Ethnic-specific predictors of neurotoxicity among patients with pediatric acute lymphoblastic leukemia after high-dose methotrexate.

BACKGROUND: High-dose methotrexate (HD-MTX; 5000 mg/m2 ) is an important component of curative therapy in many treatment regimens for high-risk pediatric acute lymphoblastic leukemia (ALL). However, methotrexate therapy can result in dose-limiting neurotoxicity, which may disproportionately affect Latino children. This study evaluated risk factors for neurotoxicity after HD-MTX in an ethnically diverse population of patients with ALL. METHODS: The authors retrospectively reviewed the medical records of patients who were diagnosed with ALL and treated with HD-MTX at Texas Children's Cancer Center (2010-2017). Methotrexate neurotoxicity was defined as a neurologic episode (e.g., seizures or stroke-like symptoms) occurring within 21 days of HD-MTX that resulted in methotrexate treatment modifications. Mixed effects multivariable logistic regression was used to estimate the odds ratio (OR) and corresponding 95% confidence interval (CI) for the association between clinical factors and neurotoxicity. RESULTS: Overall, 351 patients (58.1% Latino) who received 1183 HD-MTX infusions were evaluated. Thirty-five patients (10%) experienced neurotoxicity, 71% of whom were Latino. After adjusting for clinical risk factors, the authors observed that serum creatinine elevations ≥50% of baseline were associated with a three-fold increased odds (OR, 3.32; 95% CI, 0.98-11.21; p = .05) for neurotoxicity compared with creatinine elevation <25%. Notably, predictors of neurotoxicity differed by ethnicity. Specifically, Latino children experienced a nearly six-fold increase in neurotoxicity odds (OR, 5.80; 95% CI, 1.39-24.17; p = .02) with serum creatinine elevation ≥50% compared with creatinine elevation <25%. CONCLUSIONS: The current findings indicate that serum creatinine elevations ≥50% may be associated with an increased risk for neurotoxicity among Latino children with ALL and may identify potential candidates for therapeutic or supportive care interventions.

Impact of undernutrition on the pharmacokinetics of chemotherapy in children with cancer: A systematic review.

OBJECTIVES: This systematic review provides an overview of the effect of undernutrition on the pharmacokinetics of chemotherapy in children with cancer. METHODS: PubMed, Embase, and Cochrane were searched to identify eligible studies. This study uses the definition for undernutrition from the World Health Organization and the Gomez classification. RESULTS: Four studies with a total of 668 children with cancer were included and n = 121 (18%) were undernourished. Significant decreased clearance rates were found for vincristine in undernourished children compared to children with a normal nutritional status. CONCLUSION: Presenting outcomes only show significant changes in the pharmacokinetics of vincristine in undernourished children with cancer. However, data are scarce, groups were small, and none of the studies included severely undernourished children. In order to improve outcomes for (severely) undernourished children with cancer, more pharmacokinetic research is needed. The ultimate goal would be to develop subgroups, and ultimately individualized drug dosing in order to improve outcomes for children with cancer worldwide.

Distribution and frequency of tyrosine kinase inhibitor-associated long-term complications in children with chronic myeloid leukemia.

BACKGROUND: Tyrosine kinase inhibitors (TKIs) improve outcomes for pediatric malignancies characterized by specific gene rearrangements and mutations; however, little is known about the long-term impact of TKI exposure. Our objective was to assess the incidence and type of late-onset TKI-related toxicities in children with chronic myeloid leukemia (CML). METHODS: We reviewed medical records from patients diagnosed with CML between 2006 and 2019 at <21 years of age and prescribed one or more TKIs. Patients treated with stem cell transplant were excluded. Outcomes were captured beginning at 1 year after CML diagnosis. Outcome incidence was described overall and stratified by TKI exposure during the data-capture period. RESULTS: Twenty-two eligible TKI-exposed patients with CML were identified. The median follow-up was 6.0 years (range: 2.2-14.3). All pericardial (n = 3) or pleural (n = 3) effusion outcomes occurred in patients treated with TKIs during the data-capture period. Other outcomes included hypertension (n = 2), ectopy on electrocardiogram (n = 2), and gastrointestinal bleed (n = 1). All outcomes were graded as mild to moderate: some resulted in a temporary discontinuation of TKI, but none led to a change in TKI. No differences were noted in outcome incidence by type of TKI exposure. CONCLUSIONS: TKIs have substantially improved prognosis for subsets of childhood leukemia, but there are limited long-term data to inform exposure-based risk for late-onset complications and screening. Our results suggest that TKI-exposed survivors may be at risk for long-term outcomes that extend well into survivorship.

