A Phase Ib Case Series Using Bortezomib With Chemotherapy in Atypical Teratoid/Rhabdoid Tumors Authors: Wan‐Ling Ho, Ming‐Yun Hsieh, Yu‐Mei Liao, Shyh‐Shin Chiou, Her‐Shyong Shiah, Chang‐Hsu Chiang, Shiann‐Tarng Jou, Hsiu‐Ju Yen, Yi‐Yen Lee, Muh‐Lii Liang, Ting‐Chi Yeh, Yu‐Cheng Chou, Fang‐Liang Huang, Shu‐Huey Chen, Hsin‐Lun Lee, Chia‐Chun Kuo, Yi‐Shan Yang, Shu‐Mei Chen, Min‐Lan Tsai, Hsi Chang, Kevin Li‐Chun Hsieh, Chia‐Lang Fang, Chia‐Yau Chang, Jinn‐Li Wang, Chu‐Chin Chen, Joon‐Khim Loh, Pei‐Chin Lin, Tzu‐Chin Lin, Shih‐Chung Wang, Yi‐Fang Tu, Chao‐Neng Cheng, Jimmy Ming‐Jung Chuang, Huy Minh Tran, Shian‐Ying Sung, Irving Lai, Kuo‐Sheng Wu, James S. Miser, Yen‐Lin Liu, Tai‐Tong Wong Published: 2026 ABSTRACT ABSTRACTBackgroundAtypical teratoid/rhabdoid tumors (ATRTs) are aggressive pediatric brain tumors that are sensitive to proteasome inhibitors in preclinical models. We evaluated the feasibility of combining the proteasome inhibitor, bortezomib, with chemotherapy.MethodsSix patients aged 0.8–13 years with newly diagnosed ATRT were enrolled from January 2022 through December 2023. Bortezomib was given intravenously at the dose of 1.3 mg/m2/day on days 1, 4, 8, and 11 of a 21‐day chemotherapy cycle for 6 cycles. Radiation therapy was given before initiation of chemotherapy in 4 patients or at the end of chemotherapy in one patient. The primary endpoint was grade 3 or higher non‐hematologic toxicities.ResultsThirty cycles of bortezomib with chemotherapy were evaluated. No grade 3 or higher toxicities were considered bortezomib‐related. Four patients completed 6 cycles of bortezomib; 2 patients discontinued treatment due to progressive disease after 2 and 4 cycles, respectively. An additional 6 cycles of bortezomib were evaluated after high‐dose chemotherapy in three patients without grade 4 bortezomib‐related toxicity. Five patients had a near‐total resection; one patient had a subtotal resection. Four patients were evaluable for response with 2 partial responses, 1 stable disease, and 1 progressive disease. Two patients without measurable tumor after surgery were not evaluable for response. Two patients remain alive without disease at 37 and 40 months, and one patient died without disease at 19 months after diagnosis. Three patients died with progressive disease.ConclusionsThe toxicities were no more than expected with chemotherapy alone. Adding bortezomib to treatment of ATRT is feasible.Trial RegistrationNCT06853080 Introduction Atypical teratoid/rhabdoid tumors (ATRTs) are highly aggressive pediatric malignancies of the central nervous system (CNS) with a dismal outcome. They account for 1%–2% of all CNS tumors in children aged 0–14 years and mainly affect children below 3 years of age. In addition, ATRTs are the most common malignant CNS tumors in infants less than 1 year of age [1]. The incidence of CNS ATRT in Taiwan from 1993 to 2014 was 2.9 newly‐diagnosed patients per year [2]. There is no widely accepted standard of care for ATRT. Multimodal treatment strategies for ATRT include a combination of conventional chemotherapy, intrathecal chemotherapy, radiotherapy after tumor resection, and high dose chemotherapy with stem cell rescue [3, 4, 5]. The survival outcomes, even with aggressive treatment, remain poor with 2‐year survival rate of 32.6%–44.6% [6]. Therefore, there is a pressing need for new treatment including targeted therapies for ATRT. Bortezomib (Velcade) is a first‐generation proteasome inhibitor that reversibly inhibits chymotrypsin‐like activity of 26S proteasome in eukaryotic cells and was approved by the U.S. Food and Drug Administration in 2003 for the treatment of adult patients with multiple myeloma and mantle cell lymphoma [7, 8, 9]. In children, the safety of bortezomib in combination with chemotherapy has been established in phase I clinical trials of patients with relapsed acute lymphoblastic leukemia [10], refractory or recurrent solid tumors [11], and relapsed/refractory high‐risk neuroblastoma [12]. The efficacy of adding bortezomib to chemotherapy in children has been shown in the phase II trial of high‐risk acute lymphoblastic leukemia in first relapse and in the phase III trial of newly diagnosed T‐cell acute lymphoblastic leukemia and lymphoma [13, 14]. Recent studies of ATRT cell lines and patient‐derived xenografts in intracranial murine models have shown the treatment potential of proteasome inhibition in ATRT [15, 16]. The rationale for combining bortezomib with chemotherapy in recurrent or refractory malignancies is based on the ability of bortezomib to inhibit the ubiquitin‐proteasome pathway (UPP), thereby enhancing apoptotic signaling and sensitizing malignant cells to cytotoxic agents [17, 18]. Bortezomib has also demonstrated synergistic interactions with various chemotherapeutic agents [10, 11, 12, 18, 19, 20]. However, despite encouraging preclinical data, clinical trials evaluating bortezomib as a single agent have shown limited activity in both adult [21] and pediatric [22] acute leukemia and certain solid tumors, underscoring the importance of combination strategies to achieve optimal therapeutic efficacy [21, 22, 23, 24]. Osteopontin is a bone matrix glycoprotein that modulates mineralization and bone resorption [25], and plays an important role in tumorigenicity and metastasis [26, 27, 28]. Osteopontin levels in plasma and cerebrospinal fluid (CSF) are overexpressed in ATRTs and correlate with overall survival [29], and were used as a biomarker in this study. Our previous work demonstrated that bortezomib inhibited proliferation, induced apoptosis in ATRT cell lines, and prolonged survival in orthotopic xenograft mouse models [16]. To translate these preclinical data to patients with ATRT, we conducted the current phase 1b clinical trial that demonstrated the safety and feasibility of combining bortezomib with chemotherapy in patients with CNS ATRT. Patients and Methods This is a multicenter, open‐label, single‐arm study testing the feasibility of adding bortezomib to standard and high‐dose chemotherapy in CNS ATRT. The clinical trial was registered on ClinicalTrials.gov (identifier NCT06853080). Eight hospitals treating childhood cancers in Taiwan participated in this trial. Patients aged 0–20 years with newly diagnosed CNS ATRT by histopathology were eligible to be enrolled in this trial from January 2022 through December 2023. The exclusion criteria included patients who required concomitant use of strong CYP3A4 interfering agents, patients with extra‐CNS rhabdoid tumors without CNS ATRT, and patients with grade 2 or higher peripheral neuropathy assessed by Common Terminology Criteria of Adverse Events, version 5.0 (CTCAE v5.0) [30]. No patient was excluded. The backbone chemotherapy regimen used in this trial was modified from the Medical University of Vienna (MUV)‐ATRT regimen (Figure 1) [4]. Following surgery, patients received bortezomib in combination with three blocks of induction chemotherapy. Each induction block consisted of three combination regimens: vincristine, doxorubicin, and cyclophosphamide (VDC); cisplatin, etoposide, and ifosfamide (PEI); and high‐dose methotrexate with vincristine (VM). This was augmented with four doses of intrathecal chemotherapy (topotecan or cytarabine). Peripheral blood stem cell harvesting was performed after the VDC regimen of the second block of induction chemotherapy. Upon completion of induction chemotherapy, patients underwent high‐dose chemotherapy (carboplatin 500 mg/m2 from day −8 to day −6, thiotepa 300 mg/m2 from day −5 to day −3, and etoposide 250 mg/m2 from day −5 to day −3), followed by autologous hematopoietic stem cell reinfusion. Bortezomib was administered intravenously at 1.3 mg/m2 on days 1, 4, 8, and 11 of a 21‐day cycle. The treatment was given in combination with VDC and PEI regimens during induction, as well as following hematopoietic stem cell transplantation (HSCT) (Figure 1 and Supporting Information). Schema of bortezomib add‐on therapy for patients with newly diagnosed ATRT. Patients must have adequate organ and marrow function as defined below at enrollment and on (or within 1 week prior to) day 1 of each cycle: Absolute neutrophil count (ANC) ≥ 1000/μL, platelets ≥ 100,000/μL, hemoglobin ≥ 8 g/dL (red blood cell transfusion allowed), total bilirubin ≤ 1.5 times institutional upper limit of normal (ULN), aspartate aminotransferase (AST) and alanine aminotransferase (ALT) ≤ 2.5 times institutional ULN (4× ULN if liver involvement), and creatinine ≤ upper limits by age (0–5 years old: 0.8 mg/dL; 6–9 years old: 1.0 mg/dL; 10–12 years old: 1.2 mg/dL; 13–15 years old: Male, 1.5 mg/dL and female, 1.4 mg/dL; 16 years or older: Male, 1.7 mg/dL and female, 1.4 mg/dL). If the patient develops sepsis, grade 3 or higher non‐hematological toxicities, or grade 4 hematological toxicities that