CoChat

The Medulloblastoma Meta-Analysis Portal Enables Integrative Exploration of Clinical and Genomic Data from Patients

Open access

Published version of a preprint

The DOI matches its Crossref record; 30 of the 30 references it cites were found in a registry or an index; 16 of the 20 sources its claims cite were checked for support.

Abstract

Abstract Medulloblastoma is a prevalent, highly heterogeneous malignant pediatric brain cancer. In this study, we developed the Medulloblastoma Meta-Analysis portal, a web-based data analysis and visualization platform that consolidates clinical, molecular, and survival data of 898 patients from three major clinical trials (ACNS0331, ACNS0332, and SJMB03). The portal integrated clinical data with genomic mutations and DNA methylomes, enabling in-depth analyses of tumor subgroups and survival outcomes. Dynamic exploration and custom stratification of these data were made possible by the portal’s user-friendly visualization tools powered by the ProteinPaint framework, allowing users to select from a wide range of variables or customize their own. The Cox regression analysis tool in the portal demonstrated utility in examining the prognostic impact of variables across a heterogeneous dataset. Additionally, an in-depth exploration of medulloblastomas that harbor KBTBD4 mutations and their corresponding methylomes generated leads that could help improve molecular classification. These use cases underscore the portal’s utility in enhancing risk stratification, informing tailored therapeutic strategies, and revealing the complexities of molecular heterogeneity in patients with medulloblastoma. The portal, along with links to the tutorials and live figures, is freely available at https://viz.stjude.cloud/st-jude-childrens-research-hospital/visualization/medulloblastoma-integrative-analysis-portal~2882. Significance: The MB-meta portal facilitates risk stratification of patients using molecular, genomic, clinical, and demographic data, representing a significant step in neuro-oncology data sharing by offering an intuitive web-based platform for complex analyses.

Overview

CoChat Summary

TL;DR

The MB-meta portal integrates clinical, molecular, and survival data from 898 medulloblastoma patients across three clinical trials into a web-based analysis platform. Using it, the authors confirm MYC amplification, G3 group, and metastatic disease as poor PFS indicators in G3/G4 tumors and refine KBTBD4 mutation patterns across methylation subgroups.

Key findings

  • The portal hosts 898 patients ages 3–22 from ACNS0331, ACNS0332, and SJMB03, with molecular groups WNT (n = 131), SHH (n = 151), G3 (n = 220), and G4 (n = 396); methylation data are available for all samples, sequencing for 754, and paired somatic/germline sequencing for 145.
  • In a G3/G4 PFS Cox analysis restricted to standard- or high-dose craniospinal irradiation and excluding low-dose or carboplatin-containing regimens, poorer outcomes were observed for MYC amplifications (univariate HR = 5.94; 95% CI, 3.46–10.2; multivariate HR = 5.39; 95% CI, 3.1–9.38), G3 group (univariate HR = 2.48; 95% CI, 1.75–3.513; multivariate HR = 2.186; 95% CI, 1.464–3.263), and metastatic disease M+ (univariate HR = 2.9; 95% CI, 2.05–4.1; multivariate HR = 2.58; 95% CI, 1.8–3.69); iso17, age group, and gender did not significantly affect PFS.
  • Manual review of nearly 700 tumors rescued KBTBD4 mutation calls in three samples and corrected mutation alleles for 21 tumors; 26 of 29 KBTBD4-mutated cases fell into two clusters, with 13 of 16 (81%) in the G3/G4-2 cluster harboring R313insPRR and 5 of 10 (50%) in the G3/G4-7 cluster harboring R312_R313insP.
  • MYC amplifications were not exclusive to G3: most occurred in G3/G4-2 and G3/G4-3, but a few were classified as G3/G4-5 and G3/G4-6 (G4), and one MYC-amplified G3 case was unclassified.