Prospective patient-reported symptom profiles associated with pediatric acute lymphoblastic leukemia relapse.

PURPOSE: Despite improvements in frontline pediatric acute lymphoblastic leukemia (ALL) treatment, relapse remains a concern. Research in adult cancer patients suggests that patient-reported symptoms may predict survival, but the relationship between symptoms and relapse for pediatric ALL has received little attention. METHODS: Pediatric patients with ALL (age 2-18 years) and/or their primary caregivers completed symptom surveys at the end of induction, start of delayed intensification (DI), start of maintenance cycle 1 (MC1), and start of maintenance cycle 2 (MC2). Symptom clusters for co-occurring fatigue, pain, sleep disruptions, and nausea were defined using latent profile analysis. Hazard ratios (HR) and 95% confidence intervals (CI) for the association between symptom clusters, individual symptoms, and subsequent relapse were calculated using multivariable Cox proportional hazards models, adjusting for clinical and demographic factors. RESULTS: Eligible patients (n = 208) were followed an average of 2.6 years for the incidence of relapse (n = 22). Associations between relapse and symptoms were identified for fatigue at DI (HR = 1.83, 95%CI 1.23-2.73) and MC1 (HR = 2.14, 95%CI 1.62-2.84), pain at DI (HR = 1.80, 95%CI 1.19-2.72), nausea at the end of induction (HR = 1.19, 95%CI 1.01-1.39), and sleep disturbances at the end of induction (HR = 2.00, 95%CI 1.11-3.62), DI (HR = 1.73, 95%CI 1.01-2.96), and MC1 (HR = 2.19, 95%CI 1.10-4.35). Symptom clusters comprised of individuals with a higher average symptom burden at DI were significantly (p < 0.05) associated with relapse. CONCLUSION: Patient-reported symptoms may provide prognostic information to aid in the identification of pediatric ALL patients at increased risk of relapse.

Blinatumomab use in pediatric ALL: Taking a BiTE out of preparation, administration and toxicity challenges.

Blinatumomab is the first in its class bispecific T-cell engager monoclonal antibody, which binds to CD19 expressed on B-cells and CD3 expressed on T-cells, resulting in lysis of CD19-positive cells common in B-cell malignancies. Blinatumomab is Food and Drug Administration (FDA) approved for the treatment of adults and children with relapsed/refractory or minimal residual disease (MRD) positive precursor B-cell ALL (B-ALL). Despite impressive efficacy for the approved indications and favorable toxicity profile compared to standard-of-care chemotherapy, blinatumomab presents unique health-system challenges related to preparation, administration, toxicity monitoring and medication error prevention. Blinatumomab delivery also offers plethora of opportunities for interdisciplinary planning and collaboration. The purpose of this paper is to discuss practical considerations for safe blinatumomab delivery from the pharmacy and nursing perspectives.

Effect of Postreinduction Therapy Consolidation With Blinatumomab vs Chemotherapy on Disease-Free Survival in Children, Adolescents, and Young Adults With First Relapse of B-Cell Acute Lymphoblastic Leukemia: A Randomized Clinical Trial.

