Holdout Method Research Articles

Prediction of dual survival following fetoscopic laser photocoagulation for twin-twin transfusion syndrome.

To develop a model based on factors available at the time of diagnosis of twin-twin transfusion syndrome (TTTS) for predicting the probability of dual twin survival following fetoscopic laser photocoagulation (FLP) using a machine-learning algorithm. This was a retrospective study of data collected at two university-affiliated tertiary fetal centers between 2012 and 2021. The cohort included monochorionic diamniotic twin pregnancies complicated by TTTS that underwent FLP. Data were stratified based on survival 30 days after delivery, and cases with dual survival were compared to those without dual survival. A random forest machine-learning algorithm was used to construct a prediction model, and the relative importance value was calculated for each parameter that presented a statistically significant difference between the two study groups and was included in the model. The holdout method was applied to check overfitting of the random forest algorithm. A prediction model for dual twin survival 30 days after delivery was presented based on the test set. The study included 537 women with monochorionic diamniotic twin pregnancy, of whom 346 (64.4%) had dual twin survival at 30 days after delivery and were compared with 191 (35.6%) cases that had one or no survivors. Univariate analysis demonstrated no differences in demographic parameters between the groups. At the time of diagnosis, the dual-survival group had lower rates of estimated fetal weight (EFW) < 10th centile for gestational age in the donor twin (56.4% vs 69.4%; P = 0.004), intertwin EFW discordance > 25% (40.8% vs 56.5%; P = 0.001) and anterior placenta (40.5% vs 50.0%; P = 0.034). Comparison of Doppler findings between the two groups demonstrated significant differences in the donor twin, with a lower rate of pulsatility index (PI) > 95th centile in the umbilical artery and ductus venosus and a lower rate of PI < 5th centile in the fetal middle cerebral artery in the dual-survival group. Relative importance values for each of these six parameters were calculated, allowing the construction of a prediction model with an area under the receiver-operating-characteristics curve of 0.916 (95% CI, 0.887-0.946). We developed a predictive model for dual survival in monochorionic twin pregnancies following FLP for TTTS, which incorporates six variables obtained at the time of diagnosis of TTTS, including donor EFW < 10th centile, intertwin EFW discordance > 25%, anterior placenta and abnormal PI in the umbilical artery, ductus venosus and middle cerebral artery of the donor twin. This clinically applicable tool may improve treatment planning and patient counseling. © 2022 International Society of Ultrasound in Obstetrics and Gynecology.

Hold-out estimates of prediction models for Markov processes

We consider the selection of prediction models for Markovian time series. For this purpose, we study the theoretical properties of the hold-out method. In the econometrics literature, the hold-out method is called ‘out-of-sample’ and is the main method to select a suitable time series model. This method consists of estimating models on a learning set and picking up the model with minimal empirical error on a validation set of future observations. Hold-out estimates are well studied in the independent case, but, as far as we know, this is not the case when the validation set is not independent of the learning set. In this paper, assuming uniform ergodicity of the Markov chain, we state generalization bounds and oracle inequalities for such method; in particular, we show that the ‘out-of-sample’ selection method is adaptive to noise condition.

An Intelligent System for Heart Disease Diagnosis using Regularized Deep Neural Network

Heart disease is one of the deadly diseases. Timely detection of the disease can prevent mortality. In this paper, an intelligent system is proposed for the diagnosis of heart disease using clinical parameters at early stages. The system is developed using the Regularized Deep Neural Network model (Reg-DNN). Cleveland heart disease dataset has been used for training the model. Regularization has been achieved by using dropout and L2 regularization. Efficiency of Reg-DNN was evaluated by using hold-out validation method.70% data was used for training the model and 30% data was used for testing the model. Results indicate that Reg-DNN provided better performance than conventional DNN. Regularization has helped to overcome overfitting. Reg-DNN has achieved an accuracy of 94.79%. Results achieved are quite promising as compared to existing systems in the literature. Authors developed a system containing a graphical user interface. So, the system can be easily used by anyone to diagnose heart disease using the clinical parameters.

