Embryonal tumor with multilayered rosettes (ETMR) diagnosis was previously based on pathological appearance and included tumors such as embryonal tumor with abundant neuropil and true rosettes (ETANTR), medulloepithelioma, and ependymoblastoma.¹ In the recent WHO classification of tumors of the CNS, ETMR, which is characterized by a chromosome 19 microRNA cluster (C19MC) that is a specific molecular alteration of locus 19q13.42, has been newly defined as an embryonal tumor² amplification, “ETMR, C19MC-altered,” to avoid misdiagnosis as one of the molecularly different types of brain tumors that are morphologically similar to ETMR.

Clinically, the tumor occurs through the neuroaxis and most commonly develops in children younger than 4 years, and its prognostic outcome is poor because of the aggressive behavior of the tumor. There are several reported studies that included extensive molecular analyses of ETMR;^3,4 however, the authors’ conclusions from past reports about the treatment of patients with these tumors are varied due to the suboptimal inclusion criteria for analysis without molecular confirmation of the tumor diagnosis.⁵

Methods

In this study, we evaluated 4 cases of ETMR with different clinical/morphological/radiological features in patients at our institution.

In addition, we performed a multivariate prognostic analysis of previously reported cases, including cases at our institution, in the English-language literature.

Results

In our evaluation of reported cases of ETMR, including those in patients treated at our institution, we recognized that multimodal treatment approaches that included aggressive surgery, high-dose chemotherapy (HDC), and local irradiation may be therapeutic factors that improve the clinical outcome of patients with ETMR.

Case Series

In the cohort of patients treated at our institution, there were 3 girls and 1 boy with a median age at diagnosis of 2.8 years (range 1.3–3.1 years; Table 1). One patient (case 2 in Table 1) has already been described elsewhere.⁶ The 3 remaining patients (cases 1, 3, and 4) each developed a supratentorial tumor without metastasis, and one of these patients (case 2) had infratentorial lesions with spinal dissemination. In pathological analyses, 3 of these cases presented morphological features of ETANTR; however, in case 1 there was also a case of perivascular pseudorosette without multilayered rosettes (Fig. 1), which was primarily diagnosed as a CNS ganglioneuroblastoma (GNB). In all cases, both positive immunohistochemical staining for Lin-28 homolog A (LIN28A) and locus 19q13.42 amplification were detected by fluorescence in situ hybridization (FISH).⁶ In terms of surgery, 2 patients underwent gross-total resection (GTR). One patient received upfront local irradiation, and the other underwent both craniospinal irradiation (CSI) and local radiation therapy (RT) in the primary treatment. There were 2 long-term survivors (cases 1 and 4 in Table 1) who underwent GTR and HDC with or without local irradiation. However, 2 patients with recurrence died of the disease 2 years after the primary diagnosis despite undergoing CSI or HDC after relapse of the tumor (cases 2 and 3 in Table 1).

TABLE 1.

Characteristics of ETANTR patients treated at our institution

Case No.	Age, yrs/Sex	Location	Pathology	LIN28A IHC	Metastasis	Op	Primary Tx	EFS, mos	OS, mos	Outcome
1	2.8/F	Frontal lobe	CNS GNB	+	No	GTR	Chemo/HDC/local RT	42	42	NED
2	1.3/M	Cerebellum, brainstem	ETANTR	+	Yes	Biopsy	Chemo	5	8	DOD
3	3.1/F	Occipital & parietal lobes	ETANTR	+	No	STR	Chemo/CSI/local RT	8	15	DOD
4	2.7/F	Temporal lobe	ETANTR	+	No	GTR	Chemo/HDC	141	141	NED

Chemo = chemotherapy; DOD = died of disease; IHC = immunohistochemistry; NED = no evidence of disease; STR = subtotal resection; Tx = treatment; + = present.

Pathological features of case 1. A: In the pathological slides with H&E staining, there are a number of perivascular pseudorosettes, but multilayered rosettes cannot be observed. B: LIN28 immunostaining showing positivity. C: FISH analysis revealing amplification of locus 19q13.42. Figure is available in color online only.