Importance: Standard chemotherapy for first relapse of B-cell acute lymphoblastic leukemia (B-ALL) in children, adolescents, and young adults is associated with high rates of severe toxicities, subsequent relapse, and death, especially for patients with early relapse (high risk) or late relapse with residual disease after reinduction chemotherapy (intermediate risk). Blinatumomab, a bispecific CD3 to CD19 T cell-engaging antibody construct, is efficacious in relapsed/refractory B-ALL and has a favorable toxicity profile. Objective: To determine whether substituting blinatumomab for intensive chemotherapy in consolidation therapy would improve survival in children, adolescents, and young adults with high- and intermediate-risk first relapse of B-ALL. Design, Setting, and Participants: This trial was a randomized phase 3 clinical trial conducted by the Children's Oncology Group at 155 hospitals in the US, Canada, Australia, and New Zealand with enrollment from December 2014 to September 2019 and follow-up until September 30, 2020. Eligible patients included those aged 1 to 30 years with B-ALL first relapse, excluding those with Down syndrome, Philadelphia chromosome-positive ALL, prior hematopoietic stem cell transplant, or prior blinatumomab treatment (n = 669). Interventions: All patients received a 4-week reinduction chemotherapy course, followed by randomized assignment to receive 2 cycles of blinatumomab (n = 105) or 2 cycles of multiagent chemotherapy (n = 103), each followed by transplant. Main Outcome and Measures: The primary end point was disease-free survival and the secondary end point was overall survival, both from the time of randomization. The threshold for statistical significance was set at a 1-sided P <.025. Results: Among 208 randomized patients (median age, 9 years; 97 [47%] females), 118 (57%) completed the randomized therapy. Randomization was terminated at the recommendation of the data and safety monitoring committee without meeting stopping rules for efficacy or futility; at that point, 80 of 131 planned events occurred. With 2.9 years of median follow-up, 2-year disease-free survival was 54.4% for the blinatumomab group vs 39.0% for the chemotherapy group (hazard ratio for disease progression or mortality, 0.70 [95% CI, 0.47-1.03]); 1-sided P = .03). Two-year overall survival was 71.3% for the blinatumomab group vs 58.4% for the chemotherapy group (hazard ratio for mortality, 0.62 [95% CI, 0.39-0.98]; 1-sided P = .02). Rates of notable serious adverse events included infection (15%), febrile neutropenia (5%), sepsis (2%), and mucositis (1%) for the blinatumomab group and infection (65%), febrile neutropenia (58%), sepsis (27%), and mucositis (28%) for the chemotherapy group. Conclusions and Relevance: Among children, adolescents, and young adults with high- and intermediate-risk first relapse of B-ALL, postreinduction treatment with blinatumomab compared with chemotherapy, followed by transplant, did not result in a statistically significant difference in disease-free survival. However, study interpretation is limited by early termination with possible underpowering for the primary end point. Trial Registration: ClinicalTrials.gov Identifier: NCT02101853.

Review of nutritional status, body composition, and effects of antineoplastic drug disposition.

The overall survival for children with cancer in high income countries is excellent. However, there are many disparities that may negatively affect survival, which are particularly problematic in low income countries, such as nutritional status at diagnosis and throughout therapy. Nutritional status as well as concomitant foods, supplements, and medications may play a role in overall exposure and response to chemotherapy. Emerging science around the microbiome may also play a role and should be further explored as a contributor to disease progression and therapeutic response. This article highlights some of these issues and proposes additional areas of research relevant to nutritional status and pharmacology that are needed in pediatric oncology.

Risk factors associated with delayed methotrexate clearance and increased toxicity in pediatric patients with osteosarcoma.

High-dose methotrexate (HD-MTX; 12 g/m2 ) is part of standard therapy for pediatric osteosarcoma (OS). Risk factors associated with MTX toxicity in children with OS are not well defined. We investigated the association between peak MTX levels (four-hour) and delayed MTX clearance or treatment toxicity. Information was retrieved from electronic medical records of 33 OS patients treated with HD-MTX at Texas Children's Hospital from 2008 to 2015. We found that the four-hour MTX level did not contribute to toxicity or delayed MTX clearance. We demonstrated that certain demographic characteristics are associated with delayed clearance and increased toxicity.

A phase 1 study of cabozantinib in children and adolescents with recurrent or refractory solid tumors, including CNS tumors: Trial ADVL1211, a report from the Children's Oncology Group.

BACKGROUND: We conducted a phase 1 trial to determine the maximum tolerated dose (MTD), toxicity profile, pharmacokinetics (PK), pharmacodynamics (PD), and preliminary activity of cabozantinib in children with refractory or relapsed solid tumors. METHODS: Patients received cabozantinib tablets on a continuous dosing schedule in a rolling-six escalating phase 1 trial design. PK and PD studies were performed. RESULTS: Forty-one patients, median (range) age 13 (4-18) years, received cabozantinib to achieve a weekly cumulative dose equivalent to 30 (n = 6), 40 (n = 23). or 55 (n = 12) mg/m2 /day. At 40 mg/m2 /d, dose-limiting toxicities (DLTs) were palmar-plantar erythrodysesthesia syndrome, mucositis, and elevated alanine aminotransferase, lipase, and bilirubin. At 55 mg/m2 /d, hypertension, reversible posterior leukoencephalopathy syndrome, headache, fatigue, and proteinuria were DLTs. Frequent non-DLTs included diarrhea, hypothyroidism, fatigue, nausea, vomiting, elevated hepatic transaminases, and proteinuria. In subsequent cycles, DLTs occurred at all dose levels. Across all dose levels, the steady-state exposure and peak cabozantinib concentrations were similar. Four patients experienced a confirmed partial response: medullary thyroid cancer (MTC; n = 2), Wilms tumor, and clear cell sarcoma. Stable disease (>6 cycles) was seen in seven patients (MTC [n = 2], Ewing sarcoma, synovial sarcoma, alveolar soft part sarcoma, paraganglioma, and ependymoma). CONCLUSIONS: A protocol-defined MTD was not reached; DLTs and dose reductions for toxicity occurred in the first and subsequent cycles at all dose levels. Based on the toxicity profile, pharmacokinetics, and responses, the recommended dose of cabozantinib in pediatric patients with refractory solid tumors is 40 mg/m2 /day. A phase 2 study of cabozantinib is being conducted.

