Medulloblastoma in children with Fanconi anemia: Association with FA-D1/FA-N, SHH type and poor survival independent of treatment strategies

Abstract

Background

The outcome of children with medulloblastoma (MB) and Fanconi Anemia (FA), an inherited DNA repair deficiency, has not been described systematically. Treatment is complicated by high vulnerability to treatment-associated side effects, yet structured data are lacking. This study aims to give a comprehensive overview of clinical and molecular characteristics of pediatric FA MB patients.

Methods

Clinical data including detailed information on the treatment and toxicities of 6 previously unreported FA MB patients were supplemented with data of 16 published cases.

Results

We identified 22 cases of children with FA and MB with clinical data available. All MBs with subgroup reporting were SHH-activated (n = 9), confirmed by methylation profiling in 5 patients. FA MB patients exclusively belonged to complementation groups FA-D1 (n = 16) or FA-N (n = 3). Patients were treated with postoperative chemotherapy only (50%) or radiotherapy (RT) ± chemotherapy (27%). Of 23% did not receive adjuvant therapy. Excessive treatment-related toxicities were frequent. Severe hematological toxicity occurred in 91% of patients treated with alkylating chemotherapy, while non-alkylating agents and RT were less toxic. Median overall survival (OS) was 1 year (95%CI: 0.3–1.8). 1-year-progression-free-survival (PFS) was 26.3% ± 10.1% and 1-year-OS was 42.1% ± 11.3%. Adjuvant therapy prolonged survival (1y-OS/1y-PFS 0%/0% without adjuvant therapy vs. 53.3% ± 12.9%/33.3 ± 12.2% with adjuvant therapy, P = .006/P = .086).

Conclusions

MB in FA patients is strongly associated with SHH activation and FA-D1/FA-N. Despite the dismal prognosis, adjuvant therapy may prolong survival. Non-alkylating chemotherapy and RT are feasible in selected patients with careful monitoring of toxicities and dose adjustments. Curative therapy for FA MB-SHH remains an unmet medical need.

Graphical Abstract

Graphical Abstract.

Key Points.

• Children with Fanconi anemia and medulloblastoma have a very poor prognosis.

• Adjuvant therapy prolongs survival but frequently leads to significant toxicity.

• Antineoplastic therapy is feasible but must respect the high risk of toxicities.

Importance of the Study.

This is the first structured overview of detailed clinical and molecular characteristics of children with Fanconi anemia (FA) diagnosed with medulloblastoma (MB). Severe treatment-associated side effects were reported, yet no structured data on survival, treatment strategies, and treatment-related toxicities are available. We show that pediatric FA patients develop exclusively SHH-activated MB and observe that the increased risk of MB development is restricted to FA patients with biallelic BRCA2 or PALB2 germline mutations (FA complementation groups FA-D1 or FA-N). Severe treatment-associated toxicities are mainly hematological and pronounced after alkylating chemotherapy. Non-alkylating chemotherapy and radiotherapy are feasible and better tolerated. Unlike in standard MB protocols, focal radiotherapy may be considered in FA MB patients. Outcome of FA MB patients is still poor. Adjuvant therapy provides a survival benefit despite toxicity, with no significant difference between radio- and chemotherapy. Our data provide a basis for future treatment decisions for pediatric FA MB patients.

Fanconia anemia (FA) is a rare inherited syndromic disease caused by chromosomal instability due to deficiency in the DNA repair system.¹ It is characterized by congenital abnormalities, progressive bone marrow failure, and cancer predisposition.^1,2 Mutations in at least 22 genes (FANCA to FANCW) have been identified to cause FA, constituting the FA pathway and the corresponding FA complementation groups FA-A to -W.^1,3,4 The inheritance pattern is autosomal recessive¹ except for FA-B, which is linked to the X-chromosome,⁵ and FA-R, being autosomal dominant.⁶ The different complementation groups differ in frequency and type of congenital malformations⁴ as well as cancer type and cancer risk.^1,7 The genes involved in the FA pathway play a key role in removing and repairing interstrand crosslinks (ICLs) from DNA.¹ Disruption of this pathway renders cells prone to cross-linker induced stress, eg,by DNA cross-linking chemotherapeutics and radiotherapy (RT).⁸ Excessive increase of chromosomal breakage upon exposition to cross-linking agents such as diepoxybutane or mitomycin C is used in functional assays to confirm FA,⁹ enabling the diagnosis of FA without the knowledge of the underlying genetic cause.⁷

