REAME.md

---
language: en
license: cc-by-nc-4.0
tags:
- medical
- oncology
- genomics
- synthetic-data
- ehr
- electronic-health-records
pretty_name: Clinical EHR & Somatic Genomics Dataset (Breast Cancer Cohort)
size_categories:
- 10K<n<100K
task_categories:
- text-generation
- token-classification
---

# 🧬 EHR & Somatic Genomics Synthetic Oncology Dataset (AnodeAI)

This dataset contains highly structured, multi-modal synthetic clinical oncology records coupled with Next-Generation Sequencing (NGS) somatic variant profiles and longitudinal electronic health records (EHR). 

Unlike standard stochastic mock data, this dataset is engineered via a **deterministic clinical logic engine** that guarantees 100% biological and pharmacological integrity across all fields.

---

## 💡 Key Architectural Highlights
* **Zero Hallucinations:** Strict internal constraints prevent medical contradictions (e.g., Triple-Negative profiles will never mistakenly receive HER2-targeted therapies).
* **Multi-Modal Synchronization:** Seamlessly unifies three typically isolated clinical data silos: clinical staging/phenotypes, digital molecular diagnostics (IHC/NGS), and longitudinal real-world evidence (RWE).
* **Mathematical Outcome Correlation:** Progression-Free Survival (PFS) metrics are statistically correlated with Response Evaluation Criteria in Solid Tumors (RECIST 1.1) values to preserve downstream machine learning utility.

---

## 📊 Schema & Clinical Logic Deep Dive

### 1. Patient Demographics
Contains basic baseline parameters for tracking patient cohorts.
* `patient_id`: Unique tracking identifier prefixed for institutional indexing.
* `age` / `gender` / `ethnicity`: Demographically balanced variables tailored to replicate true epidemiological distributions in oncology.

### 2. Clinical Oncology Profile
Establishes the specific breast cancer phenotypic anchor, which strictly drives the rest of the record:
* **Hormone Receptor Positive (HR+/HER2-)**: Characterized by elevated ER/PR percentages and negative HER2 IHC statuses.
* **HER2 Positive (HER2+)**: Configured with amplified HER2 expressions (3+ by IHC) and varying hormone receptor profiles.
* **Triple-Negative Breast Cancer (TNBC)**: Enforces absolute zero/negative values for ER, PR, and HER2 receptors.

### 3. Genomic Sequencing Data
Simulates a targeted Next-Generation Sequencing (NGS) assay reporting somatic variants matching the patient's biological subtype:
* **HR+/HER2- Pool**: Features highly relevant driver alterations such as *PIK3CA* (e.g., `p.E545K`, `p.H1047R`) and *ESR1* (`p.D538G`).
* **HER2+ Pool**: Maps copy-number variations like *ERBB2* amplifications alongside *TP53* mutations.
* **TNBC Pool**: Integrates pathogenic tumor-suppressor variants in *BRCA1*, *BRCA2*, and *TP53*.
* `variant_allele_frequency_vaf`: Synthesized as a continuous float value ($0.05 \le \text{VAF} \le 0.65$) indicating structural clone density.

### 4. Longitudinal Treatment & RWE
Tracks the active first-line therapeutic journey, mapping outcomes precisely to standard oncology reporting frameworks:
* `regimen_administered`: Pairs standard-of-care protocols seamlessly to biomarkers (e.g., CDK4/6 inhibitors for HR+, anti-HER2 monoclonal antibodies for HER2+, checkpoint inhibitors for TNBC).
* `adverse_events_ctcae_v5`: Captures severe toxicities (Grades 2-4) linked explicitly to the administered drug toxicological profile, including exact clinical intervention steps (`action_taken`).
* `treatment_response_recist_1_1`: Classifies patient response into standard categories (`CR`, `PR`, `SD`, `PD`).
* `progression_free_survival_months`: Employs hard-coded survival boundaries to match the clinical response (e.g., restricting progressive disease to rapid relapse ranges while scaling complete responses past 2-5 years).

---

> ## ⚡ COMMERCIAL LICENSING & ENTERPRISE ACCESS
> 
> **IMPORTANT NOTICE:** The dataset version hosted on Hugging Face is published under the **Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)** license. It is strictly limited to academic research, open-source benchmarking, and local sandbox prototyping.
>
> ### 🚀 Looking for Commercial Scale or Custom Disease Cohorts?
> If your organization requires high-fidelity synthetic clinical data for commercial production, enterprise AI training pipelines, or software dashboard validation, we offer scalable production engines.
> 
> * **Production Volumes:** Instantly generate from 10,000 to 1,000,000+ perfectly clean, fully validated clinical records.
> * **Expanded Pathologies:** Custom logic modules available for **Non-Small Cell Lung Cancer (NSCLC)**, **Colorectal Cancer (CRC)**, **Melanoma**, and more.
> * **Standardized Medical Ontologies:** Optional integration of explicit medical coding systems including **ICD-10-CM**, **SNOMED CT**, **RxNorm**, **HGVS**, and **CTCAE numerical codes** for direct production pipeline compatibility.
> 
> 📧 **Acquire a Commercial License:** To unlock unrestricted commercial usage rights, purchase full enterprise cohorts, or request bespoke oncology schema adaptations, please contact the engineering team directly at **anodeai.corp@gmail.com** (or via your platform's official contact portal).