Efficacy of liposomal bupivacaine in pediatric patients undergoing spine surgery.

BACKGROUND: Liposomal bupivacaine may be an option for reducing opioid utilization in pediatric scoliosis surgery. The use of liposomal bupivacaine in this patient population has not been previously described. METHODS: Patients who underwent posterior spinal fusion surgery at our institution from 2011-2016 were identified. We performed a retrospective matched cohort study, matching patients who received intraoperative liposomal bupivacaine by age, gender, and extent of surgery to patients who did not. The primary endpoint was the use of morphine equivalents in the first 72 hours after surgery. Data collection included demographic and surgical data, pain medication utilization, and pain scores. Area under the curve (AUC) for pain scores was calculated. Descriptive statistical methods and univariable analysis were used to compare patients who received liposomal bupivacaine to patients who did not. RESULTS: One hundred and forty-one patients met study criteria; 47 patients who received liposomal bupivacaine were matched to 94 control patients who did not receive liposomal bupivacaine. No significant differences were noted in the patient population with the patients requiring a median of 11 segments (range 10-13 segments) fused. Patients received a mean of 56.6 ± 37.4 mg/kg of intravenous acetaminophen, a mean of 3.4 ± 2.1 mg/kg of intravenous ketorolac, and 1.9 ± 0.93 mg/kg of morphine equivalents in the first 72 hours after surgery. On univariable analysis, no differences were noted in intravenous acetaminophen use, pain score AUC, intravenous ketorolac use, or morphine equivalents (2.0 ± 98 vs 1.8 ± 0.82) in patients who did not receive liposomal bupivacaine as compared to those patient who did received liposomal bupivacaine. CONCLUSION: Liposomal bupivacaine was not associated with reductions in postoperative opioid use in pediatric spinal surgery.

Using a Bedside Algorithm to Individually Dose High-dose Methotrexate for Patients at Risk for Toxicity.

We developed a bedside algorithm for individually adjusting the high-dose methotrexate (HDMTX) dose (5 g/m) given to patients with acute lymphoblastic leukemia at high risk for methotrexate toxicity. Data were reviewed for 8 patients receiving 21 cycles of HDMTX as per our algorithm. Eleven cycles began with 5 g/m, 10 cycles began with a preinfusion 20% to 25% dose reduction. Neither mean MTX AUC (2320.5±179.1 vs. 2080.4±161.7 µmol×h/L), mean Cpss (64.3±7.9 vs. 60.8±6.1 µM), nor toxicities were statistically different between groups. Our algorithm allowed the safe administration of HDMTX to patients at risk of MTX toxicities and obviated the need for preinfusion dose reduction.

Algorithmic encoding of protected characteristics in chest X-ray disease detection models.

BACKGROUND: It has been rightfully emphasized that the use of AI for clinical decision making could amplify health disparities. An algorithm may encode protected characteristics, and then use this information for making predictions due to undesirable correlations in the (historical) training data. It remains unclear how we can establish whether such information is actually used. Besides the scarcity of data from underserved populations, very little is known about how dataset biases manifest in predictive models and how this may result in disparate performance. This article aims to shed some light on these issues by exploring methodology for subgroup analysis in image-based disease detection models. METHODS: We utilize two publicly available chest X-ray datasets, CheXpert and MIMIC-CXR, to study performance disparities across race and biological sex in deep learning models. We explore test set resampling, transfer learning, multitask learning, and model inspection to assess the relationship between the encoding of protected characteristics and disease detection performance across subgroups. FINDINGS: We confirm subgroup disparities in terms of shifted true and false positive rates which are partially removed after correcting for population and prevalence shifts in the test sets. We find that transfer learning alone is insufficient for establishing whether specific patient information is used for making predictions. The proposed combination of test-set resampling, multitask learning, and model inspection reveals valuable insights about the way protected characteristics are encoded in the feature representations of deep neural networks. INTERPRETATION: Subgroup analysis is key for identifying performance disparities of AI models, but statistical differences across subgroups need to be taken into account when analyzing potential biases in disease detection. The proposed methodology provides a comprehensive framework for subgroup analysis enabling further research into the underlying causes of disparities. FUNDING: European Research Council Horizon 2020, UK Research and Innovation.

