Search

Purpose: Ceritinib is an orally bioavailable, small-molecule inhibitor of anaplastic lympoma kinase (ALK), insulin-like growth factor 1 receptor (IGFR1), and focal adhesion kinase (FAK), which are highly expressed in glioblastoma and many brain metastases. Preclinical and clinical studies indicate that ceritinib has antitumor activity in central nervous system (CNS) malignancies. This phase 0 trial measured the tumor pharmacokinetics (PK) and pharmacodynamics (PD) of ceritinib in patients with brain metastasis or recurrent glioblastoma. Patients and Methods: Preoperative patients with brain tumors demonstrating high expression of pSTAT5b/pFAK/pIGFR1 were administered ceritinib for 10 days prior to tumor resection. Plasma, tumor, and cerebrospinal fluid (CSF) samples were collected at predefined timepoints following the final dose. Total and unbound drug concentrations were determined using LC-MS/MS. In treated tumor and matched archival tissues, tumor PD was quantified through IHC analysis of pALK, pSTAT5b, pFAK, pIGFR1, and pIRS1. Results: Ten patients (3 brain metastasis, 7 glioblastoma) were enrolled and no dose-limiting toxicities were observed. Ceritinib was highly bound to human plasma protein [median fraction unbound (Fu), 1.4%] and to brain tumor tissue (median Fu, 0.051% and 0.045% in gadolinium-enhancing and -nonenhancing regions respectively). Median unbound concentrations in enhancing and nonenhancing tumor were 0.048 and 0.006 μmol/L, respectively. Median unbound tumor-to-plasma ratios were 2.86 and 0.33 in enhancing and nonenhancing tumor, respectively. No changes in PD biomarkers were observed in the treated tumor samples as compared to matched archival tumor tissue. Conclusions: Ceritinib is highly bound to plasma proteins and tumor tissues. Unbound drug concentrations achieved in brain metastases and patients with recurrent glioblastoma were insufficient for target modulation.

BACKGROUND mTOR activation is a mechanism of resistance in CDK4/6 targeting. We evaluated tumor pharmacokinetics (PK) and tumor pharmacodynamics (PD) of combined CDK4/6 and mTOR inhibition in recurrent high-grade glioma (HGG) patients. METHODS Recurrent HGG patients with (1) intact RB, (2) CDKN2A/B deletion or CDK4/6 amplification, and (3) PTEN loss or PIK3CA mutations receive five days of presurgical ribociclib plus everolimus prior to resection at 2, 8 or 24 hours after the final dose. Beginning at 400mg QD ribociclib plus 2.5mg QD everolimus, six dose-escalations summit at 600mg QD plus 60mg QW. Gadolinium [Gd]-enhancing and nonenhancing tumor regions, CSF, and plasma are collected. Total and unbound drug concentrations are determined using validated LC-MS/MS methods. RB and S6 phosphorylation are compared to matched archival tissue. To select patients for a therapeutic expansion phase of combined drug therapy, the protocol includes a PK ‘trigger’ (i.e., for each drug, unbound concentration in Gd-nonenhancing tumor > 5-fold biochemical IC50) and a PD ‘trigger’ (i.e., for each tumor, > 30% decrease in pRB and pS6). RESULTS 21 patients with WHO Grade III (n=2) and IV (n=19) gliomas were enrolled into the Phase 0 component of the study. No dose-limiting toxicities were observed. In Gd-nonenhancing tumor regions, the median unbound concentration of ribociclib was 719 nM, whereas unbound everolimus tumor concentrations were undetectable. Across all dose-levels, 62% (13/21) and 22% (5/21) of tumors demonstrated decreased tumor RB and S6 phosphorylation, respectively. Tumor proliferation (MIB-1) was decreased in 67% (14/21) of all patients. No patients qualified for the therapeutic expansion phase. CONCLUSION In adult HGG, ribociclib achieves pharmacologically-relevant concentrations in Gd-nonenhancing tumor whereas everolimus exhibits no meaningful tumor penetration. These findings support further clinical development of ribociclib, but not everolimus, for the treatment of high-grade glioma patients.

