Lessons learned from annotation of VAERS reports on adverse events following influenza vaccination and related to Guillain-Barré syndrome.

BACKGROUND: Vaccine Adverse Events ReportingSystem (VAERS) is a promising resource of tracking adverse events following immunization. Medical Dictionary for Regulatory Activities (MedDRA) terminology used for coding adverse events in VAERS reports has several limitations. We focus on developing an automated system for semantic extraction of adverse events following vaccination and their temporal relationships for a better understanding of VAERS data and its integration into other applications. The aim of the present studyis to summarize the lessons learned during the initial phase of this project in annotating adverse events following influenza vaccination and related to Guillain-Barré syndrome (GBS). We emphasize on identifying the limitations of VAERS and MedDRA. RESULTS: We collected 282 VAERS reports documented between 1990 and 2016 and shortlisted those with at least 1,100 characters in the report. We used a subset of 50 reports for the preliminary investigation and annotated all adverse events following influenza vaccination by mapping to representative MedDRA terms. Associated time expressions were annotated when available. We used 16 System Organ Class (SOC) level MedDRA terms to map GBS related adverse events and expanded some SOC terms to Lowest Level Terms (LLT) for granular representation. We annotated three broad categories of events such as problems, clinical investigations, and treatments/procedures. The inter-annotator agreement of events achieved was 86%. Incomplete reports, typographical errors, lack of clarity and coherence, repeated texts, unavailability of associated temporal information, difficulty to interpret due to incorrect grammar, use of generalized terms to describe adverse events / symptoms, uncommon abbreviations, difficulty annotating multiple events with a conjunction / common phrase, irrelevant historical events and coexisting events were some of the challenges encountered. Some of the limitations we noted are in agreement with previous reports. CONCLUSIONS: We reported the challenges encountered and lessons learned during annotation of adverse events in VAERS reports following influenza vaccination and related to GBS. Though the challenges may be due to the inevitable limitations of public reporting systems and widely reported limitations of MedDRA, we emphasize the need to understand these limitations and extraction of other supportive information for a better understanding of adverse events following vaccination.

Machine learning-based donor permission extraction from informed consent documents.

BACKGROUND: With more clinical trials are offering optional participation in the collection of bio-specimens for biobanking comes the increasing complexity of requirements of informed consent forms. The aim of this study is to develop an automatic natural language processing (NLP) tool to annotate informed consent documents to promote biorepository data regulation, sharing, and decision support. We collected informed consent documents from several publicly available sources, then manually annotated them, covering sentences containing permission information about the sharing of either bio-specimens or donor data, or conducting genetic research or future research using bio-specimens or donor data. RESULTS: We evaluated a variety of machine learning algorithms including random forest (RF) and support vector machine (SVM) for the automatic identification of these sentences. 120 informed consent documents containing 29,204 sentences were annotated, of which 1250 sentences (4.28%) provide answers to a permission question. A support vector machine (SVM) model achieved a F-1 score of 0.95 on classifying the sentences when using a gold standard, which is a prefiltered corpus containing all relevant sentences. CONCLUSIONS: This study provides the feasibility of using machine learning tools to classify permission-related sentences in informed consent documents.

Extracting postmarketing adverse events from safety reports in the vaccine adverse event reporting system (VAERS) using deep learning.

OBJECTIVE: Automated analysis of vaccine postmarketing surveillance narrative reports is important to understand the progression of rare but severe vaccine adverse events (AEs). This study implemented and evaluated state-of-the-art deep learning algorithms for named entity recognition to extract nervous system disorder-related events from vaccine safety reports. MATERIALS AND METHODS: We collected Guillain-Barré syndrome (GBS) related influenza vaccine safety reports from the Vaccine Adverse Event Reporting System (VAERS) from 1990 to 2016. VAERS reports were selected and manually annotated with major entities related to nervous system disorders, including, investigation, nervous_AE, other_AE, procedure, social_circumstance, and temporal_expression. A variety of conventional machine learning and deep learning algorithms were then evaluated for the extraction of the above entities. We further pretrained domain-specific BERT (Bidirectional Encoder Representations from Transformers) using VAERS reports (VAERS BERT) and compared its performance with existing models. RESULTS AND CONCLUSIONS: Ninety-one VAERS reports were annotated, resulting in 2512 entities. The corpus was made publicly available to promote community efforts on vaccine AEs identification. Deep learning-based methods (eg, bi-long short-term memory and BERT models) outperformed conventional machine learning-based methods (ie, conditional random fields with extensive features). The BioBERT large model achieved the highest exact match F-1 scores on nervous_AE, procedure, social_circumstance, and temporal_expression; while VAERS BERT large models achieved the highest exact match F-1 scores on investigation and other_AE. An ensemble of these 2 models achieved the highest exact match microaveraged F-1 score at 0.6802 and the second highest lenient match microaveraged F-1 score at 0.8078 among peer models.

MRI characterization of cobalt dichloride-N-acetyl cysteine (C4) contrast agent marker for prostate brachytherapy.