Clinical Characterization and Prognostic Analysis of Reported Cases of ETMR With C19MC Amplification

We reviewed ETMR cases reported in the literature and analyzed their data to detect clinical indicators that influenced the survival of patients. We used PubMed to search the published peer-reviewed literature. Search terms included ETANTR, ETMR, "embryonal tumor with multilayered rosettes," "embryonal tumors with abundant neuropil and true rosettes," C19MC, and "19q13.42." To reduce the bias from molecular heterogeneity, this analysis only included cases in which the presence of C19MC amplification was confirmed by FISH or various molecular copy number analyses (Table 2). Regarding treatment, we only included treatment options for the primary tumor as a variable in the prognostic analysis. For statistical analysis, overall survival (OS) was defined as the probability of survival, with only death as the event. Event-free survival (EFS) was defined as the probability of being alive and free of progression or relapse. Survival curves were plotted using the Kaplan-Meier method. The log-rank test and Cox proportional hazards model were used to detect differences in survival between the different groups of patients. Two-sided tests were used for all analyses, and p values < 0.05 were considered significant. All statistical analyses were performed with EZR (Saitama Medical Center, Jichi Medical University), which is a graphical user interface for R (The R Foundation for Statistical Computing). More precisely, it is a modified version of R commander designed to add statistical functions frequently used in biostatistics.⁷

In total, there were 43 cases from 12 articles, including the 4 cases reported in this study. Up to 88% of the patients (n = 38) developed tumors at the age of 3 years or younger.^8–19 The majority of tumors were located in the cerebrum (34 cases, 79%). There were very few metastatic cases in this cohort (3 of 43 cases, 7.0%). In terms of surgery, GTR of the tumor was achieved in 24 cases (56%). Fourteen cases (33%) were treated with HDC and 13 cases (30%) received upfront local irradiation. No patient underwent upfront CSI other than the one reported in our institution, probably due to the young age at onset. The median EFS and OS of 43 patients were 8.0 months and 12.5 months, respectively. Both EFS and OS were significantly longer for ETMR patients who underwent GTR, HDC, and upfront local RT. The multivariate analysis revealed that both HDC and upfront local RT correlated significantly with better EFS (HR 0.17, p = 0.0087; HR 0.17, p = 0.010) and OS (HR 0.29, p = 0.023; HR 0.28, p = 0.019), respectively. GTR was also correlated with better OS (HR 0.40, p = 0.039) (Figs. 2 and 3 and Table 3).

EFS and OS of 43 cases, including 39 cases from the literature and 4 cases from our hospital.

EFS and OS of 43 cases depending on the variables, including GTR (A and B), HDC (C and D), and local RT (E and F). Figure is available in color online only.

Discussion

ETMR is pathologically characterized by the presence of a multilayer rosette in which the lumen is either filled with eosinophilic debris or empty.²⁰ The morphological features of ETMR have mainly been regarded as characterizing three different pathological entities, as follows: ETANTR, ependymoblastoma, and medulloepithelioma.² After diagnostic integration detecting C19MC amplification and peculiar DNA methylation profiling of ETMRs, the pathological varieties reported among the tumors were different from the three subtypes described above.²⁰ The patient in case 1 of this study presented with several neuroblastoma-like cells without a multilayered rosette, which seemed to be CNS GNB. This case was also reminiscent of the histological diversity of ETMR and the clinical importance of molecular diagnosis among embryonal tumors, not otherwise specified, particularly during infancy, when most ETMRs occur.

For the diagnosis of ETMR, LIN28A immunostaining is a useful marker for screening to detect the tumors.² All cases in this report were positive for LIN28A. It should also be noted that positivity of the immunostaining has been identified in other childhood brain tumors including atypical teratoid/rhabdoid tumor.²¹

Amplification of the 19q13.42 locus including microRNA clusters, called C19MC, was primarily found in a subset of "CNS-primitive neuroectodermal tumor (PNET)."²² Afterward, Korshunov et al. reported that genetic amplification actually occurs in a variety of childhood embryonal tumors, including ETANTR, ependymoblastoma, and medulloepithelioma,²³ indicating that the tumors with C19MC are involved in a single tumor entity regardless of the pathological features. Amplified C19MC is a diagnostic hallmark of ETMRs.

