Kossisoroyce's picture
Upload README.md with huggingface_hub
06323d7 verified
---
language:
- en
tags:
- systemic-inflammation
- NLR
- lymphocytes
- CRP
- ESR
- breast-cancer
- sub-saharan-africa
license: cc-by-nc-4.0
pretty_name: SSA Breast Systemic Immune Markers Dataset (Women, Multi-ancestry)
task_categories:
- other
size_categories:
- 1K<n<10K
---
# SSA Breast Systemic Immune Markers Dataset (Women, Multi-ancestry, Synthetic)
## Dataset summary
This dataset provides a **synthetic cohort of invasive breast cancers** in women across multiple ancestry groups, with emphasis on **sub-Saharan Africa (SSA)** and comparable reference populations.
Each tumour is linked to **systemic immune and inflammation markers** derived from complete blood counts and standard laboratory tests:
- **Neutrophil-to-lymphocyte ratio (NLR)** – category and continuous value.
- **Lymphocyte counts** – category and continuous value (x10⁹/L).
- **Neutrophil counts** – continuous value (x10⁹/L) to maintain internal consistency with NLR.
- **C-reactive protein (CRP)** – category and continuous value (mg/L).
- **Erythrocyte sedimentation rate (ESR)** – category and continuous value (mm/hr).
Distributions are qualitatively anchored to global and SSA-focused literature on NLR, lymphopenia, CRP, and ESR in cancer patients, while ensuring that **all records are fully synthetic** and non-identifiable.
## Cohort design
### Sample size and populations
- **Total N**: 10,000 synthetic invasive breast cancers.
- **Populations**:
- `SSA_West`: 2,000
- `SSA_East`: 2,000
- `SSA_Central`: 1,500
- `SSA_Southern`: 1,500
- `AAW` (African American women): 1,500
- `EUR` (European reference): 1,000
- `EAS` (East Asian reference): 500
- **Sex**:
- Predominantly `Female`, with a small fraction of male breast cancers (~1%).
- **Age**:
- 18–90 years.
- Older mean age in EUR/EAS/AAW vs somewhat younger in SSA cohorts, consistent with registry patterns.
## Systemic immune markers
### Neutrophil-to-lymphocyte ratio (NLR)
Variables:
- `nlr_category``Low`, `Intermediate`, `High`.
- `nlr_value` – continuous NLR value.
- Derived: `high_nlr``True` if `nlr_category == "High"` or `nlr_value ≥ 4.0`.
Anchoring:
- Meta-analyses in breast cancer identify high NLR (cut-offs ≈3–4) in **~25–40%** of patients, associated with worse outcomes.
- In this dataset:
- SSA and AAW populations have higher fractions of `High` NLR (~27–31%).
- EUR/EAS references have lower `High` NLR (~15%).
NLR is constructed from separately sampled neutrophil and lymphocyte counts but constrained to match the target category distribution.
### Lymphocyte counts
Variables:
- `lymphocyte_category``Low`, `Normal`, `High`.
- `lymphocyte_count_x10e9_per_L` – continuous lymphocyte count.
- Derived: `lymphopenia``True` if `lymphocyte_category == "Low"` or `lymphocyte_count_x10e9_per_L < 1.0`.
Anchoring:
- Lymphopenia (low lymphocyte count) is observed in **~15–25%** of oncology patients at diagnosis or during treatment.
- SSA and AAW cohorts have slightly **higher `Low` lymphocyte fractions** (~21–24%) than EUR/EAS (~15–16%).
Counts are sampled from category-specific distributions in the approximate reference ranges of adult lymphocyte counts.
### Neutrophil counts
Variable:
- `neutrophil_count_x10e9_per_L` – continuous neutrophil count.
Anchoring:
- Elevated neutrophils occur in approximately **20–30%** of cancer patients with systemic inflammation.
- Neutrophil counts are drawn to produce realistic NLR ranges while remaining within plausible haematology intervals.
### C-reactive protein (CRP)
Variables:
- `crp_category``Normal`, `Mildly_elevated`, `Markedly_elevated`.
- `crp_mg_per_L` – continuous CRP value.
Anchoring:
- Elevated CRP (>10 mg/L) occurs in **~25–40%** of breast and other cancer patients.
- SSA and AAW populations have more `Markedly_elevated` CRP (~26–32%) vs EUR/EAS (~15%).
Continuous CRP values are drawn from category-specific distributions, e.g. Normal ~0–5 mg/L, Mildly elevated ~3–20 mg/L, Markedly elevated ~10–200 mg/L.
