Background: Numerous applications of artificial intelligence have been applied in radiological imaging ranging from computer-aided diagnosis based on machine learning to deep learning using convolutional neural networks. One of the nuclear medicine imaging tests being commonly performed today is bone scan. The use of deep learning methods through convolutional neural networks in bone scintigrams has not been fully explored. Very few studies have been published on its diagnostic capability of convolutional neural networks in assessing osseous metastasis. Objective: The aim of our study is to assess the classification performance of the pre-trained convolutional neural networks in the diagnosis of bone metastasis from whole body bone scintigrams of a local institutional dataset. Methods: Bone scintigrams from all types of cancer were retrospectively reviewed during the period 2019-2020 at the University of Perpetual Help Medical Center in Las Pinas City, Metro Manila. The study was approved by the Institutional Ethical Review Board and Technical Review Board of the medical center. Bone scan studies should be mainly for metastasis screening. The pre-processing techniques consisting of image normalization, image augmentation, data shuffling, and train-test split (testing at 30% and the rest (70%) was split 85% for training and 15% for validation) were applied to image dataset. Three pre-trained architectures (ResNet50, VGG19, DenseNet121) were applied to the processed dataset. Performance metrics such as accuracy, recall (sensitivity), precision (positive predictive value), and F1-scores were obtained. Results: A total of 570 bone scan images with dimension 220 x 646 pixel sizes in .tif file format were included in this study with 40% classified with bone metastasis while 60% were classified as without bone metastasis. DenseNet121 yielded the highest performance metrics with an accuracy rate of 83%, 76% recall, 86% precision, and 81% F1-score. ResNet50 and VGG19 had similar performance with each other across all metrics but generally lower predictive capability as compared to DenseNet121. Conclusion A bone metastasis machine learning classification study using three pre-trained convolutional neural networks was performed on a local medical center bone scan dataset via transfer learning. DenseNet121 generated the highest performance metrics with 83% accuracy, 76% recall, 86% precision and 81% F1-score. Our simulation experiments generated promising outcomes and potentially could lead to its deployment in the clinical practice of nuclear medicine physicians. The use of deep learning techniques through convolutional neural networks has the potential to improve diagnostic capability of nuclear medicine physicians using bone scans for the assessment of metastasis.
Deep Learning ; Machine Learning

OBJECTIVES

This study evaluated the accuracy of an artificial intelligence (AI) model in identifying retinal lesions, validated its performance on a Filipino population dataset, and evaluated the impact of dataset diversity on AI analysis accuracy.

This cross-sectional, analytical, institutional study analyzed standardized macula-centered fundus photos taken with the Zeiss Visucam®. The AI model’s output was compared with manual readings by trained retina specialists.

A total of 215 eyes from 109 patients were included in the study. Human graders identified 109 eyes (50.7%) with retinal abnormalities. The AI model demonstrated an overall accuracy of 73.0% (95% CI 66.6% – 78.8%) in detecting abnormal retinas, with a sensitivity of 54.1% (95% CI 44.3% – 63.7%) and specificity of 92.5% (95% CI 85.7% – 96.7%).

The availability and sources of AI training datasets can introduce biases into AI algorithms. In our dataset, racial differences in retinal morphology, such as differences in retinal pigmentation, affected the accuracy of AI image-based analysis. More diverse datasets and external validation on different populations are needed to mitigate these biases.

Background: Worldwide, coronary artery disease (CAD) is a leading cause of mortality and morbidity and remains to be a top health priority in many countries. A non-invasive imaging modality for diagnosis of CAD such as single photon emission computed tomography-myocardial perfusion imaging (SPECT-MPI) is usually requested by cardiologists as it displays radiotracer distribution in the heart reflecting myocardial perfusion. The interpretation of SPECT-MPI is done visually by a nuclear medicine physician and is largely dependent on his clinical experience and showing significant inter-observer variability. Objective: The aim of the study is to apply a deep learning approach in the classification of SPECT-MPI for perfusion abnormalities using convolutional neural networks (CNN). Methods: A publicly available anonymized SPECT-MPI from a machine learning repository (https://www.kaggle.com/ selcankaplan/spect-mpi) was used in this study involving 192 patients who underwent stress-test-rest Tc99m MPI. An exploratory approach of CNN hyperparameter selection to search for optimum neural network model was utilized with particular focus on various dropouts (0.2, 0.5, 0.7), batch sizes (8, 16, 32, 64), and number of dense nodes (32, 64, 128, 256). The base CNN model was also compared with the commonly used pre-trained CNNs in medical images such as VGG16, InceptionV3, DenseNet121 and ResNet50. All simulations experiments were performed in Kaggle using TensorFlow 2.6.0., Keras 2.6.0, and Python language 3.7.10. Results: The best performing base CNN model with parameters consisting of 0.7 dropout, batch size 8, and 32 dense nodes generated the highest normalized Matthews Correlation Coefficient at 0.909 and obtained 93.75% accuracy, 96.00% sensitivity, 96.00% precision, and 96.00% F1-score. It also obtained higher classification performance as compared to the pre-trained architectures. Conclusions The results suggest that deep learning approaches through the use of CNN models can be deployed by nuclear medicine physicians in their clinical practice to further augment their decision skills in the interpretation of SPECT-MPI tests. These CNN models can also be used as a dependable and valid second opinion that can aid physicians as a decision-support tool as well as serve as teaching or learning materials for the less-experienced physicians particularly those still in their training career. These highlights the clinical utility of deep learning approaches through CNN models in the practice of nuclear cardiology.
Coronary Artery Disease ; Deep Learning