Clinical covariates that improve the description of high dose methotrexate pharmacokinetics in a diverse population to inform MTXPK.org.

The MTXPK.org webtool was launched in December 2019 and was developed to facilitate model-informed supportive care and optimal use of glucarpidase following the administration of high-dose methotrexate (HDMTX). One limitation identified during the original development of the MTXPK.org tool was the perceived generalizability because the modeled population comprised solely of Nordic pediatric patients receiving 24-h infusions for the treatment of acute lymphoblastic leukemia. The goal of our study is to describe the pharmacokinetics of HDMTX from a diverse patient population (e.g., races, ethnicity, indications for methotrexate, and variable infusion durations) and identify meaningful factors that account for methotrexate variability and improve the model's performance. To do this, retrospectively analyzed pharmacokinetic and toxicity data from pediatric and adolescent young adult patients who were receiving HDMTX (>0.5 g/m2 ) for the treatment of a cancer diagnosis from three pediatric medical centers. We performed population pharmacokinetic modeling referencing the original MTXPK.org NONMEM model (includes body surface area and serum creatinine as covariates) on 1668 patients, 7506 administrations of HDMTX, and 30,250 concentrations. Our results support the parameterizations of short infusion duration (<8 h) and the presence of Down syndrome on methotrexate clearance, the parameterization of severe hypoalbuminemia (<2.5 g/dL) on the intercompartmental clearance (Q2 and Q3), and the parameterization of pleural effusion on the volume of distribution (V1 and V2). These novel parameterizations will increase the generalizability of the MTXPK.org model once they are added to the webtool.

Children's Oncology Group AALL1331: Phase III Trial of Blinatumomab in Children, Adolescents, and Young Adults With Low-Risk B-Cell ALL in First Relapse.

PURPOSE: Blinatumomab, a bispecific T-cell engager immunotherapy, is efficacious in relapsed/refractory B-cell ALL (B-ALL) and has a favorable toxicity profile. One aim of the Children's Oncology Group AALL1331 study was to compare survival of patients with low-risk (LR) first relapse of B-ALL treated with chemotherapy alone or chemotherapy plus blinatumomab. PATIENTS AND METHODS: After block 1 reinduction, patients age 1-30 years with LR first relapse of B-ALL were randomly assigned to block 2/block 3/two continuation chemotherapy cycles/maintenance (arm C) or block 2/two cycles of continuation chemotherapy intercalated with three blinatumomab blocks/maintenance (arm D). Patients with CNS leukemia received 18 Gy cranial radiation during maintenance and intensified intrathecal chemotherapy. The primary and secondary end points were disease-free survival (DFS) and overall survival (OS). RESULTS: The 4-year DFS/OS for the 255 LR patients accrued between December 2014 and September 2019 were 61.2% ± 5.0%/90.4% ± 3.0% for blinatumomab versus 49.5% ± 5.2%/79.6% ± 4.3% for chemotherapy (P = .089/P = .11). For bone marrow (BM) ± extramedullary (EM) (BM ± EM; n = 174) relapses, 4-year DFS/OS were 72.7% ± 5.8%/97.1% ± 2.1% for blinatumomab versus 53.7% ± 6.7%/84.8% ± 4.8% for chemotherapy (P = .015/P = .020). For isolated EM (IEM; n = 81) relapses, 4-year DFS/OS were 36.6% ± 8.2%/76.5% ± 7.5% for blinatumomab versus 38.8% ± 8.0%/68.8% ± 8.6% for chemotherapy (P = .62/P = .53). Blinatumomab was well tolerated and patients had low adverse event rates. CONCLUSION: For children, adolescents, and young adults with B-ALL in LR first relapse, there was no statistically significant difference in DFS or OS between the blinatumomab and standard chemotherapy arms overall. However, blinatumomab significantly improved DFS and OS for the two thirds of patients with BM ± EM relapse, establishing a new standard of care for this population. By contrast, similar outcomes and poor DFS for both arms were observed in the one third of patients with IEM; new treatment approaches are needed for these patients (ClinicalTrials.gov identifier: NCT02101853).