### Erythrocyte sedimentation rate (ESR)
Variables:
- `esr_category``Normal`, `Mildly_elevated`, `Markedly_elevated`.
- `esr_mm_per_hr` – continuous ESR value.
Anchoring:
- High ESR (>40–50 mm/hr) is reported in **~25–40%** of oncology patients.
- SSA and AAW cohorts are modeled with higher `Markedly_elevated` ESR fractions (≈26–31%) vs EUR/EAS (~15%).
ESR values are generated using category-specific distributions spanning normal (~0–25 mm/hr), mildly elevated (~15–60 mm/hr), and markedly elevated (~30–120 mm/hr) ranges.
### Composite inflammatory burden
Variable:
- `high_inflammatory_burden``True` if **any** of the following holds:
- `high_nlr` is `True`.
- `crp_category == "Markedly_elevated"`.
- `esr_category == "Markedly_elevated"`.
This flag approximates a **high systemic inflammatory state** for risk stratification and modeling.
## File and schema
### `systemic_immune_markers_data.parquet` / `systemic_immune_markers_data.csv`
Each row represents a synthetic breast cancer case with demographics and systemic markers:
- **Demographics**
- `sample_id`
- `population`
- `region`
- `is_SSA`
- `is_reference_panel`
- `sex`
- `age`
- **NLR and counts**
- `nlr_category`, `nlr_value`, `high_nlr`
- `lymphocyte_category`, `lymphocyte_count_x10e9_per_L`, `lymphopenia`
- `neutrophil_count_x10e9_per_L`
- **CRP and ESR**
- `crp_category`, `crp_mg_per_L`
- `esr_category`, `esr_mm_per_hr`
- **Composite marker**
- `high_inflammatory_burden`
## Generation
The dataset is generated using:
- `systemic_immune_markers/scripts/generate_systemic_immune_markers.py`
with configuration in:
- `systemic_immune_markers/configs/systemic_immune_markers_config.yaml`
and literature inventory in:
- `systemic_immune_markers/docs/LITERATURE_INVENTORY.csv`
Key steps:
1. **Cohort sampling** – multi-ancestry invasive breast cancer cohort with age/sex by population.
2. **Lymphocyte assignment** – sample `lymphocyte_category` by population and draw continuous counts.
3. **NLR assignment** – sample `nlr_category` by population; draw neutrophil counts and compute `nlr_value` consistent with the category.
4. **CRP and ESR assignment** – sample categories by population and generate continuous values within plausible clinical ranges.
5. **Derived flags** – compute `high_nlr`, `lymphopenia`, and `high_inflammatory_burden`.
## Validation
Validation is performed with:
- `systemic_immune_markers/scripts/validate_systemic_immune_markers.py`
and summarized in:
- `systemic_immune_markers/output/validation_report.md`
Checks include:
- **C01–C02** – Sample size and population counts vs config.
- **C03** – NLR category distributions by population.
- **C04** – Lymphocyte category distributions by population.
- **C05** – CRP category distributions by population.
- **C06** – ESR category distributions by population.
- **C07** – Missingness across demographics, NLR/lymphocytes, CRP/ESR, and composite flags.
The released version is configured to stay within a **10% absolute deviation tolerance** for categorical distributions, with an **overall validation status of `PASS`**.
## Intended use
This dataset is intended for:
- **Risk stratification and prognostic modeling** using systemic inflammatory markers.
- **Integration** with other Electric Sheep Africa synthetic modules (pathology, IHC, immune profiles, comorbidities, environmental exposures) to build multi-modal models.
- **Educational use** for teaching relationships between NLR, lymphocyte counts, CRP/ESR, and cancer outcomes across ancestries.
It is **not intended** for:
- Estimating true prevalence of high NLR or elevated CRP/ESR in any specific population.
- Direct clinical decision-making or triage.
## Ethical considerations
- No real patient data are used; all cohorts and markers are simulated.
- Differences in systemic inflammation between populations reflect literature-informed trends and must not be used to stigmatize or essentialize groups.
- Users should interpret modeled patterns alongside high-quality epidemiological data and local clinical context.
## License
- License: **CC BY-NC 4.0**.
- Free for non-commercial research, method development, and education with attribution.
## Citation
If you use this dataset, please cite:
> Electric Sheep Africa. "SSA Breast Systemic Immune Markers Dataset (Women, Multi-ancestry, Synthetic)." Hugging Face Datasets.
and, as appropriate, key literature on NLR, lymphocyte counts, CRP, and ESR in breast cancer and other solid tumours, including studies from Sub-Saharan Africa.