Short-term outcomes associated with temozolomide or PCV chemotherapy for 1p/19q-codeleted WHO grade 3 oligodendrogliomas: A national evaluation

Abstract

Background

The optimal chemotherapy regimen between temozolomide and procarbazine, lomustine, and vincristine (PCV) remains uncertain for WHO grade 3 oligodendroglioma (Olig3) patients. We therefore investigated this question using national data.

Methods

Patients diagnosed with radiotherapy-treated 1p/19q-codeleted Olig3 between 2010 and 2018 were identified from the National Cancer Database. The overall survival (OS) associated with first-line single-agent temozolomide vs multi-agent PCV was estimated by Kaplan-Meier techniques and evaluated by multivariable Cox regression.

Results

One thousand five hundred ninety-six radiotherapy-treated 1p/19q-codeleted Olig3 patients were identified: 88.6% (n = 1414) treated with temozolomide and 11.4% (n = 182) with PCV (from 5.4% in 2010 to 12.0% in 2018) in the first-line setting. The median follow-up was 35.5 months (interquartile range [IQR] 20.7-60.6 months) with 63.3% of patients alive at the time of analysis. There was a significant difference in unadjusted OS between temozolomide (5-year OS 58.9%, 95%CI: 55.6-62.0) and PCV (5-year OS 65.1%, 95%CI: 54.8-73.5; P = .04). However, a significant OS difference between temozolomide and PCV was not observed in the Cox regression analysis adjusted by age and extent of resection (PCV vs temozolomide HR 0.81, 95%CI: 0.59-1.11, P = .18). PCV was more frequently used for younger Olig3s but otherwise was not associated with patient’s insurance status or care setting.

Conclusions

In a national analysis of Olig3s, first-line PCV chemotherapy was associated with a slightly improved unadjusted short-term OS compared to temozolomide; but not following adjustment by patient age and extent of resection. There has been an increase in PCV utilization since 2010. These findings provide preliminary data while we await the definitive results from the CODEL trial.

WHO grade 3 oligodendrogliomas (Olig3s; traditionally referred to as “grade III anaplastic”) are a rare subset of infiltrative gliomas—comprising about 6% of all gliomas—defined by the presence of both an IDH mutation and 1p/19q chromosomal arms codeletion.^1–3 The 1p/19q codeletion is a strong prognostic marker associated with increased responsiveness to chemotherapy and superior outcomes in patients with diffuse gliomas.^4,5 However, despite recent advances in understanding its distinct molecular profile, the optimal therapeutic regimen for newly diagnosed Olig3s remains unclear and controversial.^6,7 The NCCN guidelines (version 2.2021) currently recommend maximal safe surgical resection followed by radiotherapy and chemotherapy for Olig3s.^5,8 The addition of chemotherapy to radiotherapy in patients with Olig3 stems from 2 major randomized controlled trials that demonstrated improved survival among patients harboring a 1p/19q codeletion who received procarbazine, lomustine (ie, CCNU), and vincristine (PCV) chemotherapy with radiotherapy, as compared to those who received radiotherapy alone.^9,10 In the past decade, extrapolating from the positive results of the EORTC 26981-22981/NCIC CE3 trial in the glioblastoma setting, in addition to the lower toxicity profile and more convenient administration of temozolomide than PCV, temozolomide has become an alternative first-line chemotherapy for the treatment of Olig3s.^5,11–13

Initial results from the CODEL phase III trial (NCT00887146) confirmed the superior progression-free survival (PFS) associated with radiotherapy and radiotherapy with concomitant and adjuvant temozolomide for Olig3 patients, compared to temozolomide monotherapy.¹⁴ CODEL has since been redesigned, randomizing newly diagnosed oligodendroglioma patients to radiotherapy with either adjuvant PCV or with concurrent temozolomide (followed by adjuvant temozolomide), in order to identify the optimal chemotherapeutic approach¹⁵; however, until its results become available, there is no clear consensus among the neuro-oncology community as to which regimen to use in the management of patients with Olig3s.⁶ Therefore, we sought to compare the early survival outcomes for newly diagnosed Olig3 patients following temozolomide or PCV in the first-line setting using a large National Cancer Database (NCDB). As the final results of the CODEL trial are years away, our study has the potential to offer preliminary clinical guidance about the effectiveness of these regimens in the management of patients with newly diagnosed Olig3.

