Feasibility and efficacy of MEMMAT-like regimen in heavily pretreated adult patients with recurrent malignant embryonal brain tumors: A series of 6 cases

Abstract

Background

The prognosis of adult patients with recurrent malignant embryonal brain tumors remains poor due to the lack of effective and validated treatment. A metronomic and antiangiogenic chemotherapy regimen called MEMMAT was developed in children, with promising results. Additional data on feasibility, tolerance, and efficacy in adults are necessary.

Methods

This retrospective observational case series included adult patients with relapsing medulloblastoma (MB) or pineoblastoma (PB) treated with MEMMAT-adapted protocol. Treatment consisted of daily oral thalidomide, fenofibrate, and celecoxib, and alternating 21-day cycles of metronomic etoposide and cyclophosphamide associated with bevacizumab and personalized intraventricular chemotherapy.

Results

Four SHH-activated MB and 2 PB were included. Median duration of treatment was 12 months (range 3–21). Significant partial response occurred in 2/4 MB patients and 2/2 PB. The best responses were observed on leptomeningeal lesions. Main grade 3–4 toxicity was neutropenia in all patients (no febrile neutropenia) and lymphopenia in all but one (2 opportunistic infections). Dose adjustments in chemotherapy and thalidomide for hematotoxicity were necessary in all patients within the first 3 months. Cumulative neurotoxicity from thalidomide affected the 4 patients with prolonged treatment (1 grade 3, 3 grade 2). Rechallenge was tried in PB patients and successful (duration: 9 months).

Conclusions

MEMMAT is feasible in adult patients and can lead to a significant and sustained response in recurrent malignant embryonal brain tumors. Hematotoxicity is progressive and manageable. The withdrawal of thalidomide and starting dose adjustments on chemotherapy might be discussed.

Key Points.

• First report of MEMMAT feasibility in adult

• Signs of efficacy in adult medulloblastoma and pineoblastoma

Importance of the Study.

This is the first report of MEMMAT, a metronomic chemotherapy-based regimen, in adult patients with embryonal brain tumors. As in pediatric studies, stability or responses were assessed in some of our patients whereas they had progressive disease under conventional chemotherapy. Meningeal responses were particularly noted. Main toxicity was progressive neutropenia and remained manageable. No febrile neutropenia occurred. MEMMAT could be a valuable option in recurrent embryonal brain tumors in adult patients.

Medulloblastoma (MB) is an embryonal neoplasm arising in the cerebellum or dorsal brainstem. Its incidence decreases with age from 6.6 cases per million in childhood to 0.5 in adulthood, representing less than 1% off all brain tumors^.1 The Sonic Hedgehog (SHH)-activated subgroup is predominant in adults.^2–5MYCN amplification is rare among adult SHH-MB, whereas telomerase reverse transcriptase (TERT) promoter mutations are common.⁶ SHH activation in adult MB is usually secondary to protein patched homolog 1 (PTCH1) or smoothened, frizzled class receptor (SMO) alterations, occurring in approximately 80% of cases.

MB relapses are generally incurable due to intrinsic pathological and molecular characteristics, incomplete resection, metastatic disease, and limited therapeutic options. The latter includes challenges in radiotherapy, particularly due to previous cranio-spinal irradiation (CSI), and chemotherapy in heavily pretreated patients.⁷ Similar challenges are observed in the rare cases of recurrent pineoblastoma (PB). Given the limited efficacy of cytotoxic agents or targeted therapies in these patients, alternative approaches such as metronomic chemotherapy (MC) have been explored. MC involves the frequent administration of anticancer agents at relatively low, minimally toxic doses with nonprolonged drug-free breaks. The first significant multidrug metronomic and antiangiogenic regimen for recurrent/refractory brain tumors was developed in children in 2005.⁸ This regimen combined alternating oral continuous low-dose etoposide and cyclophosphamide with continuous oral administration of thalidomide and celecoxib, establishing a backbone for further regimens. It was subsequently enhanced by adding oral fenofibrate, a peroxisome proliferator-activated receptor alpha (PPARα) agonist with antiangiogenic and neuroprotective properties.⁹ Ultimately, this 5-drug regimen was complemented with intravenous bevacizumab and intraventricular chemotherapy, resulting in long-term survival for a subset of patients^.10,11

Additional data on tolerance and efficacy in adult patients are required. Here, we report a retrospective case series of adult patients with recurrent SHH-MB or PB treated with a MEMMAT-like strategy, discussing the feasibility, efficacy, and side effects of this innovative therapeutic approach in these patients.

