Outcomes after first relapse of children with intracranial ependymoma treated on the second l’Associazione Italiana di Ematologia e Oncologia Pediatrica (AIEOP) protocol

A new paper¹ from this Italian group in the current issue of Neuro-Oncology looks at the 64 children who had a first relapse out of the 160 originally treated on the previously published 2nd AIEOP protocol.² Of these 64 discussed in the current paper,¹ there were only 17 surviving at a median follow-up of 84-month post-relapse. This confirms the poor overall prognosis post-first relapse in childhood ependymoma seen by others.^3,4 The paper¹ selects for further analysis a cohort of 55 of these 64 children, excluding 4 who received no therapy at relapse and 5 with both tumor at the primary site and metastatic disease at relapse.

This cohort of 55 children¹ had 5-year EFS (event-free survival)/OS (overall survival) rates of 26.2%/30.8%, suggesting a potential for salvage in a minority, although potential for recurrence remains even after 5 years. This is a predominantly male cohort with 44 males and only 11 females, reflecting not only a male predominance in ependymoma² but also a higher relapse rate in males compared to females.¹ In the detailed analysis girls had a significantly favorable prognosis compared to boys, but only if they were asymptomatic at relapse. Girls presenting with symptoms did very badly.

There was no molecular subgrouping or characterization done on the relapsed samples.¹ There is a paucity of data on biological changes between initial presentation and relapse. Paired diagnostic and relapse samples might be extremely helpful in this regard and could give clues about innovative therapeutic approaches at relapse. A reasonable portion of the relapsed patients had molecular subgrouping at diagnosis, and this confirms that the relapsed population (26/32) is mainly posterior fossa group A (PFA) and that some had chromosome 1q gain (1q+) at diagnosis. Interestingly, although 1q+ was associated with metastatic disease at relapse,¹ it was not prognostic in this study,¹ as it was diagnosed in a prior study.⁵ The importance of the biological studies in PFA such as the interaction of EZHIP and loss of H27 trimethylation were emphasized in a recent review.⁶ and these should be pursued at presentation and at first relapse.

Therapeutic and prognostic variables were looked at in this study. However, the therapy was not standardized and varied over time and presumably between AIEOP’s 15 centers. A large proportion, 47/55, had surgery, at relapse.¹ Of the 37 with relapses at the primary site only, 26 had complete resections. Of the 6 with single metastasis only, all had complete resections. It is encouraging that those with a single metastatic nodule could be completely resected and that there are survivors in this group. Of those with complete resection, the 5-year EFS/OS were 36.3% and 41.2%, while there were no survivors whose surgery was incomplete.

This paper¹ therefore confirms starkly the need for complete resection if there is to be a chance of salvage at first relapse.¹ It would be interesting for the study authors to review the pre- and postoperative scans and operative notes of those with disease restricted to the primary site, who did not achieve a complete resection. Ten out of 11 of these patients, who did not achieve a complete resection, were in the posterior fossa speaking to the difficulty of surgery in this area suggesting that such patients be referred to surgeons with special expertise.

Radiation has become widely used at first relapse although there remains considerable controversy over whether this should be focal⁷ or craniospinal⁸ and whether it should be hypofractionated or standard fractionation. In this cohort, it was mainly hypofractionated and varied in dosing although some received standard fractionation. The use of craniospinal radiation was rare. The variable therapy and the small numbers of patients on this study¹ preclude comment on whether any method of radiation at first relapse was advantageous. However, it is worthwhile to note that of the 51 patients re-radiated that only one developed radionecrosis, a reassuringly low rate given that the median time from radiation at diagnosis to reradiation at relapse was only 17 months. It is, however, possible given potential variation in timing of scans, as these were not standardized post-reradiation, that some cases of radionecrosis may have been missed.

Only 18 patients were treated with chemotherapy with no standardization.¹ No responses were seen in evaluable disease. This supports the probability that standard chemotherapy may be futile in children with residual disease.⁹ However, its utility in those with complete resections at diagnosis has been suggested in a presentation of the preliminary results of the Children’s Oncology Study ACNS031¹⁰ and should be assessed in a randomized study of relapsed patients with complete resections.

The most important message of this paper is that complete resection should be the primary aim of the care team at first relapse in childhood ependymoma in those with localized single-site disease.¹ Second look surgery should be considered.⁴ Additionally, this paper shows that though the prognosis at first relapse is poor, there are long-term survivors. Also, routine scanning after initial presentation is worthwhile given that relapses in asymptomatic patients, discovered by scanning, are potentially salvageable, while those presenting with symptoms have very poor outcomes.

It is also possible, given the unknowns about therapy at relapse that better therapy may increase the number of survivors. As highlighted in the discussion, when frontline trials are designed for ependymoma, thought should be given to designing formal randomized relapse trials for those failing primary therapy. This would give a relapse population to study whose initial therapy was standardized. It would enable questions such as the right radiation approach, whether focal or craniospinal, whether standard fraction or hypofractionated, to be answered. The utility of chemotherapy in those with a complete resection at first relapse can only be answered in a randomized trial. The utility of chemotherapy studies at relapse, that use response of evaluable disease, is limited and may not reflect efficacy in those with complete resections. The need for this type of study is emphasized by the authors who state rightly: “There are still more questions than answers regarding salvage treatment for ependymoma”.¹ Randomized studies at relapse as well as at presentation are needed in childhood ependymoma.