Explainable artificial intelligence for LDL cholesterol prediction and classification.

INTRODUCTION: Monitoring LDL-C levels is essential in clinical practice because there is a direct relation between low-density lipoprotein cholesterol (LDL-C) levels and atherosclerotic heart disease risk. Therefore, measurement or estimate of LDL-C is critical. The present study aims to evaluate Artificial Intelligence (AI) and Explainable AI (XAI) methodologies in predicting LDL-C levels while emphasizing the interpretability of these predictions. MATERIALS AND METHODS: We retrospectively reviewed data from the Laboratory Information System (LIS) of Ankara Etlik City Hospital (AECH). We included 60.217 patients with standard lipid profiles (total cholesterol [TC], high-density lipoprotein cholesterol, and triglycerides) paired with same-day direct LDL-C results. AI methodologies, such as Gradient Boosting (GB), Random Forests (RF), Support Vector Machines (SVM), and Decision Trees (DT), were used to predict LDL-C and compared directly measured and calculated LDL-C with formulas. XAI techniques such as Shapley additive annotation (SHAP) and locally interpretable model-agnostic explanation (LIME) were used to interpret AI models and improve their explainability. RESULTS: Predicted LDL-C values using AI, especially RF or GB, showed a stronger correlation with direct measurement LDL-C values than calculated LDL-C values with formulas. TC was shown to be the most influential factor in LDL-C prediction using SHAP and LIME. The agreement between the treatment groups based on NCEP ATPIII guidelines according to measured LDL-C and the LDL-C groups obtained with AI was higher than that obtained with formulas. CONCLUSIONS: It can be concluded that AI is not only a reliable method but also an explainable method for LDL-C estimation and classification.

PToPI: A Comprehensive Review, Analysis, and Knowledge Representation of Binary Classification Performance Measures/Metrics.

Although few performance evaluation instruments have been used conventionally in different machine learning-based classification problem domains, there are numerous ones defined in the literature. This study reviews and describes performance instruments via formally defined novel concepts and clarifies the terminology. The study first highlights the issues in performance evaluation via a survey of 78 mobile-malware classification studies and reviews terminology. Based on three research questions, it proposes novel concepts to identify characteristics, similarities, and differences of instruments that are categorized into 'performance measures' and 'performance metrics' in the classification context for the first time. The concepts reflecting the intrinsic properties of instruments such as canonical form, geometry, duality, complementation, dependency, and leveling, aim to reveal similarities and differences of numerous instruments, such as redundancy and ground-truth versus prediction focuses. As an application of knowledge representation, we introduced a new exploratory table called PToPI (Periodic Table of Performance Instruments) for 29 measures and 28 metrics (69 instruments including variant and parametric ones). Visualizing proposed concepts, PToPI provides a new relational structure for the instruments including graphical, probabilistic, and entropic ones to see their properties and dependencies all in one place. Applications of the exploratory table in six examples from different domains in the literature have shown that PToPI aids overall instrument analysis and selection of the proper performance metrics according to the specific requirements of a classification problem. We expect that the proposed concepts and PToPI will help researchers comprehend and use the instruments and follow a systematic approach to classification performance evaluation and publication.

The first case of glioma detected by an artificial intelligence algorithm running on real-time data in neurosurgery: illustrative case.

BACKGROUND: The aim of this paper is to report one of the significant applications of artificial intelligence (AI) and how it affects everyday clinical practice in neurosurgery. The authors present a case in which a patient was diagnosed via an AI algorithm during ongoing magnetic resonance imaging (MRI). According to this algorithm, the corresponding physicians were immediately warned, and the patient received prompt appropriate treatment. OBSERVATIONS: A 46-year-old female presenting with nonspecific headache was admitted to undergo MRI. Scanning revealed an intraparenchymal mass that was detected by an AI algorithm running on real-time patient data while the patient was still in the MRI scanner. The day after MRI, a stereotactic biopsy was performed. The pathology report confirmed an isocitrate dehydrogenase wild-type diffuse glioma. The patient was referred to the oncology department for evaluation and immediate treatment. LESSONS: This is the first report of a glioma diagnosed by an AI algorithm and a subsequent prompt operation in the literature-the first of many and an example of how AI will enhance clinical practice.

Utilizing Deep Convolutional Generative Adversarial Networks for Automatic Segmentation of Gliomas: An Artificial Intelligence Study.

