README.md

---
task_categories:
- tabular-classification
tags:
- tabular
- cancer
- immunotherapy
- immune-checkpoint-blockade
- biology
- health
pretty_name: "Chowell et al. 2021 - ICB Efficacy Prediction"
size_categories:
- 1K<n<10K
dataset_info:
  - config_name: default
    features:
      - name: SAMPLE_ID
        dtype: int64
      - name: Cancer_type_grouped_2
        dtype: string
      - name: Cancer_Type2
        dtype: int64
      - name: Cancer_Type
        dtype: string
      - name: "Chemo_before_IO (1:Yes; 0:No)"
        dtype: int64
      - name: Age
        dtype: float64
      - name: "Sex (1:Male; 0:Female)"
        dtype: int64
      - name: BMI
        dtype: float64
      - name: "Stage (1:IV; 0:I-III)"
        dtype: int64
      - name: Stage at IO start
        dtype: string
      - name: NLR
        dtype: float64
      - name: Platelets
        dtype: int64
      - name: HGB
        dtype: float64
      - name: Albumin
        dtype: float64
      - name: "Drug (1:Combo; 0:PD1/PDL1orCTLA4)"
        dtype: int64
      - name: Drug_class
        dtype: string
      - name: TMB
        dtype: float64
      - name: FCNA
        dtype: float64
      - name: HED
        dtype: float64
      - name: HLA_LOH
        dtype: int64
      - name: "MSI (1:Unstable; 0:Stable_Indeterminate)"
        dtype: int64
      - name: MSI_SCORE
        dtype: float64
      - name: "Response (1:Responder; 0:Non-responder)"
        dtype: int64
      - name: OS_Event
        dtype: int64
      - name: OS_Months
        dtype: float64
      - name: PFS_Event
        dtype: int64
      - name: PFS_Months
        dtype: float64
      - name: RF16_prob
        dtype: float64
    splits:
      - name: train
        num_examples: 1184
      - name: test
        num_examples: 295
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train.parquet
      - split: test
        path: data/test.parquet
---

# Improved prediction of immune checkpoint blockade efficacy across multiple cancer types

## Description

This dataset contains genomic, molecular, demographic, and clinical data from 1,479 patients treated with immune checkpoint blockade (ICB) across 16 different cancer types. The data was curated from the MSK-IMPACT cohort and used to develop a machine learning model for predicting ICB treatment response.

The dataset is provided as the supplementary data (Supplementary Table 3) from:

> Chowell, D., Yoo, S.K., Valero, C. et al. Improved prediction of immune checkpoint blockade efficacy across multiple cancer types. *Nat Biotechnol* **40**, 499--506 (2022). [https://doi.org/10.1038/s41587-021-01070-8](https://doi.org/10.1038/s41587-021-01070-8)

The original XLSX file (`41587_2021_1070_MOESM3_ESM.xlsx`) is included in this repository and is also available directly from the publisher at [Supplementary Table 3](https://static-content.springer.com/esm/art%3A10.1038%2Fs41587-021-01070-8/MediaObjects/41587_2021_1070_MOESM3_ESM.xlsx). The parquet files in `data/` were converted from the "Training" and "Test" sheets in the XLSX file.

## Abstract

Only a fraction of patients with cancer respond to immune checkpoint blockade (ICB) treatment, but current decision-making procedures have limited accuracy. In this study, we developed a machine learning model to predict ICB response by integrating genomic, molecular, demographic and clinical data from a comprehensively curated cohort (MSK-IMPACT) with 1,479 patients treated with ICB across 16 different cancer types. In a retrospective analysis, the model achieved high sensitivity and specificity in predicting clinical response to immunotherapy and predicted both overall survival and progression-free survival in the test data across different cancer types. Our model significantly outperformed predictions based on tumor mutational burden, which was recently approved by the U.S. Food and Drug Administration for this purpose. Additionally, the model provides quantitative assessments of the model features that are most salient for the predictions. We anticipate that this approach will substantially improve clinical decision-making in immunotherapy and inform future interventions.

