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

Ophthalmology Ophthalmology

A deep learning approach in diagnosing fungal keratitis based on corneal photographs.

In Scientific reports ; h5-index 158.0

Fungal keratitis (FK) is the most devastating and vision-threatening microbial keratitis, but clinical diagnosis a great challenge. This study aimed to develop and verify a deep learning (DL)-based corneal photograph model for diagnosing FK. Corneal photos of laboratory-confirmed microbial keratitis were consecutively collected from a single referral center. A DL framework with DenseNet architecture was used to automatically recognize FK from the photo. The diagnoses of FK via corneal photograph for comparing DL-based models were made in the Expert and NCS-Oph group through a majority decision of three non-corneal specialty ophthalmologist and three corneal specialists, respectively. The average percentage of sensitivity, specificity, positive predictive value, and negative predictive value was approximately 71, 68, 60, and 78. The sensitivity was higher than that of the NCS-Oph (52%, P < .01), whereas the specificity was lower than that of the NCS-Oph (83%, P < .01). The average accuracy of around 70% was comparable with that of the NCS-Oph. Therefore, the sensitive DL-based diagnostic model is a promising tool for improving first-line medical care at rural area in early identification of FK.

Kuo Ming-Tse, Hsu Benny Wei-Yun, Yin Yu-Kai, Fang Po-Chiung, Lai Hung-Yin, Chen Alexander, Yu Meng-Shan, Tseng Vincent S


General General

AI-based prediction for the risk of coronary heart disease among patients with type 2 diabetes mellitus.

In Scientific reports ; h5-index 158.0

Type 2 diabetes mellitus (T2DM) is one common chronic disease caused by insulin secretion disorder that often leads to severe outcomes and even death due to complications, among which coronary heart disease (CHD) represents the most common and severe one. Given a huge number of T2DM patients, it is thus increasingly important to identify the ones with high risks of CHD complication but the quantitative method is still not available. Here, we first curated a dataset of 1,273 T2DM patients including 304 and 969 ones with or without CHD, respectively. We then trained an artificial intelligence (AI) model using randomly selected 4/5 of the dataset and use the rest data to validate the performance of the model. The result showed that the model achieved an AUC of 0.77 (fivefold cross-validation) on the training dataset and 0.80 on the testing dataset. To further confirm the performance of the presented model, we recruited 1,253 new T2DM patients as totally independent testing dataset including 200 and 1,053 ones with or without CHD. And the model achieved an AUC of 0.71. In addition, we implemented a model to quantitatively evaluate the risk contribution of each feature, which is thus able to present personalized guidance for specific individuals. Finally, an online web server for the model was built. This study presented an AI model to determine the risk of T2DM patients to develop to CHD, which has potential value in providing early warning personalized guidance of CHD risk for both T2DM patients and clinicians.

Fan Rui, Zhang Ning, Yang Longyan, Ke Jing, Zhao Dong, Cui Qinghua


General General

[Application of scattering microscopy for evaluation of iPS cell and its differentiated cells].

In Nihon yakurigaku zasshi. Folia pharmacologica Japonica

Various artificial cells and artificial tissues can be generated from induced pluripotent stem cells (iPS cells). There is now an urgent need to standardize the quality evaluation and management of iPS cells. Recently, artificial intelligence (AI) technology such as machine learning is providing evaluation method for the quality of iPS cells and iPS cell-derived somatic cells based on optical microscopy. Light, which is the principle of optical microscopy, has an interesting and important feature. There are various kinds of interaction between light and molecule, and the scattered light includes internal information of the molecule. Raman scattering inheres all the vibration mode of molecular bonds composing a molecule, and second harmonic generation (SHG) light, which is one of second-order non-linear scattering light, is derived from electric polarizations in the molecule, in other words, carries structural information within the protein. While states of a cell are usually defined by protein/gene expression patterns, we have proposed to apply Raman spectra for cellular fingerprinting as an alternative for identifying the cell state, and now succeeded in predicting gene-expression of antibiotic resistant bacteria in combination with machine learning technology. Meanwhile, SHG microscopy has been used to visualize fiber structures in living specimens, such as collagen, and microtubules as a label-free modality. By utilizing the feature that SHG senses protein structure change, we developed a new method to measure actomyosin activity in cardiac cells. The most important advantage of the use of the scattering light is their non-labeling and non-invasive capability.