Methods

Data Collection

This retrospective observational case series includes adult patients with recurrent embryonal tumors treated with MC according to MEMMAT protocol or inspired by it («MEMMAT-like»), as described below. Data were collected by local investigators, including date of diagnosis, age, gender, disease staging, pathology based on WHO 2021 classification, available molecular data, previous treatments, date of relapses, adverse effects associated with treatment, and treatments administered post MEMMAT. Imaging and clinical outcomes were collected. Tumor tissues were reviewed by local pathologists.

All patients underwent physical examination, performance status evaluation, repeated routine laboratory tests, and magnetic resonance imaging (MRI) scans at least every 3 months. Patients receiving intraventricular therapy had CSF samples before injections for cytological analysis at least once a month. Toxicity was evaluated according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0 criteria. Radiological response was assessed locally according to Response Assessment in Neuro-Oncology criteria. Complete response (CR) was defined as complete radiological response and negative CSF cytological analysis.

MEMMAT-Like Treatment

The treatment regimen consisted of an oral 5-drug protocol, including daily thalidomide (3 mg/kg), daily fenofibrate (90 mg/m²), twice daily celecoxib (400 mg × 2), and alternating continuous 21-day cycles of low-dose etoposide (35–50 mg/m²) and cyclophosphamide (2.5 mg/kg, maximum 100 mg). Doses were rounded down if necessary. Bevacizumab (10 mg/kg) was administered intravenously every 2 weeks. Intraventricular therapy consisted of alternating etoposide (0.5 mg/day for 5 consecutive days) in week 1 with bi-weekly infusions of aqueous cytarabine (30 mg) in week 3 and 4 of 28-day cycles, administered via an Ommaya reservoir. Adjustments or omissions of intraventricular therapy were performed as described below.

Dose reductions were implemented to avoid treatment interruption due to toxicity. In case of treatment-induced neutropenia, most of these adjustments can be summarized in Table 1. All patients received prophylaxis for opportunistic infections with trimethoprime-sulfamethoxazole, and some patients received valaciclovir in case of chronic grade 3 lymphopenia. Concomitant additional radiotherapy to focal residues was allowed.

Table 1.

Treatment Dose Adjustments of Oral Therapies in Case of Neutropenia

	Grade 3 Neutropenia	Grade 4 Neutropenia
Target dose Thalidomide 3 mg/kg Etoposide 35–50 mg/m² Cyclophosphamide 2.5mg/kg	No interruption and DA1	Interruption of oral treatment until grade 2 and DA1
Dose Adjustment (DA) 1 Thalidomide 2 mg/kg Etoposide 50 mg flat dose Cyclophosphamide 50 mg flat dose	No interruption and DA2	Interruption of oral treatment until grade 2 and DA2
Dose Adjustment (DA) 2 Stop Thalidomide Etoposide 50 mg flat dose Cyclophosphamide 50 mg flat dose	Interruption of oral treatment until grade 2 Consider definitive interruption if exposure < 15 days/months	Interruption of oral treatment

Patients were treated “off-label” under the responsibility of their physician, following patient information and the validation by the ANOCEF-AYA national tumor board (Association of the Neuro-oncologists of French Expression). The study was conducted in accordance with the Declaration of Helsinki and approved by the French “Commission Nationale de l’Informatique et des Libertés” based on the MR-004 declaration of conformity (R201-004-217).

Results

Six recurrent patients were included in the study: 4 with SHH-MB (patients 1–4) and 2 with PB (patients 5 and 6). All patients with SHH-MB were staged as high-risk at diagnosis, 3 presented with metastatic disease (patients 1–3), and 1 had >1.5 cm² residual tumor after surgery (patient 4). Moreover, patient 2 had a TP53-mutated MB. Initial management relied on a combination of chemotherapy and CSI. Notably, 3 patients received 2 courses of high-dose chemotherapy followed by hematopoietic stem cell rescue according to a pediatric protocol before CSI.¹² First-line treatments are summarized in Table 2. From MEMMAT start, mPFS was 12 months (range: 3–21) and mOS was 20 months (range: 4.5–46).