Materials and Methods

Data Source and Study Design

The NCDB is a nationwide database comprising more than 85% of newly diagnosed malignant brain cancers from more than 1500 Commission on Cancer-accredited institutions in the United States.^16,17 Patients were identified with histologically confirmed infiltrative gliomas between 2010 (1p/19q codeletion data are only incorporated in the NCDB as of 2010) and 2018, using ICD-O-3 histological codes (ie, 9380-9382, 9400-9401, 9420, 9440-9442, 9450-9451, 9460), a malignant behavior code (ie, 3), and a brain site code (ie, 71.0-71.9).¹⁸ Cases were defined using the primary brain-specific variables for WHO grade 3 and for loss of heterozygosity/deletion of both chromosome arms 1p and 19q.¹⁹ A subset of cases with 1p/19q codeletion were encoded in the NCDB as WHO grade 4—likely reflecting the outmoded “glioblastoma with oligodendroglial features”—and herein were reclassified as WHO Olig3s in accordance with the 2016 and 2021 WHO classifications.^2,3 Patients were excluded if they were <18 years old at diagnosis, were diagnosed at an index institution and treated entirely elsewhere, or if they were not 1p/19q-codeleted—resulting in 2308 patients. Given the survival benefits associated with radiotherapy for Olig3 patients, we only analyzed Olig3 patients who received radiotherapy (n = 1748). Of these, 1596 patients received either single-agent or multi-agent chemotherapy in the first line and were included for analysis.

Variables and Statistical Analyses

The NCDB only reported a patient’s first course of chemotherapy treatment, defined as that received prior to disease progression or recurrence. Chemotherapy is encoded in the NCDB as single-agent or multi-agent, without further details. In order to confirm that single-agent referred to temozolomide and multi-agent to PCV, we reviewed all encoded chemotherapy data for Olig3s from registry-submitted data from three Commission on Cancer-accredited institutions (Brigham and Women’s Hospital, Dana-Farber Cancer Institute, and Massachusetts General Hospital) between 2000 and 2017 and validated their encoding using the medical records. Extent of resection (EOR) was categorized as biopsy-only (either excisional biopsy surgery or diagnostic biopsy of primary tumor site), subtotal resection (STR), or gross total resection (GTR).

The primary outcome of interest was overall survival (OS), measured from pathological diagnosis to the date of death with patients censored at the date of most recent follow-up.^19,20 OS was estimated by Kaplan-Meier methods and compared by log-rank tests and Cox regression analysis. Multivariable Cox proportional hazards regression was used to evaluate the association between chemotherapy regimen and OS, excluding patients with <1 month of follow-up. Due to limited follow-up, the NCDB excludes OS data for patients diagnosed in the final year of the dataset, which for this dataset was 2018. For the exploratory analyses of factors associated with PCV vs temozolomide receipt, categorical variables were compared using Fisher’s exact test (or χ ² approximation) and continuous variables were compared using Wilcoxon’s rank sum test. The NCDB only reports Karnofsky Performance Scale (KPS) and MGMT promoter methylation data in a subset of patients. Insurance status was categorized as uninsured, privately insured, Medicare, Medicaid, or other governmental insured. Patients’ managing institution type (only available for patients 40 or older) was categorized by Commission on Cancer classification: community, comprehensive community, academic/NCI-comprehensive cancer center, or integrated network cancer program. STATA (v15.1, StataCorp) was used for statistical analyses, with 2-sided P-values <.05 specified as significant. This study was approved by the Mass General Brigham institutional review board (#2015P002352).

Results

One thousand five hundred ninety-six patients with radiotherapy-treated 1p/19q-codeleted WHO Olig3s met inclusion and exclusion criteria; of whom 88.6% (n = 1414) received single-agent chemotherapy (from 94.6% in 2010 to 86.2% in 2013 and 88.0% in 2018) and 11.4% (n = 182) received multi-agent chemotherapy (from 5.4% in 2010 to 13.8% in 2013 to 12.0% in 2018). The cohort represented patients treated at 445 centers.