Background and Objectives: The Philippines faces challenges in the screening of tuberculosis (TB), one of them being the shortage in the health workforce who are skilled and allowed to screen TB. Deep learning neural networks (DLNNs) have shown potential in the TB screening process utilizing chest radiographs (CXRs). However, local studies on AIbased TB screening are limited. This study evaluated qXR3.0 technology's diagnostic performance for TB screening in Filipino adults aged 15 and older. Specifically, we evaluated the specificity and sensitivity of qXR3.0 compared to radiologists' impressions and determined whether it meets the World Health Organization (WHO) standards. Methods: A prospective cohort design was used to perform a study on comparing screening and diagnostic accuracies of qXR3.0 and two radiologist gradings in accordance with the Standards for Reporting Diagnostic Accuracy (STARD). Subjects from two clinics in Metro Manila which had qXR 3.0 seeking consultation at the time of study were invited to participate to have CXRs and sputum collected. Radiologists' and qXR3.0 readings and impressions were compared with respect to the reference standard Xpert MTB/RiF assay. Diagnostic accuracy measures were calculated. Results: With 82 participants, qXR3.0 demonstrated 100% sensitivity and 72.7% specificity with respect to the reference standard. There was a strong agreement between qXR3.0 and radiologists' readings as exhibited by the 0.7895 (between qXR 3.0 and CXRs read by at least one radiologist), 0.9362 (qXR 3.0 and CXRs read by both radiologists), and 0.9403 (qXR 3.0 and CXRs read as not suggestive of TB by at least one radiologist) concordance indices. Conclusions qXR3.0 demonstrated high sensitivity to identify presence of TB among patients, and meets the WHO standard of at least 70% specificity for detecting true TB infection. This shows an immense potential for the tool to supplement the shortage of radiologists for TB screening in the country. Future research directions may consider larger sample sizes to confirm these findings and explore the economic value of mainstream adoption of qXR 3.0 for TB screening.
Tuberculosis ; Diagnostic Imaging ; Deep Learning

Radiotherapy is one of the main treatments for tumor with increasingly high request for technique precision and the equipment stability. Machine learning may bring radiotherapy simplicity, individualization and precision, and may improve the automatic level of planning and quality assurance. Based on the process of radiotherapy, this paper reviews the applications and researches on machine learning, with an emphasis on deep learning, and proposes the prospects in the following aspects: segmentation of normal tissue and tumor, planning, treatment delivery, quality assurance and prognosis prediction.
Deep Learning ; Humans ; Machine Learning ; Neoplasms ; radiotherapy

In recent years, deep learning has provided a new method for cancer prognosis analysis. The literatures related to the application of deep learning in the prognosis of cancer are summarized and their advantages and disadvantages are analyzed, which can be provided for in-depth research. Based on this, this paper systematically reviewed the latest research progress of deep learning in the construction of cancer prognosis model, and made an analysis on the strengths and weaknesses of relevant methods. Firstly, the construction idea and performance evaluation index of deep learning cancer prognosis model were clarified. Secondly, the basic network structure was introduced, and the data type, data amount, and specific network structures and their merits and demerits were discussed. Then, the mainstream method of establishing deep learning cancer prognosis model was verified and the experimental results were analyzed. Finally, the challenges and future research directions in this field were summarized and expected. Compared with the previous models, the deep learning cancer prognosis model can better improve the prognosis prediction ability of cancer patients. In the future, we should continue to explore the research of deep learning in cancer recurrence rate, cancer treatment program and drug efficacy evaluation, and fully explore the application value and potential of deep learning in cancer prognosis model, so as to establish an efficient and accurate cancer prognosis model and realize the goal of precision medicine.
Deep Learning ; Humans ; Neoplasms ; Precision Medicine ; Prognosis