have not recovered by day 1 of each cycle, treatment should be withheld for 7 days or until recovery. Dose adjustment of bortezomib should be implemented in the subsequent cycle (Supporting Information). The symbol "☢" indicates that radiotherapy was given before initiation of chemotherapy or at the end of protocol chemotherapy with bortezomib. BIW, twice weekly; CTE, carboplatin, thiotepa, etoposide; HSCT, hematopoietic stem cell transplantation; PEI, cisplatin, etoposide, ifosfamide; VDC, vincristine, doxorubicin, cyclophosphamide; VM, vincristine, methotrexate. Four patients who had not experienced disease progression after block 3 of induction chemotherapy and had no greater than grade 2 toxicities proceeded to high‐dose chemotherapy. Radiation therapy was given before initiation of chemotherapy or at the end of protocol chemotherapy with bortezomib (Figure 1). All toxicities were assessed by the CTCAE v5.0 criteria. For the measurement of osteopontin levels in plasma, sampling was done before chemotherapy cycles. The concentration of soluble osteopontin in plasma was determined using the Human Osteopontin ELISA (enzyme‐linked immunosorbent assay) EIA (enzyme immunoassay) kit (code No. 17158, Immuno‐Biological Laboratories, Gumma, Japan), which has an inter‐assay coefficient of variation varying from 0.7% to 2% and an intra‐assay coefficient of variation varying from 3.7% to 4.7%, depending on the level of the marker. The ELISA was designed to detect human osteopontin in plasma with a detection level of 5 ng/mL or more. The developed reaction was quantified by reading at 490 nm (SpectraMax 190, Molecular devices). Each sample was analyzed in triplicate. Data were presented as the mean ± standard deviation (SD). Statistical analysis was performed using the R 4.3.1 software (R Foundation for Statistical Computing, Vienna, Austria). For statistical testing, a two‐sided p value < 0.05 was considered statistically significant. The paired t‐test was used to compare differences between the two consecutive time points. The primary endpoint was the number of cycles showing grade 3 or higher non‐hematologic toxicities combining bortezomib with chemotherapy. Secondary endpoints included overall response rate (ORR), duration of response (DOR), progression‐free survival (PFS), and overall survival (OS). Intracranial response to treatment was assessed using Response Assessment in Neuro‐Oncology (RANO) criteria based on serial magnetic resonance images (MRIs) [31, 32]. The ORR was defined as the percentage of patients who achieved complete response (CR) or partial response (PR). The DOR was defined as the time from the onset of CR or PR to progression or death due to any reason, whichever occurred earlier [33]. The prognostic role of plasma osteopontin levels was also evaluated. This study was conducted in compliance with the protocol approved by the Taipei Medical University‐Joint Institutional Review Board (TMU‐JIRB), and in accordance with the standard of good clinical practice. Written informed consent was provided by all patients or by parents or guardians of minor patients before the initiation of any study‐related procedures. Results Patient Characteristics A total of six patients (five females and one male) were enrolled from four of the eight participating institutions. The median age at diagnosis was 2.2 years (range, 0.8–13.8 years). The primary tumor site was infratentorial in four, all in the cerebellum, and supratentorial in two, one in the parietal lobe and one in the pineal region. Two patients had metastatic disease at diagnosis. The clinical characteristics of patients are summarized in Table 1 [34, 35]. TABLE 1Patient characteristics.Patient no.Age at initial diagnosis (year)GenderTumor siteSurgery [34]Stage [35]Radiation therapy (cGy)Cycles of bortezomibStatus before bortezomibBest RANO responsePFS (month)Outcome13.1M Cerebellar mid‐lower vermis NTRM3 CSI 2,400 Focal 5,280 Spinal mets 4380 (Upfront) 8M3PR10.7Died after progression at 19.4 month21.9FCerebellopontine angleSTRM0 a Focal 4,480 (Upfront) 4M0 a SD6.8DoD at 9.1 month30.8FCerebellar vermis, extensiveNTRM0 a Focal 5,600 (Upfront; PRT)8M0 a NE37NED413.8FLeft parietal lobeNTRM0 CSI 3,600 Focal 5,600 (Upfront) 6M0PR, with tiny residual19.4Died of sepsis after HSCT52.2FCerebellar vermisNTRM0 Focal 5,400 (End of therapy) 8M0NE40.1NED62.1FPineal regionNTRM1 CSI 2,080 (Salvage PRT) 2M1PD1.9DoD at 5.3 