The FA complementation groups D1 and N (FA-D1 and FA-N) are characterized by biallelic alterations in FANCD1/BRCA2 and FANCN/PALB2, respectively.¹ FA-D1 constitutes approximately 3% of FA patients⁷ and about 1 % of FA patients belong to complementation group FA-N.^1,7 Both are associated with frequent phenotypical abnormalities⁴ and the highest cancer risk among all FA patients.⁷ In a recently published German cohort study, FA-D1 patients had the highest risk of developing cancer among all FA patients, with a chance of 76.1 % of developing cancer until their 18th birthday.⁷ While in the more prevalent FA complementation groups, hematologic disorders, namely myelodysplastic syndrome (MDS), and acute myeloid leukemia (AML), as well as squamous cell carcinoma prevail,^1,3,7 in contrast, the spectrum of malignancies in patients with FA-D1 and FA-N is broader. Besides hematologic malignancies, medulloblastoma (MB) is the most frequently observed solid tumor, but patients also develop other embryonal tumors including Wilms tumor, hepatoblastoma, and neuroblastoma.^10–12 The distinct association of FA-D1 and FA-N to brain tumors, primarily MB, as well as to Wilms tumor and neuroblastoma was even designated as “D1-N associated cancers” by some authors.¹⁰

MB is one of the most common malignant brain tumors in children and adolescents.¹³ It comprises 4 genetically defined groups: Wingless/int-1 (WNT)-activated, Sonic Hedgehog (SHH)-activated TP53-wild type, SHH-activated TP53-mutant and non-WNT/non-SHH MB.¹⁴ Therapy is risk- and age-adapted. Broadly speaking, after surgery, older children are treated with adjuvant craniospinal irradiation (CSI) and chemotherapy, while in younger children, CSI-avoiding/-delaying chemotherapy concepts are primarily pursued.^15,16 Treating children with FA and MB has turned out to be challenging. Intensive conventional (radio-) chemotherapy concepts frequently are not feasible in FA patients due to increased incidence and extent of treatment-related toxicities such as prolonged bone marrow aplasia with consecutive infectious complications.^17–20

Since the first case report of a patient with FA and MB in 1985,²¹ several other case reports have been published that describe the occurrence of MB in FA patients and its management. Unlike epidemiologic and genetic findings,^7,20,22 structured data on therapy-related toxicities in FA patients with MB are not available. For this report, we collected published and unpublished reports on treatment in FA-associated MB aiming to assess effectiveness and associated toxicities.

Methods

Unpublished patients with FA and MB were identified within the HIT2000 study (NCT00303810; n = 2) or from clinical consultation requests (n = 4). The project was approved by the responsible ethics committee. Informed consent was obtained from patients or legal guardians before data collection. The histological MB diagnosis was confirmed at the Brain Tumor Reference Center of the German Society for Neuropathology and Neuroanatomy at the University of Bonn, Germany in all patients. FA diagnosis was confirmed functionally (sensitivity towards cross-linking agents) and/or genetically at the German reference laboratory for FA at the Department of Human Genetics, University of Würzburg, Germany in all patients but case 4 (genetic diagnosis by local center).

The DKFZ Molecular Neuropathology classifier, version 12.5, was used to analyze DNA methylation profiles.²³ For t-SNE visualization, all MB cases from Capper et al.²³ and Cavalli et al.²⁴ were used as references. Raw IDAT files were processed and visualized using “R Studio” (v4.1.3), after loading the files using the minfi package (v1.40.0). Data from 450k and EPIC arrays were combined and normalized using the Noob function. All probes targeting the sex chromosomes and all probes containing SNPs within 2 nucleotides around a CpG site were removed. Probes not explicitly targeting the human genome were removed from analysis as well. Batch effects introduced due to different array types were corrected using ComBat function from sva R package (v3.42.0). Finally, the 10 000 CpG sites with the largest standard deviation across all samples were chosen for t-SNE analysis, which was performed using the R package Rtsne (v0.16). The values were calculated by defining the “eigenvalue k” as “k=50,” and 100 random numbers were created using “set.seed(100).” The settings for Rtsne function were “max_iter=2,000” and “perplexity=20.”

Methylation-derived copy number variation analysis was performed using raw IDAT files resulting from DNA methylation profiling. The files were processed and visualized using “R Studio” (v4.3.2). The individual copy number profiles were calculated based on raw methylated/unmethylated signals using the conumee R package (v1.36.0). Ratios were normalized against reference samples.

Published case reports on FA patients were identified via a literature review performed on Pubmed on July 14th, 2023 using the search term “(FA medulloblastoma) OR (FA brain tumor) OR ([FA BRCA2] OR [FA PALB2] AND [Case Reports{Publication Type}]).” Only reports on pediatric patients with FA and MB with information on treatment were extracted (for details see Supplementary Figure S1).

Statistical analyses were performed using R v4.3.1 (R packages survival v.3.5.7 and survminer v.0.4.9) and IBM © SPSS ® Version 28.0.1.1.

Results

Cohort Composition

We identified 6 unpublished and 16 published^{17–21,25–35} pediatric patients with FA-associated MB and information on treatment available (Table 1, cohort assembly visualized in Figure 1A). Detailed description of the unpublished cases is available in the supplements (Supplementary Case Descriptions). There is no overlap between cohorts. Literature cases were published, and previously unreported cases were diagnosed between 1985 and October 2022.

Table 1.

