Machine learning to predict notes for chart review in the oncology setting: a proof of concept strategy for improving clinician note-writing.

OBJECTIVE: Leverage electronic health record (EHR) audit logs to develop a machine learning (ML) model that predicts which notes a clinician wants to review when seeing oncology patients. MATERIALS AND METHODS: We trained logistic regression models using note metadata and a Term Frequency Inverse Document Frequency (TF-IDF) text representation. We evaluated performance with precision, recall, F1, AUC, and a clinical qualitative assessment. RESULTS: The metadata only model achieved an AUC 0.930 and the metadata and TF-IDF model an AUC 0.937. Qualitative assessment revealed a need for better text representation and to further customize predictions for the user. DISCUSSION: Our model effectively surfaces the top 10 notes a clinician wants to review when seeing an oncology patient. Further studies can characterize different types of clinician users and better tailor the task for different care settings. CONCLUSION: EHR audit logs can provide important relevance data for training ML models that assist with note-writing in the oncology setting.

Association of Maternal Serum Ischemia Modified Albumin (IMA) with Placental Histopathological Changes and Fetomaternal Outcome: A Prospective Case Control Study in Normotensive and Pre-eclamptic Women.

Introduction: Ischemia and oxidative stress leads to generation of hydroxyl free radicals and modification of 'N-terminus' of human serum albumin. This modified albumin molecule, known as Ischemia Modified Albumin (IMA), is elevated in early stages of ischemia. It has recently been approved by US Food and Drug Administration (US FDA) for its clinical use, as early marker of myocardial ischemia in cardiology. IMA is a novel marker of ischemia and is elevated in other clinical conditions associated with ischemia like pulmonary embolism, uncontrolled type II diabetes mellitus, acute decompensated heart failure, preeclampsia, recurrent pregnancy losses and IUGR. Role of IMA in birth asphyxia in perinatology is of current interest and needs further research. Methodology: A prospective case control study was conducted in a tertiary center in North India for one year. Total 80 pregnant women between 34 and 40 weeks were recruited and allocated in two groups. Case group comprised of 40 pre-eclamptic pregnant women and control group comprised of 40 normotensive pregnant women. Comparison and association of maternal serum IMA levels with fetomaternal outcome and number and types of placental histopathological changes was done in two groups. Results: In preeclampsia group mean serum IMA (115.23 ± 49.51) was significantly higher as compared to the normotensive group (79.21 ± 14.35). The optimum cut off value of IMA to detect a case was 94.5 IU/ml (sensitivity 65%, specificity 87.5%, PPV 83.9%, NPV 71.4% and diagnostic accuracy of 76.3). Pre-eclamptic women, had significantly higher incidence of PTVD, lower fetal birth weight and placental histopathological changes as compared to normotensive group. 83.8% of the women with raised IMA levels were pre-eclamptic. Raised IMA levels were significantly associated with higher incidence of PTVD, birth weight ≤ 2 kg and hypoxic histopathological lesions of chorangiosis, intervillous fibrin and hyalinization. Conclusion: Determination of maternal serum IMA levels early in pregnancy can predict preeclampsia and avoid future severe preeclampsia related complications. It might be useful to optimize both maternal and fetal/neonatal outcomes.

Cervical gestational trophoblastic neoplasia: A rare form of gestational trophoblastic disease.

Gestational trophoblastic neoplasia (GTN) is a rare disease and is characterized by an abnormal proliferation of trophoblastic cells of placenta. Since majority of them are chemotherapy sensitive, they are one of the highly curable cancers. However, due to its varied presentation, clinicians still face many challenges in its diagnosis and management. We present a case of 26-year-old woman, P0 + 3 (previous 3 abortions) who presented to us with a history of vaginal bleeding postuterine evacuation after 2 months of incomplete abortion. After clinical, radiological assessment and high human chorionic gonadotropin (hCG) titers, the patient was diagnosed as a case of cervical GTN. Risk assessment by the WHO prognostic scoring system was done and "Low Risk" was assigned to her. She was given total nine cycles of single agent chemotherapy including two consolidation cycles after normalization of hCG titers which patient tolerated well and remained asymptomatic.

Altered pro-inflammatory and anti-inflammatory plasma cytokines levels in children with Down's syndrome: A meta-analysis.