BACKGROUND This dual-drug Phase 0 study (NCT04391595) evaluates the tumor pharmacokinetics (PK) and tumor pharmacodynamics (PD) of abemaciclib, a selective CDK4/6-inhibitor, plus LY3214996, a selective ERK1/2 inhibitor, in recurrent GBM patients. METHODS Adult recurrent GBM patients (n=10) with intact RB expression, > 30% pERK expression, and CDKN2A/B deletion or CDK4/6 amplification received six days of abemaciclib (150mg BID) plus LY3214996 (200mg QD) prior to a planned resection at 3-5 or 7-9 hour time interval after the final drug dose in a Time-Escalation Arm. Tumor tissue (gadolinium [Gd]-enhancing and nonenhancing regions), cerebrospinal fluid (CSF), and plasma were collected. Total and unbound drug concentrations were measured using validated LC-MS/MS methods. Tumor PD effects, including RB and RSK phosphorylation, were compared to matched archival or pre-treatment biopsied tissue. A PK ‘trigger’ (i.e., unbound concentration > 5x biochemical IC50) was set for each drug. Gd-nonenhancing tumor tissue exhibiting abemaciclib and LY3214996 concentrations in excess of their trigger threshold qualified patients for postoperative dual-drug therapy. RESULTS No dose-limiting toxicities were observed. In Gd-nonenhancing tumor regions, median unbound concentrations of abemaciclib were 31.2 nM (3-5 hour cohort) and 25.1 nM (7-9 hour cohort) while median unbound concentrations of LY3214996 were 52.0 nM (3-5 hour cohort) and 10.2 nM (7-9 hour cohort). Tumor RB and RSK phosphorylation decreased in 6/10 and 2/10 patients, respectively. Tumor proliferation (MIB-1) was decreased in 8/10 patients. 5/10 patients exceeded PK thresholds for both abemaciclib (12 nM) and LY3214996 (25 nM), thereby entering the study’s therapeutic expansion phase. CONCLUSION Abemaciclib and LY3214996 achieve pharmacologically-relevant concentrations in Gd-non-enhancing GBM tissue and are associated with suppression of RB pathway and tumor proliferation. The Optimal Time Interval (OTI) for tissue sampling was 3-5 hours after the final drug dose. Based on this interim analysis, the trial will accrue an additional 25 patients at this OTI.

Cell-cycle deregulation is at the crux of all malignancies, including glioblastoma (GBM). Aurora Kinase A (AURKA) plays a central role in G2/M transition and faithful chromosome segregation. In this study, we evaluated the pharmacokinetics, pharmacodynamics, and radiation sensitization properties of LY3295668, a highly specific AURKA inhibitor, in orthotopic patient-derived xenograft (PDX) models of GBM. Mice with intracranial tumors were randomized to 50 mg/kg LY3295668 PO BID x 4 days vs. placebo. LY3295668 levels in plasma and contrast-enhancing tumor tissue were measured by liquid chromatography tandem mass spectrometry (LC-MS/MS). Unbound fractions were determined by equilibrium dialysis. Immunohistochemistry was performed to assess levels of pAURKA (T288), phospho-Histone H3 (pHH3), and cleaved caspase 3 (CC3). For survival studies, mice with intracranial tumors were randomized to four cohorts – vehicle, radiotherapy, LY3295668 monotherapy, and LY3295668 plus radiotherapy. The median unbound concentration of LY3295668 was 270.88 nmol/L and 22.33 nmol/kg in plasma and tumor tissue, respectively – significantly higher than the biochemical IC50 of LY3295668 for AURKA inhibition (0.8 nM). A decrease in pHH3(+) cells (0.8% vs. 6.4%, p=0.036) indicated drug-induced mitotic arrest and was accompanied by an increase in CC3(+) cells (6.4% vs. 8.0%, p=0.67). Combination of LY3295668 with radiotherapy prolonged survival compared to either therapy alone in orthotopic GBM PDX models. LY3295668 is well tolerated, achieves pharmacologically-relevant unbound concentrations in GBM PDX models, and is associated with significant target modulation. Preclinical combination of LY3295668 with radiation therapy leads to synergistic effects and supports future clinical study of this multimodal strategy in glioblastoma patients.