Cohort Characteristics

Characteristics	Unpublished cases n = 6	Literature cohort n = 16
Median age at MB dx in years (range)	4.1 (0.6–6.9)	3.2 (1.3–6.0)
Sex
Female	2	8
Male	4	8
Histology
DMB/MB with desmoplastic components	5	2
CMB	1	1
Other/not specified	0	13
SHH pathway
Activated	6	3
N/R	0	13
Extent of resection
GTR	3	8
R ≤ 1.5 cm2	2	0
R > 1.5 cm2	1	0
No information available or N/Aa	0	8
Initial M status (MRI)
M0	5	5
M+	1	1
No information available	0	10
DNA methylation analysis
SHH MB	5	0
N/C or N/R	1	16
FA complementation group
FA-D1	5	11
FA-N	1	2
N/R	0	3
Therapy group
No adjuvant therapy	0	5b
Chemotherapy only	3	8
Radiotherapy (± Chemotherapy)	3	3
Occurrence of other malignancies
Yes	5 → 5 patients developed 7 malignancies: 2 simultaneously with MB dx 2 before MB dx 3 after MB dx	9 → 9 patients developed 13 malignancies: 3 simultaneously with MB dx 8 before MB dx 2 after MB dx
No	1	7
Relapse/PD
Yes	6c	9
No	0	5 (of which 4 died shortly after MB dx)
N/R	0	2
Death
Yes	6	14
No, alive at LFU	0	1
N/R	0	1
Reason of death
Primary tumor or PD/relapse	5	11 (2—primary tumor; 9—PD/relapse)d
Other	1 (sAML)	3 (therapy-associated toxicity)d 1 (medulloblastoma or hepatoblastoma or MDS)e
N/A—death N/R or alive	0	2 (1—N/R; 1- alive)
Median PFS in years (range)	0.55 (0.2–1.4), n = 6	0.3 (0.0–6.4), n = 13
Median OS in years (range)	0.8 (0.2–1.8), n = 6	1.0 (0.0–6.4), n = 13

^a Extent of resection not applicable in three patients who did not undergo surgery before death.

^b Two patients who died before the initiation of adjuvant therapy were excluded in treatment-dependent survival analyses.

^c In case 3, progression was most probable, but not securely diagnosed (Supplementary Table 4).

^d In case 14, both PD and therapy-associated sepsis were stated as the cause of death.

^e One patient (case 21) died shortly after diagnosis of 3 simultaneous malignancies. The exact cause of death remains unclear.

MB, medulloblastoma; dx, diagnosis; DMB, desmoplastic medulloblastoma; CMB, classic medulloblastoma; SHH, sonic hedgehog; N/R, not reported; N/A, not applicable; N/C, not conducted; FA, Fanconi anemia; PD, progressive disease; LFU, last follow-up; sAML, secondary acute myeloid leukemia; MDS, myelodysplastic syndrome; PFS, progression-free survival; OS, overall survival.

Figure 1.

Clinical courses and molecular information of the study cohort. (A) Overview of cohort assembly, clinical courses, and outcomes. (B) Explorative t-SNE plot of unpublished cases with methylation data available, all MB reference cases from Capper et al., 2018 (23) and Cavalli et al., 2017 (24). (C) Explorative t-SNE plot of unpublished cases with methylation data available, only SHH-MB reference cases from Capper et al., 2018 (23) and Cavalli et al., 2017 (24).

Clinical Characteristics of FA MB Patients

Age at MB diagnosis ranged from 0.6–6.9 (median 3.7 ± 1.6) years. Information on extent of resection and dissemination was available in 14 and 12 cases, respectively. Gross total resection or tumor residue of ≤ 1.5 cm² could be achieved in 13/14 (92.9%) cases. MB was localized (M0) in 10/12 (83.3%) and disseminated (M+) in 2/12 (16.7%) cases (Table 1, Supplementary Table S1). Bone marrow biopsy as part of initial MB staging was performed in 6 patients, three of whom had MDS (Supplementary Table S2). One patient (Case 13) was diagnosed with MB due to MRI screening. In another patient, a cMRI was conducted for hepatoblastoma diagnosis due to family history (case 21). In the remaining, clinical symptoms led to MB diagnosis (89.5%, 17/19 cases with information available), including one patient who took part in an institutional screening program but was diagnosed in between the scheduled MRI scans (Supplementary Table S3).

Mutational and Molecular Features of FA MB Patients

FA complementation group allocation was available in 19/22 cases. All patients with FA-associated MB were either FA-D1 (n = 16) or FA-N (n = 3). In one patient, the complementation group was determined by loss of BRCA2 protein. In the other 18 (94.7%) patients, compound heterozygous (13/18, 72.2%) or homozygous (5/18, 27.8%) pathogenic or likely pathogenic germline mutations in BRCA2 or PALB2 were detected (Supplementary Table S4).

MB diagnosis preceded FA diagnosis in 9/22 (40.9%) cases, while in 12/22 (54.5%) patients, FA was already known at the time of MB diagnosis. In one patient, no information on the relation of MB and FA diagnosis was available (Supplementary Table S1).