Methods

  • Retrospective cohort from three medulloblastoma clinical trials, including patients 3–22 years old; tumors were molecularly classified into WNT, SHH, G3, and G4 using DNA methylation profiles overlaid with next-generation sequencing mutations, with endpoints OS, EFS, and PFS.
  • The MB-meta portal provides a data dictionary and interactive visualizations; analyses used t-SNE clustering, Fisher exact tests, Kaplan–Meier/log-rank survival comparisons, and Cox proportional hazards models, with structural modeling via AlphaFold Colab and UCSF Chimera.
  • For the G3/G4 PFS example, the cohort was filtered to patients receiving standard- or high-dose craniospinal irradiation, excluding low-dose craniospinal irradiation and carboplatin-containing regimens; a custom iso17 variable was created from 17p loss and 17q gain.

Limitations

  • Stated: KBTBD4 findings should be experimentally validated; binding affinities of newly identified KBTBD4 cases remain to be experimentally validated; the effect of a frameshift mutation on transcription cannot be ascertained because of a lack of RNA sequencing data; and further investigation is required to determine whether the homozygous KBTBD4 mutation relates to M+ metastatic status.
  • Stated/evident: Some cases were unclassified or had low confidence scores (PARPDA CS = 0.47; SJMB067030 CS = 0.69; a G3/G4-3 case CS = 0.6; an overlapping unclassified case CS < 0.5), and one rescued case clustered closely with G3/G4-3 despite being classified as G3/G4-2.
  • Evident from text: the analyses are secondary analyses of existing trial cohorts, and no independent validation cohort is described in the provided text.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1158/0008-5472.can-24-4976