Identifying pregnancy in cows using ovarian ultrasound images and convolutional neural networks - a proof-of-concept study

Visual analysis of ovarian structures by ultrasound in cows is a relevant support tool for improving reproductive performance in livestock. However, with human visual analysis, subjectiveness is a limiting factor, and computer vision technologies are a way to overcome this. We thus aimed to evaluate the use of convolutional neural networks (CNNs) for identifying pregnancy in cows using ovarian ultrasound images obtained 30 days after artificial insemination. The dataset consisted of 510 images from individual independent functional ovaries of 238 pregnant and 272 non-pregnant Nellore cows. All the images were collected from the same commercial farm. To evaluate the dependency of CNN performance on image quality, the images were classified by two independent veterinarians with significant experience in ultrasound evaluation as good, regular, or bad for visual identification of ovarian structures. Four CNN architectures were evaluated, namely ResNet50, ResNeXt50, InceptionResNetV2, and DenseNet121. All the CNNs were evaluated using the complete dataset (ALL; n = 510) and a subset of good and regular images (GR; n = 462). Ten evaluations were conducted for each architecture and dataset combination. In each run, the datasets were randomly divided into 70 % for training and 30 % for testing according to the holdout method. A regularization-based method was used during the training phase using dropout, data augmentation, and a dynamic learning rate. The model’s performance was evaluated in terms of accuracy, precision, sensitivity, and specificity, which ranged from 46.5 to 78.8, 34.4–80.0, 28.3–75.4, and 11.2–82.2 %, respectively. The DenseNet121 architecture performed better based on higher accuracy, sensitivity, and specificity, and both lower SD and variation due to dataset. ResNet50 and DenseNet121 models performed better when using the GR dataset. InceptionResNetV2 had a more precise architecture when using the ALL dataset but had lower performance for the GR dataset. We concluded that CNNs are a promising tool to identify the pregnancy status of cows in ovarian images collected 30 days after artificial insemination. The effect of image quality on the performance depends on the CNN architecture.

A locally weighted linear ridge regression framework for spatial interpolation of monthly precipitation over an orographically complex area

AbstractThis article aimed to investigate the potential of using three large‐scale precipitation products (PPs), including TRMM, ERA5‐Land, and MSWEP, and two land surface characteristics, including NDVI and Land Surface Temperature (LST), to improve the accuracy of an elevation‐based spatial non‐stationary, Locally Weighted Linear Regression (LWLR) method. A two‐step approach consisting of downscaling and merging was proposed and assessed across an orographically complex region in Iran to construct monthly precipitation maps for the 2001–2015 period. In the first step, three large‐scale PPs were downscaled to 1 km spatial resolution via the Area To Point Kriging (ATPK) method to address the spatial resolution discrepancy between the coarse pixels of PPs and pointwise gauge data. In the second step, five regression models with different combinations of predictors were developed and compared with the benchmark precipitation‐elevation regression model. The L2 regularization method was adopted to overcome the potential multicollinearity issue. The proposed framework could successfully diagnose the multicollinearity issues and dynamically estimate the regularization parameter in space and time. We utilized the holdout method to validate and compare the developed models, in which 50% of gauges were used for fitting the regression functions, and the remaining gauges were assigned to the validation dataset. The validation results showed that the overall monthly accuracy of the benchmark univariate elevation‐based model was improved where the MAE metric was decreased by 16.5%, and CC and KGE were increased by 3.4 and 5.8%, respectively. Also, the rBias was enhanced from −5.72 to 1.24%. Mean monthly precipitation maps of the 2001–2015 period generated by the model with the best combination of predictors and the benchmark model displayed the same spatial pattern consistent with the topography. Further analysis under different validation scenarios revealed that the contribution of PPs and auxiliary variables is the greatest when the training gauge network is sparse. Also, the developed multivariate model has a higher extrapolation ability and is more robust than the benchmark model.

