Long-term outcomes from the second l’Associazione Italiana di Ematologia e Oncologia Pediatrica (AIEOP) protocol

See the article by Massimino et al in this issue, pp. 848–857.

The article in this issue by Massimino et al. provides a 10-year follow-up of the second AIEOP protocol (designed as AIEOP 2 in this commentary) of pediatric intracranial ependymoma, which was first reported in 2016 with a 5-year outcome.^1,2 There is no other comparable prospective trial with such long follow-up. The 2016 paper from AIEOP 2 reported a 5-year PFS (progression-free survival) of 65.4% and an OS (overall survival) of 81.1%, for this trial of 160 patients between 2012 and 2016, while the current paper shows a 10-year PFS of 58.5% and an OS of 73.3%.^1,2

Poorer results than AIEOP 2 are reported from retrospective data, although these are often collected over long periods of time with potential confounding temporal improvements in therapy.^3,4 Interestingly, 5-year results reported from COG (Children’s Oncology Group) trial ACNS0121 had a PFS of 62.7% and 83.8%, similar to the 5-year results of the AIEOP trial.⁵ If the ACNS0121 results hold up as well as AIEOP 2, this would suggest a relatively good outcome for those treated on trial. On AIEOP 2, 15% of all relapses occurred after 5 years and did not impact overall survival.¹ It is possible that a complete understanding of PFS and OS data in ependymoma trials may take longer follow-up than 10 years.

It should be noted that ependymoma trials may exclude high-risk populations. The AIEOP 2 trial excluded those under the age of 3 years before an amendment and the COG0121 trial excluded those under age 1, which are considered high-risk populations.^1,5 The results from from the St. Jude trial (SJYC07) of children younger than 3 years with ependymoma, with a PFS of 75.1% and an OS of 92.6% at a 4-year follow up, suggest caution in the assignment of higher risk to the very young.⁶

There is general agreement on the importance of a complete resection in ependymoma, seen again in AIEOP 2, in contrast to doubts about the efficacy of chemotherapy.^1,7 In ACNS0121, those children with a gross total resection/near total resection received no chemotherapy.⁵ On AIEOP 2, those with Grade III tumors, and no evidence of disease, received 4 chemotherapy courses, while those with Grade II tumors did not.¹ On AIEOP 2, the results for those with Grade III tumors were 58.3% PFS at 5 years and 50.9% at 10 years.¹ On ACNS0121 the PFS for Grade III tumors was 60.7% at 5 years and 49.9% at 10 years, with a shorter follow-up.⁵ These similar results do not lend support to the effectiveness of chemotherapy in Grade III tumors. However, the place of chemotherapy in ependymoma can only be truly assessed with the results from the randomized recently closed ACNS0831 trial (NCT01096368) and the still accruing randomized SIOP-EP-II study (NCT02265770). Results were significantly poorer for those with Grade III tumors compared to Grade II on both the AIEOP 2 and ACNS0121 trials.^1,5 The impact of Grade in these 2 large trials is noteworthy given prior negative reports of the value of histological grading, including in the SJYC07 trial and the lack of agreement between pathologists on grading in ependymoma.^6,8

The long-term outcome data of AIEOP 2 are improved on that of the earlier report by an updated exploration of molecular factors in outcome. Outcomes for posterior fossa group B (PFB) were excellent, with 100% survival, although one had a late relapse.¹ For the 41 patients identified in AIEOP 2 as having posterior fossa group A (PFA) tumors the 10-year PFS was 43% confirming, with ACNS0121, that this group of tumors has a high relapse rate.^1,5 Surprisingly on AIEOP 2, the 10-year OS for PFA tumors was 64.5% given the expectation most with PFA tumors, who relapsed, would have died. There was a relatively low number of samples, 68 out of 95 samples, in which were robust molecular subgrouping was successful, out of a total number of 160 children on the trial. The authors offer reassurance that were no major differences in clinical characteristics between patients for whom molecular analyses were feasible and those for whom they were not.¹ In the analysis of future trials, immunohistochemical assay for H3K27me3 is recommended to improve definition of the PFA group who, unlike PFB, have global reduction in staining.⁹ The AIEOP 2 trial did not show any significant adverse impact of RELA fusions on the prognosis of supratentorial ependymoma which is similar to the findings on ACNS0121.^1,5

The presence of 1q+ in PFA tumors was associated in AIEOP 2 with a worse, but nonsignificant, outcome.¹ A significant adverse association in PFA with 1q+ was seen in ACNS0121 and in a paper looking at several European trials.^5,10 However, PFA 1q+ was associated with a significant risk of distant relapses in AIEOP 2 which led the authors to speculate about the utility of craniospinal for PFA 1q+. An analysis of either 1q gain or CDKN2A loss on AIEOP 2 showed a nonsignificant worse outcome but was significantly associated with tumor dissemination at recurrence.¹ It should be noted that those with CDKN2A loss in AIEOP 2 were all supratentorial tumors and that a faction of the 1q+-positive tumors were supratentorial, a location where the significance of 1q+ is unknown.¹ The combined risk of 1q gain or CDKN2A loss in PFA would be an interesting analysis to perform in other prospective trials. The relevance of loss of chromosome 6q has been advanced recently as a marker of ultra high-risk PFA.¹¹ This was not looked at in AIEOP 2 but would potentially be evaluable given the strength of this study’s copy number analysis.

AIEOP 2 is a valuable addition to the field as it is the only trial with published long-term outcome data¹ and with the addition of molecular data.¹ The 10-year outcomes are better than expected from retrospective studies although the lack of a plateau in the PFS curve shown is worrying. More than half the children with the PFA tumors relapsed by 10 years and this group continues to need better therapy. AIEOP 2 shows chemotherapy feasibility post-radiation in a high-risk ependymoma cohort but its efficacy needs published results from randomized trials.

Conflict of interest statement. There was no funding for this work. The author declares no conflict of interest.