BACKGROUND In neuro-oncology clinical trials, pharmacodynamic (PD) analysis of tumor tissue before and after experimental therapy is challenging. Today, Phase 0 studies in brain tumor patients typically employ archival tissue from a prior resection as the baseline PD comparator. However, the time-interval between tissue acquisitions can be months to years and often includes confounding exposure to other therapies. Here, we demonstrate the feasibility of a pre-treatment, single-pass stereotactic needle biopsy to support CLIA-certified genetic screening and tumor PD analysis in a Phase 0/2 clinical trial. METHODS In support of an ongoing Phase 0/2 ‘trigger’ trial testing CDK4/6 plus ERK1/2 inhibition in recurrent glioblastoma (GBM), a single-pass stereotactic biopsy was completed in select patients prior to enrollment. Each biopsy yields six tissue cores for immunohistochemistry (IHC) and genetic characterization. Using CLIA-certified protocols, genomic DNA (gDNA) was extracted from two cores, followed by library preparation and next-generation sequencing of a customized panel of 37 genes (IvySeq). In the four remaining cores, IHC staining was performed for RB, pRB, pERK, pS6, CDKN2A, MIB-1, ClCas-3 and pH2AX. RESULTS Single-pass stereotactic needle biopsies were collected from five recurrent GBM patients. No intraoperative, perioperative, or radiographic evidence of morbidity was observed. 800-2400ng of gDNA was isolated from each stereotactic biopsy. IvySeq analysis detected CDKN2A deletion, ATRX loss, and mutations in IDH, PTEN and TP53 genes. Based on IHC and genetic analyses of the biopsied samples, three patients were enrolled into the Phase 0/2 clinical trial and baseline value were used for PD analysis. CONCLUSION Single-pass stereotactic needle biopsy is a feasible and safe strategy to collect pre-treatment tissue in support of a Phase 0 clinical trial, enabling CLIA-certified genetic analysis for trial screening and tumor PD analysis.

BACKGROUND The RB-CDK4/6 and mTOR signaling pathways are deregulated in high-grade glioma (HGG) and mTOR activation is a potential mechanism of resistance to CDK4/6 inhibition. This study evaluates the tumor pharmacokinetics (PK) and tumor pharmacodynamics (PD) of combined CDK4/6 and mTOR inhibition in recurrent HGG patients. MATERIAL AND METHODS Eligible patients had recurrent HGG with (1) intact RB expression, (2) CDKN2A/B deletion or CDK4/6 amplification, and (3) PTEN loss or PIK3CA mutations. Six patients received five days of presurgical ribociclib (400mg QD) plus everolimus (2.5mg QD) and then underwent tumor resection at 2, 8 or 24 hours following the last dose. Five subsequent dose-escalation cohorts each enrolled three additional patients, reaching a maximum dose-level of ribociclib (600mg QD) plus everolimus (60mg QW). Tumor tissue (gadolinium [Gd]-enhancing and nonenhancing regions), CSF, and plasma were collected. Total and unbound drug concentrations were determined using validated LC-MS/MS methods. Tumor PD effects, including RB and S6 phosphorylation, were compared to matched archival tissue. A PK ‘trigger’ (i.e., unbound concentration > 5-fold biochemical IC50) and a PD ‘trigger’ (>30% decrease in both pRB and pS6) were set for each drug. Gd-nonenhancing tissue exhibiting both PK and PD effects in excess of these thresholds qualified patients for postoperative combination therapy. RESULTS 21 patients with WHO Grade III (n=2) and WHO Grade IV (n=19) gliomas were enrolled. No dose-limiting toxicities were observed. Following presurgical drug, all patients demonstrated marked decrease in Gd-enhancement on preoperative MRI. In Gd-nonenhancing tumor regions, the median unbound concentration of ribociclib was 719 nM (i.e., > 5-fold biochemical IC50 for CDK4/6 inhibition), whereas the unbound everolimus tumor concentrations in all patients were below the lower limit of quantitation (i.e., < 0.2 nM). The median total concentrations of everolimus in tumors at dose-levels 0 to 5 were 2.9, 8.8, 10.3, 5.0, 15.7, and 13.7 nM, respectively. Across all dose-levels, 62% (13/21) and 22% (5/21) of tumors demonstrated decreased tumor RB and S6 phosphorylation, respectively. Tumor proliferation (MIB-1) was decreased in 67% (14/21) of all patients. CONCLUSION In adult HGG, ribociclib achieves pharmacologically-relevant concentrations in Gd-nonenhancing tumor, consistent with the observed tumor PD effects. Everolimus exhibits very limited penetration into human glioma tissue. Our study supports further development of ribociclib, but not everolimus, for the treatment of glioma patients.

