Evidence level: B
导言
中医辨证论治的核心在于”证候”——这一综合性的生理病理状态描述体系,与现代医学以解剖-病理为基准的ICD疾病分类体系之间,存在结构性差异。然而,两者若能实现可计算的互映射,将为中西医结合临床研究、流行病学调查、健康保险结算以及人工智能辅助诊断提供关键基础设施。
课题#78旨在建立一个证候-ICD互检索API协议,实现以下核心功能:给定一个中医证候,查询最相关的ICD分类条目;反之,给定一个ICD诊断,检索最匹配的中医证候体系。本文综合现有文献证据,阐述这一映射协议的必要性、技术路径与设计要点。
证候与ICD分类的结构性鸿沟
中医证候(如”肝气郁结””心脾两虚”)是对患者整体功能状态的动态评估,包含寒热、虚实、表里、气血等多个维度,每个维度均可取多级严重程度。这种多维连续谱与传统ICD的离散分类(单一疾病编码)之间并非简单的一一对应关系,而是呈现多对多网状重叠结构。
Ji等(PMID: 26778224)在七种恶性肿瘤的证候研究中发现,同一西医疾病在不同证候分型下的生物学指标谱差异显著,提示疾病-证候映射不能依赖单一标签。Sun等(PMID: 22675378)进一步指出,证候客观化需要多技术融合,这一原则同样适用于映射协议设计。
标准化映射协议的技术路径
第一步:证候本体构建
以国家中医药管理局发布的《中医证候诊断标准》为基础,融合WHO传统医学分类建议(ICD-11 MMS Section MG2),建立标准化的证候本体库。本体以”证候-维度-要素”三级结构组织,支持维度间的交叉检索。
第二步:ICD映射规则
对每个标准证候条目,基于以下策略分配候选ICD编码集合:
- 表型重叠法:提取证候的客观指标(如血瘀证→血流变学异常→ICD-10 I73.9外周血管疾病),建立表型-ICD比值矩阵
- Delphi专家共识法:组织多中心中西医专家就核心证候-ICD对应关系进行多轮投票,形成高置信度映射对(参考Song等 PMID: 33128537 的Delphi方法)
- 数据驱动验证法:利用真实世界临床数据(中医医院电子病历),基于共现频率和机器学习聚类,量化各映射的支撑强度
第三步:互检索API设计
API协议需支持双向查询:
GET /api/v1/syndrome-to-icd?syndrome={证候名称}&confidence_threshold=0.7
GET /api/v1/icd-to-syndrome?icd_code={ICD编码}&confidence_threshold=0.7
返回结构包含:映射目标、置信度评分(0-1)、映射依据(表型重叠/专家共识/数据驱动)、上游证候维度分解。Guan等(PMID: 41830874)提出的LLM-SDaT框架为证候判定的自动化提供了技术参照,可作为API底层判定引擎的组件。
临床效用与风险
正向价值:使中医临床数据得以纳入ICD编码的健康统计体系,支撑真实世界研究和医保结算。FuQ等(PMID: 40944470)的过敏性鼻炎国际指南已示范了将证候分型嵌入标准化临床路径的可行路径。
局限与风险:证候-ICD映射的本质是两种不同认知体系的近似翻译,存在不可避免的信息损耗。映射不应被解读为”中医证候等同于某ICD疾病”,而应明确标注为”功能状态重叠参考”,以避免临床误用。
证据与局限
本文综合了6项相关研究(PMID: 40944470, 26778224, 41830874, 33128537, 22675378, 31623396),涵盖Delphi共识方法、证候客观化技术、LLM辅助判定及真实世界临床数据应用。现有证据显示,构建标准化映射协议具有方法学基础,但大规模多中心验证研究仍属空白。本设计属于数据基础设施级别,临床落地需前瞻性验证映射准确率和临床安全性。
相关阅读
免责说明:本文属传统知识与科学文献的整理、对照与研究设计讨论,仅供理解与参考,不构成医疗建议,不能替代专业诊断与治疗。文中涉及的效应量、频率区间与暴露结论多为研究报告层面的推断,个体情况差异很大;如有身体不适或健康疑虑,请就医并由专业医师判断。
Standardization Mapping Protocol and Inter-retrieval API Design for Traditional Chinese Medicine Syndrome-ICD
Evidence level: B
Introduction
The core of TCM syndrome differentiation and treatment lies in the “Syndrome” (证候)—a comprehensive descriptive system of physiological and pathological states. Between this system and the modern medical ICD disease classification framework, which is based on anatomical-pathological benchmarks, there exists a structural disparity. However, if a computable inter-mapping is achieved between the two, it will provide critical infrastructure for integrated TCM-Western clinical research, epidemiological surveys, health insurance settlement, and AI-assisted diagnosis.