Table 2.

Tumor Characteristics and First-Line Treatments

Patients	Age at diagnosis/gender	Pathology	Molecular biology	Stage at diagnosis13	Initial management	Recurrence (months)	Stage at first recurrence13	OS from diagnosis (months)	OS from first recurrence (months)
1	44/M	SHH-MB	MYC – TP53wt pTERT mut PTCH1 mut	R0 M3	2 Etoposide-Carboplatin CSI 36/54 Gy 2 Etoposide-Carboplatin	34	M3	86	52
2	26/F	SHH-MB Anaplastic at relapsea	MYC – TP53 mutb pTERT mut PTCH1 mutc	R0 M3	2 Etoposide-Carboplatin 2 HD TTP CSI 36/54 Gy	26	M2	44	18
3	28/M	SHH-MB	MYC – TP53wt pTERT mut PTCH1 altd	R0 M3	2 Etoposide-Carboplatin -MTX CSI 36/54 Gy 12 TMZ	192	M2	202	10
4	33/F	SHH-MB	MYC – TP53wt pTERT mut PTCH1 mut	R + M0	2 Etoposide-Carboplatin CSI 23.4/54 Gy 2 Etoposide-Carboplatin	32	M2	75	44
5	19/F	PB	DICER 1 mute	R + M0	2 Etoposide-Carboplatin 2 HD TTP Focal RT 54 Gy 6 maintenance TMZ	32	M3	147	104
6	18/F	PB	PTCH1 mute	R + M0	3 Etoposide-Carboplatin 2 HD TTP Focal RT 54 Gy 6 maintenance TMZ	36	M3	101	65

^aDesmoplastic at diagnosis and anaplastic at relapse.

^b(c.1024C>T) TP53 mutation at diagnosis while (c.328C>T) TP53 mutation at relapse.

^c(c.3471_3478dup) PTCH1 alteration at diagnosis while (c.1792del) PTCH1 alteration at relapse.

^dPTCH1 inactivation by partial duplication and GLI2 amplification.

^eConstitutional mutation.

PB: pineoblastoma; wt: wild type; pTERT: TERT promotor; mut: mutated; alt: altered; HD TTP: high-dose thiotepa; MTX: methotrexate; TMZ: temozolomide; RT: radiotherapy; OS: overall survival; CSI: cranio-spinal irradiation.

Context of MEMMAT Initiation and Results ( Table 2 and Figure 1)

Figure 1.

Clinical course of the patients.

All patients experienced metastatic recurrences. Karnofsky Performance Status (KPS) was greater than 70% in every patient at MEMMAT initiation.

Patient 1 presented with diffuse leptomeningeal recurrence and received carboplatin and etoposide then focal radiotherapy. At second relapse, the patient received 9 cycles of temozolomide, then radiotherapy and bevacizumab maintenance. At the third relapse, the patient received vismodegib, which led to stable disease for almost 1 year but was discontinued due to tumor progression. At the fourth relapse, he experienced radiological tumor progression in the conus medullaris and cauda equina, leading to MEMMAT treatment without intraventricular therapies due to geographic remoteness. First evaluation at M3 showed a CR on the largest lesion (L1; Figure 2A and B) and stable disease on a second lesion (L3). Clinical stabilization was achieved. Bevacizumab was discontinued due to Covid-19 pandemic. Interestingly, at M6, the next evaluation under oral therapies only showed a partial response on the second lesion (−40% from M3 and baseline) and a persistent CR in L1. Reintroduction of bevacizumab did not bring clear benefit. The M12 evaluation showed disease progression on all lesions without clinical worsening. Treatment was continued until loss of clinical benefit.

Figure 2.

Tumoral response on target lesion in patients 1 and 5.