Recently, DNA methylation analysis has emerged as a promising tool for the molecular diagnosis of ETMR because the profiling may reflect the cell of origin.²⁴ An analysis of the DNA methylation profiles revealed that there are ETMRs without C19MC amplification but with methylation features similar to those with the genetic alteration.³ The authors of this study reported that the majority of C19MC–wild-type ETMRs actually have a DICER1 mutation or amplification of the miR-17–92 miRNA cluster, indicating that there may be biological and/or clinical heterogeneities among the tumors.³ Furthermore, Lambo et al. suggested that ETMRs without C19MC amplification may have a worse prognosis than those with C19MC amplification.⁴

Treatment strategies for ETMRs, which have a dismal prognosis, have not been established because of its rarity. There are several retrospective analyses of ETMR cases in which the authors discuss its clinical and prognostic factors.^5,25,26 Horwitz et al. retrospectively analyzed the treatment results of 38 ETMR, ETANTR, and medulloepithelioma cases.⁵ They concluded that complete surgical resection, RT, and HDC were significantly associated with a better OS. In their study, however, Horwitz et al. analyzed only 12 cases (32%) for C19MC status.⁵ Manjila et al. described 2 pathologically diagnosed ETANTR cases.²⁵ In one of these cases the patient underwent radical multimodal therapy and lived longer than 7 years. On the contrary, another patient treated with HDC showed a poor clinical outcome. Neither of the patients in these 2 cases were checked for C19MC status. Another study by Spence et al. on 54 ETMR, ETANTR, and medulloepithelioma cases suggested that chemotherapy and/or RT may benefit a subset of patients.²⁶ In this study, Spence et al. checked 51 of the 54 patients for C19MC status; however, all of the reports mentioned here included pathologically diagnosed ETMR regardless of C19MC status, which may have provided only suboptimal information for elucidating the best treatment option for the tumors.

It may be possible to involve molecularly different tumors if ETMR is diagnosed using only conventional pathological tools other than FISH to detect amplified C19MC. These studies reported varied prognostic factors, including GTR and early local RT.^5,25 Our analysis included only patients with ETMR with C19MC amplification and revealed the significant clinical impact of multimodal treatment for these patients. Regarding RT, it could be reasonable to spare irradiation in the treatment of the primary tumor to avoid severe neurological sequelae because the tumor develops more frequently in patients 3 years of age or younger with a high risk for neurotoxicity from irradiation. Antonelli et al. reported that a long-term survivor received HDC without irradiation,¹⁴ as seen in our illustrative case 4 (Table 1). However, our data suggested the prognostic importance of local RT as a part of the primary treatment, which is concordant with a recent report by Mayr et al.¹⁹ Furthermore, salvage irradiation after recurrence may not be effective even with CSI, as seen in our case, in which the patient suffered a relapse.

In terms of the optimal drug combination for the tumors, Hanson et al. reported that their ETMR cases were successfully treated with a modified Third Intergroup Rhabdomyosarcoma Study (IRS-III) protocol, including doxorubicin and actinomycin D, which are known to be susceptible to ETMR cell line/xenograft.^18,27 We were unable to reach any conclusion through our analysis of drug selection or combination for the tumor due to the limited information available in the literature. Novel therapies for ETMR, such as radioimmunotherapy, have also been reported.¹⁷

The drawbacks of our study include its small sample size, which prevented the inclusion of other variables than those selected into the multivariate analysis for prognostic investigation. Future prospective clinical trials should be performed to confirm the validity of our findings since our data originated from retrospective reviews. Further research to seek the best therapeutic approach for the rare and prognostically dismal ETMR through international collaboration is warranted.

Conclusions

This case series highlights the pathological and clinical heterogeneity among ETMRs and the diagnostic importance of the molecular genetic approach among infant embryonal tumors. This retrospective analysis revealed that a multimodal approach had a great impact on the prognosis of patients with C19MC-altered ETMR.

Acknowledgments

This work is dedicated to the memory of Dr. Hiroshi Kishimoto.

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Fukuoka, Sugawa. Acquisition of data: Fukuoka, Sugawa, Tanami, Nobusawa, Hirato, Nakazawa, Kurihara. Analysis and interpretation of data: Fukuoka, Sugawa. Drafting the article: Sugawa. Critically revising the article: Fukuoka, Sugawa, Mori, Arakawa, Nakazawa, Koh. Reviewed submitted version of manuscript: Fukuoka. Approved the final version of the manuscript on behalf of all authors: Fukuoka. Statistical analysis: Sugawa. Study supervision: Fukuoka, Kurihara, Koh.

Supplemental Information