## Dataset Structure

The dataset is split into training (1,184 samples) and test (295 samples) sets, preserving the original splits from the publication.

### Features

| Feature | Type | Description |
|---|---|---|
| `SAMPLE_ID` | int | Unique sample identifier |
| `Cancer_type_grouped_2` | string | Cancer type (16 categories: Bladder, Breast, Colorectal, Endometrial, Esophageal, Gastric, Head & Neck, Hepatobiliary, Melanoma, Mesothelioma, NSCLC, Ovarian, Pancreatic, Renal, Sarcoma, SCLC) |
| `Cancer_Type2` | int | Encoded cancer type (0, 1, 2) |
| `Cancer_Type` | string | Grouped cancer type (Melanoma, NSCLC, Others) |
| `Chemo_before_IO (1:Yes; 0:No)` | int | Whether chemotherapy was administered before immunotherapy |
| `Age` | float | Patient age |
| `Sex (1:Male; 0:Female)` | int | Patient sex |
| `BMI` | float | Body mass index |
| `Stage (1:IV; 0:I-III)` | int | Cancer stage |
| `Stage at IO start` | string | Detailed stage at immunotherapy start |
| `NLR` | float | Neutrophil-to-lymphocyte ratio |
| `Platelets` | int | Platelet count |
| `HGB` | float | Hemoglobin level |
| `Albumin` | float | Albumin level |
| `Drug (1:Combo; 0:PD1/PDL1orCTLA4)` | int | Drug type (combination vs. single agent) |
| `Drug_class` | string | Drug class description |
| `TMB` | float | Tumor mutational burden |
| `FCNA` | float | Fraction of copy number alterations |
| `HED` | float | HLA evolutionary divergence |
| `HLA_LOH` | int | HLA loss of heterozygosity |
| `MSI (1:Unstable; 0:Stable_Indeterminate)` | int | Microsatellite instability status |
| `MSI_SCORE` | float | Microsatellite instability score |
| `Response (1:Responder; 0:Non-responder)` | int | Clinical response to immunotherapy |
| `OS_Event` | int | Overall survival event indicator |
| `OS_Months` | float | Overall survival in months |
| `PFS_Event` | int | Progression-free survival event indicator |
| `PFS_Months` | float | Progression-free survival in months |
| `RF16_prob` | float | Random forest model predicted probability |

## Source

- **Publication**: [Nature Biotechnology](https://www.nature.com/articles/s41587-021-01070-8)
- **Supplementary Data**: [Supplementary Table 3 (XLSX)](https://static-content.springer.com/esm/art%3A10.1038%2Fs41587-021-01070-8/MediaObjects/41587_2021_1070_MOESM3_ESM.xlsx)
- **PubMed**: [34725502](https://pubmed.ncbi.nlm.nih.gov/34725502/)
- **PMC**: [PMC9363980](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9363980/)

## Citation

```bibtex
@article{chowell2022improved,
  title={Improved prediction of immune checkpoint blockade efficacy across multiple cancer types},
  author={Chowell, Diego and Yoo, Seong-Keun and Valero, Cristina and Pastore, Alessandro and Krishna, Chirag and Lee, Mark and Hoen, Douglas and Shi, Hongyu and Kelly, Daniel W and Patel, Neal and Makarov, Vladimir and Ma, Xiaoxiao and Vuong, Lynda and Sabio, Erich Y and Weiss, Kate and Kuo, Fengshen and Lenz, Tobias L and Samstein, Robert M and Riaz, Nadeem and Adusumilli, Prasad S and Balachandran, Vinod P and Plitas, George and Hakimi, A Ari and Abdel-Wahab, Omar and Shoushtari, Alexander N and Postow, Michael A and Motzer, Robert J and Ladanyi, Marc and Zehir, Ahmet and Berger, Michael F and G{\"o}nen, Mithat and Morris, Luc G T and Weinhold, Nils and Chan, Timothy A},
  journal={Nature Biotechnology},
  volume={40},
  number={4},
  pages={499--506},
  year={2022},
  publisher={Nature Publishing Group},
  doi={10.1038/s41587-021-01070-8}
}
```