Project #78 aims to establish a Syndrome-ICD Inter-retrieval API Protocol, realizing the following core functions: given a TCM syndrome, query the most relevant ICD classification entry; conversely, given an ICD diagnosis, retrieve the most matching TCM syndrome system. This paper synthesizes existing literature evidence to elucidate the necessity, technical pathways, and design considerations of this mapping protocol.
Structural Chasm Between Syndromes and ICD Classification
TCM syndromes (e.g., “Liver Qi Stagnation,” “Heart-Spleen Deficiency”) represent a dynamic assessment of the patient’s overall functional state, encompassing multiple dimensions such as cold-heat, deficiency-excess, exterior-interior, and qi-blood. Each dimension can assume multiple levels of severity. This multi-dimensional continuum does not exhibit a simple one-to-one correspondence with the discrete classification of traditional ICD (single disease codes), but rather presents a many-to-many meshed overlap structure.
Ji et al. (PMID: 26778224) found in their syndrome research on seven types of malignant tumors that biological indicator spectra varied significantly under different syndrome subtypes for the same Western disease, suggesting that disease-syndrome mapping cannot rely on a single label. Sun et al. (PMID: 22675378) further indicated that the objectification of syndromes requires multi-technology fusion, a principle equally applicable to mapping protocol design.
Technical Pathways for Standardized Mapping Protocols
Step 1: Syndrome Ontology Construction
Based on the “TCM Syndrome Diagnosis Standards” issued by the National Administration of Traditional Chinese Medicine, integrating WHO Traditional Medicine Classification recommendations (ICD-11 MMS Section MG2), a standardized Syndrome Ontology Repository shall be established. The ontology is organized in a three-level structure of “Syndrome-Dimension-Element,” supporting cross-retrieval between dimensions.
Step 2: ICD Mapping Rules
For each standard syndrome entry, candidate ICD code sets shall be allocated based on the following strategies:
- Phenotypic Overlap Method: Extract objective indicators of the syndrome (e.g., Blood Stasis Syndrome → Hemorheological abnormalities → ICD-10 I73.9 Peripheral vascular diseases) to establish a phenotype-ICD ratio matrix.
- Delphi Expert Consensus Method: Organize multi-center TCM and Western medicine experts to conduct multi-round voting on core syndrome-ICD correspondence relationships, forming high-confidence mapping pairs (referencing the Delphi method by Song et al., PMID: 33128537).
- Data-driven Verification Method: Utilize real-world clinical data (electronic medical records from TCM hospitals) to quantify the support strength of each mapping based on co-occurrence frequency and machine learning clustering.
Step 3: Inter-retrieval API Design
The API protocol must support bidirectional queries:
GET /api/v1/syndrome-to-icd?syndrome={Syndrome Name}&confidence_threshold=0.7
GET /api/v1/icd-to-syndrome?icd_code={ICD Code}&confidence_threshold=0.7
The return structure includes: mapping target, confidence score (0-1), mapping basis (phenotypic overlap/expert consensus/data-driven), and upstream syndrome dimension decomposition. The LLM-SDaT framework proposed by Guan et al. (PMID: 41830874) provides a technical reference for the automation of syndrome determination and can serve as a component for the underlying judgment engine of the API.
Clinical Utility and Risks
Positive Value: Enables TCM clinical data to be incorporated into the ICD-coded health statistics system, supporting real-world research and insurance settlement. FuQ et al. (PMID: 40944470) demonstrated a feasible pathway for embedding syndrome subtypes into standardized clinical paths via the International Guideline for Allergic Rhinitis.
Limitations and Risks: The essence of syndrome-ICD mapping is an approximate translation between two distinct cognitive systems, involving inevitable information loss. The mapping should not be interpreted as “TCM syndrome equals a specific ICD disease,” but rather explicitly labeled as “Functional State Overlap Reference” to prevent clinical misuse.
Evidence and Limitations
This paper synthesizes six related studies (PMID: 40944470, 26778224, 41830874, 33128537, 22675378, 31623396), covering Delphi consensus methods, syndrome objectification technologies, LLM-assisted determination, and real-world clinical data applications. Current evidence indicates that constructing a standardized mapping protocol possesses methodological foundations, yet large-scale multi-center validation studies remain absent. This design belongs to the data infrastructure level; prospective validation of mapping accuracy and clinical safety is required before clinical implementation.
Further Reading
- Syndrome-Disease Overlap Spectrum Analysis
- Pulse-Harmonic Syndrome Independent Correlation Verification
- HRI Spectrum Cardiovascular Event Prediction Value Independent Verification
Disclaimer: This article compiles traditional knowledge and scientific literature for comparison and research-design discussion only. It is not medical advice and cannot replace professional diagnosis or treatment. Effect sizes, frequency ranges and exposure conclusions cited here are largely inferences at the research-report level, and individual circumstances vary widely. If you have any discomfort or health concerns, please consult a physician.
Leave a Reply