Brachytherapy, a radiotherapy technique for treating prostate cancer, involves the implantation of numerous radioactive seeds into the prostate. While the implanted seeds can be easily identified on a computed tomography image, distinguishing the prostate and surrounding soft tissues is not as straightforward. Magnetic resonance imaging (MRI) offers superior anatomical delineation, but the seeds appear as dark voids and are difficult to identify, thus creating a conundrum. Cobalt dichloride-N-acetyl-cysteine (C4) has previously been shown to be promising as an encapsulated contrast agent marker. We performed spin-lattice relaxation time (T1) and spin-spin relaxation time (T2) measurements of C4 solutions with varying cobalt dichloride concentrations to determine the corresponding relaxivities, r1 and r2. These relaxation parameters were investigated at different field strengths, temperatures and orientations. T1 measurements obtained at 1.5 and 3.0 T, as well as at room and body temperature, showed that r1 is field-independent and temperature-independent. Conversely, the T2 values at 3.0 T were shorter than at 1.5 T, while the T2 values at body temperature were slightly higher than at room temperature. By examining the relaxivities with the C4 vials aligned in three different planes, we found no orientation-dependence. With these relaxation characteristics, we aim to develop pulse sequences that will enhance the C4 signal against prostatic stroma. Ultimately, the use of C4 as a positive contrast agent marker will encourage the use of MRI to obtain an accurate representation of the radiation dose delivered to the prostate and surrounding normal anatomical structures.

Metabolic shifts induced by fatty acid synthase inhibitor orlistat in non-small cell lung carcinoma cells provide novel pharmacodynamic biomarkers for positron emission tomography and magnetic resonance spectroscopy.

PURPOSE: Abnormal fatty acid (FA) synthesis is one of the common features of cancer. Fatty acid synthase (FASN), a multifunctional enzyme playing a key role in biosynthesis of FA, is up-regulated in prostate, breast, and lung carcinomas. Orlistat is a FDA-approved anti-obesity drug that inhibits the thioesterase domain of FASN, interferes with cellular FA synthesis, can arrest tumor cell proliferation, and induces tumor cell apoptosis. The current study was aimed to investigate the metabolic changes associated with FASN inhibition by orlistat and to understand the molecular mechanisms behind the observed metabolic changes in non-small cell lung carcinoma (NSCLC) cell lines. PROCEDURES: Changes in metabolite pools in four NSCLC cell lines (H441, H1975, H3255, and PC14) with different mutational profiles were studied using NMR spectroscopy before and after in vitro incubation with sub-toxic concentration of orlistat and [1-(13)C]D-glucose or [1,2-(13)C2]choline. In vitro radiotracer accumulation assays in cells were performed with [(3)H]acetate, [(14)C]fluoroacetate, and 2-deoxy-2-[(18)F]fluoro-D-glucose. In parallel, microarray profiling of genes involved in the regulation of carbohydrate and lipid metabolism was performed. RESULTS: In orlistat-treated NSCLC cells, FASN inhibition results in characteristic changes in intermediary metabolites (FAs, choline, phospholipids, and TCA cycle metabolites) as observed by magnetic resonance spectroscopy. Further, FASN inhibition by orlistat induces multiple adaptive changes in FA synthetic pathway and associated metabolic pathways, including induction of ketone metabolism and glutaminolysis, as well as the up-regulation of 5' adenosine monophosphate-activated protein kinase. CONCLUSIONS: These observed changes in metabolic pools in orlistat-treated cells demonstrate the critical role of fatty acid de novo synthesis and metabolism for cellular energy production, especially in tumor cells with low glycolytic activity, which goes beyond the widely accepted concept that FA synthesis is important for cell membrane biosynthesis in rapidly proliferating tumor cells.

Macromolecular dynamic contrast-enhanced (DCE)-MRI detects reduced vascular permeability in a prostate cancer bone metastasis model following anti-platelet-derived growth factor receptor (PDGFR) therapy, indicating a drop in vascular endothelial growth factor receptor (VEGFR) activation.

The antivascular function of the platelet-derived growth factor receptor (PDGFR) inhibitor imatinib combined with paclitaxel has been demonstrated by invasive immunohistochemistry. The purpose of this study was to 1) noninvasively monitor the response to anti-PDGFR treatment, and 2) understand the underlying mechanism of this response. Thus, response to treatment was studied in a prostate cancer bone metastasis model using macromolecular (biotin-bovine serum albumin [BSA]-Gd-diethylene triamine pentaacetic acid [GdDTPA]) dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). Human prostate cancer (PC-3MM2) bone metastases that caused osteolysis and grew in neighboring muscle showed a high blood-volume fraction (fBV) and vascular permeability (PS) at the tumor periphery compared to muscle tissue and intraosseous tumor. Imatinib alone or with paclitaxel significantly reduced PS by 35% (one-tailed paired t-test, P = 0.045) and 40% (P = 0.0003), respectively, whereas paclitaxel alone or no treatment had no effect. Based on changes in PS, we hypothesized that imatinib interferes with the signaling pathway of vascular endothelial growth factor (VEGF). This mechanism was verified by immunohistochemistry. It demonstrated reduced activation of both PDGFR-beta and VEGF receptor 2 (VEGFR2) in imatinib-treated mice. Our study therefore demonstrates that macromolecular DCE-MRI can be used to detect early vascular effects associated with response to therapy targeted to PDGFR, and provides insight into the role played by VEGF in anti-PDGFR therapy.