The majority of MB were desmoplastic or contained at least desmoplastic components (7/10 with data available, 70%; Table 1, Supplementary Table S1). Immunohistochemistry showed activation of SHH pathway in all unpublished cases (6/6, 100%) and in 3 published cases (cases 18, 19, and 22; Table 1). In half of the tumors with TP53 somatic status available, TP53 was mutant (3/6); in the other 50%, only wild-type alleles were detected. Interestingly, 1 patient (case 5) had 2 simultaneous MB lesions (L1 and L2), 1 classified as TP53-wild type (L1), and the other as TP53-mutant (L2; Supplementary Tables S1 and S5). In case 1, no sequencing was conducted, yet the lack of nuclear p53 accumulation makes a mutation unlikely. Panel sequencing of tumor tissue conducted in 4 patients identified a pathogenic PTCH1 mutation in 1 patient (Supplementary Table S5). BRCA2 mutation was the only recurrent somatic pathogenic gene mutation in these 4 patients.

Methylation profiling was available for 6 tumors of 5 unpublished cases (cases 1–2, cases 4–6; L1, and L2 of case 5) and confirmed all tumors as “MB, SHH activated” with a score of ≥0.9. With regard to the SHH subtypes, 1 tumor was classified as “SHH subtype 1” with a score of ≥0.9. Of the other tumors with scores for SHH subtypes <0.9, 3/5 were classified as “SHH subtype 3” (score range 0.55–79), 1/5 as “SHH subtype 4“ (score 0.61), and one tumor had equally high scores for “SHH subtype 3” (score 0.41) and “SHH subtype 1” (score 0.39; Figures 1B and 1C, Supplementary Table S1).

Copy number variation profiles derived from methylation profiling showed various chromosomal gains and losses, without uniform changes (Supplementary Figure S2) and suggested MYCN amplification in 2 cases (case 2, L2 of case 5), and no MYC amplification in any tumor.

Treatment Strategies and Toxicities

Five patients (22.7%) received no adjuvant treatment (surgery at maximum), 11 patients (50.0%) were treated with chemotherapy only and 6 patients (27.3%) received RT with (n = 5/6) or without (n = 1/6) chemotherapy. No patient tolerated MB therapy according to standard protocols. Either drug doses were reduced, therapy elements like RT omitted or therapy was discontinued due to toxicity, FA diagnosis, PD, or second (syn- and metachronous) malignancy (see Table 2 for treatment details of unpublished cases and Supplementary Table S6 of published cases).

Table 2.

