Radiotherapy Plus Nivolumab or Temozolomide for Patients With Newly Diagnosed MGMT-Unmethylated Glioblastoma

The end-of-the-road for immunotherapies in GBM or new opportunities for more nuanced/aggressive approaches? CheckMate 498

In March 2011, on the basis of survival improvement, the FDA approved ipilimumab for treating patients with metastatic melanoma, representing the first new treatment for advanced melanoma in more than a decade and the first major approval of an immune checkpoint inhibitor (ICI). In the intervening decade, 16 immunomodulators, including 9 ICIs, have been approved by the FDA for more than a dozen major cancer types, including non–small cell lung cancer, small cell lung cancer, melanoma, renal cell carcinoma, breast cancer, head and neck cancer, and several others.

The one glaring gap in this success story is the paucity of data on ICIs providing similar clinical benefit in either the recurrent or newly diagnosed glioblastoma (GBM) setting. Preclinical rodent studies, especially those that have used hypofractionated radiation courses, have demonstrated robust efficacy of this approach in GBM, and, based on the biological principles of tumor analysis used in other disease sites, there appeared to be a potential role for this therapeutic approach in GBM.

The current study, CheckMate 498, represents a large, randomized effort in 560 patients with MGMT-unmethylated GBM in order to establish whether inhibition of the programmed cell death pathway with nivolumab (NIVO) would yield a survival advantage over chemotherapy. Patients were randomized 1:1 to receive either standard radiotherapy (RT; 60 Gy) with concurrent and maintenance temozolomide (TMZ) using the so-called “Stupp-regimen” or the same RT, without TMZ, but with the addition of 240 mg of NIVO every 2 weeks for eight cycles and then 480 mg every 4 weeks. TMZ was excluded in the experimental arm for two key reasons; first, the collective body of pre-existing data suggests that the survival benefit from the addition of TMZ to RT for MGMT-unmethylated GBM is minimal to non-existent. Secondly, ICIs work best in a setting wherein lymphocytopenia is avoided or minimized, and TMZ is a well-known lymphocytocidal agent. In this context, it is worth noting that the original clinical trials evaluating the activity of ipilimumab in melanoma brain metastases and follow-up trials with ipilimumab and NIVO identified that the use of steroids was also associated with considerable minimization of the efficacy of these particular ICIs. The majority of patients with GBM are usually on dexamethasone, and the lympho-depleting effect of this is a major concern regarding the value of ICIs in GBM. Precisely for this reason, CheckMate 498 mandated that, at randomization, patients must have been receiving ≤20 mg prednisone or ≤3 mg dexamethasone (or equivalent), and biopsy-only patients, who require large steroid doses were excluded. In spite of this robust effort, of the overall patients enrolled in the trial, 28% of those in the RT-NIVO and 34% of those in the RT-TMZ arm were on steroids, with almost 6% and 8% in each arm receiving >3 mg/day of a dexamethasone equivalent. Although, it is important to note that treatment in the study did not start until patients were within protocol-defined steroid dosing schedules.

From March 1, 2016, to October 25, 2018, 560 patients at 124 sites across 19 countries were randomized, 280 patients in each arm. More than 98% of patients received the assigned therapy, and there were no significant inter-arm imbalances in baseline characteristics. One interesting feature of this trial was the baseline assessment of PD-L1 expression in tumor tissue, and, crucially, this was <1% in 55.4% of RT-TMZ and 62.2% of RT-NIVO patients. Although considerable debate still rages regarding the precise role and predictive value of this biomarker, as well as optimal thresholds for enriching patient cohorts, such a high level of lack of expression of a key mechanistic molecule is clearly worrisome.

The final analysis was performed after 462 overall survival (OS) events had occurred. The median OS was 13.4 months in the RT-NIVO arm and 14.9 months in the RT-TMZ arm (HR, 1.31; P = .0037). The 24-month OS rates were 10.3% in the RT-NIVO arm and 21.2% in the RT-TMZ arm. This was clearly both disappointing and stunning; disappointing in that the addition of NIVO produced no discernible survival benefit and stunning in that the use of TMZ in patients with MGMT-unmethylated GBM was clearly associated with superior survival compared with those in the arm without TMZ, implying that withholding TMZ in these patients should be attempted with a great degree of caution. What could be a possible reason for this? Clearly, methylation is not an “on-off” switch but is “more on a spectrum,” implying that, depending on the test used, a certain proportion of “somewhat methylated” patients with GBM are likely going to be labeled “unmethylated,” and this could produce the kind of results observed in CheckMate 498.

