Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

General General

Influence of medical domain knowledge on deep learning for Alzheimer's disease prediction.

In Computer methods and programs in biomedicine

BACKGROUND AND OBJECTIVE : Alzheimer's disease (AD) is the most common type of dementia that can seriously affect a person's ability to perform daily activities. Estimates indicate that AD may rank third as a cause of death for older people, after heart disease and cancer. Identification of individuals at risk for developing AD is imperative for testing therapeutic interventions. The objective of the study was to determine could diagnostics of AD from EMR data alone (without relying on diagnostic imaging) be significantly improved by applying clinical domain knowledge in data preprocessing and positive dataset selection rather than setting naïve filters.

METHODS : Data were extracted from the repository of heterogeneous ambulatory EMR data, collected from primary care medical offices all over the U.S. Medical domain knowledge was applied to build a positive dataset from data relevant to AD. Selected Clinically Relevant Positive (SCRP) datasets were used as inputs to a Long-Short-Term Memory (LSTM) Recurrent Neural Network (RNN) deep learning model to predict will the patient develop AD.

RESULTS : Risk scores prediction of AD using the drugs domain information in an SCRP AD dataset of 2,324 patients achieved high out-of-sample score - 0.98-0.99 Area Under the Precision-Recall Curve (AUPRC) when using 90% of SCRP dataset for training. AUPRC dropped to 0.89 when training the model using less than 1,500 cases from the SCRP dataset. The model was still significantly better than when using naïve dataset selection.

CONCLUSION : The LSTM RNN method that used data relevant to AD performed significantly better when learning from the SCRP dataset than when datasets were selected naïvely. The integration of qualitative medical knowledge for dataset selection and deep learning technology provided a mechanism for significant improvement of AD prediction. Accurate and early prediction of AD is significant in the identification of patients for clinical trials, which can possibly result in the discovery of new drugs for treatments of AD. Also, the contribution of the proposed predictions of AD is a better selection of patients who need imaging diagnostics for differential diagnosis of AD from other degenerative brain disorders.

Ljubic Branimir, Roychoudhury Shoumik, Cao Xi Hang, Pavlovski Martin, Obradovic Stefan, Nair Richard, Glass Lucas, Obradovic Zoran


“Alzheimers disease prediction”, Cognitive impairment, Deep learning, Electronic medical records, Recurrent Neural Networks

Cardiology Cardiology

Prediction of incident myocardial infarction using machine learning applied to harmonized electronic health record data.

In BMC medical informatics and decision making ; h5-index 38.0

BACKGROUND : With cardiovascular disease increasing, substantial research has focused on the development of prediction tools. We compare deep learning and machine learning models to a baseline logistic regression using only 'known' risk factors in predicting incident myocardial infarction (MI) from harmonized EHR data.

METHODS : Large-scale case-control study with outcome of 6-month incident MI, conducted using the top 800, from an initial 52 k procedures, diagnoses, and medications within the UCHealth system, harmonized to the Observational Medical Outcomes Partnership common data model, performed on 2.27 million patients. We compared several over- and under- sampling techniques to address the imbalance in the dataset. We compared regularized logistics regression, random forest, boosted gradient machines, and shallow and deep neural networks. A baseline model for comparison was a logistic regression using a limited set of 'known' risk factors for MI. Hyper-parameters were identified using 10-fold cross-validation.

RESULTS : Twenty thousand Five hundred and ninety-one patients were diagnosed with MI compared with 2.25 million who did not. A deep neural network with random undersampling provided superior classification compared with other methods. However, the benefit of the deep neural network was only moderate, showing an F1 Score of 0.092 and AUC of 0.835, compared to a logistic regression model using only 'known' risk factors. Calibration for all models was poor despite adequate discrimination, due to overfitting from low frequency of the event of interest.

CONCLUSIONS : Our study suggests that DNN may not offer substantial benefit when trained on harmonized data, compared to traditional methods using established risk factors for MI.