BACKGROUND: Down syndrome (DS) is the commonest chromosomal anomalies at birth. DS is portrayed by the event of extra complete/deficient duplicate of chromosome number 21 (trisomy 21). Around the world, this disordered influencing roughly 1 out of 1000 infants. Pro-inflammatory and anti-inflammatory cytokines engaged with a few physiological procedures involving the guideline of inflammatory reactions. In DS kids, the creation of few important inflammatory and anti-inflammatory cytokines is altered. Different investigations shows that the cytokines are dysregulated in patients with DS. In this study, we led a meta-analysis to evaluate the connections of pro-inflammatory and anti-inflammatory cytokine changes in youngsters with DS patients. METHODOLOGY: We searched PubMed, Google and Web of Science for studies in exploring the association of pro-inflammatory and anti-inflammatory serum level with DS patients. Total 10 studies were included in the meta-analysis. The random effects were used to analyze the pooled data. All statistical tests were two-sided. RESULTS: High circulating level of serum MCP-1 was significantly associated with DS [Cohen's d = 143.91 95% confidence interval (CI) =110.38-177.43]. However, the other circulating cytokines IL-2 and IL-17 level were lower whereas IL-13 level was higher but not significantly different in DS as contrasted to healthy controls. The heterogeneity level was higher in IL-2, IL-13 and IL-17 cytokines. CONCLUSION: This meta-analysis shows that the higher circulating level of MCP-1 was associated with DS.

Automated NLP Extraction of Clinical Rationale for Treatment Discontinuation in Breast Cancer.

PURPOSE: Key oncology end points are not routinely encoded into electronic medical records (EMRs). We assessed whether natural language processing (NLP) can abstract treatment discontinuation rationale from unstructured EMR notes to estimate toxicity incidence and progression-free survival (PFS). METHODS: We constructed a retrospective cohort of 6,115 patients with early-stage and 701 patients with metastatic breast cancer initiating care at Memorial Sloan Kettering Cancer Center from 2008 to 2019. Each cohort was divided into training (70%), validation (15%), and test (15%) subsets. Human abstractors identified the clinical rationale associated with treatment discontinuation events. Concatenated EMR notes were used to train high-dimensional logistic regression and convolutional neural network models. Kaplan-Meier analyses were used to compare toxicity incidence and PFS estimated by our NLP models to estimates generated by manual labeling and time-to-treatment discontinuation (TTD). RESULTS: Our best high-dimensional logistic regression models identified toxicity events in early-stage patients with an area under the curve of the receiver-operator characteristic of 0.857 ± 0.014 (standard deviation) and progression events in metastatic patients with an area under the curve of 0.752 ± 0.027 (standard deviation). NLP-extracted toxicity incidence and PFS curves were not significantly different from manually extracted curves (P = .95 and P = .67, respectively). By contrast, TTD overestimated toxicity in early-stage patients (P < .001) and underestimated PFS in metastatic patients (P < .001). Additionally, we tested an extrapolation approach in which 20% of the metastatic cohort were labeled manually, and NLP algorithms were used to abstract the remaining 80%. This extrapolated outcomes approach resolved PFS differences between receptor subtypes (P < .001 for hormone receptor+/human epidermal growth factor receptor 2- v human epidermal growth factor receptor 2+ v triple-negative) that could not be resolved with TTD. CONCLUSION: NLP models are capable of abstracting treatment discontinuation rationale with minimal manual labeling.

MTHFR C677T, Prothrombin G20210A, and Factor V Leiden (G1691A) Polymorphism and Beta-Thalassemia Risk: A Meta-Analysis.

Background Beta (ß)-thalassemia major patients frequently suffer from many vascular problems. Thrombophilia is a blood disorder that comprises imbalances in the blood coagulating factor due to ecological and hereditary components. Previous evidence shows that thrombosis is the commonest risk in beta-thalassemia patients. Several studies have examined that MTHFR C677T, prothrombin G20210A (PT G20210A), and Factor V Leiden G1691A (FVL G1691A) polymorphism play a crucial role in the development of ß-thalassemia major, yet the result was questionable and uncertain. Therefore, in this study, we executed the correlation between these gene polymorphisms with ß-thalassemia major patients. Methods Suitable keywords were used to search related articles in PubMed, Google Scholar, and Web of Science. In this random-effects meta-analysis, we analyzed the odds ratio (OR) for the estimation of risk. Results A total of nine research articles with 645 ß-thalassemia major patients and 989 healthy controls were incorporated in this meta-analysis. The pooled OR was assessed in MTHFR C677T, PT G20210A, and FVL G1691A polymorphism. This random-effects meta-analysis demonstrated that MTHFR C677T, PT G20210A, and FVL G1691A gene polymorphism did not significantly associate with ß-thalassemia major. Moreover, the heterogeneity was significantly found in genotype CC vs CT+TT C677T (I2=61%) and allele C vs T (I2=71%) of MTHFR and genotype GG vs GA (I2=95%), GG vs GA+AA (I2=95%), GA vs GG+AA (I2=95%), and allele G vs A (I2=93%) of FVL G1691A. Conclusion The results of this meta-analysis show that MTHFR C677T, prothrombin G20210A, and Factor V Leiden (G1691A) gene polymorphism are not a risk factor for ß-thalassemia major.

A demographic prevalence of ß Thalassemia carrier and other hemoglobinopathies in adolescent of Tharu population.

