Automated Classification of Radiology Reports to Facilitate Retrospective Study in Radiology
Automated Classification of Radiology Reports to Facilitate Retrospective Study in Radiology
复制标题
DOI：
10.1007/s10278-014-9708-x
复制
发表时间：
2014-12-01
影响因子：
4.4
通讯作者：
Wippold, Franz J.
中科院分区：
文献类型：
--
作者：
Zhou, Yihua;Amundson, Per K.;Wippold, Franz J.
回顾性研究是放射学的重要工具。从非结构化放射学报告中识别适合给定研究问题的影像检查非常有用，但需要大量人力。使用 LingPipe [1] 中实现的机器学习文本挖掘方法，我们评估了动态语言模型（DLM）和朴素贝叶斯（NB）分类器在对放射学报告进行分类时的性能，以促进研究项目的放射学检查的识别。训练数据集由 11,432 份放射学报告中的 14,325 个句子组成，这些句子是从包含所有放射学学科的 5,104,594 份报告的数据库中随机选择的。训练句子被手动分为六类（阳性、差异、治疗后、阴性、正常和历史）。使用 10 倍交叉验证 [2] 来评估模型的性能，并在鞍区或鞍上肿块和胶样囊肿病例的放射学报告分类中进行测试。 DLM 和 NB 分类器的平均准确度分别为 88.5%（95% 置信区间 (CI) 为 1.9%）和 85.9%（95% CI 为 2.0%）。 DLM 的表现稍好一些，用于对包含关键词“鞍区或鞍上肿块”或“胶样囊肿”的 1,397 份放射学报告进行分类。对于 959 份包含“鞍区或鞍上肿块”的报告，DLM 模型的准确度为 88.2%（95% CI 为 2.1%）；对于 437 份“胶样囊肿”的报告，DLM 模型的准确度为 86.3%（95% CI 为 2.5%）。我们的结论是，使用机器学习技术对放射学报告进行自动分类可以有效地促进识别适合回顾性研究的病例。
Retrospective research is an import tool in radiology. Identifying imaging examinations appropriate for a given research question from the unstructured radiology reports is extremely useful, but labor-intensive. Using the machine learning text-mining methods implemented in LingPipe [1], we evaluated the performance of the dynamic language model (DLM) and the Na < ve Bayesian (NB) classifiers in classifying radiology reports to facilitate identification of radiological examinations for research projects. The training dataset consisted of 14,325 sentences from 11,432 radiology reports randomly selected from a database of 5,104,594 reports in all disciplines of radiology. The training sentences were categorized manually into six categories (Positive, Differential, Post Treatment, Negative, Normal, and History). A 10-fold cross-validation [2] was used to evaluate the performance of the models, which were tested in classification of radiology reports for cases of sellar or suprasellar masses and colloid cysts. The average accuracies for the DLM and NB classifiers were 88.5 % with 95 % confidence interval (CI) of 1.9 % and 85.9 % with 95 % CI of 2.0 %, respectively. The DLM performed slightly better and was used to classify 1,397 radiology reports containing the keywords "sellar or suprasellar mass", or "colloid cyst". The DLM model produced an accuracy of 88.2 % with 95 % CI of 2.1 % for 959 reports that contain "sellar or suprasellar mass" and an accuracy of 86.3 % with 95 % CI of 2.5 % for 437 reports of "colloid cyst". We conclude that automated classification of radiology reports using machine learning techniques can effectively facilitate the identification of cases suitable for retrospective research.