OpenDeID Pipeline for Unstructured Electronic Health Record Text Notes Based on Rules and Transformers: Deidentification Algorithm Development and Validation Study.

BACKGROUND: Electronic health records (EHRs) in unstructured formats are valuable sources of information for research in both the clinical and biomedical domains. However, before such records can be used for research purposes, sensitive health information (SHI) must be removed in several cases to protect patient privacy. Rule-based and machine learning-based methods have been shown to be effective in deidentification. However, very few studies investigated the combination of transformer-based language models and rules. OBJECTIVE: The objective of this study is to develop a hybrid deidentification pipeline for Australian EHR text notes using rules and transformers. The study also aims to investigate the impact of pretrained word embedding and transformer-based language models. METHODS: In this study, we present a hybrid deidentification pipeline called OpenDeID, which is developed using an Australian multicenter EHR-based corpus called OpenDeID Corpus. The OpenDeID corpus consists of 2100 pathology reports with 38,414 SHI entities from 1833 patients. The OpenDeID pipeline incorporates a hybrid approach of associative rules, supervised deep learning, and pretrained language models. RESULTS: The OpenDeID achieved a best F1-score of 0.9659 by fine-tuning the Discharge Summary BioBERT model and incorporating various preprocessing and postprocessing rules. The OpenDeID pipeline has been deployed at a large tertiary teaching hospital and has processed over 8000 unstructured EHR text notes in real time. CONCLUSIONS: The OpenDeID pipeline is a hybrid deidentification pipeline to deidentify SHI entities in unstructured EHR text notes. The pipeline has been evaluated on a large multicenter corpus. External validation will be undertaken as part of our future work to evaluate the effectiveness of the OpenDeID pipeline.

Simultaneous quantification of exosomal MMP14 expression and proteolysis activity on a spherical dual-probe-based fluorescent nanosensor.

Among various exosomal proteins, matrix metalloproteinases (MMPs) are a family of membrane associated endopeptidases and have been considered as potential biomarkers in liquid biopsy owing to their multiple roles in pathological processes. However, the potential of MMP14 expression (MMP14-E) and MMP14 proteolytic activity (MMP14-A) in clinical diagnosis is still not clear due to the lack of sensitive and simultaneous detection techniques. Herein, we propose a fluorescent nanosensor for the simultaneous detection of MMP14-E and MMP14-A using a spherical aptamer/peptide dual-probe strategy. The aptamer and peptide probes were sequentially immobilized on Fe3O4 magnetic nanoparticles coated with gold nanoparticles (m-AuNPs) using a disulfide linker. MMP14 can be specifically recognized by the aptamer, while the proteolytic-active MMP14 can cleave the peptide probe. While achieving simultaneous detection, the proposed sensor provides better analytical performances than traditional MMP14 sensors owing to the m-AuNP-based spherical dual-probe strategy. This sensor has been successfully applied for the detection of exosomal MMP14 from cell culture media and real serum samples. Levels of both MMP14-E and MMP14-A increase in serum from cancer patients, indicating their potential applications as biomarkers in liquid biopsy for disease diagnosis and real-time surveillance.

Efficient exosome subpopulation isolation and proteomic profiling using a Sub-ExoProfile chip towards cancer diagnosis and treatment.

Exosomes have been extensively studied as liquid biopsy biomarkers in the past decade. However, the origin and molecular heterogeneity of exosomes hinder the research development moving from proof-of-concept to clinical applications. Herein, we report an integrated microfluidic platform termed Sub-ExoProfile chip, to achieve the selective isolation and subsequent proteomic profiling of multiplex exosome subpopulations simultaneously. The Sub-ExoProfile chip comprises three cylindrical self-assembled nanopillars, on which specific exosome capture antibodies (CD81, EpCAM, HER2) were immobilized to capture and sort different exosome subpopulations. It is worth noting that the 3D porous nanopillars afford enhanced interface binding efficiency; thus, a tumor-specific exosome subpopulation with lowly-expressed surface marker was still isolated with satisfactory capture efficiency. Moreover, the amphiphilic mesoporous silica nanoparticle self-assembled nanopillars also served as a nanoreactor for the enrichment and in situ digestion of exosomal proteins, providing improved performance for the mass-spectrometry based molecular characterization of exosome subpopulations. The Sub-ExoProfile chip was investigated on standard exosome samples from different breast cancer cell lines. The isolation and quantitative detection of exosome subpopulations were in line with the molecular subtype of breast cancer cell lines, and the molecular makeup was confirmed using classic microplate ELISA. Clinical samples from HER2-positive and triple-negative breast cancer patients were also examined using the Sub-ExoProfile chip. The quantitative results of three exosome subpopulations distinguished the three subtypes of breast cancer significantly. Most importantly, the molecular characterization of three exosome subpopulations revealed that the distinct exosome subpopulation participated in a different signaling pathway and performed distinct biological functions. It is envisioned that the analysis of the exosome subpopulation on the Sub-ExoProfile chip will facilitate the screening of tumor-specific exosomal biomarkers and open a new avenue for exosome-based liquid biopsy.

The OpenDeID corpus for patient de-identification.

For research purposes, protected health information is often redacted from unstructured electronic health records to preserve patient privacy and confidentiality. The OpenDeID corpus is designed to assist development of automatic methods to redact sensitive information from unstructured electronic health records. We retrieved 4548 unstructured surgical pathology reports from four urban Australian hospitals. The corpus was developed by two annotators under three different experimental settings. The quality of the annotations was evaluated for each setting. Specifically, we employed serial annotations, parallel annotations, and pre-annotations. Our results suggest that the pre-annotations approach is not reliable in terms of quality when compared to the serial annotations but can drastically reduce annotation time. The OpenDeID corpus comprises 2,100 pathology reports from 1,833 cancer patients with an average of 737.49 tokens and 7.35 protected health information entities annotated per report. The overall inter annotator agreement and deviation scores are 0.9464 and 0.9726, respectively. Realistic surrogates are also generated to make the corpus suitable for distribution to other researchers.

Crowdsourcing digital health measures to predict Parkinson's disease severity: the Parkinson's Disease Digital Biomarker DREAM Challenge.

Consumer wearables and sensors are a rich source of data about patients' daily disease and symptom burden, particularly in the case of movement disorders like Parkinson's disease (PD). However, interpreting these complex data into so-called digital biomarkers requires complicated analytical approaches, and validating these biomarkers requires sufficient data and unbiased evaluation methods. Here we describe the use of crowdsourcing to specifically evaluate and benchmark features derived from accelerometer and gyroscope data in two different datasets to predict the presence of PD and severity of three PD symptoms: tremor, dyskinesia, and bradykinesia. Forty teams from around the world submitted features, and achieved drastically improved predictive performance for PD status (best AUROC = 0.87), as well as tremor- (best AUPR = 0.75), dyskinesia- (best AUPR = 0.48) and bradykinesia-severity (best AUPR = 0.95).

Generation of Surrogates for De-Identification of Electronic Health Records.

Unstructured electronic health records are valuable resources for research. Before they are shared with researchers, protected health information needs to be removed from these unstructured documents to protect patient privacy. The main steps involved in removing protected health information are accurately identifying sensitive information in the documents and removing the identified information. To keep the documents as realistic as possible, the step of omitting sensitive information is often followed by replacement of identified sensitive information with surrogates. In this study, we present an algorithm to generate surrogates for unstructured electronic health records. We used this algorithm to generate realistic surrogates on a Health Science Alliance corpus, which is constructed specifically for the use of development of automated de-identification systems.