From 3 NCDB-reporting institutions, n = 73 1p/19q-codeleted Olig3s treated with chemotherapy between 2000 and 2017 were abstracted and submitted to registries. Of these, 84.9% (n = 62) and 15.1% (n = 11) were encoded as first-line single-agent and multi-agent chemotherapy and were validated by medical chart review to represent temozolomide and PCV, respectively, in all cases. Although these institutional data may not be wholly representative of chemotherapy coding practices nationwide, they provide some reassurance for the validity of the assumptions made in the design of the methodology.

In the national data, Olig3 patients presented at a median of 51 years of age (interquartile range [IQR] 40-61). By EOR, 18.0% of patients had biopsy-only, 35.7% STR, and 46.4% GTR. Radiotherapy began a median of 37 days (IQR 29-48) after definitive resection, with a median of 59.4 Gy (IQR 59.4-60.0) in 30 fractions (IQR 30-33), including a median of 59.5 Gy (IQR 59.4-60.0) in a median of 30 fractions (IQR 30-33) for single-agent temozolomide patients, and a median of 59.4 Gy (IQR 59.4-60.0) in a median of 33 doses (IQR 30-33) for multi-agent PCV patients. Single-agent temozolomide chemotherapy often began on the same day as radiotherapy (median difference 0 days after start of radiotherapy, IQR: 0-0; likely representing concurrent temozolomide followed by adjuvant temozolomide), whereas multi-agent PCV chemotherapy often began following radiotherapy (median difference 70 days after start of radiotherapy, IQR: 0-84). Olig3s treated with multi-agent PCV were younger (median 48 years, IQR 36-59) as compared to temozolomide-treated Olig3s (median 52 years, IQR 40-62, P = .008). There was no difference in PCV vs temozolomide utilization by MGMT promoter methylation status (P = .17; only available for n = 692). The median KPS for temozolomide- and PCV-treated patients was 90 (IQR 80-90; only available for n = 243). Furthermore, there were no differences in PCV vs temozolomide utilization by patients’ insurance status (P = .24) or managing cancer program type (P = .74).

The median follow-up from the time of diagnosis was 35.5 months (IQR 20.7-60.6 months). At the time of analysis, 63.3% of Olig3 patients in the national dataset were alive, with 36.7% of patients reaching the endpoint of mortality. Overall, the unadjusted OS difference between single-agent temozolomide (5-year OS 58.9%, 95%CI: 55.6-62.0) and multi-agent PCV chemotherapy (5-year OS 65.1%, 95%CI: 54.8-73.5) reached significance (log-rank P = .04; Figure 1). In multivariable Cox regression analysis adjusted for age and EOR—factors were previously shown to have crucial prognostic value in Olig3s—a significant difference in OS between first-line PCV (adjusted HR 0.81, 95%CI: 0.59-1.11, P = .18) and temozolomide chemotherapy was no longer detected (Table 1).

Figure 1.

Unadjusted overall survival (OS) associated with first-line single-agent temozolomide vs multi-agent PCV for radiotherapy-treated WHO grade 3 oligodendroglioma patients. Kaplan-Meier OS estimates for radiotherapy-treated Olig3 patients who received either single-agent TMZ (solid line) or multi-agent PCV (dashed line) treatment. Log-rank test P = .04. OS is measured from the date of surgical diagnosis. Patients with <1 month of follow-up were excluded from the analysis. Abbreviations: PCV, procarbazine, lomustine, and vincristine; TMZ, temozolomide.

Table 1.

Multivariable Cox Regression Analysis of Overall Survival in Olig3s

	OS Measured From Pathological Diagnosis
	HR	95% CI	P value
Chemotherapy
Temozolomide	Referent
PCV	0.81	(0.59-1.11)	.18
Age at diagnosis	1.06/yr	(1.05-1.07)
Extent of resection
Biopsy-only	Referent
Subtotal resection	0.83	(0.65-1.05)
Gross total resection	0.82	(0.65-1.04)

Abbreviations: CI, confidence interval; HR, hazard ratio; OS, overall survival; PCV, procarbazine, lomustine, and vincristine.