month Note: M0: No evidence of gross subarachnoid or hematogenous metastasis. M1: Microscopic tumor cells found in the cerebrospinal fluid. M3: Gross nodular seeding in the spinal subarachnoid space.Abbreviations: CSI, craniospinal irradiation; DoD, died of disease; HSCT, hematopoietic stem cell transplantation; mets, metastases; NE, not evaluable for response; NED, no evidence of disease; NTR, near total resection; PD, progressive disease; PFS, progression‐free survival; PR, partial response; PRT, proton radiotherapy; RANO, Response Assessment in Neuro‐Oncology; SD, stable disease; STR, subtotal resection. a CSF not evaluated in Patients No. 2 and 3. Patient characteristics. Cerebellar mid‐lower vermis CSI 2,400 Focal 5,280 Spinal mets 4380 (Upfront) Focal 4,480 (Upfront) CSI 3,600 Focal 5,600 (Upfront) Focal 5,400 (End of therapy) CSI 2,080 (Salvage PRT) Note: M0: No evidence of gross subarachnoid or hematogenous metastasis. M1: Microscopic tumor cells found in the cerebrospinal fluid. M3: Gross nodular seeding in the spinal subarachnoid space. Abbreviations: CSI, craniospinal irradiation; DoD, died of disease; HSCT, hematopoietic stem cell transplantation; mets, metastases; NE, not evaluable for response; NED, no evidence of disease; NTR, near total resection; PD, progressive disease; PFS, progression‐free survival; PR, partial response; PRT, proton radiotherapy; RANO, Response Assessment in Neuro‐Oncology; SD, stable disease; STR, subtotal resection. CSF not evaluated in Patients No. 2 and 3. Treatment Five patients had a near‐total resection (residual tumor < 1.5 cm2); one patient had a subtotal resection (residual tumor ≥ 1.5 cm2) [34]. Radiation therapy was delivered immediately after tumor resection in 4 patients and after completion of protocol chemotherapy with bortezomib in one patient. One patient progressed prior to radiation therapy and received proton therapy as salvage therapy. Bortezomib was administered with chemotherapy and post hematopoietic stem cell transplant (HSCT): 15 cycles were given concomitantly with vincristine, doxorubicin, and cyclophosphamide (VDC), 15 cycles were given concomitantly with cisplatin, etoposide, and ifosfamide (PEI), and 6 cycles were given following HSCT in 3 patients (Table 1). Four patients completed 6 cycles of bortezomib treatment; 2 patients discontinued treatment due to progressive disease after 2 and 4 cycles, respectively. Six cycles of bortezomib were evaluated after high‐dose chemotherapy in three patients without bortezomib‐related toxicity. Toxicities Thirty cycles of bortezomib were given with chemotherapy. No grade 3 or higher toxicities were considered bortezomib‐related (Table 2 and Figure 2). The number of cycles with grade 3 or higher hematologic toxicities included: 28 (93%) with neutropenia, 28 (93%) with leukopenia, 12 (40%) with anemia, and 22 (74%) with thrombocytopenia. Number of cycles with grade 3 or higher infectious and non‐hematologic toxicities included: 2 (7%) with lung infection, 7 (23%) with urinary tract infection, 3 (10%) with sepsis, and 20 (67%) with febrile neutropenia; 1 (3%) with AST/ALT elevation, 2 (7%) with hypokalemia, 2 (7%) with subdural hemorrhage, 2 (7%) with vomiting, and 3 (10%) with oral mucositis. The majority of treatment‐emergent adverse events (AEs) were hematologic. Three episodes of subdural hemorrhage developed in 3 patients and 2 required surgical intervention. Two episodes of sepsis occurred after VDC chemotherapy and one episode of sepsis occurred after PEI chemotherapy. TABLE 2Treatment‐emergent adverse events of bortezomib with standard chemotherapy (No. of cycles with bortezomib = 30).Adverse eventGrade 3Grade 4Any gradeNo. of cycles (%)No. of cycles (%)No. of cycles (%)HematologicWBC decreased0 (0%)28 (93%)30 (100%)Neutrophil count decreased0 (0%)28 (93%)29 (97%)Anemia12 (40%)0 (0%)30 (100%)Platelet count decreased11 (37%)11 (37%)30 (100%)InfectiousLung infection2 (7%)0 (0%)2 (7%)Urinary tract infection7 (23%)0 (0%)7 (23%)Sepsis3 (10%)0 (0%)3 (10%)Febrile neutropenia20 (67%)0 (0%)20 (67%)Non‐hematologicAST increased1 (3%)0 (0%)14 (47%)ALT increased1 (3%)0 (0%)20 (67%)Hypokalemia2 (7%)0 (0%)10 (33%)Subdural hemorrhage2 (7%)0 (0%)3 (10%)Vomiting2 (7%)0 (0%)8 (27%)Oral mucositis3 (10%)0 (0%)7 (23%) Note: All toxicities were assessed by the CTCAE v5.0 criteria.Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; WBC, white blood cell. Treatment‐emergent adverse events of bortezomib with standard chemotherapy (No. of cycles with bortezomib = 30). Note: All toxicities were assessed by the CTCAE v5.0 criteria. Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; WBC, white blood cell. Treatment‐emergent adverse events of bortezomib in combination with standard chemotherapy (No. of cycles with bortezomib = 30). All toxicities were assessed by the CTCAE v5.0 criteria. Abd, abdominal; Alk‐P, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; dec., decreased; GGT, γ‐glutamyl transferase; PLT, platelet; WBC, white blood cell. Six additional cycles of bortezomib were evaluated after recovery from HCST in three patients. No grade 4 toxicities were seen after recovery from HSCT. The number of cycles with grade 3 or higher toxicities included: 1 (17%) with leukopenia, 1 (17%) with thrombocytopenia, 1 (17%) with sensory and motor peripheral neuropathy in one patient with spinal metastases, and 1 (17%) with urinary tract infection (Table S1). Response and Survival Outcomes Of the 6 patients, 4 patients had measurable disease before treatment with bortezomib and chemotherapy and were evaluable for response. Two achieved a partial response (PR), one achieved stable disease (SD), and one had progressive disease (PD), resulting in an ORR of 50% (Table 1). Two patients without measurable tumor after surgery were not evaluable (NE) for response to bortezomib with chemotherapy and survive. Two patients remain alive without disease at 37 and 40 months, and one patient died without disease at 19 months after diagnosis. Three patients died with progressive disease. Median PFS is 15.1 months (range, 1.9–40.1 months) at a median follow‐up of 19.4 months (range, 5.3–40.1 months). The OS is 33% at a median follow‐up of 19.4 months (range, 5.3–40.1 months). Four patients died: two due to progressive disease before high‐dose chemotherapy, one due to progressive disease after high‐dose chemotherapy and post‐HSCT bortezomib, and one due to sepsis immediately after high‐dose chemotherapy in remission before post‐HSCT bortezomib. Among the four patients who received high‐dose chemotherapy, one died of sepsis, and one died of disease progression. Plasma Osteopontin Levels as a Biomarker of Treatment Response Five patients were evaluated for plasma osteopontin levels during protocol therapy. The sampling was inconsistent with the plan to obtain samples before each chemotherapy cycle. The plasma osteopontin level appears to have decreased after treatment with bortezomib and chemotherapy in patients No. 2 and 6 (Figure 3). Plasma osteopontin levels as a biomarker of treatment response. Data are presented as mean ± SD from three independent experiments. Error bar = SD. The cycle numbers indicate the cycle with bortezomib before which the plasma sample was obtained. *p < 0.05, **p < 0.01, ***p < 0.001. OPN conc., osteopontin concentration. Discussion Despite recent advances in anticancer therapies for brain tumors, the progress in treatment of ATRT has been relatively slow due to molecular heterogeneity, poor blood–brain barrier penetration of therapeutic agents, and lack of effective neoadjuvant regimens prior to local therapy. Histopathologically, ATRTs are embryonal tumors that contain a rhabdoid component as well as areas with primitive neuroectodermal, mesenchymal, and epithelial features [36]. The genetic cause underlying the pathogenesis of ATRTs is inactivating bi‐allelic mutations in SMARCB1 or rarely in SMARCA4. Loss of function mutations in components of the SWItch/Sucrose Non‐Fermentable (SWI/SNF) chromatin remodeling complex result in loss of the respective proteins. Therefore, loss of immunohistochemical staining of INI1 or BRG1 has been used as an important diagnostic tool for SMARCB1‐ and SMARCA4‐mutated ATRTs, respectively. The SMARCB1 mutated ATRTs have been divided into three distinct molecular subgroups: ATRT‐SHH (Group 1), ATRT‐TYR (Group 2A), and ATRT‐MYC (Group 2B) [37, 38]. DNA methylation profiling and RNA sequencing revealed that ATRT‐SMARCA4 (0.5%–2% of ATRT) should be regarded as a separate molecular subgroup distinct from ATRT‐SHH, ATRT‐TYR, and ATRT‐MYC [39]. Torchia et al. revealed that ATRT subgroups are associated with a distinct epigenomic landscape and sensitivity to inhibitors of NOTCH, BMP, PDGFRB, and epigenetic signaling [40]. The subgroups differ not only in the expression of distinct genes and