Detailed Overview About Treatment of and Toxicities in the Unpublished Cases

Case No.	Adjuvant treatment group	Previous cancer treatment or presence of MDS/AML at MB treatment initiation	Detailed description of adjuvant MB treatment	Toxicity
1	Chemotherapy only	Yes: Chemotherapy according to modified dose-reduced NHL BFM 90 protocol for B-NHL	Five blocks of modified SKK chemotherapy: -1 × carboplatin 160 mg/m2/d + etoposide 120 mg/m2/d on day 1–3, w/o i.vent. MTX - 2 × HD-MTX 5 g/m2/d + VCR 1.5 mg/m2/d on day 1, with i.vent. MTX on day 1 + 2, with elevated leucovorin rescue -NFDA on last 2 blocks → Discontinued due to diagnosis of sAML	After block 1: Pancytopenia with the need to delay next block for >2 weeks, fever in neutropenia After blocks 2–4: Episodes of fever, blocks better tolerated than the first block After block 5/upon sAML diagnosis: Prolonged pancytopenia
2	Radiotherapy (± Chemotherapy)	Yes: MDS-RAEB with blasts, no previous treatment conducted	Conventional RT (photons): CSI 35.2 Gy + PF boost up to 55 Gy + spinal Met boost up to 44.8 Gy w/o concomitant VCR At PD/diagnosis of sAML: no specific AML treatment initiated, 6 blocks of i.vent. MTX (4 consecutive days each, in a 4 week interval) At further PD: Irradiation of 2 metastases with stereotactic RT (1 treated 1-stage and the other with 10 fractions of RT)	RT: Pancytopenia with the need for platelet transfusions, otherwise moderate acute toxicity (skin toxicity, mucositis, nausea); toxicity led to therapy interruption of 4 days i.vent. MTX: Ulceration and necrosis of the skin above Rickham reservoir upon sAML diagnosis: Progredient bone marrow failure → repeated localized bacterial and viral infections in the ear-nose-throat and gastrointestinal area, mucositis
3	Radiotherapy (± Chemotherapy)	No	9 blocks of modified SKK chemotherapy: -1 × cyclophosphamide 800 mg/m2/d on day 1–3, VCR 1.5 mg/m2/d on day 1, with i.vent. MTX day 1–4 - 8 × HD-MTX (dose finding in first blocks: 0.5 -> 1 -> 0.75 mg/m2/d) + VCR 1.5 mg/m2/d on day 1, with i.vent./.i.th. MTX (i.th. after 3rd block due to leaking Rickham Reservoir) At suspicion of early local relapse: Focal conventional RT (photons; tumor boost 54 Gy, PF boost 45 Gy)	After block 1: Prolonged pancytopenia with fever in neutropenia and the need for transfusion of erythrocytes and platelets → change to exclusively HD-MTX + VCR blocks After block 2–9: Mild granulocytopenia, no severe toxicity RT: no toxicity
4	Chemotherapy only	No	1 block of modified SKK chemotherapy: - 1 × HD-MTX 1 g/m2/d + VCR 0.6 mg/m2/d on day 1, with i.vent. MTX on day 1 + 2 → Discontinued due to diagnosis of retroperitoneal tumor (most likely nephroblastoma)	After block 1: Diaper dermatitis, no severe toxicity
5	Chemotherapy only	Yes: MDS with transition into sAML, no antineoplastic treatment	1 block of modified SKK chemotherapy: - 1 x HD-MTX 5 g/m2/d + VCR 1.5 mg/m2/d on day 1, with i.vent. MTX on day 1 + 2 → Discontinued due to diagnosis of sAML Palliative chemotherapy with i.v. vinblastine 3 mg/m2 weekly	After block 1: Hepatobiliary toxicity (GOT > 2.5 x ULN, GPT > 5 x ULN, GGT > 5 x ULN), mucositis Vinblastine chemotherapy: Well-tolerated, no relevant bone marrow aplasia
6	Radiotherapy (±chemotherapy)	No	Six blocks of modified SKK chemotherapy: - 6 × HD-MTX (dose finding in first blocks: 0.5 -> 1 -> 2 -> 5 g/m2/d) + VCR 1.5 mg/m2/d on day 1, with i.vent. MTX on day 1 + 2 Focal conventional RT (protons): Tumor boost 54 Gy Maintenance chemotherapy (8 months): - i.v. vinblastine 3 m/m2 weekly - p.o. valproate (target serum level: 75–100 µg/m) - 8 × i.v. cyclophosphamide: 4-weekly, starting dose 500 mg/m2 (application 1–4) -> 1000 mg/m2 (application 5–8)	SKK Chemotherapy: Well-tolerated, no relevant toxicity RT: well-tolerated, no relevant toxicity Maintenance chemotherapy: After first application of cyclophosphamide SIRS with detection of norovirus, bone marrow depression and need for thrombocyte transfusions, prompt clinical reconstitution; episodes of fever after each application of cyclophosphamide

No., number; MDS, myelodysplastic syndrome; (s)AML, (secondary) acute myeloid leukemia; MB, medulloblastoma; NHL, non-Hodgkin lymphoma; BFM, Berlin-Frankfurt-Münster study group; SKK, infants and young children (German: Säuglinge und Kleinkinder); HD, high-dose; MTX, methotrexate; VCR, vincristine; i.vent., intraventricular; NFDA, no further details available; RAEB, refractory anemia with excess blasts; RT, radiotherapy; CSI, craniospinal irradiation; PF, posterior fossa; Met, metastases; w/o, without; PD, progressive disease; i.th., intrathecal; GOT, glutamic oxaloacetic transaminase; ULN, upper limit of normal; GPT, glutamic pyruvic transaminase; GGT, gamma-glutamyl transferase; SIRS, systemic inflammatory response syndrome.

No Adjuvant Therapy

Two patients died before surgery. In 3 patients, physicians and/or parents made an active decision against adjuvant therapy: One patient did not receive adjuvant therapy in the presence of 3 synchronous malignant processes; in 1 patient, parents refused adjuvant therapy and in the third, adjuvant therapy was omitted due to extensive MB dissemination, presence of MDS and previous antineoplastic treatment.

Chemotherapy

Chemotherapy approaches consisted primarily of intravenous cyclophosphamide (CPM), vincristine (VCR), etoposide, carboplatin, and methotrexate (MTX), with or without intraventricular MTX, as used in SKK chemotherapy (SKK chemotherapy previously described in³⁶). One patient was treated with CPM, VCR, etoposide and cisplatin, and myeloablative chemotherapy with thiotepa and carboplatin according to CCG99703.^35,37

Radiotherapy

RT was focal to the tumor bed (boost dose 54 Gy) in 2 patients. Two patients received high-risk CSI (35 and 35.2 Gy, respectively), of which one patient received additional stereotactic RT to 2 metastases at relapse. For 2 patients, no further details on RT were reported.

Toxicity

Information on toxicity was available in 16/17 patients who received adjuvant therapy (Figure 2). 6/16 (37.5%) had previously received chemotherapy or had MDS/AML at MB therapy initiation. No patient had received RT prior to MB diagnosis. Hematological toxicity and infectious complications were very frequent and especially prominent after the use of alkylating agents and anthracyclines, occurring in 10/11 (90.9%) patients, of which 3 had received chemotherapy previously (Table 2, Supplementary Table S6).

Figure 2.

