A systematic review of the clinicopathological features and prognostic outcomes of DICER1-mutant malignant brain neoplasms

Abstract

OBJECTIVE

DICER1-mutant malignant brain neoplasms are very rare tumors, and published data have relied on case reports or small case series. In this review, the authors aimed to systematically summarize the types and distribution patterns of DICER1 mutations, clinicopathological characteristics, and prognostic outcomes of these tumors.

METHODS

The authors searched PubMed and Web of Science for relevant studies. They included studies if they provided individual patient data of primary malignant brain tumors carrying DICER1 mutations.

RESULTS

The authors found 16 studies consisting of 9 embryonal tumors with multilayered rosettes (ETMRs), 30 pineoblastomas, 52 primary intracranial sarcomas, and 27 pituitary blastomas. Pineoblastoma, ETMR, and pituitary blastoma were more likely to carry DICER1 germline mutations, while only a small subset of primary intracranial sarcomas harbored these mutations (p < 0.001). Nearly 80% of tumors with germline mutations also had another somatic mutation in DICER1. ETMR and primary intracranial sarcoma were associated with an increased risk for tumor progression and relapse compared with pituitary blastoma and pineoblastoma (p = 0.0025), but overall survival (OS) was not significantly different. Gross-total resection (GTR) and radiotherapy administration were associated with prolonged OS.

CONCLUSIONS

ETMR, pineoblastoma, primary intracranial sarcoma, and pituitary blastoma should be considered rare phenotypes of the DICER1 syndrome, and families should be counseled and screened for associated tumors. ETMR and primary intracranial sarcoma had a higher risk of relapse. GTR and radiotherapy appeared to improve the OS of patients with DICER1-mutant malignant intracranial tumors.

The RNase III endoribonuclease is encoded by DICER1, resulting in the activation of the RNA-induced silencing complex, which is mandatory for the cleavage of double-stranded RNA and premicroRNA into small interfering RNA and microRNA.¹ Mutations in the DICER1 gene, either germline or somatic, are known to increase susceptibility to diverse benign and malignant tumors.^1–3 DICER1 syndrome is a rare autosomal dominant condition mainly seen in childhood and is caused by inherited germline DICER1 variants, with the most common neoplasm being pleuropulmonary blastoma.^2,4 In the latest WHO 2021 classification of tumors of the central nervous system, new DICER1-associated neoplasms have been recognized as distinct tumor types, including DICER1-mutant primary intracranial sarcoma and pituitary blastoma.⁵

In addition to pituitary blastoma and primary intracranial sarcoma, a subset of pineoblastomas and embryonal tumors with multilayered rosettes (ETMRs) is associated with DICER1 mutations.^6,7 These uncommon tumors are considered WHO grade 4 because of their aggressiveness and malignant behavior. The clinicopathological characteristics and prognostic outcomes of these tumors remain investigational because of their rarity. In this review, we aimed to systematically summarize the distributions of DICER1 mutations, clinicopathological characteristics, and survival outcomes of these rare malignant brain tumors.

Methods

Literature Search

A registration protocol for this review was not available. We conducted a comprehensive search in two electronic databases (PubMed and Web of Science) in February 2022 using the following search term: (DICER OR DICER1 OR DICER-1) AND (brain OR central nervous system OR intracranial OR CNS). Our study protocol strictly followed the recommendation of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.⁸

Selection Criteria

The title and abstract of the articles were screened by two independent reviewers (H.G.V. and M.K.L.). We used the following criteria to select studies: 1) observational studies on malignant brain tumors harboring DICER1 mutations, and 2) availability of individual patient data.

Data Mining

Two reviewers (H.G.V. and M.K.L.) independently reviewed the full text of potential studies, and data were extracted into a standardized spreadsheet. The following individual patient data were extracted from the articles: authors, institution, country, year of publication, study period, patient identification number, variants of DICER1 mutations, patient age and sex, histopathological features, Ki-67 labeling index, extent of resection (EOR), receipt of radiotherapy and or chemotherapy, progression-free survival (PFS) status, PFS time, overall survival (OS) status, OS time, and associated genetic events.

Statistical Analysis

For studies with potentially overlapping patients,^7,9–13 we carefully double-checked the patient identification number and demographic information, as well as DICER1 variants to avoid duplications. Chi-square and Fisher’s exact tests were used to compare categorical covariates, while ANOVA was used for comparison of continuous variables between tumor groups. Kaplan-Meier analysis and a Cox proportional hazards model were computed to assess PFS and OS. Two-sided p < 0.05 was considered statistically significant. All analyses were performed using R software (version 4.1.1, The R Foundation for Statistical Computing).