In an attempt to identify whether there were any subgroups with potential survival benefit, an analysis based on PD-L1 expression was undertaken. Unfortunately, this, too, was a further disappointment. Among patients with baseline PD-L1 expression ≥1%, median OS was 12.6 months (n = 104) in the RT-NIVO arm and 15.5 months (n = 125) in the RT-TMZ arm (HR, 1.4). In a further attempt to understand whether any therapeutic benefit could be nested in a more selective subgroup, OS by baseline PD-L1 expression ≥5% was analyzed. Unfortunately, no significant survival improvement was identified, including in patients undergoing complete resection. In fact, no patient- or biomarker-specific subgroup could be identified that benefited more from NIVO compared with TMZ. Secondarily, a tiny proportion of patients were actually radiographic responders to both therapies, 7.2% to TMZ and 7.8% to NIVO. High grade (3/4) toxicities were observed in about a quarter of patients across both arms.

So, what should we make of these data? In 2020, in one of the earliest randomized trials on this theme, CheckMate 143, 369 patients with recurrent GBM were randomized to receive bevacizumab or NIVO, and no survival benefit was identified (Reardon, JAMA Oncology, 2020).¹ In 2021, results of the CheckMate 548 trial, randomizing 716 patients with newly diagnosed methylated GBM to the standard “Stupp regimen” plus either placebo or NIVO, were published and did not reveal any survival improvement from the addition of NIVO (Weller, Neuro-Oncology, 2021).² The CheckMate 498 trial represents the third negative trial in this “series.” Is this the end of the road for immunotherapies in GBM? Or does it represent an opportunity for more nuanced or aggressive approaches?

Given the rather promising preclinical results from these approaches, this is not the time to abandon this line of research. There are several opportunities for optimizing these approaches further. As a first pass, the combination of CTLA-4 and PD-1 inhibitors has shown more promise than single-agent therapy in most settings, and perhaps, in a relatively immunoprivileged site such as the brain, it might require dual-agent targeting before any significant impact is observed. That approach is being tested in the ongoing NRB-BN007 trial (NCT04396860). Perhaps, much shorter and dose-intensified RT courses are required in order to launch a radiation-induced antigen-release storm, and, additionally, to avoid the lymphocytopenia associated with long-course RT. This approach is being tested in an MSKCCC-led multi-institutional trial in recurrent, methylated GBM. Perhaps, we need to explore alternative enhancements to single-agent immunomodulation, such as some of the ongoing IDO inhibitor plus PD-1 inhibitor strategies. NRG-BN010 trial (NCT04729959) also represents a further line of exploration in terms of such a combination approach, investigating hypofractionated stereotactic RT with tocilizumab and atezolizumab in recurrent GBM, based on the expectation that PD-L1 blockade (atezolizumab) alone is most likely inadequate, and that the addition of the IL-6R antagonist tocilizumab would overcome the microenvironmental immunosuppression mediated by IL-6. Perhaps, we need to be far more selective and enroll genetically selected patients who are likely to have greater tumor mutational burden or mismatch-repair, which is being explored in an ongoing study led by the Alliance consortium (NCT04145115). We are just scratching the surface here, as multiple such ideas are currently under study, with the hope that, one of these days, we will find another drug with survival benefit in GBM, the first since TMZ was approved almost 2 decades ago.

References

1. DA Reardon, AA Brandes, A Omuro, et al. Effect of Nivolumab vs Bevacizumab in Patients With Recurrent Glioblastoma: The CheckMate 143 Phase 3 Randomized Clinical Trial. JAMA Oncol. 2020;6(7):1003-1010. https://jamanetwork.com/journals/jamaoncology/issue/
2. M Weller, M Lim, A Idbaih, et al. A Randomized Phase 3 Study of Nivolumab or Placebo Combined With Radiotherapy Plus Temozolomide in Patients With Newly Diagnosed Glioblastoma With Methylated MGMT Promoter: Checkmate 548. Neuro-Oncology. 2021;23(Suppl_6):vi55–vi56. https://academic.oup.com/neuro-oncology/article-abstract/