Complete multivariable data were available for n = 1299 radiotherapy-treated Olig3s. Patients with <1 month of follow-up were excluded, as measured from pathological diagnosis. P value is only shown for the primary association of interest.

Discussion

In this national analysis of 1596 patients with radiotherapy-treated, 1p/19q-codeleted WHO Olig3, we found that first-line multi-agent (PCV-based) chemotherapy was associated with improved short-term unadjusted OS compared to single-agent (temozolomide-based) chemotherapy. However, following adjustment for key prognostic variables of age at diagnosis and EOR, a significant difference was no longer observed between the 2 chemotherapy regimens. This observation likely reflected the baseline differences in Olig3 patients that received PCV vs temozolomide: for instance, PCV-treated Olig3s were associated with younger ages than temozolomide-treated Olig3s.

Two landmark trials established the chemosensitivity of 1p/19q-codeleted oligodendrogliomas and defined a role for PCV in the treatment of Olig3s: RTOG 9402 and EORTC 26951 both showed that among patients with 1p/19q-codeleted tumors, treatment with radiotherapy and PCV resulted in significant improvements in OS as compared to treatment with adjuvant radiotherapy alone.^9,10 However, despite class 1 evidence being only available for adjuvant PCV with radiotherapy, temozolomide was the more commonly used regimen in the United States. This likely reflects several factors that lead to providers favoring temozolomide over PCV in the adjuvant setting (for instance, fewer side effects and ease of administration).⁶ Specifically, lomustine can cause myelosuppression that may be prolonged and progressive. In addition, vincristine is not thought to achieve appreciable penetration of the blood-brain barrier, potentially limiting the overall efficacy of PCV in the treatment of an intracranial tumor.²¹ In light of patients’ poor tolerance of and ability to complete treatment with PCV, temozolomide has also been explored as a potential alternative in the treatment of Olig3s. In the national data, we observed that younger Olig3 patients were more likely to receive PCV than older patients, which may reflect providers’ concerns over toxicity and tolerability of PCV in older patients. Unfortunately, the NCDB did not report details regarding the clinical decision making behind regimen selection. In light of patients’ poor tolerance of PCV, temozolomide has also been explored as a potential alternative in the treatment of Olig3s. Our findings suggest that this approach does not impact short-term survival. Longer-term real-world data—as well as class 1 evidence—are still warranted to address whether temozolomide and PCV are equivalent in this setting.

Numerous retrospective studies have demonstrated that temozolomide may be a safe and efficacious alternative for Olig3 treatment, achieving similar response rates as chemotherapy with PCV. In a clinical trial of 20 patients with Olig3, temozolomide was given for up to 24 weeks following surgical resection: results demonstrated a high response to temozolomide, with clinical improvement reported in 60% of patients and 2 reports of grade 3/4 toxicity.²² Similarly, another phase II study assessed the use of upfront temozolomide in the treatment of newly diagnosed Olig3 and found a median PFS of 62 months, with limited high-grade toxicities despite a dose-intense schedule.²³ Robust tumor responsiveness and well-managed toxicity following temozolomide for Olig3 have been consistently observed in retrospective studies.^24,25 It should be noted, however, that none of the above studies included radiotherapy, so it is possible that response rates might have been even higher had radiotherapy been added to temozolomide therapy, although toxicity profiles may also have been more severe.

