Yuan Luo

Evaluating the State-of-the-Art in Automatic De-identification

To facilitate and survey studies in automatic de-identification, as a part of the i2b2 (Informatics for Integrating Biology to the Bedside) project, authors organized a Natural Language Processing (NLP) challenge on automatically removing private health information (PHI) from medical discharge records. This manuscript provides an overview of this de-identification challenge, describes the data and the annotation process, explains the evaluation metrics, discusses the nature of the systems that addressed the challenge, analyzes the results of received system runs, and identifies directions for future research. The de-indentification challenge data consisted of discharge summaries drawn from the Partners Healthcare system. Authors prepared this data for the challenge by replacing authentic PHI with synthesized surrogates. To focus the challenge on non-dictionary-based de-identification methods, the data was enriched with out-of-vocabulary PHI surrogates, i.e., made up names. The data also included some PHI surrogates that were ambiguous with medical non-PHI terms. A total of seven teams participated in the challenge. Each team submitted up to three system runs, for a total of sixteen submissions. The authors used precision, recall, and F-measure to evaluate the submitted system runs based on their token-level and instance-level performance on the ground truth. The systems with the best performance scored above 98% in F-measure for all categories of PHI. Most out-of-vocabulary PHI could be identified accurately. However, identifying ambiguous PHI proved challenging. The performance of systems on the test data set is encouraging. Future evaluations of these systems will involve larger data sets from more heterogeneous sources.

A de-identifier for medical discharge summaries

OBJECTIVE Clinical records contain significant medical information that can be useful to researchers in various disciplines. However, these records also contain personal health information (PHI) whose presence limits the use of the records outside of hospitals. The goal of de-identification is to remove all PHI from clinical records. This is a challenging task because many records contain foreign and misspelled PHI; they also contain PHI that are ambiguous with non-PHI. These complications are compounded by the linguistic characteristics of clinical records. For example, medical discharge summaries, which are studied in this paper, are characterized by fragmented, incomplete utterances and domain-specific language; they cannot be fully processed by tools designed for lay language. METHODS AND RESULTS In this paper, we show that we can de-identify medical discharge summaries using a de-identifier, Stat De-id, based on support vector machines and local context (F-measure=97% on PHI). Our representation of local context aids de-identification even when PHI include out-of-vocabulary words and even when PHI are ambiguous with non-PHI within the same corpus. Comparison of Stat De-id with a rule-based approach shows that local context contributes more to de-identification than dictionaries combined with hand-tailored heuristics (F-measure=85%). Comparison with two well-known named entity recognition (NER) systems, SNoW (F-measure=94%) and IdentiFinder (F-measure=36%), on five representative corpora show that when the language of documents is fragmented, a system with a relatively thorough representation of local context can be a more effective de-identifier than systems that combine (relatively simpler) local context with global context. Comparison with a Conditional Random Field De-identifier (CRFD), which utilizes global context in addition to the local context of Stat De-id, confirms this finding (F-measure=88%) and establishes that strengthening the representation of local context may be more beneficial for de-identification than complementing local with global context.

Automatic lymphoma classification with sentence subgraph mining from pathology reports.

OBJECTIVE Pathology reports are rich in narrative statements that encode a complex web of relations among medical concepts. These relations are routinely used by doctors to reason on diagnoses, but often require hand-crafted rules or supervised learning to extract into prespecified forms for computational disease modeling. We aim to automatically capture relations from narrative text without supervision. METHODS We design a novel framework that translates sentences into graph representations, automatically mines sentence subgraphs, reduces redundancy in mined subgraphs, and automatically generates subgraph features for subsequent classification tasks. To ensure meaningful interpretations over the sentence graphs, we use the Unified Medical Language System Metathesaurus to map token subsequences to concepts, and in turn sentence graph nodes. We test our system with multiple lymphoma classification tasks that together mimic the differential diagnosis by a pathologist. To this end, we prevent our classifiers from looking at explicit mentions or synonyms of lymphomas in the text. RESULTS AND CONCLUSIONS We compare our system with three baseline classifiers using standard n-grams, full MetaMap concepts, and filtered MetaMap concepts. Our system achieves high F-measures on multiple binary classifications of lymphoma (Burkitt lymphoma, 0.8; diffuse large B-cell lymphoma, 0.909; follicular lymphoma, 0.84; Hodgkin lymphoma, 0.912). Significance tests show that our system outperforms all three baselines. Moreover, feature analysis identifies subgraph features that contribute to improved performance; these features agree with the state-of-the-art knowledge about lymphoma classification. We also highlight how these unsupervised relation features may provide meaningful insights into lymphoma classification.