Results

We found 406 articles after searching two electronic databases, 35 of which were included for full-text reading. Subsequently, we included 16 studies with 118 DICER1-mutant malignant brain tumors comprising 9 ETMRs, 30 pineoblastomas, 52 primary intracranial sarcomas, and 27 pituitary blastomas for final analyses (Fig. S1 and Table S1).^7,9–23

Incidences of DICER1 Mutations Among Intracranial Malignant Neoplasms

Table 1 shows the incidences of DICER1 mutations among ETMR, pineoblastoma, primary intracranial sarcoma, and pituitary blastoma. Overall, DICER1 mutation accounted for most primary intracranial sarcomas and pituitary blastomas, with incidences from 93% to 100% of analyzed cases. However, these mutations were only found in 26%–50% of pineoblastomas and 4% of ETMRs.

TABLE 1.

Incidences of DICER1 mutations in each tumor type

Authors & Year	Tumor Type	No. of Cases	Incidence of DICER1 Alterations Among Analyzed Cases (%)
Diaz Coronado et al., 202210	PIS	70	26/27 (96)
de Kock et al., 201416	PineoB	21	9/21 (43)
de Kock et al., 201417	PitB	13	11/11 (100)
Kamihara et al., 202013	PIS	6	6/6 (100)
Koelsche et al., 201811	PIS	22	21/22 (95)
Lambo et al., 20197	ETMR	193	8/193 (4)
Lee et al., 201918	PineoB	4	2/4 (50)
Lee et al., 201919	PIS	3	3/3 (100)
Li et al., 202020	PineoB	93	16/61 (26)
Liu et al., 202112	PitB	17	13/14 (93)
Pfaff et al., 202039	PineoB	23	9/23 (39)

PineoB = pineoblastoma; PIS = primary intracranial sarcoma; PitB = pituitary blastoma.

Differences in Types and Distribution Patterns of DICER1 Mutation Variants Across the DICER1-Mutant Malignant Intracranial Tumors

Among the cases with a germline mutation, 78.3% of cases concomitantly carried another DICER1 somatic mutation. Most ETMRs, pineoblastomas, and pituitary blastomas had at least one DICER1 germline mutation, whereas 84% of primary intracranial sarcomas only harbored somatic mutations. Frameshift and/or nonsense mutations were the predominant mutations in ETMR, pineoblastoma, and pituitary blastoma, while missense mutations were commonly found in primary intracranial sarcoma (Table 2).

TABLE 2.

Associations of DICER1-mutant malignant brain tumors with types of DICER1 mutations

	ETMR	PineoB	PIS	PitB	p Value
DICER1 germline mutation					<0.001
≥1 germline mutation	8 (100)	14 (70.0)	5 (16.1)	22 (91.7)
No germline mutation	0 (0)	6 (30.0)	26 (83.9)	2 (8.33)
Frameshift/nonsense mutation					<0.001
Only missense mutation	1 (11.1)	3 (11.1)	38 (73.1)	8 (29.6)
≥1 frameshift/nonsense	8 (88.9)	24 (88.9)	14 (26.9)	19 (70.4)

Values are given as number (%) unless otherwise indicated.

Within-sequence distributions of DICER1 mutations in ETMR, pineoblastoma, primary intracranial sarcoma, and pituitary blastoma are summarized in Fig. 1. Overall, the RNase III domain was the most common site of mutation in the 4 tumors. In pineoblastoma, somatic mutations appeared to be sporadically distributed across exons 24 and 25 of DICER1, whereas germline mutations were more concentrated within the RNase III domain. ETMR and pituitary blastoma expressed the opposite patterns, with broader distribution of germline mutations and the aggregation of all somatic mutations in the RNase III domain in DICER1-mutated cases. Loss of heterozygosity was uncommon among all tumor types, occurring in a total of 8 pineoblastomas and pituitary blastomas (7.8%). This genetic event was likely to occur with germline mutation (5/8 cases).

FIG. 1.