Fatty acid synthase inhibition results in a magnetic resonance-detectable drop in phosphocholine.

Expression of fatty acid synthase (FASN), the key enzyme in de novo synthesis of long-chain fatty acids, is normally low but increases in cancer. Consequently, FASN is a novel target for cancer therapy. However, because FASN inhibitors can lead to tumor stasis rather than shrinkage, noninvasive methods for assessing FASN inhibition are needed. To this end, we combined (1)H, (31)P, and (13)C magnetic resonance spectroscopy (MRS) (a) to monitor the metabolic consequences of FASN inhibition and (b) to identify MRS-detectable metabolic biomarkers of response. Treatment of PC-3 cells with the FASN inhibitor Orlistat for up to 48 h resulted in inhibition of FASN activity by 70%, correlating with 74% inhibition of fatty acid synthesis. Furthermore, we have determined that FASN inhibition results not only in lower phosphatidylcholine levels but also in a 59% drop in the phospholipid precursor phosphocholine (PCho). This drop resulted from inhibition in PCho synthesis as a result of a reduction in the cellular activity of its synthetic enzyme choline kinase. The drop in PCho levels following FASN inhibition was confirmed in SKOV-3 ovarian cancer cells treated with Orlistat and in MCF-7 breast cancer cells treated with Orlistat as well as cerulenin. Combining data from all treated cells, the drop in PCho significantly correlated with the drop in de novo synthesized fatty acid levels, identifying PCho as a potential noninvasive MRS-detectable biomarker of FASN inhibition in vivo.

Monitoring histone deacetylase inhibition in vivo: noninvasive magnetic resonance spectroscopy method.

Histone deacetylase inhibitors (HDACis) are emerging as promising and selective antitumor agents. However, HDACis can lead to tumor stasis rather than shrinkage, in which case, traditional imaging methods are not adequate to monitor response. Consequently, novel approaches are needed. We have shown in cells that (19)F magnetic resonance spectroscopy (MRS)-detectable levels of the HDAC substrate Boc-Lys-TFA-OH (BLT) are inversely correlated with HDAC activity. We extended our investigations to a tumor xenograft model. Following intraperitoneal injection of BLT, its accumulation within the tumor was monitored by in vivo (19)F MRS. In animals treated with the HDACi suberoylanilide hydroxamic acid (SAHA), tumoral BLT levels were higher by 77% and 132% on days 2 and 7 of treatment compared with pretreatment levels (n = 6; p < .05). In contrast, tumoral BLT levels remained unchanged in control animals and in normal tissue. Thus, (19)F MRS of BLT detected the effect of HDACi treatment as early as day 2 of treatment. Importantly, tumor size confirmed that SAHA treatment leads to inhibition of tumor growth. However, difference in tumor size reached significance only on day 6 of treatment. Thus, this work identifies BLT as a potential molecular imaging agent for the early noninvasive MRS detection of HDAC inhibition in vivo.

Detection of histone deacetylase inhibition by noninvasive magnetic resonance spectroscopy.

Histone deacetylase (HDAC) inhibitors are new and promising antineoplastic agents. Current methods for monitoring early response rely on invasive biopsies or indirect blood-derived markers. Our goal was to develop a magnetic resonance spectroscopy (MRS)-based method to detect HDAC inhibition. The fluorinated lysine derivative Boc-Lys-(Tfa)-OH (BLT) was investigated as a (19)F MRS molecular marker of HDAC activity together with (31)P MRS of endogenous metabolites. In silico modeling of the BLT-HDAC interaction and in vitro MRS studies of BLT cleavage by HDAC confirmed BLT as a HDAC substrate. BLT did not affect cell viability or HDAC activity in PC3 prostate cancer cells. PC3 cells were treated, in the presence of BLT, with the HDAC inhibitor p-fluoro-suberoylanilide hydroxamic acid (FSAHA) over the range of 0 to 10 micromol/L, and HDAC activity and MRS spectra were monitored. Following FSAHA treatment, HDAC activity dropped, reaching 53% of control at 10 micromol/L FSAHA. In parallel, a steady increase in intracellular BLT from 14 to 32 fmol/cell was observed. BLT levels negatively correlated with HDAC activity consistent with higher levels of uncleaved BLT in cells with inhibited HDAC. Phosphocholine, detected by (31)P MRS, increased from 7 to 16 fmol/cell following treatment with FSAHA and also negatively correlated with HDAC activity. Increased phosphocholine is probably due to heat shock protein 90 inhibition as indicated by depletion of client proteins. In summary, (19)F MRS of BLT, combined with (31)P MRS, can be used to monitor HDAC activity in cells. In principle, this could be applied in vivo to noninvasively monitor HDAC activity.