Usefulness of muscle ultrasound in appendicular skeletal muscle mass estimation for sarcopenia assessment.

The measurement of skeletal muscle mass is essential for the diagnosis of sarcopenia. Muscle ultrasonography has emerged as a useful tool for evaluating sarcopenia because it can be used to assess muscle quality and quantity. This study investigated whether muscle ultrasonography is effective for estimating appendicular skeletal muscle mass (ASM) and screening for sarcopenia. This study prospectively enrolled 212 healthy volunteers aged 40-80 years. ASM was measured using the bioelectrical impedance analysis. Muscle thickness (MT) and echo-intensity (EI) were measured in four muscles (biceps brachii, BB; triceps brachii, TB; rectus femoris, RF; biceps femoris, BF) on the dominant hand. A hold-out cross-validation method was used to develop and validate the ASM prediction equation. In the model development group, the ASM prediction equations were deduced as follows: estimated ASM for men (kg) = 0.167 × weight (kg) + 0.228 × height (cm) + 0.143 × MT of BF (mm)- 0.822 × EI to MT ratio of BB- 28.187 (R2 = 0.830) and estimated ASM for women (kg) = 0.115 × weight + 0.215 × height (cm) + 0.139 × MT of RF-0.638 × EI to MT ratio of BB- 23.502 (R2 = 0.859). In the cross-validation group, the estimated ASM did not significantly differ from the measured ASM in both men (p = 0.775; intraclass correlation coefficient [ICC] = 0.948) and women (p = 0.516; ICC = 0.973). In addition, multiple logistic regression analysis revealed that the ratios of EI to MT in the BF and RF muscles in men and MT in the BB muscle in women could be valuable parameters for sarcopenia screening. Therefore, our study suggests that muscle ultrasound could be an effective tool for estimating ASM and screening sarcopenia.

Automatic Marine Debris Inspection

Plastic trash can be found anywhere, around the marina, beaches, and coastal areas in recent times. This study proposes a trash dataset called HAIDA and a trash detector that uses a YOLOv4-based object detection algorithm to monitor coastal trash pollution efficiently. Model selection, model evaluation, and hyperparameter tuning were applied to obtain the best model for the lowest generalization error in the real world. Comparison of the state-of-the-art object detectors based on YOLOv3, YOLOv4, and Scaled-YOLOv4 that used hyperparameter tuning, the three-way holdout method, and k-fold cross-validation have been presented. An unmanned aerial vehicle (UAV) was also employed to detect trash in coastal areas using the proposed method. The performance on image classification was satisfactory.

Siamese Convolutional Neural Network for Heartbeat Classification Using Limited 12-Lead ECG Datasets

The Electrocardiogram (ECG) is a low-cost exam commonly used to diagnose abnormalities in the cardiac cycle. Over the years, the scientific community has investigated the automatic classification of ECG signals driven by advanced Machine Learning (ML) techniques. Despite recent scientific advances, annotating large and diverse datasets to support the training of ML techniques is still very time-consuming and error-prone. Indeed, ML techniques whose training does not require extensive and well-annotated datasets are becoming even more prominent. Therefore, it is possible to correctly identify and classify abnormalities in the cardiac cycle (e.g., rare cardiologic disturbs) using limited data available in ECG datasets. However, the classification of heartbeats from digital tracings of ECG signals containing 12 leads from imbalanced datasets is challenging due to many existing heart diseases. This study investigates the few-shot learning paradigm based on Siamese Convolutional Neural Networks (SCNN), popular in imaging classification problems, to classify 12-Lead ECG heartbeats using a few training samples with supervised information. The proposed SCNN model presented an accuracy of up to 95% in a public dataset based on the hold-out validation method, implemented for different combinations of similarity and loss functions. Besides, using the 7-fold cross-validation method, the model presented a mean area under the curve of 89%. We also compared the class-by-class classification results with those of similar methods available in the literature, obtaining the same or better results based on performance metrics such as accuracy, precision, recall, and specificity.