At first focal relapse in posterior fossa, patient 2 received 2 cycles of chemotherapy combining topotecan and temozolomide (TOTEM), followed by subtotal surgery, focal reirradiation, and adjuvant TOTEM chemotherapy. The best response was stable disease before surgery. Significantly newly appeared leptomeningeal tumor progression was documented after 3 adjuvant cycles. While treated with TOTEM chemotherapy, she experienced a multimetastatic intraventricular recurrence associated with anterograde amnesia. The MEMMAT regimen was performed with adapted intraventricular therapy, avoiding cytarabine injections to reduce repeated transports. Clinical and radiological infiltrative progression in limbic system led to early discontinuation. Vismodegib was initiated and resulted in a radiological stability at M2, associated with a clinical worsening. A new MRI at M4 confirmed tumor progression.

Patient 3 experienced a very late documented distant relapse 16 years after initial diagnosis, presenting with an infiltrating lesion in the left cerebellopontine angle (CPA). This patient received chemotherapy combining carboplatin and etoposide. The best response was stable disease after 2 cycles, followed by local progression and leptomeningeal dissemination after 4 cycles. He received a complete MEMMAT regimen. At M2, stable disease (−17%) was observed on the posterior fossa lesion, associated with CR on the leptomeningeal spinal lesions and cerebrospinal fluid tumor cells disappearance. At M4, only the parenchymal target lesion was progressive whereas cytological analysis remained negative. MEMMAT was switched for temozolomide and focal reirradiation. The patient died from rapid progression of the CPA lesion (+45% in 1 month after MEMMAT cessation).

At first focal relapse, patient 4 received 4 cycles of chemotherapy combining carboplatin and etoposide, followed by focal reirradiation and 6 cycles of TOTEM. CR was achieved before radiotherapy. Leptomeningeal focal relapse was diagnosed 8 months after the end of treatment. At the second focal relapse, occurring 32 months after initial diagnosis, he received MEMMAT regimen with adapted intraventricular etoposide scheme (1 mg/day for 3 days of each 28-day cycle). Several temporary discontinuations and doses adjustments of oral therapies were necessary to manage side effects (Table 3). At M6, patient remained in stable disease and was concurrently focally irradiated (35 Gy in 5 fractions), achieving a CR. At M9, due to numerous side effects, asthenia, and persistent CR, treatment was discontinued. Three months later, MRI showed a major parenchymal relapse at the former nodule location, along with mild cognitive decline. Treatment with sonidegib was initiated, with a significant but short response (PR at M3, PD at M5) and then axitinib (PD).

Table 3.

Tolerance and Efficacy of MEMMAT Treatment

Patients	Immediate previous treatment	Time to prior relapse (months)	Age at MEMMAT start (y)	Line	Adverse events (grade)	First dose adjustment (weeks from initiation)	Best response on target lesion/Time (months)	PFS (months)	OS from MEMMAT start (months)/status
1	Vismodegib	11.5	49	5th	Neutropenia (3) Lymphopenia (3) Neurotoxicity (2)	Thalidomide (9) Etoposide (9) Cyclophosphamide (9)	PR (−90%)/M9	12	16/DOD
2	TOTEM	4.5	29	3rd	Neutropenia (3) Lymphopenia (3)	Thalidomide (6) Etoposide (6) Cyclophosphamide (6)	PD (+381%)/M3	3	8/DOD
3	Etoposide-Carboplatin	3	45	3rd	Neutropenia (3) Epigastralgia (3) Lymphopenia (3)	Thalidomide (2) Etoposide (8) Celecoxib (8)	SD (−17%)/M2 Complete cytological response on leptomeningeal disease	3	4.5/DOD
4	TMZ	4.6	37	3rd	Neutropenia (3) Lymphopenia (4) Esophageal candidosis (3) Arachnoiditis (2) Neurotoxicity (2)	Thalidomide (12) Etoposide (12)	CR (in combination with RT)	12	24/DOD
5	TEMIRI	60	28	3rd	Neutropenia (4) Neurotoxicity (3) Bacterial meningitis (3)	Thalidomide (14) Etoposide (14) Cyclophosphamide (15)	CR	15 (1st c.) 9 (2nd c.)	46+/AWD
6	CSI	22	23	3rd	Neutropenia (3) Lymphopenia (4) Pneumocystis pneumonia (3) Neurotoxicity (2)	Thalidomide (4) Etoposide (12) Cyclophosphamide (8)	PR (−58%)/M6	21 (1st c.) 9 months (2nd c.)	43/DOD

TOTEM: topotecan-temozolomide; TMZ: temozolomide; TEMIRI: temozolomide-irinotecan; CSI: cranio-spinal irradiation; PR: partial response; PD: progressive disease; SD: stable disease; CR: complete response; DOD: died of disease; AWD: alive with disease; c.: course.

