跳转至

trial — 患者到临床试验的匹配

TrialGPT 三阶段流程(Jin et al. Nat Commun 15, 9074, 2024)的端到端桩实现:

  1. 检索(Retrieval) —— 基于关键词重叠,从患者摘要中匹配候选临床试验。
  2. 匹配(Matching) —— 按入排标准逐条评估,并给出解释。
  3. 排序(Ranking) —— 将逐条标准得分聚合为单个试验级别的排序分。

用法

mrnavax trial --patient mrnavax/examples/patient_summary.txt \
              --trials mrnavax/examples/trials.jsonl \
              --top-k 5 --backend mock

输入

  • 患者摘要(Patient summary) —— 自由文本。纯英文临床记录效果最佳。
  • 临床试验(Trials) —— JSONL,每行一条试验记录:
{
  "nct_id": "NCT05933577",
  "title": "INTerpath-001: Personalized mRNA-4157 + Pembrolizumab",
  "condition": "Stage IIB-IV melanoma",
  "phase": "3",
  "inclusion": ["Completely resected melanoma", "ECOG 0 or 1"],
  "exclusion": ["Active autoimmune disease"],
  "biomarkers": ["BRAF V600E", "BRAF V600K"]
}

输出 schema

{
  "ranked": [
    {
      "nct_id": "NCT05933577",
      "title": "INTerpath-001: ...",
      "score": 1.0,
      "eligibility_pct": 100.0,
      "n_met": 4,
      "n_total": 4,
      "reasons": ["Completely resected melanoma", "ECOG 0 or 1"]
    }
  ],
  "n_candidates_screened": 5
}

score = eligibility_pct / 100 − 0.2 × (met exclusion criteria)。 得分最高的试验即为推荐匹配项。