Machine Learning Models for Predicting the Type and Outcome of Ureteral Stones Treatments

AbstractBackground:Kidney and ureter stones are the third pathologies in urological diseases. Less invasive treatments such as transureteral lithotripsy and extracorporeal shock wave lithotripsy are used to treat ureteral stones. Data mining has provided the possibility of improving decision-making in choosing the optimal treatment. In this paper predictive models for the detection of ureter stone treatment (first model) and its outcome (second model) is developed based on the patient’s demographic, clinical, and laboratory factors.Methods and Material:In this cross-sectional study a questionnaire was used to identify the most effective features in the predictive models, and Information on 440 patients was collected. The models were constructed using machine learning techniques (Multilayer perceptron, Classification, and regression tree, k-nearest neighbors, Support vector machine, Naïve Bayes classifier, Random Forest, and AdaBoost) in the Bigpro1 analytical system.Results:Among the Holdout and K-fold cross-validation methods used, the Holdout method showed better performance. From the data-based balancing methods used in the second model, the Synthetic Minority oversampling technique showed better performance. Also, the AdaBoost algorithm had the best performance. In this algorithm, accuracy, sensitivity, specificity, precision, F- measure, and Area under the carve in the first model were 89%, 87%, 91%, 90%, 89%, and 94% respectively, and in the second model were 81%, 81%, 82%, 84%, 82%, and 85% respectively.Conclusions:The results were promising and showed that the data mining techniques could be a powerful assistant for urologists to predict a surgical outcome and also to choose an appropriate surgical treatment for removing ureter stones.

Third molar maturity index (I3M) assessment according to different geographical zones: a large multi-ethnic study sample

Identification of living undocumented individuals highlights the need for accurate, precise, and reproducible age estimation methods, especially in those cases involving minors. However, when their country of origin is unknown, or it can be only roughly estimated, it is extremely difficult to apply assessment policies, procedures, and practices that are accurate and child-sensitive. The main aim of this research is to optimize the correct classification of adults and minors by establishing new cut-off values for four different continents (Africa, America, Asia, and Europe). For this purpose, a vast sample of 10,701 orthopantomographs (OPTs) from four continents was evaluated. For determination and subsequent validation of the new third molar maturity index (I3M) cut-off values by world regions, a cross-validation by holdout method was used and contingency tables (confusion matrices) were generated. The lower third molar maturity indexes, from both left and right side (I3ML and I3MR) and the combination of both sides (I3ML_I3MR) were calculated. The new cut-off values, that aim to differentiate between a minor and an adult, with more than 74.00% accuracy for all populations were as follows (I3ML; I3MR; I3ML_I3MR, respectively): Africa = (0.10; 0.10; 0.10), America = (0.10; 0.09; 0.09), Asia = (0.15; 0.17; 0.14), and Europe = (0.09; 0.09; 0.09). The higher sensitivity (Se) was detected for the I3ML for male African people (91%) and the higher specificity (Sp) of all the parameters (I3ML; I3MR; I3ML_I3MR) for Europeans both male and female (> 91%). The original cut-off value (0.08) is still useful, especially in discriminating individuals younger than 18years old which is the goal of the forensic methods used for justice.