2005 Background: The RB-CDK4/6 and MAPK signaling pathways are dysregulated in glioblastoma (GBM). Our recent phase 0 study of ribociclib in recurrent GBM patients suggested that CDK4/6 inhibitor monotherapy is not durable. In this ongoing, dual-drug phase 0 study (NCT04391595), we evaluate the tumor pharmacokinetics (PK) and tumor pharmacodynamics (PD) of abemaciclib, a selective CDK4/6-inhibitor, plus LY3214996, a selective ERK1/2 inhibitor, in recurrent GBM patients. Methods: Adult patients eligible for this open-label, multi-center phase 0 protocol had recurrent GBM with (1) intact RB expression, (2) >30% pERK expression, and (3) CDKN2A/B deletion or CDK4/6 amplification. Prior to a planned resection, patients received six days of abemaciclib (150mg BID) plus LY3214996 (200mg QD). In a Time-Escalation Arm, ten patients were assigned to 3-5 hour or 7-9 hour intervals from final drug dose to tumor removal. Tumor tissue (gadolinium [Gd]-enhancing and nonenhancing regions), cerebrospinal fluid (CSF), and plasma were collected. Total and unbound drug concentrations were measured using validated LC-MS/MS methods. Tumor PD effects, including RB and RSK phosphorylation, were compared to matched archival or pre-treatment biopsied tissue. A PK ‘trigger’ ( i.e., unbound concentration > 5x biochemical IC50) was set for each drug. Gd-nonenhancing tumor tissue exhibiting abemaciclib and LY3214996 concentrations in excess of their trigger threshold qualified patients for postoperative dual-drug therapy. Results: In this interim analysis, no dose-limiting toxicities were observed. In Gd-nonenhancing tumor regions, median unbound concentrations of abemaciclib (including its equipotent M2 and M20 metabolites) were 31.2 nM (3-5 hour cohort) and 25.1 nM (7-9 hour cohort). In the same tissue, median unbound concentrations of LY3214996 were 52.0 nM (3-5 hour cohort) and 10.2 nM (7-9 hour cohort). Tumor RB and RSK phosphorylation decreased in 6/10 and 2/10 patients, respectively. Gd-enhancing tumor proliferation (MIB-1) was decreased in 8/10 patients. 5/10 patients exceeded PK thresholds for both abemaciclib (12 nM) and LY3214996 (25 nM), thereby entering the study’s therapeutic expansion phase. Conclusion: Abemaciclib and LY3214996 achieve pharmacologically-relevant concentrations in Gd-non-enhancing GBM tissue and are associated with suppression of the RB pathway and tumor proliferation. Following 6 days of presurgical drug exposure, the Optimal Time Interval (OTI) for tissue sampling was 3-5 hours after the final drug dose. Based on this interim analysis, the trial will accrue an additional 25 patients at this OTI. Clinical trial information: NCT04391595.

BACKGROUND New approaches are urgently needed for aggressive meningiomas, which remain largely incurable. Forkhead Box M1 (FOXM1) has been identified as a master transcription factor in aggressive meningiomas and Cyclin D-dependent Kinases (CDK) are positive regulators of cell-cycle entry, promoting tumorigenesis through FOXM1 activation. We evaluated the tumor pharmacokinetics (PK), tumor pharmacodynamics (PD), and preliminary clinical response of ribociclib, a selective CDK4/6-inhibitor, in aggressive meningioma patients. MATERIAL AND METHODS Eight aggressive WHO Grade II/III meningioma patients with intact RB expression were enrolled and administered oral ribociclib daily (900mg) for 5 days prior to tumor resection. Plasma, tumor, and cerebrospinal fluid (CSF) samples were collected at 2, 8, or 24 h after the last dose. Total and unbound drug concentrations were determined using a validated LC-MS/MS method. PD effects, including RB and FoxM1 phosphorylation, were compared to matched archival tissue. Patients with PK and PD responses in tumor tissue, defined as unbound ribociclib concentration > 5-fold in vitro IC50 (0.04µM) and >20% decrease in pRB levels, respectively, were enrolled into an exploratory Phase 2 cohort. RESULTS The median CSF concentration of ribociclib was 0.25 µM. In tumor tissue, the median unbound ribociclib concentration was 1.36 µM and the median unbound tumor-to-plasma ratio was 5.34. Suppression of G1-to-S phase was inferred in tumors with declining FoxM1 phosphorylation (50%), RB phosphorylation (38%), and cellular proliferation (75%). Four patients demonstrated concurrent PK and PD responses and were graduated to continuous ribociclib therapy. At one year, two of these patients (one Grade II and one Grade III) demonstrate partial responses per RANO criteria. CONCLUSION Ribociclib achieves pharmacologically-active concentrations in aggressive meningioma tissue. Target modulation was demonstrated by a decrease in FOXM1-mediated tumor proliferation. Further investigation of ribociclib as a therapeutic strategy for aggressive meningiomas is warranted.

