Mitotic index, microvascular proliferation, and necrosis define 3 pathological subgroups of prognostic relevance among 1p/19q co-deleted anaplastic oligodendrogliomas

Anaplastic oligodendrogliomas (AOs) are diffuse infiltrative gliomas occurring mainly in adults and classified as grade III in the current World Health Organization (WHO) classification.¹ In 2008, a dedicated program was set up in France to homogenize the management of de novo adult AOs (POLA network–Prise en charge des Oligodendrogliomes Anaplasiques), with the aim, inter alia, to provide a centralized pathological review of the cases and centralized molecular analysis.

In a previous paper,² we reported the histomolecular features of a series of 203 AOs with classic histopathological features and showed that despite careful pathological criteria, 1p/19q co-deletion (which remains a major prognostic factor) was recorded in 80% of these tumors. Following the Haarlem meeting,³ it was therefore decided to establish the entity “oligodendroglioma (and anaplastic oligodendroglioma) IDH-mutant and 1p/19q co-deleted” in the future WHO classification.

In our previous paper,² we classified 1p/19q co-deleted AO into 3 pathological subgroups: group 1 (>5 mitoses per 10-high-power fields [HPF] but no microvascular proliferation [MVP] and no necrosis), group 2 (MVP and no necrosis), and group 3 (MVP and necrosis). The threshold value for mitotic index was previously set up by Giannini and coworkers.⁴ At the time of publication, the short clinical follow-up of the cohort did not allow drawing definitive and robust conclusions in terms of prognostic value of these 3 pathological subgroups. Here, thanks to patient monitoring conducted within the POLA network, we could narrow down our results.

Briefly, 157 IDH-mutated, 1p/19q co-deleted patients were included in the present study. 1p/19q co-deletion was assessed by genomic-array analysis and IDH status by IDH1^R132H immunohistochemistry and by IDH1 and IDH2 direct sequencing for cases lacking IDH1^R132H immunostaining. Clinical characteristics of this cohort, as well as postoperative treatment, have been described in our previous study.² The following variables were searched for prognostic significance: age at diagnosis, sex, extent of surgical resection (biopsy and partial resection vs total and subtotal resection), preoperative KPS, pathological subgroups, MVP, necrosis, number of mitoses, Ki67 labelling index, and postoperative treatment (radiotherapy only, chemotherapy only, adjuvant Procarbazine, CCNU, and Vincristine (PCV) radiochemotherapy, or temoradiation and adjuvant temozolomide). Age at diagnosis, and pathological subgroup were used to build the multivariate Cox proportional hazards backward models. All statistical tests were 2-sided, and the threshold for statistical significance was P= .05. Analyses were conducted using PASW Statistics version 17.02 (IBM SPSS Inc.).

Median follow-up was 48.7 months (range: 2.8–77 mo), but median overall survival (OS) and progression-free survival (PFS) were not reached. The 5-year OS and PFS rates were 82% (SE = 4.3%) and 42% (SE = 8.7%), respectively. Among the clinical variables analyzed with univariate analysis (Fig 1A), age at diagnosis >50 years was a negative prognostic factor (P= .30 for OS and P= .031 for PFS), and preoperative KPS ≥ 80 was predictive of a longer PFS (P= .013). Postoperative treatment was predictive of PFS (P= .008) but was not predictive of OS. The worst prognosis was observed for patients treated with adjuvant temozolomide radiochemotherapy, as previously described.² Pathological subgroups were predictive of OS (P= .025, Fig 1B) and PFS (P= .025, Fig 1C), with the better prognosis being observed in group 1 patients.

Fig. 1.

Results of univariate analysis for the 157 1p/19q co-deleted AO. (A). The 3 pathological subgroups (group 1 (>5 mitoses per 10-HPF but no MVP and no necrosis), group 2 (MVP and no necrosis) and group 3 (MVP and necrosis)) are predictive of OS (B) and PFS (C). Abbreviations: AO, anaplastic oligodendroglioma; MVP, microvascular proliferation; OS, overall survival; PFS, progression-free survival.

On multivariate analysis, pathological subgroups adjusted by age at diagnosis remained of prognostic significance regarding OS and PFS, with patients belonging to group 1 having a lower tendency to recur (PFS: P= .014, HR = 1.87; 95%CI[1.14–3.09]) and patients belonging to group 3 displaying the worst prognosis (OS: P= .032, HR = 2.46; 95%CI[1.08–5.61]).

In AO, postoperative treatment has been shown to be associated with survival, as recently demonstrated by 2 phase 3 clinical trials.^5,6 However, these 2 clinical trials have also shown that impact of treatments in this tumor group need a median follow-up > 60 months to draw reliable conclusions.^7,8 Indeed, the major finding that 1p/19q co-deleted AOs benefit from early chemotherapy (reported in the 2013 paper) contradicted the 2006 reports (a feature linked to follow-up that was two short in the early papers). Since the median follow-up in our series is even shorter (48.7 mo), it appears premature to analyze these data at this stage. However, our results emphasized the usefulness of mitotic index, MVP, and necrosis to predict prognosis of “anaplastic oligodendroglioma IDH-mutant and 1p/19q co-deleted.” Whatever the occurrence of MVP or necrosis, AO remains grade III. This is in contrast with the criteria of grading used for astrocytomas.¹ In a previous study, we reported that the prognostic impact of necrosis in high-grade diffuse gliomas was radically different according to molecular subgroups,⁹ with the best prognosis being observed in IDH-mutated 1p/19q co-deleted gliomas, the worst prognosis arising in IDH-wild-type gliomas, and IDH-mutated 1p/19q non–co-deleted gliomas being of intermediate prognosis. This favors keeping grade III for AOs even if they display necrosis. However, we learned from the present study that OS of patients with AO lacking MVP was excellent, although these tumors have a significant tendency to recur. This raises the question of assessing grade III in tumor samples showing strong mitotic activity but lacking MVP; to address this question, large cohorts of patients in randomized assays and long follow-up are mandatory.

Funding

This work was supported by grants from Institut National du Cancer (grant INCa-DGOS-Inserm 6038). This work is funded by the French Institut National du Cancer (INCa, POLA network) and is part of the national program Cartes d′Identité des Tumeurs (CIT) (http://cit.ligue-cancer.net/), which is funded and developed by the Ligue nationale contre le cancer.

Conflict of interest statement. None declared.