Visualization of the frequency of treatment-associated toxicities in children with MB and FA. Treatment-associated toxicities are depicted according to the type of treatment administered. Type of toxicities are described on the x-axis while the fraction of patients experiencing those toxicities upon administration of the respective type of treatment is given on the y-axis. Organ-specific damages were: mucositis/dermatitis (n = 3 after i.v. chemotherapy with alkylating agents/anthracyclines, n = 2 after i.v. chemotherapy w/o alkylating agents/anthracyclines, n = 1 after radiotherapy), cholangitis (n = 1 after i.v. chemotherapy with alkylating agents/anthracyclines), hepatobiliary toxicity (n = 1 after i.v. chemotherapy w/o alkylating agents/anthracyclines), impairment of renal function (n = 1 after i.v. chemotherapy with alkylating agents/anthracyclines), and leukoencephalopathy (n = 1 after i.v. chemotherapy w/o alkylating agents/anthracyclines).

Chemotherapy cycles with HD-MTX and VCR were better tolerated, even in full doses. Toxicity included mucositis and usually milder hematological toxicity. However, severe myelosuppression (case 15) also occurred after MTX administration. Intraventricular MTX was given in all unpublished and in 3/16 published cases. It was well-tolerated by 7 patients and had to be discontinued in 2 patients due to MTX-related toxicity (cases 15 and 19; Table 2, Supplementary Table S6).

After RT, no relevant toxicities were reported in 4/6 patients, who received either 35/35.2 Gy CSI with posterior fossa boost up to 55 Gy (2/4) or focal RT with tumor boost of 54 Gy (2/4). One of these patients (case 2) was diagnosed with MDS prior to MB diagnosis and no patient had been treated with chemotherapy before. RT had to be discontinued due to pancytopenia in case 14. This patient had been treated for Wilms tumor and AML prior to MB diagnosis and he died shortly after RT discontinuation of sepsis and PD. Another patient survived several years after RT; however, no detailed information on tolerance, irradiation field, or dose of RT was reported.

Three patients died from treatment-related hematological toxicities upon treatment with alkylating agents (cases 11 and 18) and chemotherapy with consecutive RT (case 14). Two of these patients received chemotherapy for other malignancies (including AML in one patient) prior to MB diagnosis (cases 14 and 18).

Other Malignancies

Fourteen patients (63.6%, n = 14/22) developed a total of 20 non-MB malignancies, of which 10 occurred before, 5 simultaneously with and 5after MB diagnosis. Wilms tumor was the most frequent malignancy, occurring in 8/14 patients. Malignancies prior to MB were treated with different chemotherapy regimens in varying intensity in 6/8 patients. Five patients had MDS or AML prior to or at the time of MB diagnosis. Of those, 2 also received antineoplastic therapy prior to MB diagnosis. Two patients developed early secondary malignancies (sAML in both cases) during or shortly after initial MB therapy (Supplementary Table S7). 10/14 patients developing other malignancies were not previously treated with RT.

Survival of FA MB Patients and Reason of Death

Outcome was reported for 21/22 patients (for details see Supplementary Table S8). All but one patient died. The reason for death was primarily MB (15/20). One patient died from secondary malignancy and 3 from complications of therapy, of which one patient also had MB progression. In one patient, it remained unclear whether she died from MB, MDS, or hepatoblastoma.

Information on the duration of survival was available for 19 patients: 1y-progression-free-survival (PFS) was 26.3 ± 10.1% and estimated 1 y-overall-survival (OS) was 42.1 ± 11.3% (Figure 3A). Median PFS was 0.5 years (0.0–6.4 years). Median OS was 1 year (0.0–6.4 years). Adjuvant therapy led to enhanced OS and PFS (1y-OS/1y-PFS 0%/0% without adjuvant therapy vs. 53.3% ± 12.9%/33.3% ± 12.2% with adjuvant therapy, P = .006/P = .086, Figure 3B), with no difference between the chemotherapy only and RT ± chemotherapy group (P = .27 for OS/P = .657 for PFS, Figure 3C). There was no survival difference between the unpublished cases and the literature cohort (Supplementary Figure S3).

Figure 3.

Survival of the whole cohort (A, n = 19) and according to treatment (B and C, n = 17) depicted by Kaplan–Meier plot.

Family History of Cancer

Information about family history of cancer was available in 18/22 (81.8%) patients. Cancer in at least one of the patient’s relatives were reported in 16/18 (88.8%) cases (Supplementary Table S3).

Discussion

To the best of our knowledge, we provide the first structured overview of treatment strategies and associated toxicities in FA MB patients. While the association of FA-D1 and FA-N with MB is described in the literature,^{11,12,28,30,32,38} structured clinical data on treatment strategies, toxicities, and outcomes of this patient cohort are lacking.