Mandair Divneet, Tiwari Premanand, Simon Steven, Colborn Kathryn L, Rosenberg Michael A


Electronic health records, Machine learning, Myocardial infarction

General General

Application of ultrasound artificial intelligence in the differential diagnosis between benign and malignant breast lesions of BI-RADS 4A.

In BMC cancer

BACKGROUND : The classification of Breast Imaging Reporting and Data System 4A (BI-RADS 4A) lesions is mostly based on the personal experience of doctors and lacks specific and clear classification standards. The development of artificial intelligence (AI) provides a new method for BI-RADS categorisation. We analysed the ultrasonic morphological and texture characteristics of BI-RADS 4A benign and malignant lesions using AI, and these ultrasonic characteristics of BI-RADS 4A benign and malignant lesions were compared to examine the value of AI in the differential diagnosis of BI-RADS 4A benign and malignant lesions.

METHODS : A total of 206 lesions of BI-RADS 4A examined using ultrasonography were analysed retrospectively, including 174 benign lesions and 32 malignant lesions. All of the lesions were contoured manually, and the ultrasonic morphological and texture features of the lesions, such as circularity, height-to-width ratio, margin spicules, margin coarseness, margin indistinctness, margin lobulation, energy, entropy, grey mean, internal calcification and angle between the long axis of the lesion and skin, were calculated using grey level gradient co-occurrence matrix analysis. Differences between benign and malignant lesions of BI-RADS 4A were analysed.

RESULTS : Significant differences in margin lobulation, entropy, internal calcification and ALS were noted between the benign group and malignant group (P = 0.013, 0.045, 0.045, and 0.002, respectively). The malignant group had more margin lobulations and lower entropy compared with the benign group, and the benign group had more internal calcifications and a greater angle between the long axis of the lesion and skin compared with the malignant group. No significant differences in circularity, height-to-width ratio, margin spicules, margin coarseness, margin indistinctness, energy, and grey mean were noted between benign and malignant lesions.

CONCLUSIONS : Compared with the naked eye, AI can reveal more subtle differences between benign and malignant BI-RADS 4A lesions. These results remind us carefully observation of the margin and the internal echo is of great significance. With the help of morphological and texture information provided by AI, doctors can make a more accurate judgment on such atypical benign and malignant lesions.

Niu Sihua, Huang Jianhua, Li Jia, Liu Xueling, Wang Dan, Zhang Ruifang, Wang Yingyan, Shen Huiming, Qi Min, Xiao Yi, Guan Mengyao, Liu Haiyan, Li Diancheng, Liu Feifei, Wang Xiuming, Xiong Yu, Gao Siqi, Wang Xue, Zhu Jiaan


Artificial intelligence, BI-RADS 4A, Breast, Differential diagnosis

General General

Database: web application for visualization of the cumulated RNAseq data against the salicylic acid (SA) and methyl jasmonate (MeJA) treatment of Arabidopsis thaliana.

In BMC plant biology

BACKGROUND : Plants have adapted to survive under adverse conditions or exploit favorable conditions in response to their environment as sessile creatures. In a way of plant adaptation, plant hormones have been evolved to efficiently use limited resources. Plant hormones including auxin, jasmonic acid, salicylic acid, and ethylene have been studied to reveal their role in plant adaptation against their environment by phenotypic observation with experimental design such as mutation on hormone receptors and treatment / non-treatment of plant hormones along with other environmental conditions. With the development of Next Generation Sequencing (NGS) technology, it became possible to score the total gene expression of the sampled plants and estimate the degree of effect of plant hormones in gene expression. This allowed us to infer the signaling pathway through plant hormones, which greatly stimulated the study of functional genomics using mutants. Due to the continued development of NGS technology and analytical techniques, many plant hormone-related studies have produced and accumulated NGS-based data, especially RNAseq data have been stored in the sequence read archive represented by NCBI, EBI, and DDBJ.

DESCRIPTION : Here, hormone treatment RNAseq data of Arabidopsis (Col0), wild-type genotype, were collected with mock, SA, and MeJA treatments. The genes affected by hormones were identified through a machine learning approach. The degree of expression of the affected gene was quantified, visualized in boxplot using d3 (data-driven-document), and the database was built by Django.

