Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2025
|
| Online Access: | https://doi.org/10.1038/s43018-024-00772-7 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866901755065794560 |
|---|---|
| author | Morais Lyra Junior, Paulo Cilas |
| author_facet | Morais Lyra Junior, Paulo Cilas |
| contents | <p dir="auto"><strong>Zenodo Repository Description (final version – ready to paste)</strong></p> <p dir="auto">This dataset contains the patient-level clinical, pathological, and genomic features from the multi-cancer immune checkpoint blockade cohort originally compiled and analyzed in:</p> <p dir="auto"><strong>Chowell D, Yoo SK, Valero C, et al.</strong> Improved prediction of immune checkpoint blockade efficacy across multiple cancer types. <em>Nature Biotechnology</em> 2022 Apr;40(4):499–506. doi: 10.1038/s41587-021-01070-8 PMID: 34725502 | PMCID: PMC9363980</p> <p dir="auto">The cohort has become a standard benchmark for pan-cancer ICB response modeling and was notably used to develop and validate the LORIS predictor:</p> <p dir="auto"><strong>Chang TG, Cao Y, Sfreddo HJ, et al.</strong> LORIS robustly predicts patient outcomes with immune checkpoint blockade therapy using common clinical, pathologic and genomic features. <em>Nature Cancer</em> 2024;5:1158–1175. doi: 10.1038/s43018-024-00772-7</p> <p dir="auto"><strong>Files provided (preprocessed and ready-to-use)</strong></p> <ul> <li>Chowell_train.tsv – 515 patients (original training split)</li> <li>Chowell_test.tsv – 964 patients (original held-out test split)</li> </ul> <p dir="auto">Both files contain the same 21 predictive features (TMB, Systemic_therapy_history, Albumin, NLR, Age, and one-hot encoded CancerType1–CancerType16) plus the binary outcome column Response (1 = objective response [complete or partial] by RECIST; 0 = no response).</p> <p dir="auto"><strong>Preprocessing applied</strong> These TSV files have been uniformly preprocessed with the following reproducible steps (code publicly available at <a href="https://github.com/goeckslab/gleam_use_cases">GLEAM Use Cases</a>):</p> <ul> <li>Extraction of the exact Chowell_train and Chowell_test sheets from the original AllData.xlsx file hosted by the LORIS authors</li> <li>Selection of the 21 features used in the published LORIS model</li> <li>Conservative outlier clipping: TMB ≤ 50, Age ≤ 85, NLR ≤ 25 (identical to LORIS paper)</li> <li>Export as clean, delimiter-separated TSV files with the target column named Response</li> </ul> <p dir="auto">These are therefore the exact train/test splits employed in both the original Chowell et al. (2022) study and all subsequent LORIS benchmarking experiments, guaranteeing full reproducibility and fair model comparison.</p> <p dir="auto"><strong>License</strong> Shared for academic and non-commercial research use only, in accordance with the policies of the original publications.</p> <p dir="auto"><strong>Keywords</strong> immune checkpoint blockade, cancer immunotherapy, tumor mutational burden, pan-cancer, ICB response prediction, machine learning benchmark, LORIS, Chowell cohort, tabular data</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_1038_s43018-024-00772-7 |
| institution | Zenodo |
| language | |
| publishDate | 2025 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | GLEAM Tabular Learner Use Case: Chowell dataset Morais Lyra Junior, Paulo Cilas <p dir="auto"><strong>Zenodo Repository Description (final version – ready to paste)</strong></p> <p dir="auto">This dataset contains the patient-level clinical, pathological, and genomic features from the multi-cancer immune checkpoint blockade cohort originally compiled and analyzed in:</p> <p dir="auto"><strong>Chowell D, Yoo SK, Valero C, et al.</strong> Improved prediction of immune checkpoint blockade efficacy across multiple cancer types. <em>Nature Biotechnology</em> 2022 Apr;40(4):499–506. doi: 10.1038/s41587-021-01070-8 PMID: 34725502 | PMCID: PMC9363980</p> <p dir="auto">The cohort has become a standard benchmark for pan-cancer ICB response modeling and was notably used to develop and validate the LORIS predictor:</p> <p dir="auto"><strong>Chang TG, Cao Y, Sfreddo HJ, et al.</strong> LORIS robustly predicts patient outcomes with immune checkpoint blockade therapy using common clinical, pathologic and genomic features. <em>Nature Cancer</em> 2024;5:1158–1175. doi: 10.1038/s43018-024-00772-7</p> <p dir="auto"><strong>Files provided (preprocessed and ready-to-use)</strong></p> <ul> <li>Chowell_train.tsv – 515 patients (original training split)</li> <li>Chowell_test.tsv – 964 patients (original held-out test split)</li> </ul> <p dir="auto">Both files contain the same 21 predictive features (TMB, Systemic_therapy_history, Albumin, NLR, Age, and one-hot encoded CancerType1–CancerType16) plus the binary outcome column Response (1 = objective response [complete or partial] by RECIST; 0 = no response).</p> <p dir="auto"><strong>Preprocessing applied</strong> These TSV files have been uniformly preprocessed with the following reproducible steps (code publicly available at <a href="https://github.com/goeckslab/gleam_use_cases">GLEAM Use Cases</a>):</p> <ul> <li>Extraction of the exact Chowell_train and Chowell_test sheets from the original AllData.xlsx file hosted by the LORIS authors</li> <li>Selection of the 21 features used in the published LORIS model</li> <li>Conservative outlier clipping: TMB ≤ 50, Age ≤ 85, NLR ≤ 25 (identical to LORIS paper)</li> <li>Export as clean, delimiter-separated TSV files with the target column named Response</li> </ul> <p dir="auto">These are therefore the exact train/test splits employed in both the original Chowell et al. (2022) study and all subsequent LORIS benchmarking experiments, guaranteeing full reproducibility and fair model comparison.</p> <p dir="auto"><strong>License</strong> Shared for academic and non-commercial research use only, in accordance with the policies of the original publications.</p> <p dir="auto"><strong>Keywords</strong> immune checkpoint blockade, cancer immunotherapy, tumor mutational burden, pan-cancer, ICB response prediction, machine learning benchmark, LORIS, Chowell cohort, tabular data</p> |
| title | GLEAM Tabular Learner Use Case: Chowell dataset |
| url | https://doi.org/10.1038/s43018-024-00772-7 |