optimal use of this drug, either alone or in combination with other modalities of treatment, remains to be determined, and its impact on neurocognitive function still needs further investigation.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or material discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or mending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.