the activation of specific signaling pathways, but also in clinical phenotypes and outcomes. ATRT‐SHH (Group 1) with neurogenic signatures has superior survival; however, ATRT‐MYC (Group 2B) with mesenchymal signatures exhibits upregulation of protein synthesis and proteasome degradation and has aggressive, treatment‐resistant phenotypes and poor outcomes [16, 37, 41]. In the Children's Oncology Group (COG) trial, the 4‐year survival rate was higher in patients with ATRT‐SHH tumors [5]. In the European Rhabdoid Registry (EU‐RHAB) cohort, patients < 1 year at diagnosis with a non‐ATRT‐TYR profile have a significantly poorer prognosis [37]. ATRT‐SMARCA4 has been associated with a higher frequency of germline mutations, younger age, and an inferior prognosis compared to SMARCB1 mutated cases [39]. The COG ACNS0333 trial, a regimen comprising 2 courses of multiagent chemotherapy after surgery, followed by 3 courses of high‐dose chemotherapy with peripheral blood stem cell rescue and involved‐field radiation therapy, demonstrated a 4‐year EFS of 37% and OS of 43% [5]. The EU‐RHAB protocol using an anthracycline‐based induction and either radiotherapy or high‐dose chemotherapy demonstrated a 5‐year EFS of 30.5% and OS of 34.7% [37]. Slavc et al. proposed the MUV‐ATRT regimen, which was successful in the first patient diagnosed with ATRT in 1997 and has been the standard of treatment since 2002. For the patients treated on the MUV‐ATRT regimen from 1992 to 2012, the 5‐year OS was 100% and 5‐year EFS 88.9%. The MUV‐ATRT regimen also appeared to be efficacious in preventing early relapses in young children with M1‐M3 stage disease allowing postponement of radiotherapy until after high‐dose chemotherapy. The treatment‐related toxicities of this regimen were tolerable and manageable [4]. Therefore, the backbone chemotherapy we used in our trial was modified from the MUV‐ATRT regimen. Increasing data suggest that ATRT may be a good candidate for pathway‐specific targeted therapies. A wide array of compounds with potential therapeutic effect in ATRT have been studied, including aurora kinase A inhibitor (AURKA), EZH2, CDK4/6, and proteasome inhibitors [42, 43, 44, 45]. The proteasome inhibitor, bortezomib, is known to be effective in the treatment of multiple myeloma and mantle cell lymphoma [7, 8, 9]. Malignant plasma cells have exaggerated protein synthesis, resulting in a significant burden for the proteasome degradation system. Our previous work demonstrated that proteasome‐encoding genes are highly expressed in ATRTs compared with normal brain tissue, are essential for tumor cell survival, and are correlated with the malignant phenotype of tumor cells. We also found that bortezomib targets proteostasis, inhibits tumor growth, and induces apoptosis through p53 accumulation in three ATRT‐MYC cell lines and in mice with orthotopic xenografts of ATRT. These findings suggest that bortezomib is a promising targeted therapy for the ATRT‐MYC [16]. The dependence on the ubiquitin‐proteasome system for cell survival and the sensitivity to proteasome inhibitors have been reported in SMARCB1‐deficient cancer cells [46, 47, 48]. Hong et al. demonstrated that genes related to the ubiquitin‐proteasome system are critical for the survival of renal medullary carcinoma cells, which have a molecular similarity with SMARCB1‐deficient cancers. Moreover, they found increased sensitivity to bortezomib in renal medullary carcinoma and other SMARCB1‐deficient cells, including malignant rhabdoid tumors (MRTs) and ATRTs [47]. It has also been noted that ATRT‐MYC and MRTs share molecular similarities and are associated with distinct developmental processes [49]. Our previous work that demonstrated the efficacious antitumor effects of bortezomib in ATRT‐MYC cells and orthotopic xenograft ATRT‐MYC strongly supports the sensitivity to bortezomib in SMARCB1‐deficient cancers, particularly in the MYC subgroup [16]. This also suggests that bortezomib may provide clinical benefit in patients with MRT. To determine whether the other subgroups of ATRTs (i.e., SHH and TYR subgroups) are sensitive to bortezomib, we conducted in vitro drug testing in 2 ATRT‐SHH cell lines (CHLA‐02 and CHLA‐04). Although the CHLA‐02 and CHLA‐04 cell lines were less sensitive to bortezomib (IC50 of 15.1 (14.3–15.9) nM and 15.8 (14.5–17.3) nM, respectively) than the ATRT‐MYC cell line (IC50 of 5.8–8.7 nM), these inhibitory