Cites

30 of 30 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1. Data dictionary variables in the MB-meta cohort. A, Breakdown of dictionary variables with different data types. Numbers in genomics data variables indicate the number of tumors with that data. B, Medulloblastoma cohort summaries for major clinical variables. AR, average risk; CM, clinico molecular; CSI, craniospinal irradiation; HDCSI, high-dose craniospinal irradiation; HR, high risk; LDCSI, low-dose craniospinal irradiation; LR, low risk; rCM, revised clinico molecular; SDCSI, standard-dose craniospinal irradiation; VHR, very high risk; WCA, whole chromosome aberration; WES, whole-exome sequencing; WGS, whole-genome sequencing.
Figure 1. Data dictionary variables in the MB-meta cohort. A, Breakdown of dictionary variables with different data types. Numbers in genomics data variables indicate the number of tumors with that data. B, Medulloblastoma cohort summaries for major clinical variables. AR, average risk; CM, clinico molecular; CSI, craniospinal irradiation; HDCSI, high-dose craniospinal irradiation; HR, high risk; LDCSI, low-dose craniospinal irradiation; LR, low risk; rCM, revised clinico molecular; SDCSI, standard-dose craniospinal irradiation; VHR, very high risk; WCA, whole chromosome aberration; WES, whole-exome sequencing; WGS, whole-genome sequencing.CC BY-NC-ND · PubMed Central
Figure 2. Overview of the visualizations in the MB-meta portal. A, Summary plot showing the “age at diagnosis” for different subgroups (shown by different colors). B, Kaplan–Meier curves comparing PFS between the molecular subgroups G3/G4-2 and G3/G4-7. P values are based on all survival data. C, Cox regression analyses showing risks for different molecular groups on progression of the disease. D, Lollipop plot showing CTNNB1 mutations for the WNT group. E, Sample matrix showing demographic, clinical, mutational, and copy-number variation data. Rows indicate variables, and columns indicate samples. F, Methylome t-SNE showing molecular subgroup classification, and lassoing shows selection of groups for survival analysis. G, Sample view showing complete variable annotation for a sample, genome-wide mutation, and copy-number variation profile in the disco plot. CNV, copy-number variation; CSI, craniospinal irradiation; IFRT, involved-field radiotherapy; LC/A, large cell/anaplastic; RT, radiotherapy. H, Sample view showing genome-wide copy-number variation plot derived by DNA methylation plot and locus view.
Figure 2. Overview of the visualizations in the MB-meta portal. A, Summary plot showing the “age at diagnosis” for different subgroups (shown by different colors). B, Kaplan–Meier curves comparing PFS between the molecular subgroups G3/G4-2 and G3/G4-7. P values are based on all survival data. C, Cox regression analyses showing risks for different molecular groups on progression of the disease. D, Lollipop plot showing CTNNB1 mutations for the WNT group. E, Sample matrix showing demographic, clinical, mutational, and copy-number variation data. Rows indicate variables, and columns indicate samples. F, Methylome t-SNE showing molecular subgroup classification, and lassoing shows selection of groups for survival analysis. G, Sample view showing complete variable annotation for a sample, genome-wide mutation, and copy-number variation profile in the disco plot. CNV, copy-number variation; CSI, craniospinal irradiation; IFRT, involved-field radiotherapy; LC/A, large cell/anaplastic; RT, radiotherapy. H, Sample view showing genome-wide copy-number variation plot derived by DNA methylation plot and locus view.CC BY-NC-ND · PubMed Central
Figure 3. Using custom variables for regression and survival analyses. A, G3/G4 cases filtered for G3/G4 patients who received regimens containing standard- or high-dose craniospinal irradiation without carboplatin during radiotherapy. B, Groups are created for different combinations of presence or absence of MYC amplifications and iso17 using custom variable functionality within the portal. CNV, copy-number variation; CSI, craniospinal irradiation; LDCSI, low-dose craniospinal irradiation. C, Univariate and multivariate Cox regression analyses using custom variables and other cofactors with PFS as the outcome. D, Survival plots launched from custom variables with the presence or absence of MYC amplifications and iso17. P values are based on all survival data.
Figure 3. Using custom variables for regression and survival analyses. A, G3/G4 cases filtered for G3/G4 patients who received regimens containing standard- or high-dose craniospinal irradiation without carboplatin during radiotherapy. B, Groups are created for different combinations of presence or absence of MYC amplifications and iso17 using custom variable functionality within the portal. CNV, copy-number variation; CSI, craniospinal irradiation; LDCSI, low-dose craniospinal irradiation. C, Univariate and multivariate Cox regression analyses using custom variables and other cofactors with PFS as the outcome. D, Survival plots launched from custom variables with the presence or absence of MYC amplifications and iso17. P values are based on all survival data.CC BY-NC-ND · PubMed Central
Figure 4. KBTBD4 mutations within medulloblastoma subgroups visualized on the methylome t-SNE. A, Lollipop plot showing all samples and mutations for KBTBD4. B, Methylome t-SNE showing molecular subgroup (as shapes) and KBTBD4 mutations (as colors). The t-SNE plot is limited to G3/G4 samples with verified KBTBD4 mutation status. Circles highlight two clusters of KBTBD4-mutated samples and are zoomed in F and D. C, Lollipop plot showing KBTBD4 mutations in subgroup G3/G4-2. Arrow, position of the hotspot mutation. Each distinct mutation is assigned a number (circled) and is used to label all mutated samples in D and F. D, Subgroup G3/G4-2 KBTBD4 mutations visualized on the t-SNE plot, with samples with rescued mutations highlighted in cyan boxes and by arrows. E, Lollipop plot showing KBTBD4 mutations in subgroup G3/G4-7, with the rescued mutation highlighted in cyan. Arrow, position of the hotspot mutation. F, Subgroup G3/G4-7 KBTBD4 mutations visualized on the t-SNE plot. A sample with rescued mutation is highlighted in cyan. MB, medulloblastoma; MNP, molecular neuropathology; MYO, medulloblastoma with myogenic differentiation.
Figure 4. KBTBD4 mutations within medulloblastoma subgroups visualized on the methylome t-SNE. A, Lollipop plot showing all samples and mutations for KBTBD4. B, Methylome t-SNE showing molecular subgroup (as shapes) and KBTBD4 mutations (as colors). The t-SNE plot is limited to G3/G4 samples with verified KBTBD4 mutation status. Circles highlight two clusters of KBTBD4-mutated samples and are zoomed in F and D. C, Lollipop plot showing KBTBD4 mutations in subgroup G3/G4-2. Arrow, position of the hotspot mutation. Each distinct mutation is assigned a number (circled) and is used to label all mutated samples in D and F. D, Subgroup G3/G4-2 KBTBD4 mutations visualized on the t-SNE plot, with samples with rescued mutations highlighted in cyan boxes and by arrows. E, Lollipop plot showing KBTBD4 mutations in subgroup G3/G4-7, with the rescued mutation highlighted in cyan. Arrow, position of the hotspot mutation. F, Subgroup G3/G4-7 KBTBD4 mutations visualized on the t-SNE plot. A sample with rescued mutation is highlighted in cyan. MB, medulloblastoma; MNP, molecular neuropathology; MYO, medulloblastoma with myogenic differentiation.CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