Text Mining in Cancer Gene and Pathway Prioritization

Prioritization of cancer implicated genes has received growing attention as an effective way to reduce wet lab cost by computational analysis that ranks candidate genes according to the likelihood that experimental verifications will succeed. A multitude of gene prioritization tools have been developed, each integrating different data sources covering gene sequences, differential expressions, function annotations, gene regulations, protein domains, protein interactions, and pathways. This review places existing gene prioritization tools against the backdrop of an integrative Omic hierarchy view toward cancer and focuses on the analysis of their text mining components. We explain the relatively slow progress of text mining in gene prioritization, identify several challenges to current text mining methods, and highlight a few directions where more effective text mining algorithms may improve the overall prioritization task and where prioritizing the pathways may be more desirable than prioritizing only genes.

Subgraph Augmented Non-Negative Tensor Factorization (SANTF) for Modeling Clinical Narrative Text

OBJECTIVE Extracting medical knowledge from electronic medical records requires automated approaches to combat scalability limitations and selection biases. However, existing machine learning approaches are often regarded by clinicians as black boxes. Moreover, training data for these automated approaches at often sparsely annotated at best. The authors target unsupervised learning for modeling clinical narrative text, aiming at improving both accuracy and interpretability. METHODS The authors introduce a novel framework named subgraph augmented non-negative tensor factorization (SANTF). In addition to relying on atomic features (e.g., words in clinical narrative text), SANTF automatically mines higher-order features (e.g., relations of lymphoid cells expressing antigens) from clinical narrative text by converting sentences into a graph representation and identifying important subgraphs. The authors compose a tensor using patients, higher-order features, and atomic features as its respective modes. We then apply non-negative tensor factorization to cluster patients, and simultaneously identify latent groups of higher-order features that link to patient clusters, as in clinical guidelines where a panel of immunophenotypic features and laboratory results are used to specify diagnostic criteria. RESULTS AND CONCLUSION SANTF demonstrated over 10% improvement in averaged F-measure on patient clustering compared to widely used non-negative matrix factorization (NMF) and k-means clustering methods. Multiple baselines were established by modeling patient data using patient-by-features matrices with different feature configurations and then performing NMF or k-means to cluster patients. Feature analysis identified latent groups of higher-order features that lead to medical insights. We also found that the latent groups of atomic features help to better correlate the latent groups of higher-order features.

Natural Language Processing for EHR-Based Pharmacovigilance: A Structured Review

The goal of pharmacovigilance is to detect, monitor, characterize and prevent adverse drug events (ADEs) with pharmaceutical products. This article is a comprehensive structured review of recent advances in applying natural language processing (NLP) to electronic health record (EHR) narratives for pharmacovigilance. We review methods of varying complexity and problem focus, summarize the current state-of-the-art in methodology advancement, discuss limitations and point out several promising future directions. The ability to accurately capture both semantic and syntactic structures in clinical narratives becomes increasingly critical to enable efficient and accurate ADE detection. Significant progress has been made in algorithm development and resource construction since 2000. Since 2012, statistical analysis and machine learning methods have gained traction in automation of ADE mining from EHR narratives. Current state-of-the-art methods for NLP-based ADE detection from EHRs show promise regarding their integration into production pharmacovigilance systems. In addition, integrating multifaceted, heterogeneous data sources has shown promise in improving ADE detection and has become increasingly adopted. On the other hand, challenges and opportunities remain across the frontier of NLP application to EHR-based pharmacovigilance, including proper characterization of ADE context, differentiation between off- and on-label drug-use ADEs, recognition of the importance of polypharmacy-induced ADEs, better integration of heterogeneous data sources, creation of shared corpora, and organization of shared-task challenges to advance the state-of-the-art.