The distribution of DICER1 mutations within the protein sequence in pineoblastoma (PineoB), ETMR, primary intracranial sarcoma (PIS), and pituitary blastoma (PitB). Germline mutations (green), somatic mutations (orange), and mutations of unknown origins (blue) are illustrated. The included types of mutations are nonsense (circles), frameshift (rectangles), and missense (triangles) mutations. Loss of heterozygosity (shapes without color fill) is also presented. Dim = dimerization domain; Helic = helicase domain; PAZ = PAZ domain; ResIII = restriction enzyme III domain; Ribo3 = RNase III (a and b) domain.

Clinicopathological Features of DICER1-Mutant Malignant Brain Tumors

Pituitary blastoma and pineoblastoma solely developed in the pituitary and pineal regions, respectively. On the other hand, ETMR was primarily found in the posterior fossa, while primary intracranial sarcoma was more commonly found in the cerebral hemispheres (p < 0.001). DICER1-mutant intracranial tumors were mainly seen in pediatric patients, and we found no differences in age at diagnosis between tumor groups. A subset of primary intracranial sarcoma might occasionally occur in adults and elderly patients. There was no difference in sex distribution between patients with DICER1-mutant intracranial tumors (p = 0.577). There was no significant difference in EOR patterns between different tumor groups, but we found that pituitary blastoma patients less commonly received radiotherapy and chemotherapy in comparison with the other groups (p < 0.001) (Table 3).

TABLE 3.

Associations between DICER1-mutant malignant intracranial neoplasms and clinicopathological features

Parameter	ETMR (n = 9)	PineoB (n = 30)	PIS (n = 52)	PitB (n = 27)	All Cases (n = 118)	p Value
Median age, yrs (range)	1.00 (1.00–2.00)	5.10 (1.00–31.0)	6.50 (0–76.0)	2.00 (0.170–24.0)	6.00 (0–76.0)	0.081
Sex						0.577
Female	6 (66.7)	16 (53.3)	30 (57.7)	19 (70.4)	71 (60.2)
Male	3 (33.3)	14 (46.7)	22 (42.3)	8 (29.6)	47 (39.8)
Surgery						0.175
GTR	1 (100)	9 (52.9)	23 (65.7)	5 (31.2)	38 (55.1)
STR	0 (0)	6 (35.3)	11 (31.4)	9 (56.2)	26 (37.7)
Biopsy	0 (0)	2 (11.8)	1 (2.9)	2 (12.5)	5 (7.2)
Radiotherapy						<0.001
No	NA	3 (17.6)	4 (10.3)	14 (87.5)	21 (29.2)
Yes	NA	14 (82.4)	35 (89.7)	2 (12.5)	51 (70.8)
Chemotherapy						<0.001
No	0 (0)	1 (5.9)	7 (17.9)	10 (62.5)	18 (24.7)
Yes	1 (100)	16 (94.1)	32 (82.1)	6 (37.5)	55 (75.3)

NA = not available.

Values are given as number (%) unless otherwise indicated. Percentages are based on the number of patients with available data.

We could not compare the demographic and clinicopathological characteristics of DICER1-positive and DICER1-negative tumors for ETMR, primary intracranial sarcoma, and pituitary blastoma because of a lack of data. For pineoblastoma, there were no significant differences in patient age, sex distribution, extent of surgery, and radiotherapy and chemotherapy administration between pineoblastoma cases with and without DICER1 mutations (Table 4).

TABLE 4.

Demographic and clinicopathological features of DICER1-positive and DICER1-negative pineoblastoma

Parameter	DICER1-Negative (n = 56)	DICER1-Positive (n = 27)	p Value
Median age, yrs (range)	5.05 (0.20–60.0)	5.20 (1.00–31.0)	0.427
Sex			0.874
Female	28 (50.0)	14 (51.9)
Male	28 (50.0)	13 (48.1)
Extent of surgery			0.422
Biopsy	2 (7.4)	2 (11.8)
STR	15 (55.6)	6 (35.3)
GTR	10 (37.0)	9 (52.9)
Radiotherapy			0.521
No	8 (25.8)	3 (17.6)
Yes	23 (74.2)	14 (82.4)
Chemotherapy			0.486
No	4 (12.1)	1 (5.88)
Yes	29 (87.9)	16 (94.1)

Values are given as number (%) unless otherwise indicated. Three DICER1-positive patients were omitted from this analysis because clinical data were missing.

Prognostic Outcomes of DICER1-Mutant Malignant Brain Tumors

We found that ETMR and primary intracranial sarcoma had higher risks of tumor progression and relapse compared with pineoblastoma and pituitary blastoma (Fig. 2). Patient sex, EOR, administration of radiotherapy, and chemotherapy did not affect PFS of patients with DICER1-mutant tumors (Fig. S2).