BACKGROUND New approaches are urgently needed for aggressive meningiomas, which remain largely incurable. Forkhead Box M1 (FOXM1) has been identified as a master transcription factor in aggressive meningiomas and Cyclin D-dependent Kinases (CDK) are positive regulators of cell-cycle entry, promoting tumorigenesis through FOXM1 activation. We evaluated the tumor pharmacokinetics (PK), tumor pharmacodynamics (PD), and preliminary clinical response of ribociclib, a selective CDK4/6-inhibitor, in aggressive meningioma patients. METHODS Eight aggressive WHO Grade II/III meningioma patients with intact RB expression were enrolled and administered oral ribociclib daily for 5 days prior to tumor resection. Plasma, tumor, and cerebrospinal fluid (CSF) samples were collected at 2, 8, or 24 h after the last dose. Total and unbound drug concentrations were determined using a validated LC-MS/MS method. PD effects, including RB and FoxM1 phosphorylation, were compared to matched archival tissue. Patients with PK and PD responses in tumor tissue, defined as unbound ribociclib concentration > 5-fold in vitro IC50 (0.04µM) and >20% decrease in pRB levels, respectively, were enrolled into an expansion cohort for preliminary assessment of progression-free survival. RESULTS The median CSF concentration of ribociclib was 0.25 µM. In tumor tissue, the median unbound ribociclib concentration was 1.36 µM and the median unbound tumor-to-plasma ratio was 5.34. Suppression of G1-to-S phase was inferred in tumors with declining FoxM1 phosphorylation (50%), RB phosphorylation (38%), and cellular proliferation (75%). Four patients demonstrated concurrent PK and PD responses and were graduated to continuous ribociclib therapy. At 14 months, two of these patients (one Grade II and one Grade III) demonstrate partial responses per RANO criteria. CONCLUSION Ribociclib achieves pharmacologically-active concentrations in aggressive meningioma tissue. Target modulation was demonstrated by a decrease in FOXM1-mediated tumor proliferation. Further investigation of ribociclib as a therapeutic strategy for aggressive meningiomas is warranted.

Background New approaches are urgently needed for aggressive meningiomas, which remain largely incurable. Forkhead Box M1 (FOXM1) has been identified as a master transcription factor in aggressive meningiomas and Cyclin D-dependent Kinases (CDK) are positive regulators of cell-cycle entry, promoting tumorigenesis through FOXM1 activation. We evaluated the tumor pharmacokinetics (PK), tumor pharmacodynamics (PD), and preliminary clinical response of ribociclib, a selective CDK4/6-inhibitor, in aggressive meningioma patients. Methods Eight aggressive WHO Grade II/III meningioma patients with intact RB expression were enrolled and administered oral ribociclib daily (900mg) for 5 days prior to tumor resection. Plasma, tumor, and cerebrospinal fluid (CSF) samples were collected at 2, 8, or 24 h after the last dose. Total and unbound drug concentrations were determined using a validated LC-MS/MS method. PD effects, including RB and FoxM1 phosphorylation, were compared to matched archival tissue. Patients with PK and PD responses in tumor tissue, defined as unbound ribociclib concentration > 5-fold in vitro IC50 (0.04 µM) and >20% decrease in pRB levels, respectively, were enrolled into an exploratory phase 2 cohort. Results The median CSF concentration of ribociclib was 0.25 µM. In tumor tissue, the median unbound ribociclib concentration was 1.36 µM and the median unbound tumor-to-plasma ratio was 5.34. Suppression of G1-to-S phase was inferred in tumors with declining FoxM1 phosphorylation (50%), RB phosphorylation (38%), and cellular proliferation (75%). Four patients demonstrated concurrent PK and PD responses and were graduated to continuous ribociclib therapy. At one year, two of these patients (one Grade II and one Grade III) demonstrate partial responses per RANO criteria. Conclusions Ribociclib achieves pharmacologically-active concentrations in aggressive meningioma tissue. Target modulation was demonstrated by a decrease in FOXM1-mediated tumor proliferation. Further investigation of ribociclib as a therapeutic strategy for aggressive meningiomas is warranted. Clinical trial identification NCT02933736. Legal entity responsible for the study The authors. Funding The Ben and Catherine Ivy Foundation. Disclosure All authors have declared no conflicts of interest.