Our study supports the observation that the increased risk of MB development is restricted to FA patients with biallelic BRCA2 or PALB2 germline mutations (complementation groups FA-D1 and FA-N).¹⁰ Both BRCA2 and PALB2 are known to predispose to MB development if altered in the germline, even in a heterozygous state.³⁹ We show that FA MB patients develop SHH-activated MB, without a clear association with a specific SHH subgroup. Allocation to a specific subgroup was not successful in many patients, an observation that has been described for patients with other tumor predisposition syndromes also using a random forest algorithm.⁴⁰ Future studies must confirm if MBs in FA patients are exclusive of SHH type.

Differing from otherwise healthy children,¹⁵ the favorable prognosis of SHH-activated, TP53-wild type early childhood MB does not hold true for FA patients. Because treatment intensity is related to better outcomes,⁴¹ this might be related to the inability to tolerate high-intensity treatment in FA. Our data show that outcome of FA MB patients (2-year OS 10.5%) is inferior to other high-risk SHH-MB constellations such as SHH-MB TP53-mut (5-year OS 37.5%–41%^16,42) or Li-Fraumeni syndrome (5y-OS 22.6%).⁴³ Outcome in the presented cohort is not further influenced by the presence of other known high-risk criteria such as TP53 mutation, MYC/MYCN amplification, or dissemination. Importantly, MB is not the only life-threatening condition in our cohort, and other previous and subsequent malignancies may further reduce life expectancy. Incorporation of bone marrow biopsy as a staging procedure and screening for other solid malignancies in FA MB patients is strongly recommended. Despite the high frequency of other malignancies and treatment-associated complications, FA MB patients primarily succumbed to their MB. Other authors suggested intensive therapy concepts for FA-D1 MB patients with toxicity-adjusted dose reductions.³² While we agree that MB patients generally need intensive multimodal therapies, a curative approach does not seem feasible in FA MB patients considering our data. Therefore, the intensity of treatment schedules in FA patients with MB should be carefully weighed against the respective treatment-associated toxicities, also considering the treatment of previous malignancies and possible reduction of bone marrow reserve. This is especially relevant for children with MDS. In the following, we will discuss the pros and cons of different therapeutic approaches for FA MB patients.

Chemotherapy

We observed pronounced toxicity after the application of ICL-forming alkylating agents⁴⁴ and DNA-damaging anthracyclines, while systemic MTX-containing chemotherapy was better tolerated supporting the observations made by Ruud et al.¹⁸ Therefore, restriction of conventional chemotherapy to the use of non-alkylating agents might be beneficial.

In non-FA SHH-MB, treatment intensification by either adding intraventricular MTX or high-dose chemotherapy/autologous stem cell transplantation (HDCT/ASCT), increases cure rates without the use of CSI.^15,45

Use of intraventricular MTX was well-tolerated by most children, but caused severe neurological toxicity in 2 patients. The outcome was poor even in patients treated with intraventricular MTX and from our data, we cannot conclude if intraventricular MTX was able to prolong survival. We suggest that intraventricular MTX be administered under close neurological monitoring and after careful consideration. However, the survival benefit in non-FA MB patients^15,46,47 might support its use in selected patients.

In the only FA MB patient in our series receiving HDCT/ASCT,³⁵ treatment had to be suspended after the first cycle due to toxicity. In FA MB patients undergoing allogenic SCT for bone marrow failure, fludarabine and CPM are commonly used for conditioning.^48–50 CPM total dose used in those regimens is significantly lower compared to the dose used for HDCT in MB patients,¹⁶ as former studies have shown that conventional CPM conditioning leads to intolerable toxicities in FA patients.⁵¹ Therefore, we currently do not see a role for HDCT/ASCT in FA MB patients.

Radiotherapy

Since RT induces crosslinks in DNA,⁸ FA patients are ought to have a pronounced radiosensitivity. We are aware of a small number of FA MB patients treated with RT. Except for one heavily pretreated patient, no severe toxicities after RT including CSI were observed. Therefore, we propose that it should not be excluded as a therapy strategy a priori. Different from established standards for MB therapy, focal RT was applied to 2 FA patients to prolong local tumor control while minimizing the anticipated side effects of irradiation of large target volumes in palliative intent. Although focal RT is inferior to CSI due to early MB dissemination,⁵² we see a rationale to consider focal RT in FA patients in a palliative therapy concept. Even though CSI was well-tolerated by 2 patients in our cohort, experience with delivering RT to large target volumes in FA MB patients is very limited and mechanistically can be expected to introduce significant side effects. Therefore, we suggest reserving this modality to very selected cases. Proton beam therapy (PBT) offers anti-cancer effects comparable to conventional RT while reducing exposure to irradiation outside of target areas.⁵³ Where available, we suggest to apply PBT in FA MB patients to avoid irradiation of healthy tissue. However, even with PBT, the limitations to apply CSI mentioned above remain.

