DisTEMIST corpus: detection and normalization of disease mentions in spanish clinical cases

<strong>DisTEMIST corpus</strong> <strong>DISTEMIST-entities: complete training set (750 case reports)</strong> <strong>DISTEMIST-linking: part 1 of the training set (200 case reports)</strong> <strong>Introduction</strong> The DisTEMIST corpus is a collection of 1000 clinical cases with disease annotations linked with Snomed-CT concepts. All documents are released in the context of the BioASQ DisTEMIST track for CLEF 2022. For more information about the track and its schedule, please visit the website. <strong>File structure:</strong> The DisTEMIST corpus has been randomly divided into a training set, containing 750 clinical cases, and a test set, consisting of 250 additional cases. Participants must train their systems using the train set and submit predictions for the test set, on which they will be evaluated. The file structure of the corpus is as follows: <strong>train_set:</strong> <strong>text_files</strong>: Folder with plain text files of the clinical cases <strong>subtrack1_entities</strong>: It contains annotations in a tab-separated file (TSV) with the following columns: <em>filename</em>: document name <em>mark</em>: identifier mention id <em>label</em>: mentions type (ENFERMEDAD) <em>off0</em>: starting position of the mention in the document <em>off1</em>: ending position of the mention in the document <em>span</em>: text span <strong>subtrack2_linking: </strong>It contains annotations in a tab-separated file (TSV) with the following columns: <em>filename</em>: document name <em>mark</em>: identifier mention id <em>label</em>: mentions type (ENFERMEDAD) <em>off0</em>: starting position of the mention in the document <em>off1</em>: ending position of the mention in the document <em>span</em>: text span <em>codes</em>: List of Snomed-CT concept codes linked to the mention. If there is more than one code associated with a mention, they will be concatenated by the symbol “+”. <em>semantic relation</em>: the relationship between the assigned code and the mention. It can be EXACT, when the code corresponds exactly with the mention, or NARROW, when the mention corresponds to a narrower concept than the Snomed-CT code. For instance, the concept “Chorioretinal lacunae” does not exist in Snomed-CT. Then, it is normalized to the Snomed-CT ID 302893000 (“Chorioretinal disorder”). <strong>test_set: </strong>The 250 clinical cases that will be used to evaluate the systems will be published in accordance with the schedule of the task. <strong>Resources</strong> Web Evaluation Library <strong>DISTEMIST gazetteer</strong> <strong>DISTEMIST guidelines</strong> More resources soon**DisTEMIST语料库（DisTEMIST corpus）** **DISTEMIST-实体标注：完整训练集（750份临床病例报告）** **DISTEMIST-实体链接：训练集第一部分（200份临床病例报告）** **简介** DisTEMIST语料库是包含1000份临床病例的数据集，所有病例均带有与SNOMED CT（医学系统命名法-临床术语）概念关联的疾病标注。所有文档均为2022年跨语言评估论坛（CLEF）旗下BioASQ（生物信息学问答挑战）DisTEMIST任务的发布数据集。如需了解该任务及日程安排的更多信息，请访问官方网站。 **文件结构** DisTEMIST语料库已随机划分为训练集与测试集：训练集包含750份临床病例，测试集则包含额外250份临床病例。参赛团队需使用训练集开发模型，并针对测试集提交预测结果，系统将基于测试集进行评估。语料库的文件结构如下： **训练集（train_set）** - **text_files**：存储临床病例纯文本文件的文件夹 - **subtrack1_实体标注（subtrack1_entities）**：该目录下包含以制表符分隔文件（TSV）格式的标注数据，各列信息如下： - filename：文档名称 - mark：实体提及标识符 - label：实体提及类型（ENFERMEDAD，西班牙语中意为“疾病”） - off0：实体提及在文档中的起始位置 - off1：实体提及在文档中的结束位置 - span：实体提及文本片段 - **subtrack2_实体链接（subtrack2_linking）**：该目录下包含以制表符分隔文件（TSV）格式的标注数据，各列信息如下： - filename：文档名称 - mark：实体提及标识符 - label：实体提及类型（ENFERMEDAD，西班牙语中意为“疾病”） - off0：实体提及在文档中的起始位置 - off1：实体提及在文档中的结束位置 - span：实体提及文本片段 - codes：与该实体提及关联的SNOMED CT概念代码列表。若一个实体提及关联多个代码，则以"+"符号拼接多个代码 - semantic relation：分配的代码与实体提及之间的语义关系。可分为两种类型： - EXACT（精确匹配）：代码与实体提及完全对应 - NARROW（窄化匹配）：实体提及对应的概念比SNOMED CT代码对应的概念更具体 例如，“脉络膜视网膜隐窝（Chorioretinal lacunae）”这一概念未收录于SNOMED CT中，因此将其归一化至SNOMED CT编码302893000对应的“脉络膜视网膜疾病（Chorioretinal disorder）”。 **测试集（test_set）** 用于评估模型的250份临床病例将按照任务日程安排发布。 **相关资源** - **Web Evaluation Library（网络评估库）** - **DISTEMIST gazetteer（DISTEMIST词典）** - **DISTEMIST guidelines（DISTEMIST标注指南）** 更多资源即将上线。