FIG. 2.

Kaplan-Meier curve illustrating the PFS of DICER1-mutant malignant intracranial neoplasms.

OS data were insufficient for ETMR. The OSs of pituitary blastoma (median OS of 66 months), pineoblastoma (median OS of 76 months), and primary intracranial sarcoma (median OS of 21 months) were not statistically different (p = 0.88) (Fig. 3). Sex and chemotherapy administration were not associated with better outcomes. On the other hand, GTR, subtotal resection (STR), and administration of radiotherapy significantly improved patient OS (Fig. S3). We did not find any associations of mutation types (e.g., frameshift, nonsense, and missense) with patient PFS and OS (data not shown).

FIG. 3.

Kaplan-Meier curve illustrating the OS of DICER1-mutant malignant intracranial neoplasms.

In a multivariate Cox regression model adjusted for patient age, sex, histology, EOR, radiotherapy, and chemotherapy, only GTR and radiotherapy were associated with superior OS (Table 5).

TABLE 5.

Multivariate Cox regression analysis for OS of patients with DICER1-mutant malignant brain neoplasms

Parameter	HR (95% CI)	p Value
Age (per yr increase)	0.856 (0.727–1.008)	0.062
Sex
Female	Reference
Male	0.586 (0.187–1.835)	0.359
Histology
PineoB	Reference
PIS	2.255 (0.546–9.311)	0.261
PitB	0.171 (0.027–1.095)	0.062
Surgery
Biopsy	Reference
STR	0.328 (0.073–1.467)	0.144
GTR	0.061 (0.010–0.372)	0.002
Radiotherapy
No	Reference
Yes	0.205 (0.044–0.951)	0.042
Chemotherapy
No	Reference
Yes	0.940 (0.231–3.827)	0.931

Prognostic Differences Between DICER1-Mutant and DICER1–Wild-Type Tumors

We could not calculate the differences in outcomes of DICER1-mutant versus DICER1–wild-type ETMR, primary intracranial sarcoma, and pituitary blastoma because of data insufficiency. For pineoblastoma, there was a tendency for prolonged OS and PFS in patients with DICER1-mutant compared with those with DICER1–wild-type tumors (Figs. S4 and S5). However, the differences did not reach statistical significance.

Discussion

The increasing availability of advanced genomic profiling of human disease in the clinical setting has accelerated our understanding of how we classify and understand pediatric brain tumors.^{7,16,17,24–31} Since the first report of DICER1 in pleuropulmonary blastoma,² germline mutations have been found in a number of childhood tumors, including cystic nephroma, Sertoli-Leydig cell tumor, multinodular goiter, pineoblastoma, and pituitary blastoma,^16,17,32,33 which are considered part of the DICER1 syndrome. More recently, DICER1 mutations have been reported in less than 5% of ETMRs and a subset of primary brain sarcomas.^7,11 DICER1-associated brain tumors are typically rare, and little is known about their clinicopathological profiles, prognostic outcomes, and whether they have similar distributions of DICER1 mutations.

Our integrated data showed that recurrent DICER1 germline mutations could occur in nearly all primary intracranial sarcomas and pituitary blastomas and a subset of ETMRs and pineoblastomas. These findings demonstrate that these tumor entities, especially pituitary blastoma and primary intracranial sarcoma, should indeed be considered rare but pathognomonic manifestations of the large spectrum of tumors associated with DICER1 syndrome. Therefore, patients with these tumor entities and their families should consider genetic counseling for DICER1 syndrome. This study also highlighted the significant differences in mutational types and distribution patterns of DICER1-mutant malignant intracranial neoplasms, emphasizing that they may have unique DICER1-driven tumorigenesis pathways.

The term "embryonal tumor with multilayered rosettes" was first proposed in 2010 as a unifying entity including different histological patterns such as ependymoblastoma, medulloepithelioma, or embryonal tumor with abundant neuropil and true rosettes.³⁴ These malignant brain tumors are solely found in infants and young children and have a 1-year OS of less than 50%.³⁵ Positive prognostic factors for ETMR are complete tumor removal, radiotherapy, and high-dose chemotherapy.³⁵ It has been well established that only less than 5% of ETMRs had DICER1 mutations and all DICER1-positive ETMRs occurred in the context of no amplification/fusion of a microRNA cluster on chromosome 19 with TTYH1.^7,31 Therefore, this subset of ETMR patients should be screened and followed for other DICER1 syndrome–associated tumors.