Traditional biomedical data analysis technology faces enormous challenges in the context of the big data era. The application of deep learning technology in the field of biomedical analysis has ushered in tremendous development opportunities. In this paper, we reviewed the latest research progress of deep learning in the field of biomedical data analysis. Firstly, we introduced the deep learning method and its common framework. Then, focusing on the proposal of biomedical problems, data preprocessing method, model building method and training algorithm, we summarized the specific application of deep learning in biomedical data analysis in the past five years according to the chronological order, and emphasized the application of deep learning in medical assistant diagnosis. Finally, we gave the possible development direction of deep learning in the field of biomedical data analysis in the future.
Algorithms ; Biomedical Technology ; Data Analysis ; Deep Learning

BACKGROUND: A deep learning model (DLM) that enables non-invasive hypokalemia screening from an electrocardiogram (ECG) may improve the detection of this life-threatening condition. This study aimed to develop and evaluate the performance of a DLM for the detection of hypokalemia from the ECGs of emergency patients. METHODS: We used a total of 9908 ECG data from emergency patients who were admitted at the Second Affiliated Hospital of Nanchang University, Jiangxi, China, from September 2017 to October 2020. The DLM was trained using 12 ECG leads (lead I, II, III, aVR, aVL, aVF, and V1-6) to detect patients with serum potassium concentrations <3.5 mmol/L and was validated using retrospective data from the Jiangling branch of the Second Affiliated Hospital of Nanchang University. The blood draw was completed within 10 min before and after the ECG examination, and there was no new or ongoing infusion during this period. RESULTS: We used 6904 ECGs and 1726 ECGs as development and internal validation data sets, respectively. In addition, 1278 ECGs from the Jiangling branch of the Second Affiliated Hospital of Nanchang University were used as external validation data sets. Using 12 ECG leads (leads I, II, III, aVR, aVL, aVF, and V1-6), the area under the receiver operating characteristic curve (AUC) of the DLM was 0.80 (95% confidence interval [CI]: 0.77-0.82) for the internal validation data set. Using an optimal operating point yielded a sensitivity of 71.4% and a specificity of 77.1%. Using the same 12 ECG leads, the external validation data set resulted in an AUC for the DLM of 0.77 (95% CI: 0.75-0.79). Using an optimal operating point yielded a sensitivity of 70.0% and a specificity of 69.1%. CONCLUSIONS In this study, using 12 ECG leads, a DLM detected hypokalemia in emergency patients with an AUC of 0.77 to 0.80. Artificial intelligence could be used to analyze an ECG to quickly screen for hypokalemia.
Artificial Intelligence ; Deep Learning ; Electrocardiography ; Humans ; Hypokalemia/diagnosis* ; Retrospective Studies

The accumulation of protein sequence and structure data allows researchers to obtain large amount of descriptive information, simultaneously it poses an urgent need for researchers to extract information from existing data efficiently and apply it to downstream tasks. Protein design enables the development of novel proteins that are no longer restricted by experimental conditions, which is of great significance for drug target prediction, drug discovery, and material design. As an efficient method for data feature extraction, deep learning can be used to model protein data, and further add a priori information to design novel proteins. Therefore, protein design based on deep learning has become a promising approach despite of many challenges. This review summarizes the deep learning-based modeling and design methods of protein sequence and structure data, highlighting the strategies, principle, scope of application and case studies, with the aim to provide a valuable reference for relevant researchers.
Amino Acid Sequence ; Deep Learning ; Drug Development ; Proteins

Objective To develope a deep learning algorithm for pathological classification of chronic gastritis and assess its performance using whole-slide images (WSIs). Methods We retrospectively collected 1,250 gastric biopsy specimens (1,128 gastritis, 122 normal mucosa) from PLA General Hospital. The deep learning algorithm based on DeepLab v3 (ResNet-50) architecture was trained and validated using 1,008 WSIs and 100 WSIs, respectively. The diagnostic performance of the algorithm was tested on an independent test set of 142 WSIs, with the pathologists' consensus diagnosis as the gold standard. Results The receiver operating characteristic (ROC) curves were generated for chronic superficial gastritis (CSuG), chronic active gastritis (CAcG), and chronic atrophic gastritis (CAtG) in the test set, respectively.The areas under the ROC curves (AUCs) of the algorithm for CSuG, CAcG, and CAtG were 0.882, 0.905 and 0.910, respectively. The sensitivity and specificity of the deep learning algorithm for the classification of CSuG, CAcG, and CAtG were 0.790 and 1.000 (accuracy 0.880), 0.985 and 0.829 (accuracy 0.901), 0.952 and 0.992 (accuracy 0.986), respectively. The overall predicted accuracy for three different types of gastritis was 0.867. By flagging the suspicious regions identified by the algorithm in WSI, a more transparent and interpretable diagnosis can be generated. Conclusion The deep learning algorithm achieved high accuracy for chronic gastritis classification using WSIs. By pre-highlighting the different gastritis regions, it might be used as an auxiliary diagnostic tool to improve the work efficiency of pathologists.
Algorithms ; Deep Learning ; Gastritis/diagnosis* ; Humans ; ROC Curve ; Retrospective Studies