Despite these reports suggesting an opportunity for temozolomide in first-line treatment of Olig3, whether temozolomide can outperform PCV remains uncertain. Recent studies have shown that temozolomide chemotherapy is more likely to induce hypermutation with mismatch repair deficiency than PCV in gliomas; however, the impact of the acquired hypermutated phenotype on OS remains controversial.^26,27 There have been a limited number of retrospective studies that have attempted to compare these 2 chemotherapy regimens head-to-head; even fewer of which have incorporated the 2016 revised WHO histomolecular definitions of Olig3s. In a large multi-institutional retrospective study by Lassman et al, in which only 89 1p/19q-codeleted Olig3s were compared by chemotherapy agent, PCV-alone demonstrated improved time-to-progression (median 7.6 months compared to 3.3 months in temozolomide-alone, P = .02), but no difference in OS (P = .16).²⁸ In the Olig3s that also received radiotherapy, there was no difference in either time-to-progression or OS for PCV vs temozolomide.²⁸ In the Neuro-Oncology Working Group (NOA) phase III trial NOA-04, in which patients with anaplastic gliomas (only 69 with 1p/19q codeletion) were randomized to upfront radiotherapy or upfront chemotherapy (either PCV or temozolomide), researchers ultimately found that chemotherapy alone was not superior to radiotherapy.²⁹ In the subset (n = 33) of CpG island hypermethylated (CIMP) 1p/19q-codeleted cases, PCV-alone demonstrated improved PFS, but not OS, compared to temozolomide-alone. Notably, although NOA-04 demonstrated better tumor control with PCV, it also demonstrated worse toxicity with PCV. Among the anaplastic glioma patients treated with upfront chemotherapy, treatment was discontinued in 33% of patients treated with PCV (n = 54), all due to toxicity, while none of the temozolomide-treated patients (n = 53) required temozolomide discontinuation due to toxicity.³⁰

Although the Lassman et al’s study and NOA-04 had moderate follow-up durations of approximately 10 years, it is likely that their data had not yet reached maturity and were underpowered with respect to the question of OS for PCV vs temozolomide for Olig3s. Similar to these studies, our data only capture the short-term outcomes of PCV vs temozolomide for Olig3s, and it is possible that with much longer follow-up, one regimen may prove superior at later landmarks. To rigorously answer this question, the phase III CODEL trial (NCT00887146) was redesigned as a randomized noninferiority comparison of radiotherapy either followed by adjuvant PCV or with concomitant temozolomide followed by adjuvant temozolomide for grade 2 and 3 newly diagnosed 1p/19q-codeleted oligodendrogliomas, powered for PFS, and secondarily assessing OS, toxicity, and quality of life outcomes. Herein, the majority of first-line temozolomide was started concurrently with radiotherapy, whereas the majority of first-line PCV was administered following radiotherapy. The appropriate scheduling of chemotherapy with radiotherapy remains a key question. Additional important therapeutic considerations are under investigation, including the role of deferring radiotherapy in reducing late toxicity (POLCA trial, NCT02444000).

Because the final results of CODEL are likely years away, we evaluated the short-term OS outcomes associated with PCV or temozolomide as first-line treatment in a contemporary national cohort of patients with 1p/19q-codeleted Olig3s. Based on the recent trial results demonstrating improved efficacy associated with radiotherapy (compared to chemotherapy monotherapy), we restricted our analysis to those Olig3 patients who received radiotherapy. Although single-agent temozolomide was the most common regimen, utilization of PCV increased between 2010 and 2018—despite a concerning increase in the cost of lomustine of approximately 1500% from 2013 to 2018 in the United States.³¹ The NCDB does not report details about which chemotherapeutic agents are administered, so regimens incorporating Carmustine (BCNU) instead of lomustine would be grouped together with PCV. Notably, in our analyses, choice of single-agent temozolomide vs multi-agent PCV in the first line was not associated with patients’ insurance status or their treating hospital’s type (ie, community vs academic/NCI-comprehensive care center). This increase in multi-agent PCV utilization may reflect the timely publication of results from the RTOG 9402 and EORTC 26951 randomized controlled trials, which firmly established the survival benefit of adding PCV to radiotherapy for 1p/19q-codeleted gliomas. In line with the aforementioned clinical trials’ survival findings, we found no difference in short-term OS between radiotherapy-treated Olig3s treated with first-line PCV vs temozolomide following adjustment for age and EOR. However, in unadjusted analyses, PCV was associated with slightly improved short-term OS. Although only results from a randomized trial comparing PCV and temozolomide will be able to ultimately answer the question of differences in long-term outcomes, our findings suggest that either PCV or temozolomide may offer comparable early OS outcomes in conjunction with radiotherapy in the first-line management of newly diagnosed Olig3s when taking into account patients’ age and EOR—which supports tailoring the temozolomide vs PCV selection to each patient’s unique circumstances.