8 Valid · 4 Not valid · 1 Not checked · 45 Own results and interpretations

24 more claims

Create a free account to see all 58 claims and every source verdict.

See all claims — free account
  • It may state two separate claims as one. 22
  • Probably not a substantial claim. 10
  • The quote only partly states this claim. 1
  • Borderline: one of the tests for a substantial claim is unclear. 1

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 30 references

Cited by

See all 1

Metadata

Authors
Gangwani, K., Lu, C., Dhanda, S. K., Sioson, E., Zaldívar Peraza, A., Matt, G. Y., Paul, R., Wang, J., Reilly, C., Acić, A., Smith, K. S., Lin, T., Li, Q., Northcott, P. A., Robinson, G. W., & Zhou, X.
Affiliations
  • St. Jude Children's Research Hospital (US)
  • 1Department of Computational Biology, St. Jude Children’s Research Hospital, Memphis, Tennessee.
Corresponding author
  • Robinson, Giles W.
  • Zhou, Xin
Venue
Cancer Research
Published
2025-11-05
Volume
85
Issue
24
Pages
4890-4898
ISSN
  • 0008-5472 (ISSN-L)
  • 1538-7445
OpenAlex
W4415926276
Cited by
1
References
30
Citations by year
2026: 1
FWCI
0.15
Type
journal article
Language
en
Open access
Locations
  • Cancer Research (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
Funders
National Cancer Institute (P30CA021765)
Topics
  • Health Sciences › Medicine › Oncology › Glioma Diagnosis and Treatment
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cancer Research › Cancer Genomics and Diagnostics
  • Health Sciences › Medicine › Oncology › Neuroblastoma Research and Treatments
Keywords
medulloblastoma, pediatric brain tumors, web-based data portal, DNA methylation, tumor subtypes, survival outcomes, risk stratification, molecular classification, Cox proportional hazards model, clinical trials
MeSH terms
Cerebellar Neoplasms, Child, Humans, Medulloblastoma, Mutation, Prognosis, DNA Methylation, Genomics

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: The Medulloblastoma Meta-Analysis Portal Enables Integrative Exploration of Clinical and Genomic Data from Patients. https://app.cochat.ai/papers/10.1158/0008-5472.can-24-4976
The paper
Gangwani, K., Lu, C., Dhanda, S. K., Sioson, E., Zaldívar Peraza, A., Matt, G. Y., Paul, R., Wang, J., Reilly, C., Acić, A., Smith, K. S., Lin, T., Li, Q., Northcott, P. A., Robinson, G. W., & Zhou, X. (2025). The Medulloblastoma Meta-Analysis Portal Enables Integrative Exploration of Clinical and Genomic Data from Patients. Cancer Research. https://doi.org/10.1158/0008-5472.can-24-4976

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.