Targeted Therapy

The smoothened (SMO) inhibitor vismodegib has shown antitumor activity in a subset of pediatric patients with SHH-MB.^54,55 Mechanistically expected and in line with clinical observations, response is restricted to tumors harboring mutations upstream of SMO.^54,56 Application of vismodegib to FA MB patients has not been reported yet and the use of SMO inhibitors in young children has been discouraged because of irreversible growth plate fusions.⁵⁷ Our data indicate heterogeneity of SHH-MB in FA patients, without evidence of recurrent upstream SHH pathway mutations. Mechanistically, use of vismodegib might be considered in cases with SMO or PTCH1 somatic mutations. However, we do not find the evidence sufficient to recommend targeted therapy in FA MB patients.

In summary, administration of adjuvant radio-/chemotherapy in a toxicity-controlled manner is feasible in FA patients with MB. Adjuvant therapy might prolong survival, while therapy primarily will have to aim at palliation, not at cure of the disease, until further progress in available treatment options is made. Early integration of a palliative care team is recommendable. To ensure proper diagnosis, we suggest to screen children presented with MB and physical features including microcephaly and short stature for BRCA2 and PALB2 mutations. Regular cranial MRI screening for children with confirmed biallelic BRCA2 or PALPB2 mutations is recommendable and might allow early diagnosis and therapy, as suggested previously.⁵⁸

Our study has limitations: First, the cohort assembled through literature research and unpublished cases is small. It is likely that more FA patients with MB were treated but have not been reported or diagnosed as MB without FA. It is possible that the severity of toxicities may have prompted publication, leading to a selection bias with overestimation of treatment-associated toxicities, underestimation of survival duration, and underreporting of promising therapy approaches. Second, most published case reports focus on genetic or epidemiologic findings and frequently miss detailed clinical information, including quality-of-life data. Data availability of the unpublished cases is also restricted due to the retrospective nature of this study. This permits only limited statistical analyses and descriptive presentation of findings, including toxicities.

Despite these limitations, the overview on treatment and treatment-associated risks in MB patients with FA subtypes D1 and N might serve as a practical guide for treating physicians. Based on these observations, restriction of chemotherapy to non-alkylating agents, i.e. HD-MTX/VCR, maybe including intraventricular MTX followed by focal RT (54 Gy), preferably with protons, in non-metastatic disease, might be an option. We suggest to reserve CSI in FA patients with MB to rare exceptional circumstances, given the significant risk of toxicity. Balancing effective tumor control and reduction of treatment-associated toxicity is key in the care of these patients with MB and a good quality of life should be the overall objective. Even though the prognosis of patients with MB and FA is dismal, adjuvant therapy can prolong survival. Larger, standardized, and centrally coordinated data collection of MB patients with FA is needed to gain a better understanding of how to treat these rare patients.

Supplementary material

Supplementary material is available online at Neuro-Oncology (https://academic.oup.com/neuro-oncology).

Funding

German Childhood Cancer Foundation for HIT-MED trial center (2001.03, 2005.13, 2009.18, 2010.08 and 2016.10 to SR, 2017.10, 2019.07 and 2021.24 to SR and MM); German Childhood Cancer Foundation for reference neuropathological assessments in the HIT network to TP (2006.03, 2009.19, 2011.01 and 2014.17); German Childhood Cancer Foundation for quality assurance of radiotherapy in pediatric brain tumors in Germany to RS; Mildred-Scheel-Nachwuchszentrum Grant by German Cancer Aid to Martin Mynarek; Clinician Scientist Program CS4RARE funded by the Alliance4Rare, associated with the Berlin Institute of Health Charité to Felix Boschann; Fördergemeinschaft Kinderkrebs-Zentrum Hamburg to Ulrich Schüller.

Conflict of interest statement

PHD is the international coordinating investigator of the Sprinkle study and advisory board member for Alexion. URK is part of a data safety monitoring board or advisory board for Astra Zeneca and Merck and received support for attending meetings and/or travel from Astra Zeneca. RS receives funds from the German Childhood Cancer Foundation for quality assurance of radiotherapy in pediatric brain tumors in Germany and is part of a data safety monitoring board or advisory board for the Oscar 1 Trial (Osteosarcoma) of University Hospital Heidelberg, Germany, and for proton therapy trials of University Hospital Dresden, Germany. TP receives funds for the Brain Tumor Reference Center for reference neuropathological assessments in the HIT network by the German Childhood Cancer Foundation. SR received fees for advisory board roles from Bayer, Novartis, BMS, and Roche; for DMSC from Celegne; for assessment from the German Childhood Cancer Foundation; and from MyChildsCancer for consulting. The other authors have no competing interests to declare that are relevant to the content of this article.

Authorship statement

Design of research: M.S., M.M., and S.R.. Preparation and provision of the data: M.S., D.O.S., P.H.D., A.S., N.G., U.K., C.R., M.L., U.R.K., R.S., T.O., F.B., U.S., L.A., T.G., T.P., and M.M.. Analysis of the data: M.S. and L.A.. Interpretation of the data: M.S., D.O.S., M.M., S.R., and F.B.. Manuscript writing: M.S., D.O.S., M.M., S.R., L.A.. Final approval of the manuscript: all authors.

Data availability

Raw data is available upon reasonable request from the corresponding author, provided ethical and legal requirements are met.