xpertsystems/hconc010-sample

HC-ONC-010 — 卵巢癌合成队列（样本）是一个完全合成的上皮性卵巢癌数据集，涵盖所有6种组织类型：高级别浆液性癌（HGSOC，约70%）、低级别浆液性癌（LGSOC，约5%）、透明细胞癌（约8%）、粘液性癌（约5%）、子宫内膜样癌（约10%）和混合/其他类型（约2%）。数据集基于FIGO 2014分期（从IA到IVB子阶段），包含全面的BRCA1/2种系和体细胞分析、通过Myriad myChoice CDx进行的HRD评分、组织类型特异性分子标记（如HGSOC中的TP53标志、粘液性癌/LGSOC中的KRAS等）、叶酸受体α（FRalpha）表达（用于mirvetuximab资格评估）、细胞减灭手术（包括初次减瘤术PDS、间歇减瘤术IDS和二次手术）及其Aletti复杂性评分和残留疾病（R0/R1/R2）状态、CA-125纵向动力学（每患者121个月的月度序列，含Rustin复发标准和HEGP动力学分类）、铂类化疗（含贝伐珠单抗组合）及其RECIST 1.1反应、无铂间期（PFI）和铂敏感性分类、PARP抑制剂维持治疗（如奥拉帕利SOLO-1、尼拉帕利PRIMA等）及其锚定PFS分布、IV型mirvetuximab soravtansine（FRalpha靶向ADC）、Weibull校准的生存终点（OS、PFS、PFI、TTNT、DOR）、复发模式和继发原发性癌症监测。数据集旨在直接用于妇科肿瘤分析、建模、演示和教育，同时保持100%合成，无真实患者数据、无PHI、无重新识别风险。样本包括500名患者的主要队列和500行CA-125 121个月纵向序列，数据格式为CSV，遵循CC-BY-NC-4.0许可。

HC-ONC-010 — Ovarian Cancer Synthetic Cohort (sample) is a fully synthetic epithelial ovarian cancer dataset spanning all 6 histotypes: High-Grade Serous (HGSOC ~70%), Low-Grade Serous (LGSOC ~5%), Clear Cell (~8%), Mucinous (~5%), Endometrioid (~10%), and Mixed/Other (~2%). It incorporates FIGO 2014 staging (IA → IVB with substages), comprehensive BRCA1/2 germline and somatic profiling, HRD scoring via Myriad myChoice CDx, histotype-specific molecular markers (e.g., TP53 hallmark in HGSOC, KRAS in Mucinous/LGSOC), Folate Receptor Alpha (FRalpha) expression for mirvetuximab eligibility, cytoreductive surgery (Primary Debulking PDS / Interval Debulking IDS / Secondary) with Aletti complexity scoring and residual disease (R0/R1/R2), CA-125 longitudinal kinetics (121-month monthly series per patient with Rustin relapse criteria and HEGP kinetic classification), platinum-based chemotherapy (including bevacizumab combinations) with RECIST 1.1 response, Platinum-Free Interval (PFI) and platinum sensitivity classification, PARP inhibitor maintenance (e.g., Olaparib SOLO-1, Niraparib PRIMA) with anchored PFS distributions, IV mirvetuximab soravtansine (FRalpha-targeted ADC), Weibull-calibrated survival endpoints (OS, PFS, PFI, TTNT, DOR), relapse patterns, and secondary primary cancer surveillance. Built to be drop-in usable for gynecologic oncology analytics, modeling, demos, and education while remaining 100% synthetic — no real patient data, no PHI, no re-identification risk. The sample includes a 500-patient primary cohort and 500-row CA-125 121-month longitudinal series in CSV format, licensed under CC-BY-NC-4.0.