Watanabe Tomonobu M, Fujita Hideaki, Kaneshiro Junichi


Public Health Public Health

Proteomics and Metabolomics Approaches towards a Functional Insight onto AUTISM Spectrum Disorders: Phenotype Stratification and Biomarker Discovery.

In International journal of molecular sciences ; h5-index 102.0

Autism spectrum disorders (ASDs) are neurodevelopmental disorders characterized by behavioral alterations and currently affect about 1% of children. Significant genetic factors and mechanisms underline the causation of ASD. Indeed, many affected individuals are diagnosed with chromosomal abnormalities, submicroscopic deletions or duplications, single-gene disorders or variants. However, a range of metabolic abnormalities has been highlighted in many patients, by identifying biofluid metabolome and proteome profiles potentially usable as ASD biomarkers. Indeed, next-generation sequencing and other omics platforms, including proteomics and metabolomics, have uncovered early age disease biomarkers which may lead to novel diagnostic tools and treatment targets that may vary from patient to patient depending on the specific genomic and other omics findings. The progressive identification of new proteins and metabolites acting as biomarker candidates, combined with patient genetic and clinical data and environmental factors, including microbiota, would bring us towards advanced clinical decision support systems (CDSSs) assisted by machine learning models for advanced ASD-personalized medicine. Herein, we will discuss novel computational solutions to evaluate new proteome and metabolome ASD biomarker candidates, in terms of their recurrence in the reviewed literature and laboratory medicine feasibility. Moreover, the way to exploit CDSS, performed by artificial intelligence, is presented as an effective tool to integrate omics data to electronic health/medical records (EHR/EMR), hopefully acting as added value in the near future for the clinical management of ASD.

Ristori Maria Vittoria, Mortera Stefano Levi, Marzano Valeria, Guerrera Silvia, Vernocchi Pamela, Ianiro Gianluca, Gardini Simone, Torre Giuliano, Valeri Giovanni, Vicari Stefano, Gasbarrini Antonio, Putignani Lorenza


autism spectrum disorders (ASDs), clinical decision support systems (CDSSs), disease biomarkers, interactomics, metabolomics, proteomics

Radiology Radiology

Radiomics at a Glance: A Few Lessons Learned from Learning Approaches.

In Cancers

Processing and modeling medical images have traditionally represented complex tasks requiring multidisciplinary collaboration. The advent of radiomics has assigned a central role to quantitative data analytics targeting medical image features algorithmically extracted from large volumes of images. Apart from the ultimate goal of supporting diagnostic, prognostic, and therapeutic decisions, radiomics is computationally attractive due to specific strengths: scalability, efficiency, and precision. Optimization is achieved by highly sophisticated statistical and machine learning algorithms, but it is especially deep learning that stands out as the leading inference approach. Various types of hybrid learning can be considered when building complex integrative approaches aimed to deliver gains in accuracy for both classification and prediction tasks. This perspective reviews some selected learning methods by focusing on both their significance for radiomics and their unveiled potential.

Capobianco Enrico, Deng Jun


integrative inference approaches, machine learning, predictive modeling, radiomics

General General

Deep-Pneumonia Framework Using Deep Learning Models Based on Chest X-Ray Images.

In Diagnostics (Basel, Switzerland)

Pneumonia is a contagious disease that causes ulcers of the lungs, and is one of the main reasons for death among children and the elderly in the world. Several deep learning models for detecting pneumonia from chest X-ray images have been proposed. One of the extreme challenges has been to find an appropriate and efficient model that meets all performance metrics. Proposing efficient and powerful deep learning models for detecting and classifying pneumonia is the main purpose of this work. In this paper, four different models are developed by changing the used deep learning method; two pre-trained models, ResNet152V2 and MobileNetV2, a Convolutional Neural Network (CNN), and a Long Short-Term Memory (LSTM). The proposed models are implemented and evaluated using Python and compared with recent similar research. The results demonstrate that our proposed deep learning framework improves accuracy, precision, F1-score, recall, and Area Under the Curve (AUC) by 99.22%, 99.43%, 99.44%, 99.44%, and 99.77%, respectively. As clearly illustrated from the results, the ResNet152V2 model outperforms other recently proposed works. Moreover, the other proposed models-MobileNetV2, CNN, and LSTM-CNN-achieved results with more than 91% in accuracy, recall, F1-score, precision, and AUC, and exceed the recently introduced models in the literature.

Elshennawy Nada M, Ibrahim Dina M


CNN, LSTM, chest X-ray image, deep learning, detecting pneumonia