Repurposing Atovaquone as a Therapeutic against Acute Myeloid Leukemia (AML): Combination with Conventional Chemotherapy Is Feasible and Well Tolerated.

Survival of pediatric AML remains poor despite maximized myelosuppressive therapy. The pneumocystis jiroveci pneumonia (PJP)-treating medication atovaquone (AQ) suppresses oxidative phosphorylation (OXPHOS) and reduces AML burden in patient-derived xenograft (PDX) mouse models, making it an ideal concomitant AML therapy. Poor palatability and limited product formulations have historically limited routine use of AQ in pediatric AML patients. Patients with de novo AML were enrolled at two hospitals. Daily AQ at established PJP dosing was combined with standard AML therapy, based on the Medical Research Council backbone. AQ compliance, adverse events (AEs), ease of administration score (scale: 1 (very difficult)-5 (very easy)) and blood/marrow pharmacokinetics (PK) were collected during Induction 1. Correlative studies assessed AQ-induced apoptosis and effects on OXPHOS. PDX models were treated with AQ. A total of 26 patients enrolled (ages 7.2 months-19.7 years, median 12 years); 24 were evaluable. A total of 14 (58%) and 19 (79%) evaluable patients achieved plasma concentrations above the known anti-leukemia concentration (>10 µM) by day 11 and at the end of Induction, respectively. Seven (29%) patients achieved adequate concentrations for PJP prophylaxis (>40 µM). Mean ease of administration score was 3.8. Correlative studies with AQ in patient samples demonstrated robust apoptosis, OXPHOS suppression, and prolonged survival in PDX models. Combining AQ with chemotherapy for AML appears feasible and safe in pediatric patients during Induction 1 and shows single-agent anti-leukemic effects in PDX models. AQ appears to be an ideal concomitant AML therapeutic but may require intra-patient dose adjustment to achieve concentrations sufficient for PJP prophylaxis.

Active label cleaning for improved dataset quality under resource constraints.

Imperfections in data annotation, known as label noise, are detrimental to the training of machine learning models and have a confounding effect on the assessment of model performance. Nevertheless, employing experts to remove label noise by fully re-annotating large datasets is infeasible in resource-constrained settings, such as healthcare. This work advocates for a data-driven approach to prioritising samples for re-annotation-which we term "active label cleaning". We propose to rank instances according to estimated label correctness and labelling difficulty of each sample, and introduce a simulation framework to evaluate relabelling efficacy. Our experiments on natural images and on a specifically-devised medical imaging benchmark show that cleaning noisy labels mitigates their negative impact on model training, evaluation, and selection. Crucially, the proposed approach enables correcting labels up to 4 × more effectively than typical random selection in realistic conditions, making better use of experts' valuable time for improving dataset quality.

Comparison of the blood, bone marrow, and cerebrospinal fluid metabolomes in children with b-cell acute lymphoblastic leukemia.

Metabolomics may shed light on treatment response in childhood acute lymphoblastic leukemia (ALL), however, most assessments have analyzed bone marrow or cerebrospinal fluid (CSF), which are not collected during all phases of therapy. Blood is collected frequently and with fewer risks, but it is unclear whether findings from marrow or CSF biomarker studies may translate. We profiled end-induction plasma, marrow, and CSF from N = 10 children with B-ALL using liquid chromatography-mass spectrometry. We estimated correlations between plasma and marrow/CSF metabolite abundances detected in ≥ 3 patients using Spearman rank correlation coefficients (rs). Most marrow metabolites were detected in plasma (N = 661; 81%), and we observed moderate-to-strong correlations (median rs 0.62, interquartile range [IQR] 0.29-0.83). We detected 328 CSF metabolites in plasma (90%); plasma-CSF correlations were weaker (median rs 0.37, IQR 0.07-0.70). We observed plasma-marrow correlations for metabolites in pathways associated with end-induction residual disease (pyruvate, asparagine) and plasma-CSF correlations for a biomarker of fatigue (gamma-glutamylglutamine). There is considerable overlap between the plasma, marrow, and CSF metabolomes, and we observed strong correlations for biomarkers of clinically relevant phenotypes. Plasma may be suitable for biomarker studies in B-ALL.