Patients 5 and 6 were initially treated for a partially resected focal PB. In both cases, multifocal leptomeningeal relapse occurred during follow-up.

Patient 5 experienced a first spinal metastatic relapse 32 months after initial diagnosis, treated with chemotherapy combining temozolomide and irinotecan and CSI. She experienced a biopsy-confirmed second cerebral relapse located on the infundibulum 5 years after the end of radiotherapy. A complete MEMMAT regimen was started, leading to a CR at M8, with an early complete response on infundibulum lesion (Figure 2C and D). Unfortunately, this patient experienced bacterial meningitis at M11, requiring the temporary cessation of treatment. Treatment was restarted 1 month later without intraventricular infusions for 2 additional months. Ten months later, MRI showed a cerebral metastatic relapse, leading to the reintroduction of MEMMAT therapy for 9 additional months. Interestingly, new partial response was observed after 3 months of treatment. Finally, the treatment was definitively discontinued due to grade 3 peripheral neuropathy despite the withdrawal of thalidomide and oral chemotherapy. Treatment with rapamycin and irinotecan was then started and still ongoing (SD).

Patient 6 experienced a spinal metastatic relapse 3 years after initial diagnosis. This patient received TOTEM chemotherapy, surgery, and CSI. She experienced a multifocal leptomeningeal relapse in the posterior fossa 22 months later. A MEMMAT treatment without intrathecal therapies was initiated. The best response was a partial response at M6. After 21 months, treatment was discontinued due to hematotoxicity. Nine months later, symptomatic leptomeningeal relapse occurred. Complete MEMMAT was resumed. The best response was stability. After 10 months of MEMMAT rechallenge, 1 lesion on the right pontocerebellar angle was progressive. The lesion was resected and MEMMAT was switched to vinblastine. The patient experienced rapid symptomatic worsening due to multifocal leptomeningeal progression and died 3 months later.

Toxicity

The most frequent toxicity was grade 3/4 neutropenia in all patients, leading to temporary discontinuations and dose adjustments. It appeared progressively and was manageable. None of the patients experienced febrile neutropenia. The second most frequent toxicity was grade 3–4 lymphopenia (n = 5/6, 83%). Opportunistic infections occurred in 2 patients (33.3%); Esophageal candidiasis (patient 4) and Pneumocystis Jiroveci pneumonia (patient 6, due to poor compliance to sulfamethoxazole-trimethoprime) were both attributed to iatrogenic immunosuppression. Additionally, patient 6 suffered from recurrent sinus infections. No febrile neutropenia was observed. Hematotoxicity was attributed to oral chemotherapy and thalidomide. Initial dose adjustments occurred early, within 1 to 3 months after initiation of treatment. Progressive peripheral neurotoxicity was also common (n = 4, 66.6%; grade 3 observed in 1 patient and grade 2 in 3 patients), leading to the permanent discontinuation of thalidomide. Patient 4 experienced recurrent grade 2/3 arachnoiditis after intraventricular cytarabine administration. Patient 5 had to undergo Ommaya reservoir removal due to grade 3 iatrogenic bacterial meningitis. Complete treatment cessation due to cumulative toxicity occurred in 3 patients. The summarized toxicity data are presented in Table 3.