Background Ceritinib is an orally bioavailable, small molecule inhibitor for ALK/IGFR1/FAK, which are highly expressed in glioblastoma and brain metastases. Preclinical and clinical data suggest that ceritinib has activity in central nervous system (CNS) malignancies, but to date there is no direct evidence in patients. This study assessed the pharmacokinetics (PK) and pharmacodynamics (PD) of ceritinib in recurrent glioblastoma and brain metastasis patients. Methods Three brain metastasis and 7 glioblastoma patients with high expression of pSTAT5b/pFAK/pIGFR1 were enrolled and treated with oral ceritinib daily (750 mg) for 10 days prior to tumor resection. Plasma, tumor, and cerebrospinal fluid (CSF) samples were collected at ∼ 4 and 24 h following the last dose. Total and unbound drug concentrations were determined using LC-MS/MS. PD response was assessed by immunohistochemical analysis of pALK, pFAK, pIGFR1, and pIRS1 staining in treated tumor and matched archival tissues. Results Ceritinib was highly bound to human plasma protein (median fraction unbound (Fu), 1.4%) and to brain tumor tissue (median Fu, 0.073% and 0.14% in enhancing and non-enhancing regions respectively). There was a large interindividual variability in drug CNS penetration, with the median unbound concentrations in enhancing, non-enhancing, and CSF of 0.040, 0.006, and 0.012 µM, respectively. The median unbound tumor-to-plasma ratio was 2.44 and 0.33 in enhancing and non-enhancing areas, respectively. In one patient with brain metastasis, drug binding to enhancing tumor was significantly lower (Fu, 1.62%), resulting in a higher unbound drug tumor concentration and CSF concentration as compared to those in glioblastoma patients. In all patients, no changes in PD markers were detected. Conclusions Ceritinib is highly bound to plasma proteins and tumor tissues. Unbound drug concentrations achieved in brain metastasis and glioblastoma are unlikely sufficient for target modulation. Clinical trial identification NCT02605746. Legal entity responsible for the study The authors. Funding The Ben and Catherine Ivy Foundation. Disclosure All authors have declared no conflicts of interest.

BACKGROUND Ceritinib is an orally bioavailable, small molecule inhibitor for ALK/IGFR1/FAK, which are highly expressed in glioblastoma and brain metastases. Preclinical and clinical data suggest that ceritinib has activity in central nervous system (CNS) malignancies, but to date there is no direct evidence in patients. This study assessed the pharmacokinetics (PK) and pharmacodynamics (PD) of ceritinib in recurrent glioblastoma and brain metastasis patients. MATERIALS AND METHODS Three brain metastasis and seven glioblastoma patients with high expression of pSTAT5b/pFAK/pIGFR1 were enrolled and treated with oral ceritinib daily (750 mg) for 10 days prior to tumor resection. Plasma, tumor, and cerebrospinal fluid (CSF) samples were collected at ~ 4 and 24 h following the last dose. Total and unbound drug concentrations were determined using LC-MS/MS. PD response was assessed by immunohistochemical analysis of pALK, pFAK, pIGFR1, and pIRS1 staining in treated tumor and matched archival tissues. RESULTS Ceritinib was highly bound to human plasma protein (median fraction unbound (Fu), 1.4%) and to brain tumor tissue (median Fu, 0.073% and 0.14% in enhancing and non-enhancing regions respectively). There was a large interindividual variability in drug CNS penetration, with the median unbound concentrations in enhancing, non-enhancing, and CSF of 0.040, 0.006, and 0.012 µM, respectively. The median unbound tumor-to-plasma ratio was 2.44 and 0.33 in enhancing and non-enhancing areas, respectively. In one patient with brain metastasis, drug binding to enhancing tumor was significantly lower (Fu, 1.62%), resulting in a higher unbound drug tumor concentration and CSF concentration as compared to those in glioblastoma patients. In all patients, no changes in PD markers were detected. CONCLUSION Ceritinib is highly bound to plasma proteins and tumor tissues. Unbound drug concentrations achieved in brain metastasis and glioblastoma are unlikely sufficient for target modulation.