Tensor factorization toward precision medicine

Precision medicine initiatives come amid the rapid growth in quantity and variety of biomedical data, which exceeds the capacity of matrix-oriented data representations and many current analysis algorithms. Tensor factorizations extend the matrix view to multiple modalities and support dimensionality reduction methods that identify latent groups of data for meaningful summarization of both features and instances. In this opinion article, we analyze the modest literature on applying tensor factorization to various biomedical fields including genotyping and phenotyping. Based on the cited work including work of our own, we suggest that tensor applications could serve as an effective tool to enable frequent updating of medical knowledge based on the continually growing scientific and clinical evidence. We encourage extensive experimental studies to tackle challenges including design choice of factorizations, integrating temporality and algorithm scalability.

Using Machine Learning to Predict Laboratory Test Results

OBJECTIVES While clinical laboratories report most test results as individual numbers, findings, or observations, clinical diagnosis usually relies on the results of multiple tests. Clinical decision support that integrates multiple elements of laboratory data could be highly useful in enhancing laboratory diagnosis. METHODS Using the analyte ferritin in a proof of concept, we extracted clinical laboratory data from patient testing and applied a variety of machine-learning algorithms to predict ferritin test results using the results from other tests. We compared predicted with measured results and reviewed selected cases to assess the clinical value of predicted ferritin. RESULTS We show that patient demographics and results of other laboratory tests can discriminate normal from abnormal ferritin results with a high degree of accuracy (area under the curve as high as 0.97, held-out test data). Case review indicated that predicted ferritin results may sometimes better reflect underlying iron status than measured ferritin. CONCLUSIONS These findings highlight the substantial informational redundancy present in patient test results and offer a potential foundation for a novel type of clinical decision support aimed at integrating, interpreting, and enhancing the diagnostic value of multianalyte sets of clinical laboratory test results.

Recurrent neural networks for classifying relations in clinical notes

We proposed the first models based on recurrent neural networks (more specifically Long Short-Term Memory - LSTM) for classifying relations from clinical notes. We tested our models on the i2b2/VA relation classification challenge dataset. We showed that our segment LSTM model, with only word embedding feature and no manual feature engineering, achieved a micro-averaged f-measure of 0.661 for classifying medical problem-treatment relations, 0.800 for medical problem-test relations, and 0.683 for medical problem-medical problem relations. These results are comparable to those of the state-of-the-art systems on the i2b2/VA relation classification challenge. We compared the segment LSTM model with the sentence LSTM model, and demonstrated the benefits of exploring the difference between concept text and context text, and between different contextual parts in the sentence. We also evaluated the impact of word embedding on the performance of LSTM models and showed that medical domain word embedding help improve the relation classification. These results support the use of LSTM models for classifying relations between medical concepts, as they show comparable performance to previously published systems while requiring no manual feature engineering.

Segment convolutional neural networks (Seg-CNNs) for classifying relations in clinical notes

We propose Segment Convolutional Neural Networks (Seg-CNNs) for classifying relations from clinical notes. Seg-CNNs use only word-embedding features without manual feature engineering. Unlike typical CNN models, relations between 2 concepts are identified by simultaneously learning separate representations for text segments in a sentence: preceding, concept1, middle, concept2, and succeeding. We evaluate Seg-CNN on the i2b2/VA relation classification challenge dataset. We show that Seg-CNN achieves a state-of-the-art micro-average F-measure of 0.742 for overall evaluation, 0.686 for classifying medical problem-treatment relations, 0.820 for medical problem-test relations, and 0.702 for medical problem-medical problem relations. We demonstrate the benefits of learning segment-level representations. We show that medical domain word embeddings help improve relation classification. Seg-CNNs can be trained quickly for the i2b2/VA dataset on a graphics processing unit (GPU) platform. These results support the use of CNNs computed over segments of text for classifying medical relations, as they show state-of-the-art performance while requiring no manual feature engineering.