Automated deidentification of radiology reports combining transformer and "hide in plain sight" rule-based methods.

OBJECTIVE: To develop an automated deidentification pipeline for radiology reports that detect protected health information (PHI) entities and replaces them with realistic surrogates "hiding in plain sight." MATERIALS AND METHODS: In this retrospective study, 999 chest X-ray and CT reports collected between November 2019 and November 2020 were annotated for PHI at the token level and combined with 3001 X-rays and 2193 medical notes previously labeled, forming a large multi-institutional and cross-domain dataset of 6193 documents. Two radiology test sets, from a known and a new institution, as well as i2b2 2006 and 2014 test sets, served as an evaluation set to estimate model performance and to compare it with previously released deidentification tools. Several PHI detection models were developed based on different training datasets, fine-tuning approaches and data augmentation techniques, and a synthetic PHI generation algorithm. These models were compared using metrics such as precision, recall and F1 score, as well as paired samples Wilcoxon tests. RESULTS: Our best PHI detection model achieves 97.9 F1 score on radiology reports from a known institution, 99.6 from a new institution, 99.5 on i2b2 2006, and 98.9 on i2b2 2014. On reports from a known institution, it achieves 99.1 recall of detecting the core of each PHI span. DISCUSSION: Our model outperforms all deidentifiers it was compared to on all test sets as well as human labelers on i2b2 2014 data. It enables accurate and automatic deidentification of radiology reports. CONCLUSIONS: A transformer-based deidentification pipeline can achieve state-of-the-art performance for deidentifying radiology reports and other medical documents.

Predicting Patient Demographics From Chest Radiographs With Deep Learning.

BACKGROUND: Deep learning models are increasingly informing medical decision making, for instance, in the detection of acute intracranial hemorrhage and pulmonary embolism. However, many models are trained on medical image databases that poorly represent the diversity of the patients they serve. In turn, many artificial intelligence models may not perform as well on assisting providers with important medical decisions for underrepresented populations. PURPOSE: Assessment of the ability of deep learning models to classify the self-reported gender, age, self-reported ethnicity, and insurance status of an individual patient from a given chest radiograph. METHODS: Models were trained and tested with 55,174 radiographs in the MIMIC Chest X-ray (MIMIC-CXR) database. External validation data came from two separate databases, one from CheXpert and another from a multihospital urban health care system after institutional review board approval. Macro-averaged area under the curve (AUC) values were used to evaluate performance of models. Code used for this study is open-source and available at https://github.com/ai-bias/cxr-bias, and pixelstopatients.com/models/demographics. RESULTS: Accuracy of models to predict gender was nearly perfect, with 0.999 (95% confidence interval: 0.99-0.99) AUC on held-out test data and 0.994 (0.99-0.99) and 0.997 (0.99-0.99) on external validation data. There was high accuracy to predict age and ethnicity, ranging from 0.854 (0.80-0.91) to 0.911 (0.88-0.94) AUC, and moderate accuracy to predict insurance status, with AUC ranging from 0.705 (0.60-0.81) on held-out test data to 0.675 (0.54-0.79) on external validation data. CONCLUSIONS: Deep learning models can predict the age, self-reported gender, self-reported ethnicity, and insurance status of a patient from a chest radiograph. Visualization techniques are useful to ensure deep learning models function as intended and to demonstrate anatomical regions of interest. These models can be used to ensure that training data are diverse, thereby ensuring artificial intelligence models that work on diverse populations.

Evaluation of Automated Public De-Identification Tools on a Corpus of Radiology Reports.

PURPOSE: To evaluate publicly available de-identification tools on a large corpus of narrative-text radiology reports. MATERIALS AND METHODS: In this retrospective study, 21 categories of protected health information (PHI) in 2503 radiology reports were annotated from a large multihospital academic health system, collected between January 1, 2012 and January 8, 2019. A subset consisting of 1023 reports served as a test set; the remainder were used as domain-specific training data. The types and frequencies of PHI present within the reports were tallied. Five public de-identification tools were evaluated: MITRE Identification Scrubber Toolkit, U.S. National Library of Medicine‒Scrubber, Massachusetts Institute of Technology de-identification software, Emory Health Information DE-identification (HIDE) software, and Neuro named-entity recognition (NeuroNER). The tools were compared using metrics including recall, precision, and F1 score (the harmonic mean of recall and precision) for each category of PHI. RESULTS: The annotators identified 3528 spans of PHI text within the 2503 reports. Cohen κ for interrater agreement was 0.938. Dates accounted for the majority of PHI found in the dataset of radiology reports (n = 2755 [78%]). The two best-performing tools both used machine learning methods-NeuroNER (precision, 94.5%; recall, 92.6%; microaveraged F1 score [F1], 93.6%) and Emory HIDE (precision, 96.6%; recall, 88.2%; F1, 92.2%)-but none exceeded 50% F1 on the important patient names category. CONCLUSION: PHI appeared infrequently within the corpus of reports studied, which created difficulties for training machine learning systems. Out-of-the-box de-identification tools achieved limited performance on the corpus of radiology reports, suggesting the need for further advancements in public datasets and trained models.Supplemental material is available for this article.See also the commentary by Tenenholtz and Wood in this issue.© RSNA, 2020.

Detection of Muscle Tension Dysphonia Using Eulerian Video Magnification: A Pilot Study.

OBJECTIVE: To determine whether Eulerian Video Magnification software is useful in diagnosis of muscle tension dysphonia (MTD). STUDY DESIGN: Prospective. METHODS: Adult patients scheduled in a tertiary care laryngology practice for evaluation of dysphonia were recruited between November 2016 and March 2017. Demographic and clinical data were extracted from patient charts. Diagnosis of MTD was confirmed with videostroboscopic and physical exam and by a speech-language pathologist. Eighteen MTD patients were video recorded while at rest and with phonation. Five patients without MTD also were analyzed as controls. Videos were analyzed using Eulerian Video Magnification software (Massachusetts Institute of Technology) to assess change in blood flow at the forehead, infrahyoid muscles, and sternocleidomastoid muscles, while using the values of the background wall as a control value. RESULTS: Patients with MTD demonstrated little change in perfusion to the infrahyoid muscles of the neck while phonating (+1% ± 55%). Control subjects demonstrated an increase in perfusion to the infrahyoid muscles while phonating (+102% ± 164%), with this change being significant when comparing the two groups (P = 0.04, t = 2.189, df = 21). A change in perfusion of 0% or less to infrahyoid muscles was 75% sensitive and 70% specific for diagnosis of MTD. No differences in perfusion were found between other regions assessed. Patient age and gender did not correlate with any change in perfusion between rest and phonation. CONCLUSION: Our data suggest that Eulerian Video Magnification can be used in the diagnosis of MTD by focusing on the difference in perfusion to the infrahyoid muscles between rest and phonation.

A Prospective, Longitudinal Study of Outcomes Following Total Knee Arthroplasty Stratified by Gender.

This study aimed to compare temporal trends in outcomes between men and women following TKA. A database of 272 TKAs at 7 years follow-up were evaluated for survivorship, functional, activity, and quality of life outcomes. Evaluation of implant survivorship, KSS scores, and ROM showed no significance differences between men and women, but both functional KSS scores and LEAS showed that men had significantly higher functional scores. The SF-36 physical component was found to be significantly better in men at multiple time points, whereas, the mental component was significantly better in men at pre-operative and 2 years. Both men and women benefit from TKA, however, men have higher levels of function and activity both prior to and after TKA than women.