Discussion

In this study, we demonstrate the feasibility of MEMMAT regimen in 6 heavily pretreated adult patients with recurrent SHH-MB or PB. All but 1 patient received intraventricular therapies, some with adjustments. Toxicity was progressive, primarily hematological, and remained manageable. One patient developed grade 3 iatrogenic bacterial meningitis requiring Ommaya reservoir removal. Dose adaptations were necessary for all patients, especially for oral chemotherapies. No side effects associated with bevacizumab were observed. In patients with the longest exposure, the main side effects included cumulative neurotoxicity from thalidomide and lymphodepletion, leading to opportunistic infections such as esophageal candidiasis and Pneumocystis Jiroveci pneumonia, as previously reported.¹⁴

The initial experience with MEMMAT strategy in children with recurrent MB was detailed in a retrospective study reporting a median overall survival (OS) of 29.5 months and a subset of long-term survivors. Progression-free survival (PFS) and OS at 3 years were 42.0% and 48.3%, respectively.¹⁴ These promising results compare favorably with other therapeutic strategies. In a separate prospective phase II study, median PFS and OS were 8.5 and 25.5 months, respectively.¹¹ The OS in our study was lower, even when a sustained response was achieved. Our results are more consistent with those from a retrospective pediatric “real life” experience with the MEMMAT strategy, which reported 10% survival at 24 months in MB patients.¹⁵ Several hypotheses may explain these findings.

First, it is noteworthy that the MB populations are different. Slavc’s population was strictly pediatric (median age of ~7 years at diagnosis, range 0.4–12),¹⁴Peyrl’s population was slightly older (median age of 10 years at treatment start, range 4–17),¹¹ while Winnicki’s population was more representative of adolescent cases (median age of 14 years at treatment start, range 1–19).¹⁵ Age in MB is crucial as subtype distribution varies with age. Indeed, SHH-MB represents about 30% of pediatric MB, frequently occurring in both infants and adolescents/adults (>16 years), but is less in children.¹⁶ SHH-MB is predominant in adults^2–4,17 and may have poorer prognosis than infant SHH-MB.¹⁸ Notably, Slavc’s and Peyrl’s studies included only 1/29 (0.4 years at diagnosis) and 4/40 patients with SHH-MB (age not precised), respectively.^11,14 In our study, all MB patients had SHH-MB. SHH-MB maintains a functional blood–brain barrier, which may contribute to lower chemosensitivity and a distinct microenvironment.¹⁹

Second, MEMMAT occurred later in the overall treatment strategy in our study (second to fourth recurrence) compared to other studies (first to second recurrence), despite good KPS at initiation.^14,15Winnicki noted that MEMMAT appeared more effective at first recurrence for PFS, suggesting that MEMMAT may be less effective in rapidly progressive disease.¹⁵

Strikingly, we observed both inter-individual and intra-individual variations in tumor response among our MB patients despite disease homogeneity (metastatic SHH-MB with PTCH1 loss). This suggests that the indications for MEMMAT might need refinement or the regimen may require adaptation. Two patients experienced early failure, whereas others had extended responses despite adjustments. Efficacy within the same patient could vary across tumor sites. These observations highlight the need for a more qualitative approach to analyze dissociate responses, early failure, tumor escape mechanisms, and treatment variations.

A high intra-individual heterogeneity of responses was observed across tumor sites in our series. This heterogeneity might be related to different hypothesis:

• (i) Tumor heterogeneity. Molecular differences were evident between diagnosis and relapse in 2 patients. Patient 2 presented with desmoplastic SHH-MB at diagnosis, which transformed into anaplastic at recurrence, exhibiting different molecular alterations for PTCH1 and TP53. In patient 3, a double alteration of SHH pathway was evidenced at recurrence.

• (ii) Differences in the previous therapeutic sequences. Patient 2 received radiotherapy on the cerebellar lesion before MEMMAT initiation, which may have contributed to its slower progression compared to limbic lesions.

• (iii) Differences in microenvironment or treatment diffusion, especially in patient 3, who had a dissociate response with a complete leptomeningeal response while showing progressive on the previous treatment line.

Early failure occurred in patients 2 and 3. Although not highly sensitive to carboplatin, patient 3 experienced brief stabilization under MEMMAT before progression in the cerebellar lesion while achieving CR in leptomeningeal disease. Patient 2, with an aggressive MB phenotype (anaplastic, TP53 mutated), showed quick progression under TOTEM regimen; with early failure of MEMMAT occurring within the first months. Both patients had early failure of the previous chemotherapy line, suggesting that MEMMAT success may depend on previous chemosensitivity and/or tumor kinetics. This indicates that MEMMAT might be more suitable as a maintenance treatment, as previously proposed.¹⁵ Leptomeningeal control under MEMMAT was noteworthy in our series, as the 3 patients with isolated leptomeningeal disease demonstrated significant responses despite dose adjustments or intraventricular therapy withdrawal.

