xpertsystems/hconc004-sample

这是一个完全合成的结直肠癌队列数据集，涵盖了完整的临床路径：基于AJCC第8版T/N/M分期（涵盖结肠和直肠亚部位）、全面的分子标记（如MSI/MMR、KRAS密码子12和13、NRAS、BRAF V600E等）、手术结果（如R状态、吻合口漏、淋巴结收获等）、化疗方案（包括FOLFOX/CAPOX/FOLFIRI等）、RECIST反应及深度反应、CEA动态（基线、最低点、反应/进展标志）、生存终点（如OS/DFS/PFS/复发）和生活质量指标（如EORTC QLQ-C30）。数据集还包括一个可变长度的CEA纵向面板（在10年内18个时间点，根据总生存期截断）。该数据集旨在为分析、建模、演示和教育提供即用型数据，同时保持100%合成性，无真实患者数据、无PHI、无重新识别风险。数据集样本包括500名患者的主要队列（105列）和约3,300行的CEA面板（4列），格式为CSV，许可证为CC-BY-NC-4.0。

A fully synthetic colorectal cancer cohort spanning the complete clinical pathway: AJCC 8th Edition T/N/M staging across colon + rectum subsites, comprehensive molecular markers (MSI/MMR, KRAS codons 12 & 13, NRAS, BRAF V600E with MSI-H enrichment, HER2 IHC + amplification, PIK3CA, TP53, APC, SMAD4, TMB, PD-L1 CPS, NTRK/RET fusions, ctDNA with VAF), surgical outcomes (R-status, anastomotic leak, lymph node harvest with NCCN adequacy, operative time, EBL, ICU/LOS/readmission, CRM/DRM for rectal, stoma formation), chemotherapy regimens (MOSAIC/FIRE-3/KEYNOTE-177-era — FOLFOX/CAPOX/FOLFIRI/FOLFOXIRI, anti-EGFR cetuximab/panitumumab, anti-VEGF bevacizumab, IO with pembrolizumab/nivolumab+ipilimumab, BEACON-CRC for BRAF V600E, larotrectinib for NTRK), RECIST response with depth-of-response, CEA dynamics (baseline, nadir, response/progression flags), survival endpoints (OS/DFS/PFS/recurrence with site), QoL (EORTC QLQ-C30, LARS for rectal), and a variable-length CEA longitudinal panel (18 timepoints over 10 years, truncated by OS). Built to be drop-in usable for 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 (105 columns) and a ~3,300-row CEA panel (4 columns), in CSV format, with a CC-BY-NC-4.0 license.