concentrations are still clinically achievable (C max, 231.6–312.3 nM) [50]. In addition, we compared the antitumor effect of bortezomib with that of other targeted drugs, including the aurora kinase A inhibitor, alisertib (evaluated in a phase 2 clinical trial, NCT02114229, for patients with ATRT), the histone deacetylase inhibitor, SAHA [51], and the multiple kinase inhibitor, lenvatinib (due to the high expression of FGFR1 and RET mRNA in our PDX models). The result showed that bortezomib demonstrated the greatest in vitro activity against the three ATRT cell lines. Therefore, we hypothesized that bortezomib is potentially effective against all subgroups of CNS ATRT and undertook this clinical trial adding bortezomib to standard chemotherapy followed by high‐dose chemotherapy with autologous stem cell rescue. Regarding other proteasome inhibitors, our preclinical data (H. M. Tran et al., unpublished) suggest that the IC50 of ixozamib in ATRT‐MYC cell lines was substantially higher than that of bortezomib (101–133.6 nM vs. 5.8–8.7 nM, respectively). In addition, we could not find data supporting blood–brain barrier penetration or brain tumor accumulation of carfilzomib, therefore, we chose to evaluate bortezomib in this study. Because bortezomib has not been combined with high‐dose chemotherapy in myeloma or leukemia, and because of our concern regarding the potential risk of excessive neurotoxicity, we chose to administer bortezomib 3 weeks after high‐dose chemotherapy. In adult patients with recurrent gliomas, bortezomib monotherapy given for three doses before tumor resection has been shown to achieve measurable concentrations of bortezomib in the brain tumor tissue [52]. When bortezomib was given with chemotherapy, no grade 3 or higher toxicities were considered bortezomib‐related. Among the 30 cycles of bortezomib with standard chemotherapy, treatment‐emergent AEs consistent with the chemotherapy regimens were mainly hematologic. Two episodes of sepsis occurred after VDC chemotherapy and one episode occurred after PEI chemotherapy. These events were considered chemotherapy‐related rather than bortezomib‐related, as no episodes of sepsis were observed following post‐HSCT bortezomib cycles (Table S1). These events were manageable with blood component therapy, granulocyte‐colony stimulating factor administration and antibiotics treatment either empirically or based on the culture result. Three episodes of subdural hemorrhage developed in 3 patients with ventriculoperitoneal shunt during the nadir phase of thrombocytopenia after bortezomib given with chemotherapy; 2 of them required surgical intervention. We postulated that aggressive platelet transfusions could reduce the risk of subdural hemorrhage. We concluded that bortezomib in combination with chemotherapy had a manageable and expected safety profile. Osteopontin is a bone matrix glycoprotein that modulates mineralization and bone resorption [25]. It is involved in many cellular processes, including cell adhesion, cell survival, angiogenesis, apoptosis and inflammation [53]. Osteopontin plays an important role in tumorigenicity and metastasis and is a promising biomarker for human cancer [26, 27, 28]. In 2005, Kao et al. reported that the osteopontin level in plasma and CSF was overexpressed and correlated with overall survival in ATRT patients. Additionally, the authors observed that the plasma osteopontin level decreased after treatment and increased with relapse, suggesting that osteopontin may be a biomarker for surveillance of disease activity in ATRTs [29]. In this study, we attempted to evaluate the role of osteopontin as a biomarker in patients with ATRTs, but the sampling was incomplete. We recommend that further exploration of osteopontin as a biomarker of treatment response in a future clinical trial. Combining bortezomib with standard chemotherapy for ATRT is feasible and there were no additional toxicities when adding bortezomib to the VDC and PEI regimens. Although two of the four patients evaluable for response achieved a partial response to the chemotherapy regimen with bortezomib, and two of the six patients survived over 2 years, there was no suggestion of benefit of adding bortezomib to chemotherapy compared to historical data. There are several limitations of this study. First, the current study enrolled too few patients to evaluate the efficacy of adding bortezomib to