BACKGROUND AND AIMS: Hemoglobinopathies and thalassemias are the commonest single gene disorders in India. In Terai region of India, Hemoglobinopathies and thalassemias are the most common in the Tharu community. Therefore, in this study, we aim to evaluate the Hb variant analysis of hemoglobinopathies and thalassemias in a Tharu population in Lakhimpur Kheri Districts of Uttar Pradesh, India. MATERIALS AND METHODS: Total 493 individuals were recruited in this study. The demographic details and blood samples were collected from different location at Kheri district during mega health camp. Hb variant analysis was performed by high performance liquid chromatography (HPLC) system beta thalassemia short program in BIO-RAD VARIANT. RESULTS: Out of 493, 108 (21.9%) individual suffers with abnormal haemoglobinopathies. In which ß-thalassemia trait is the commonest haemoglobinopathy (12.98%), followed by HbE trait (7.50%), and compound heterozygous HbS/ß-Thalassemia trait (1.42%) in overall population. The HbF was significantly greater in HbS heterozygous (1.45 ± 1.41), whereas mean HbA2 was significantly greater in ß-Thalassemia trait (5.17 ± 1.36). CONCLUSION: The high incidence of hemoglobinopathies and thalassemias were observed in Tharu community in Lakhimpur Kheri districts of Uttar Pradesh, Indian.

Robustly Extracting Medical Knowledge from EHRs: A Case Study of Learning a Health Knowledge Graph.

Increasingly large electronic health records (EHRs) provide an opportunity to algorithmically learn medical knowledge. In one prominent example, a causal health knowledge graph could learn relationships between diseases and symptoms and then serve as a diagnostic tool to be refined with additional clinical input. Prior research has demonstrated the ability to construct such a graph from over 270,000 emergency department patient visits. In this work, we describe methods to evaluate a health knowledge graph for robustness. Moving beyond precision and recall, we analyze for which diseases and for which patients the graph is most accurate. We identify sample size and unmeasured confounders as major sources of error in the health knowledge graph. We introduce a method to leverage non-linear functions in building the causal graph to better understand existing model assumptions. Finally, to assess model generalizability, we extend to a larger set of complete patient visits within a hospital system. We conclude with a discussion on how to robustly extract medical knowledge from EHRs.

Does endoscopic sclerotherapy in filarial chyluria affect renal function and morphology? A prospective study using dimercaptosuccinic acid renal scan.

OBJECTIVE: To look for change in relative renal function and document renal scarring following endoscopic renal pelvic instillation sclerotherapy (RPIS) in patients with chyluria by dimercaptosuccinic acid (DMSA) renal scan. METHODS: A prospective study was performed between November 2015 and September 2016. All patients with biochemically documented chyluria who underwent RPIS using either 1%-silver nitrate or 0.1%-povidine iodine were included. Patients received either 3-, 6- or 9-doses. DMSA renal scan was performed before and 2-3 months after sclerotherapy. RESULTS: Of the 34 patients, 22 were males. Mean age was 41.08 ± 16.64 years (range, 15-70 years). Thirty-two patients (94.1%) responded to therapy while two did not respond even after 9-doses. Average follow-up was 8.94 ± 3.70 months. The mean relative renal function (pre-instillation) of normal kidney was 50.76% ± 3.55% while that of affected renal unit (side of instillation) was 49.20% ± 3.44% (range, 43.0%-61.0%). After instillation therapy, the mean relative renal function of normal side was 52.26% ± 3.57% while that of affected renal unit was 47.50% ± 3.56% (range, 41.0%-54.0%). The relative renal function did not change >5% from the baseline value in any patient except one (in which the differential function increased paradoxically by 12%). Two patients developed renal scar in post-instillation renal scan. CONCLUSION: Endoscopic sclerotherapy in chyluria is safe and effective. The relative renal function does not deteriorate by more than 5%. There is a small risk of development of renal scar. More studies involving larger number of patients are needed to answer this dilemma.

Modeling polypharmacy side effects with graph convolutional networks.

