From the Departments of Radiology (M.D.L., M.L., F.D., K.C., K.B., S.R., W.A.M., J.K.-C.)
From the Departments of Radiology (M.D.L., M.L., F.D., K.C., K.B., S.R., W.A.M., J.K.-C.).
AJNR Am J Neuroradiol. 2021 Mar;42(3):429-434. doi: 10.3174/ajnr.A6961. Epub 2020 Dec 17.
The coronavirus disease 2019 (COVID-19) pandemic has led to decreases in neuroimaging volume. Our aim was to quantify the change in acute or subacute ischemic strokes detected on CT or MR imaging during the pandemic using natural language processing of radiology reports.
We retrospectively analyzed 32,555 radiology reports from brain CTs and MRIs from a comprehensive stroke center, performed from March 1 to April 30 each year from 2017 to 2020, involving 20,414 unique patients. To detect acute or subacute ischemic stroke in free-text reports, we trained a random forest natural language processing classifier using 1987 randomly sampled radiology reports with manual annotation. Natural language processing classifier generalizability was evaluated using 1974 imaging reports from an external dataset.
The natural language processing classifier achieved a 5-fold cross-validation classification accuracy of 0.97 and an F1 score of 0.74, with a slight underestimation (-5%) of actual numbers of acute or subacute ischemic strokes in cross-validation. Importantly, cross-validation performance stratified by year was similar. Applying the classifier to the complete study cohort, we found an estimated 24% decrease in patients with acute or subacute ischemic strokes reported on CT or MR imaging from March to April 2020 compared with the average from those months in 2017-2019. Among patients with stroke-related order indications, the estimated proportion who underwent neuroimaging with acute or subacute ischemic stroke detection significantly increased from 16% during 2017-2019 to 21% in 2020 (= .01). The natural language processing classifier performed worse on external data.
Acute or subacute ischemic stroke cases detected by neuroimaging decreased during the COVID-19 pandemic, though a higher proportion of studies ordered for stroke were positive for acute or subacute ischemic strokes. Natural language processing approaches can help automatically track acute or subacute ischemic stroke numbers for epidemiologic studies, though local classifier training is important due to radiologist reporting style differences.
2019 年冠状病毒病（COVID-19）大流行导致神经影像学检查数量减少。我们的目的是使用放射学报告的自然语言处理来量化大流行期间 CT 或 MRI 上检测到的急性或亚急性缺血性卒中的变化。
我们回顾性分析了 2017 年至 2020 年 3 月 1 日至 4 月 30 日期间来自综合卒中中心的 32555 份脑 CT 和 MRI 放射学报告，涉及 20414 名独特患者。为了在自由文本报告中检测急性或亚急性缺血性卒中，我们使用 1987 份随机抽样的放射学报告和手动注释来训练随机森林自然语言处理分类器。使用来自外部数据集的 1974 份成像报告评估自然语言处理分类器的泛化能力。
自然语言处理分类器在 5 倍交叉验证中的分类准确性为 0.97，F1 得分为 0.74，略有低估（-5%）交叉验证中的实际急性或亚急性缺血性卒中数量。重要的是，按年份分层的交叉验证性能相似。将分类器应用于完整的研究队列，我们发现与 2017 年至 2019 年同期相比，2020 年 3 月至 4 月报告的 CT 或 MRI 上有急性或亚急性缺血性卒中的患者估计减少了 24%。在有卒中相关医嘱的患者中，估计在有急性或亚急性缺血性卒中检测的神经影像学检查中比例从 2017 年至 2019 年的 16%显著增加到 2020 年的 21%（=0.01）。自然语言处理分类器在外部数据上的表现较差。
COVID-19 大流行期间，通过神经影像学检测到的急性或亚急性缺血性卒中病例减少，尽管为卒中开的研究比例中急性或亚急性缺血性卒中阳性的比例更高。自然语言处理方法可帮助自动跟踪流行病学研究中的急性或亚急性缺血性卒中数量，但由于放射科医生报告风格的差异，局部分类器训练很重要。