chemotherapy. Second, there were too few patients in Taiwan to allow a comparison arm because there are only 2.9 newly diagnosed patients with CNS ATRT per year in Taiwan. Third, the plasma osteopontin levels were obtained inconsistently among the patients. Fourth, germline mutation and deletion of SMARCB1 were not evaluated in this study. In conclusion, adding bortezomib to standard chemotherapy and following high‐dose chemotherapy of ATRT is feasible. The occurrence of grade 3 or 4 toxicities was no more than what would be expected with chemotherapy alone. Giving bortezomib following high‐dose chemotherapy did not result in additional grade 4 toxicities. Author Contributions Wan‐Ling Ho: writing – original draft, writing – review and editing, investigation, formal analysis, visualization. Ming‐Yun Hsieh: investigation, writing – review and editing. Yu‐Mei Liao: writing – review and editing, investigation. Shyh‐Shin Chiou: writing – review and editing, investigation. Her‐Shyong Shiah: writing – review and editing, conceptualization. Chang‐Hsu Chiang: writing – review and editing, investigation, project administration. Shiann‐Tarng Jou: writing – review and editing, investigation. Hsiu‐Ju Yen: investigation, writing – review and editing. Yi‐Yen Lee: investigation, writing – review and editing. Muh‐Lii Liang: investigation, writing – review and editing. Ting‐Chi Yeh: investigation, writing – review and editing. Yu‐Cheng Chou: investigation, writing – review and editing. Fang‐Liang Huang: investigation, writing – review and editing. Shu‐Huey Chen: investigation, writing – review and editing. Hsin‐Lun Lee: investigation, writing – review and editing. Chia‐Chun Kuo: investigation, writing – review and editing. Yi‐Shan Yang: investigation, writing – review and editing. Shu‐Mei Chen: investigation, writing – review and editing. Min‐Lan Tsai: investigation, writing – review and editing. Hsi Chang: investigation, writing – review and editing. Kevin Li‐Chun Hsieh: investigation, writing – review and editing. Chia‐Lang Fang: investigation, writing – review and editing. Chia‐Yau Chang: investigation, writing – review and editing. Jinn‐Li Wang: investigation, writing – review and editing. Chu‐Chin Chen: investigation, writing – review and editing. Joon‐Khim Loh: investigation, writing – review and editing. Pei‐Chin Lin: investigation, writing – review and editing. Tzu‐Chin Lin: investigation, writing – review and editing. Shih‐Chung Wang: investigation, writing – review and editing. Yi‐Fang Tu: investigation, writing – review and editing. Chao‐Neng Cheng: investigation, writing – review and editing. Jimmy Ming‐Jung Chuang: investigation, writing – review and editing. Huy Minh Tran: investigation, writing – review and editing. Shian‐Ying Sung: project administration, writing – review and editing. Irving Lai: writing – review and editing, validation, investigation. Kuo‐Sheng Wu: investigation, writing – review and editing, validation. James S. Miser: conceptualization, writing – original draft, writing – review and editing. Yen‐Lin Liu: writing – review and editing, writing – original draft, conceptualization, investigation, formal analysis. Tai‐Tong Wong: writing – review and editing, investigation, formal analysis, conceptualization. Funding This study was partly supported by the Taiwan Ministry of Health and Welfare Clinical Trial Center (MOHW110TDU‐B‐212‐124007) and the 5th National Cancer Control Program (MOHW115‐TDU‐B‐222‐114013), Taipei Municipal Wanfang Hospital (IIT‐110‐19), Taipei Medical University Hospital (114TMU‐TMUH‐20), National Science and Technology Council, Taiwan (108‐2314‐B‐038‐061‐MY3 and 114‐2314‐B‐038‐031‐MY3), and the TMU Research Center of Cancer Translational Medicine (DP2‐TMU‐115‐C‐06) of Taipei Medical University from The Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education (MOE) in Taiwan. Ethics Statement This study was conducted in compliance with the protocol approved by the Taipei Medical University‐Joint Institutional Review Board (TMU‐JIRB), and in accordance with the standard of good clinical practice. Written informed consent was provided by all patients or by parents or guardians of minor patients before the initiation of any study‐related procedures. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Table S1: Treatment‐emergent adverse events of bortezomib post HSCT (No. of cycles with bortezomib = 6). Data S1: Supporting Information.