Limitations

Although our analysis encompassed a national dataset, the NCDB’s limited encoding of 1p/19q codeletion status (starting in 2010) constrained the cohort’s follow-up to a median of 36 months (IQR 21-61 months). As a result, our findings only apply to the short-term outcomes of 1p/19q-codeleted Olig3 patients and do not capture any delayed differences between PCV and temozolomide. For example, in RTOG 9402 and EORTC 26951, the survival benefits from chemotherapy were observed after approximately 6 years—suggesting that Olig3 patients that experienced an early death did not benefit from the addition of PCV. Although both trials indeed confirmed that long-term analysis is key in this population, it remained unclear whether temozolomide might outperform PCV in the short term. For instance, the lack of short-term benefit with PCV could correspond to an increased rate of toxicity-related death with this regimen. Our findings add to the previous literature and will be helpful in decision making—especially for elderly Olig3 patients with shorter life expectancy. The relatively short follow-up and a comparably small number of PCV patients also restricted the effect size that we were able to measure. Based on a 2-tailed log-rank power analysis using the Schoenfeld method and the 62.4% of temozolomide-treated patients that were censored in this cohort, this analysis had 80% power to detect a change in the 5-year OS from 58.9% (as seen for temozolomide-treated patients) to 71.2%; which was modestly larger than the change that we detected.

Furthermore, first-line chemotherapy was encoded as single-agent and multi-agent, without details about agents, doses, early and late toxicities, or duration of treatment; therefore, it is possible that some patients received multi-agent chemotherapy that included temozolomide and other investigational agents. To help address this limitation, registry-submitted data for Olig3s from 3 Commission on Cancer-accredited institutions were evaluated, in which we noted that all single-agent cases were temozolomide and all multi-agent cases were PCV. Although these institutions may not be wholly representative of all institution types nationally, NCDB coding standards are standardized across all Commission on Cancer-accredited institutions nationwide. However, there remains the possibility that some multi-agent chemotherapy represented less common regimens such as CCNU and procarbazine (without vincristine), whereas some single-agent chemotherapy may have represented CCNU monotherapy. The second or subsequent courses of therapy were not reported by the NCDB, so it could not be determined whether patients that received concurrent temozolomide with radiotherapy additionally received adjuvant temozolomide or not. IDH mutational status is not separately encoded in registry data, so all oligodendrogliomas herein were defined by their 1p/19q codeletion and thus were also likely IDH-mutant.¹⁸ Nor did the NCDB report data regarding patients’ seizure status, symptomatology, development of pseudo-progression, or radiotherapy effects. Additionally, only OS was encoded for outcomes, precluding the evaluation of PFS for PCV vs temozolomide in Olig3s.

Conclusions

In this real-world study of patients with newly diagnosed 1p/19q-codeleted Olig3 who received post-surgical radiotherapy—63% of whom were still alive at the time of analysis—we found a significant improvement in short-term unadjusted OS associated with first-line multi-agent PCV compared to single-agent temozolomide chemotherapy, which did not persist following adjustment for the key prognostic factors of age and EOR. While results from the CODEL trial will ultimately offer guidance with respect to optimal management for patients with newly diagnosed Olig3, our results offer preliminary evidence that both PCV and temozolomide chemotherapy regimens when combined with radiotherapy may be viable options in the first-line setting with respect to short-term OS outcomes after taking into account patients’ age and EOR.

Funding

No funding supported this work.

Conflict of interest statement. M.T. reports consulting or advisory role from Agios Pharmaceuticals, Integragen, and Taiho Oncology, outside the submitted work; research funding from Sanofi, outside the submitted work. The other authors report no relevant conflicts of interest.

Authorship statement. Conception and study design: J.B.I. Data collection: N.L., M.M., and V.K.K. Data analysis: J.B.I. Data interpretation and manuscript writing: all authors. Critical review and revisions: all authors.

Data availability statement. Data are available by application to the NCDB.