CONCLUSION : Using this database, we created a web application ( ) that lists hormone-related or hormone-affected genes and visualizes the boxplot of the gene expression of selected genes. This web application eventually aids the functional genomics researchers who want to gather the cases of the gene responses by the hormones.

Woo Dong U, Jeon Ho Hwi, Park Halim, Park Jin Hwa, Lee Yejin, Kang Yang Jae


Arabidopsis, Database, Jasmonic acid, RNAseq, Salicylic acid, Web-application

Cardiology Cardiology

Sustainability and Versatility of the ABCDE Protocol for Stress Echocardiography.

In Journal of clinical medicine

For the past 40 years, the methodology for stress echocardiography (SE) has remained basically unchanged. It is based on two-dimensional, black and white imaging, and is used to detect regional wall motion abnormalities (RWMA) in patients with known or suspected coronary artery disease (CAD). In the last five years much has changed and RWMA is not enough on its own to stratify patient risk and dictate therapy. Patients arriving at SE labs often have comorbidities and are undergoing full anti-ischemic therapy. The SE positivity rate based on RWMA fell from 70% in the eighties to 10% in the last decade. The understanding of CAD pathophysiology has shifted from a regional hydraulic disease to a systemic biologic disease. The conventional view of CAD encouraged the use of coronary anatomic imaging for diagnosis and the oculo-stenotic reflex for the deployment of therapy. This has led to a clinical oversimplification that ignores the lessons of pathophysiology and epidemiology, and in fact, CAD is not synonymous with ischemic heart disease. Patients with CAD may also have other vulnerabilities such as coronary plaque (step A of ABCDE-SE), alveolar-capillary membrane and pulmonary congestion (step B), preload and contractile reserve (step C), coronary microcirculation (step D) and cardiac autonomic balance (step E). The SE methodology based on two-dimensional echocardiography is now integrated with lung ultrasound (step B for B-lines), volumetric echocardiography (step C), color- and pulsed-wave Doppler (step D) and non-imaging electrocardiogram-based heart rate assessment (step E). In addition, qualitative assessment based on the naked eye has now become more quantitative, has been improved by contrast and based on cardiac strain and artificial intelligence. ABCDE-SE is now ready for large scale multicenter testing in the SE2030 study.

Picano Eugenio, Zagatina Angela, Wierzbowska-Drabik Karina, Borguezan Daros Clarissa, D’Andrea Antonello, Ciampi Quirino


coronary artery disease, functional test, heart failure, stress echo, sustainability

Pathology Pathology

Metabolomics in Sleep, Insomnia and Sleep Apnea.

In International journal of molecular sciences ; h5-index 102.0

Sleep-wake disorders are highly prevalent disorders, which can lead to negative effects on cognitive, emotional and interpersonal functioning, and can cause maladaptive metabolic changes. Recent studies support the notion that metabolic processes correlate with sleep. The study of metabolite biomarkers (metabolomics) in a large-scale manner offers unique opportunities to provide insights into the pathology of diseases by revealing alterations in metabolic pathways. This review aims to summarize the status of metabolomic analyses-based knowledge on sleep disorders and to present knowledge in understanding the metabolic role of sleep in psychiatric disorders. Overall, findings suggest that sleep-wake disorders lead to pronounced alterations in specific metabolic pathways, which might contribute to the association of sleep disorders with other psychiatric disorders and medical conditions. These alterations are mainly related to changes in the metabolism of branched-chain amino acids, as well as glucose and lipid metabolism. In insomnia, alterations in branched-chain amino acid and glucose metabolism were shown among studies. In obstructive sleep apnea, biomarkers related to lipid metabolism seem to be of special importance. Future studies are needed to examine severity, subtypes and treatment of sleep-wake disorders in the context of metabolite levels.

Humer Elke, Pieh Christoph, Brandmayr Georg


insomnia, mental disorders, metabolomics, sleep, sleep apnea