AIM: To describe a deep convolutional generative adversarial networks (DCGAN) model which learns normal brain MRI from normal subjects than finds distortions such as a glioma from a test subject while performing a segmentation at the same time. MATERIAL AND METHODS: MRIs of 300 healthy subjects were employed as training set. Additionally, test data were consisting anonymized T2-weigted MRIs of 27 healthy subjects and 27 HGG patients. Consecutive axial T2-weigted MRI slices of every subject were extracted and resized to 364x448 pixel resolution. The generative model produced random normal synthetic images and used these images for calculating residual loss to measure visual similarity between input MRIs and generated MRIs. RESULTS: The model correctly detected anomalies on 24 of 27 HGG patients? MRIs and marked them as abnormal. Besides, 25 of 27 healthy subjects? MRIs in the test dataset detected correctly as healthy MRI. The accuracy, precision, recall, and AUC were 0.907, 0.892, 0.923, and 0.907, respectively. CONCLUSION: Our proposed model demonstrates acceptable results can be achieved only by training with normal subject MRIs via using DCGAN model. This model is unique because it learns only from normal MRIs and it is able to find any abnormality which is different than the normal pattern.

Generating one biometric feature from another: faces from fingerprints.

This study presents a new approach based on artificial neural networks for generating one biometric feature (faces) from another (only fingerprints). An automatic and intelligent system was designed and developed to analyze the relationships among fingerprints and faces and also to model and to improve the existence of the relationships. The new proposed system is the first study that generates all parts of the face including eyebrows, eyes, nose, mouth, ears and face border from only fingerprints. It is also unique and different from similar studies recently presented in the literature with some superior features. The parameter settings of the system were achieved with the help of Taguchi experimental design technique. The performance and accuracy of the system have been evaluated with 10-fold cross validation technique using qualitative evaluation metrics in addition to the expanded quantitative evaluation metrics. Consequently, the results were presented on the basis of the combination of these objective and subjective metrics for illustrating the qualitative properties of the proposed methods as well as a quantitative evaluation of their performances. Experimental results have shown that one biometric feature can be determined from another. These results have once more indicated that there is a strong relationship between fingerprints and faces.

A Novel Deep Learning Algorithm for the Automatic Detection of High-Grade Gliomas on T2-Weighted Magnetic Resonance Images: A Preliminary Machine Learning Study.

AIM: To propose a convolutional neural network (CNN) for the automatic detection of high-grade gliomas (HGGs) on T2-weighted magnetic resonance imaging (MRI) scans. MATERIAL AND METHODS: A total of 3580 images obtained from 179 individuals were used for training and validation. After random rotation and vertical flip, training data was augmented by factor of 10 in each iteration. In order to increase data processing time, every single image converted into a Jpeg image which has a resolution of 320x320. Accuracy, precision and recall rates were calculated after training of the algorithm. RESULTS: Following training, CNN achieved acceptable performance ratios of 0.854 to 0.944 for accuracy, 0.812 to 0.980 for precision and 0.738 to 0.907 for recall. Also, CNN was able to detect HGG cases even though there is no apparent mass lesion in the given image. CONCLUSION: Our preliminary findings demonstrate; currently proposed CNN model achieves acceptable performance results for the automatic detection of HGGs on T2-weighted images.

Predicting coronary artery disease using different artificial neural network models.

OBJECTIVE: Eight different learning algorithms used for creating artificial neural network (ANN) models and the different ANN models in the prediction of coronary artery disease (CAD) are introduced. METHODS: This work was carried out as a retrospective case-control study. Overall, 124 consecutive patients who had been diagnosed with CAD by coronary angiography (at least 1 coronary stenosis > 50% in major epicardial arteries) were enrolled in the work. Angiographically, the 113 people (group 2) with normal coronary arteries were taken as control subjects. Multi-layered perceptrons ANN architecture were applied. The ANN models trained with different learning algorithms were performed in 237 records, divided into training (n=171) and testing (n=66) data sets. The performance of prediction was evaluated by sensitivity, specificity and accuracy values based on standard definitions. RESULTS: The results have demonstrated that ANN models trained with eight different learning algorithms are promising because of high (greater than 71%) sensitivity, specificity and accuracy values in the prediction of CAD. Accuracy, sensitivity and specificity values varied between 83.63%-100%, 86.46%-100% and 74.67%-100% for training, respectively. For testing, the values were more than 71% for sensitivity, 76% for specificity and 81% for accuracy. CONCLUSIONS: It may be proposed that the use of different learning algorithms other than backpropagation and larger sample sizes can improve the performance of prediction. The proposed ANN models trained with these learning algorithms could be used a promising approach for predicting CAD without the need for invasive diagnostic methods and could help in the prognostic clinical decision.