Motivation: The use of drug combinations, termed polypharmacy, is common to treat patients with complex diseases or co-existing conditions. However, a major consequence of polypharmacy is a much higher risk of adverse side effects for the patient. Polypharmacy side effects emerge because of drug-drug interactions, in which activity of one drug may change, favorably or unfavorably, if taken with another drug. The knowledge of drug interactions is often limited because these complex relationships are rare, and are usually not observed in relatively small clinical testing. Discovering polypharmacy side effects thus remains an important challenge with significant implications for patient mortality and morbidity. Results: Here, we present Decagon, an approach for modeling polypharmacy side effects. The approach constructs a multimodal graph of protein-protein interactions, drug-protein target interactions and the polypharmacy side effects, which are represented as drug-drug interactions, where each side effect is an edge of a different type. Decagon is developed specifically to handle such multimodal graphs with a large number of edge types. Our approach develops a new graph convolutional neural network for multirelational link prediction in multimodal networks. Unlike approaches limited to predicting simple drug-drug interaction values, Decagon can predict the exact side effect, if any, through which a given drug combination manifests clinically. Decagon accurately predicts polypharmacy side effects, outperforming baselines by up to 69%. We find that it automatically learns representations of side effects indicative of co-occurrence of polypharmacy in patients. Furthermore, Decagon models particularly well polypharmacy side effects that have a strong molecular basis, while on predominantly non-molecular side effects, it achieves good performance because of effective sharing of model parameters across edge types. Decagon opens up opportunities to use large pharmacogenomic and patient population data to flag and prioritize polypharmacy side effects for follow-up analysis via formal pharmacological studies. Availability and implementation: Source code and preprocessed datasets are at: http://snap.stanford.edu/decagon.

Large-scale analysis of disease pathways in the human interactome.

Discovering disease pathways, which can be defined as sets of proteins associated with a given disease, is an important problem that has the potential to provide clinically actionable insights for disease diagnosis, prognosis, and treatment. Computational methods aid the discovery by relying on protein-protein interaction (PPI) networks. They start with a few known disease-associated proteins and aim to find the rest of the pathway by exploring the PPI network around the known disease proteins. However, the success of such methods has been limited, and failure cases have not been well understood. Here we study the PPI network structure of 519 disease pathways. We find that 90% of pathways do not correspond to single well-connected components in the PPI network. Instead, proteins associated with a single disease tend to form many separate connected components/regions in the network. We then evaluate state-of-the-art disease pathway discovery methods and show that their performance is especially poor on diseases with disconnected pathways. Thus, we conclude that network connectivity structure alone may not be sufficient for disease pathway discovery. However, we show that higher-order network structures, such as small subgraphs of the pathway, provide a promising direction for the development of new methods.

Evaluation of Impact of Voiding Posture on Uroflowmetry Parameters in Men.

PURPOSE: To evaluate the impact of voiding position on uroflowmetry parameters and to assess its potential clinical implications. MATERIALS AND METHODS: We conducted a prospective study from 2013 to 2015 and included men between 18 and 77 years old who were either healthy volunteers with an International Prostate Symptom Score (IPSS) ≤7 or men with benign prostate enlargement that were on alpha-blocker medication and had an IPSS <10. Participants underwent uroflowmetry and post-void residual urine (PVRU) measurements twice, once in a sitting position and once in a standing position. The participants were divided into 4 groups based on age (35 years or younger, 36 to 50 years, 51 to 60 years, and older than 60 years). RESULTS: A total of 740 men with a mean age of 40.35 years were evaluated. There was no significant difference in uroflowmetry parameters until the age of 50 years between the voiding positions. However, in those older than 50 years, PVRU volume was significantly lower in the sitting position than the standing position, whereas voiding time was significantly higher in the sitting position than the standing position. Other uroflowmetry parameters, including maximal and average urine flow rates, were non-significant. CONCLUSIONS: The voiding position plays an important role in the uroflowmetry parameters of elderly men. Voiding in the sitting position was found to be optimal for elderly men, whereas the role of the voiding position in healthy young men could not be determined. More research is needed to further study this issue.

Fabrication of healthy and disease-mimicking retinal phantoms with tapered foveal pits for optical coherence tomography.

Optical coherence tomography (OCT) has become a standard tool in ophthalmology clinics for diagnosing many retinal diseases. Nonetheless, the technical and clinical communities still lack a standardized phantom that could aid in evaluating and normalizing the many protocols and systems used for diagnosis. Existing retinal phantoms are able to mimic the thickness and scattering properties of the retinal layers but are unable to model the morphology of the foveal pit, particularly the tapering of the retinal layers. This work demonstrates a new fabrication procedure that is capable of reliably and consistently replicating the shape and tapered appearance of the retinal layers near the foveal pit using a combination of spin-coating and replica molding. We characterize the effects of using different mold sizes which enable us to achieve a range of pit dimensions. We also present a modified procedure to replicate two diseased states of the retinal tissue, such as retinal detachment and dry aged-related macular degeneration. The ability to create an anatomically correct foveal pit for healthy and disease-mimicking phantoms will allow for a new standard better suited for intra- and inter-system evaluation and for improved comparison of retinal segmentation algorithms

Single-shot speckle noise reduction by interleaved optical coherence tomography.

Speckle noise is one of the dominant factors that degrade image quality in optical coherence tomography (OCT). Here, we propose a new strategy, interleaved OCT (iOCT), for spatial compounding and angular compounding. We demonstrate the efficiency of compounding with iOCT to restrain speckle noise without compromising imaging speed in phantoms and tissue samples.