IDH mutation testing in gliomas—where do we draw the line?

See the article by DeWitt et al on pages 1640–1650.

The classification of brain tumors has classically been dependent on histology. However, with the additional understanding of the genetic basis of tumorigenesis, molecular parameters have also now become integrated with histology in the 2016 World Health Organization (WHO) classification of brain tumors.¹ This concept has been integrated in the classification of the diffuse gliomas. For lower-grade gliomas (LGGs; WHO grades II and III), it is known that mutations in isocitrate dehydrogenase genes (IDH1/2) occur in high proportion, and the detection of these mutations has diagnostic, prognostic, and therapeutic implications.^2–4 Mutations in IDH1/2 lead to reduction of alpha-ketoglutarate and to the production of oncometabolite 2-hydroxyglutarate. This gain of function phenotype can be incurred via multiple different mutations. Mutations in codon 132 resulting in substation of arginine to histidine (R132H) is the most common of these mutations (termed “canonical”), accounting for up to 90% of all IDH mutations.² The development of R132H-specific anti-IDH1 antibodies allows for screening of canonical mutations by immunohistochemistry (IHC). In cases where IDH screening by IHC indicates the absence of R132H mutation, genetic sequencing of IDH may be performed to screen for noncanonical mutations. Although genetic sequencing allows for screening of both canonical and noncanonical mutations, the cost associated with testing as well as time constraints can be issues. Nevertheless, in the setting of a diffuse adult LGG that is negative by R132H-IHC, testing for a noncanonical IDH mutation is expected to conform within an integrated histomolecular diagnosis. In the absence of a complete evaluation for noncanonical IDH mutations in these cases, the “NOS” designation may be warranted for diffuse gliomas.

IDH mutations are less common in glioblastoma (GBM) but are nonetheless relevant in this tumor. Given the higher prevalence of GBM relative to LGGs, much of the IDH mutation screening in gliomas occurs in the setting of GBM. Adult LGGs tend to occur in the younger spectrum of age (below 50–55 y), while GBM tends to occur in older individuals. Added to this is the fact that after controlling for tumor grade in adult diffuse glioma (GBM vs LGG), the rate of IDH mutation varies inversely based on age. GBMs that are IDH-mutant are known to be more common in younger adults, and among LGGs an age distribution exists in the prevalence of IDH mutations, where the median age of IDH-wildtype LGGs is higher than that of IDH-mutant LGGs (55 vs 39 y according to data from The Cancer Genome Atlas). These relationships of age, tumor grade, and IDH mutation status are addressed from a practical perspective in an interesting article in this issue by Lennerz and colleagues.⁵

IHC for the R132H IDH1 mutation is efficient, rapidly obtained, inexpensive, and reliable to the point of being definitive for mutation calling when positive in most cases. R132H-antibody-negative cases represent either a true absence of IDH1 (or IDH2) mutation or alternatively noncanonical mutation. Since the detection of noncanonical IDH mutations typically adds to the time and cost of evaluation, the authors evaluated the economics and value of this testing as a function of patient age. A retrospective cohort analysis of 680 diffuse gliomas from 2010–2015 was performed at a tertiary center. IDH screening was performed by IHC (n = 198), DNA sequencing (n = 139), or both (n = 343). All tumors were reclassified according to the 2016 WHO criteria. The authors found that R132H-IHC had 100% sensitivity and specificity for detecting this mutation. The rate of noncanonical mutations in patients older than 55 was 2/253 (0.8%) compared with 24/229 (8.3%) in patients younger than 55. Both of the noncanonical mutations in patients older than 55 occurred in patients with GBM (although one of these occurred in a patient whose tumor had progressed from a previous LGG). Using in-house estimates of costs and turnaround times for tests, the authors estimate that by using the 55-year age cutoff, a total $403200 and 3642 days could be saved for their cohort, equating to 43% and 53% reduction of total costs and turnaround time, respectively. Moreover, their model estimates that the cost to sequence a noncanonical IDH variant is approximately $10000–$50000 per positive case for patients younger than 55 and approximately $250000 per positive case for patients older than 55. The authors are appropriately careful to note that all decisions for testing of noncanonical IDH mutations should be done on a case-by-case basis, with reference to the clinical situation. For example, one of the 2 over-55 positive cases was from a patient with a prior LGG, which may have prompted testing.

This study is the first to examine the economic impact of the WHO suggestion to avoid sequencing for noncanonical mutations in patients older than 55 with GBM. This study clearly shows that the costs associated with sequencing preclude its universal use, and the low rates of noncanonical mutations in patients older than 55 may present an opportunity to triage our testing accordingly. Historically, a $50000 cost per quality-adjusted life year gained has been used as a benchmark for value of care.⁶ However, it is clear that in oncology, no single benchmark exists, as the threshold for value changes depending on context. What is clear in oncology is that the benchmark for value exceeds $50000 cost per quality-adjusted life year, and may in fact even extend up to $250000. Testing for noncanonical IDH mutations is not without cost, and this cost needs to be evaluated along with the benefit of definitive information, with reference to treatment options and clinical management that may differ as a result of this information. At present, the costs for single gene sequencing are lower compared with next-generation sequencing ($420 vs $1800 in this study). In this study, the estimated cost to sequence a noncanonical IDH variant by single gene sequencing is approximately $105000 per positive case in patients older than 55.⁵ Although the metrics used here are not directly equitable to costs per quality-adjusted life year, it is clear that costs by single gene sequencing now begin to fall into the more acceptable range. As the biotechnology for sequencing continues to improve, the costs associated with testing will continue to drop, and it is possible that, despite its current costs, we will see gene sequencing become the standard of screening for IDH mutations in the future. As an additional consideration, our therapeutic approaches may evolve based on the presence of an IDH mutation, should data become available to warrant tailored management based on mutation status. These and other considerations raise the need for further study and consideration of the cost and benefits of testing in an era when molecular testing becomes increasingly available.