The first case of pituitary blastoma was described in 2008,³⁶ and its association with DICER1 mutations was established in 2011. The most common symptom of pituitary blastoma is Cushing’s syndrome.¹² These tumors were associated with a lower risk of progression compared with DICER1-mutant primary intracranial sarcoma and ETMR. Our integrated results also showed that patients with pituitary blastoma are less likely to receive radiotherapy and chemotherapy compared with patients with other DICER1-positive tumors. Previous studies showed that only GTR but not adjuvant therapy was associated with improved survival.¹² These results together with the fear of long-term radiation effects on pituitary function in very young patients could explain why adjuvant therapy is not preferred in pituitary blastoma compared with other DICER1-mutant intracranial cancers.

Germline mutations in the retinoblastoma gene RB1 lead to pineoblastoma.³⁷ Recent evidence has shown that individuals harboring germline mutations in DICER1 are predisposed to this rare brain cancer.^16,38,39 Using methylation profiling, patients with pineoblastomas with DICER1 mutations have a relatively good prognosis compared with patients with other molecular subtypes such as MYC/FOXR2-activated or RB1-altered pineoblastoma.^39,40

Primary intracranial sarcoma is a newly recognized entity in the WHO 2021 classification, which has closely similar histopathological findings to those of DICER1-altered embryonal rhabdomyosarcoma in other sites.⁵ Its incidence is particularly high in South America and is estimated to be 30-fold higher in Europe.¹⁰ Mutations in DICER1 have been found to be a major driver of alterations of primary intracranial sarcoma, in addition to mutations in TP53 or RAS.^10,11,13 These tumors typically have an aggressive clinical course. The use of adjuvant chemotherapy after surgery and prior to radiotherapy has been found to prolong patient OS compared with administering radiotherapy following surgery.¹⁰ Unfortunately, we could not confirm this finding since most included studies did not provide the timeline of treatments.

This study summarized the pathognomonic distributions and prognostic outcomes of DICER1 mutations in DICER1-mutant malignant intracranial tumors. Our findings highlighted the two-hit hypothesis with one inherited germline and another biallelic somatic alteration of DICER1-associated tumors, which has been established in previous studies.^2,4,6,7 We showed that GTR and radiotherapy could help prolong the OS of patients with DICER1-mutant brain cancers, which was further confirmed in the multivariate model, whereas chemotherapy did not add any survival advantages. Although chemotherapy has been shown to improve pineoblastoma and primary intracranial sarcoma outcomes in other series,^10,41 it should be noted that the pineoblastoma and primary intracranial sarcoma populations in those studies were confounded by other molecular subgroups with different prognoses. It is of clinical interest to know whether DICER1-positive tumors have different clinicopathological features and patient survival compared with DICER1-negative tumors in each tumor group. However, data were only sufficient to compare the PFS and OS of DICER1-positive and DICER1-negative pineoblastoma tumors. We observed a tendency of longer PFS and OS of patients with pineoblastoma carrying these mutations compared with patients with DICER1-wild type, but the differences were not statistically significant.

This study provides insights and updated information on the newly recognized entities in the 2021 WHO classification of central nervous system tumors. However, this work is constrained by certain limitations. First, given the rarity of these tumors, some of our data were based on case reports and case series, which can cause selection biases. Second, we could not estimate the prognostic differences between DICER1-mutant and DICER1-negative ETMR, primary intracranial sarcoma, and pituitary blastoma due to insufficient data. Future meta-analyses with additional data are needed to further clarify whether these mutations have any prognostic implications.

Conclusions

ETMR, pineoblastoma, primary intracranial sarcoma, and pituitary blastoma can be driven by DICER1 mutations and are rare manifestations of the DICER1 syndrome. These tumors have distinct types and distribution patterns of DICER1 mutations. ETMR, pineoblastoma, and pituitary blastoma were more likely to carry DICER1 germline mutations, whereas somatic mutations were more commonly seen in primary intracranial sarcoma. GTR and radiotherapy improved OS for patients with DICER1-mutant malignant intracranial tumors.

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: Vuong. Acquisition of data: Vuong, Le. Analysis and interpretation of data: Vuong, Le. Drafting the article: all authors. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Dunn. Statistical analysis: all authors. Administrative/technical/material support: all authors. Study supervision: Dunn, Vuong.