Tumor escape mechanisms from MEMMAT are questioning. Except for patient 1, who did not receive intraventricular therapies and experienced leptomeningeal relapse, 3 other patients had infiltrative relapse in the brain parenchyma. In patients 2 and 4, leptomeningeal seeding appeared to be the origin of subsequent parenchymal infiltration under treatment. Patient 4 was particularly illustrative: whereas its unique ventricular nodule achieved CR after MEMMAT and radiotherapy, relapse occurred in the contiguous parenchyma without intraventricular recurrence. Combined with noticeable leptomeningeal efficacy described above, it may suggest that intraventricular chemotherapies might facilitate a phenotypic switch from leptomeningeal dissemination to parenchymal infiltration in adult SHH-MB.

Considering the biology of this tumor, it can be hypothesized that parenchymal infiltration may facilitate the recovery of a more tumor-protective microenvironment, characterized by a more effective blood–brain barrier and reduced exposure to intraventricular therapy, which has limited penetration.²⁰

In 2 patients (3 and 6), cessation of MEMMAT led to rapid clinical deterioration. This suggests that discontinuing MEMMAT in patients with slowly or focally progressive high-burden disease patients should be approached with caution, depending on available therapeutic alternatives.

It is noteworthy that prolonged and significant responses were obtained in the 2 PB patients, and rechallenge after treatment interruption was effective in both cases. These responses were obtained despite various treatment adjustments. Interestingly, Winnicki did not find statistical differences between patients receiving complete regimen or not, raising questions about the optimal combination design.¹⁵ Determining the essential components of the regimen is crucial, as some drugs may be redundant or induce useless toxicities or constraints, suggesting that MEMMAT regimen could be simplified. One of the best meningeal responses was observed in patient 1, who did not receive intraventricular treatment and had a temporary pause in bevacizumab. The initial 4-drug oral metronomic regimen, combining cyclophosphamide, etoposide, celecoxib, and thalidomide, was designed as an antiangiogenic therapy.⁸ Fenofibrate was later added, followed by bevacizumab and intraventricular chemotherapy. These additions did not alter the former regimen, especially with the introduction of bevacizumab, which also has antiangiogenic and anti-inflammatory properties. For instance, thalidomide contributes to hematotoxicity in combination with chemotherapy and can cause peripheral neurotoxicity. Its antiangiogenic value is debatable when combined with bevacizumab. Similarly, thalidomide’s immunomodulatory properties overlap with those of metronomic cyclophosphamide.²¹ Hematotoxicity led to early dose adaptations or treatment discontinuations, casting doubt on the clinical benefit of thalidomide.

Another point to consider is starting cyclophosphamide, which is often 50 mg/day in most metronomic adult regimen. MEMMAT usually initiates at 100 mg/day, with subsequent dose reductions or pauses as needed. Similarly, we recommend starting etoposide at the lowest dose (35 mg/m²) to mitigate potential toxicities.

Conclusion

MEMMAT is feasible in adult patients. Hematotoxicity is progressive and manageable. Compared to children, the elimination of thalidomide and dose adjustment on chemotherapy at start might be discussed to improve long-term tolerance. This regimen showed promising results, especially in leptomeningeal disease. The MEMMAT strategy deserves consideration as an option for patients with recurrent PB. These preliminary results should warrant further investigations in a dedicated prospective study for adult patients with recurrent embryonal tumors.

• Patients 1 to 4: SHH-activated medulloblastoma

• Patients 5 and 6: Pineoblastoma

• L1 lesion in patient 1 (MB-SHH). Sagittal spine MRI; gadolinium-enhanced T1

  • a. MEMMAT initiation

  • b. Month 3

• Infundibulum lesion in patient 5 (PB). Coronal brain MRI; gadolinium-enhanced T1

  • c. MEMMAT initiation

  • d. Month 3

Funding

None declared.

Conflict of interest statement

None declared.

Author Contributions

A.M. and P.L.: Conception, writing and revision. All the authors: Data collection.

Data Availability

The data will be available upon reasonable request to the authors.