Development of a Novel Scoring System to Quantify the Severity of Incident Reports: An Exploratory Research Study

Incident reporting systems have been widely adopted to collect information about patient safety incidents. Much of the value of incident reports lies in the free-text section. Computer processing of semantic information may be helpful to analyze this. We developed a novel scoring system for decision making to assess the severity of incidents using the semantic characteristics of the text in incident reports, and compared its results with experts' opinions. We retrospectively analyzed free-text data from incident reports from January 2012 to September 2021 at Nagoya University Hospital, Aichi, Japan. The sample was allocated to training and validation datasets using the hold-out method. Morphological analysis was used to segment terms in the training dataset. We calculated a severity term score, a severity report score and severity group score, by report volume size, and compared these with conventional severity classifications by patient safety experts and reporters. We allocated 96,082 incident reports into two groups. We calculated 1,802 severity term scores from the 48,041 reports in the training dataset. There was a significant difference in severity report score between reports categorized as severe and not severe by experts (95% confidence interval [CI] -0.83 to -0.80, p < 0.001, d = 0.81). Severity group scores were positively associated with severity ratings from experts and reporters (correlation coefficients 0.73 [95% CI 0.63-0.80, p < 0.001] and 0.79 [95% CI 0.71-0.85, p < 0.001]) for all departments. Our severity scoring system could therefore contribute to better organizational patient safety.

SEMG-based technology for silent voice recognition

Silent speech recognition (SSR) is a system that implements speech communication when a sound signal is not available using surface electromyography (sEMG)-based speech recognition. Researchers have used surface electrodes to record the electrically-activated potential of human articulation muscles to recognize speech content. SSR can be used for pilot-assisted speech recognition, communication of individuals with speech impairment, private communication, and other fields. In this feasibility study, we collected sEMG data for ten single Mandarin numeric words. After reducing power frequency interference and power supply noise from the sEMG signal, short-term energy (STE) was used for voice activity detection (VAD). The power spectrum features were extracted and fed into the classifier for final identification results. We used the Hold-out method to divide the data into training and test sets on a 7-3 scale, with an average accuracy of 92.3% and a maximum of 100% using a support vector machine (SVM) classifier. Experimental results showed that the proposed method has development potential, and is effective in identifying isolated words from the sEMG signal of the articulation muscles.

Intelligent health indicator construction for prognostics of composite structures utilizing a semi-supervised deep neural network and SHM data

A health indicator (HI) is a valuable index demonstrating the health level of an engineering system or structure, which is a direct intermediate connection between raw signals collected by structural health monitoring (SHM) methods and prognostic models for remaining useful life estimation. An appropriate HI should conform to prognostic criteria, i.e., monotonicity, trendability, and prognosability, that are commonly utilized to measure the HI’s quality. However, constructing such a HI is challenging, particularly for composite structures due to their vulnerability to complex damage scenarios. Data-driven models and deep learning are powerful mathematical tools that can be employed to achieve this purpose. Yet the availability of a large dataset with labels plays a crucial role in these fields, and the data collected by SHM methods can only be labeled after the structure fails. In this respect, semi-supervised learning can incorporate unlabeled data monitored from structures that have not yet failed. In the present work, a semi-supervised deep neural network is proposed to construct HI by SHM data fusion. For the first time, the prognostic criteria are used as targets of the network rather than employing them only as a measurement tool of HI’s quality. In this regard, the acoustic emission method was used to monitor composite panels during fatigue loading, and extracted features were used to construct an intelligent HI. Finally, the proposed roadmap is evaluated by the holdout method, which shows a 77.3% improvement in the HI’s quality, and the leave-one-out cross-validation method, which indicates the generalized model has at least an 81.77% score on the prognostic criteria. This study demonstrates that even when the true HI labels are unknown but the qualified HI pattern (according to the prognostic criteria) can be recognized, a model can still be built that provides HIs aligning with the desired degradation behavior.

Classification of Cotton Genotypes with Mixed Continuous and Categorical Variables: Application of Machine Learning Models

Mixed data is a combination of continuous and categorical variables and occurs frequently in fields such as agriculture, remote sensing, biology, medical science, marketing, etc., but only limited work has been done with this type of data. In this study, data on continuous and categorical characters of 452 genotypes of cotton (Gossypium hirsutum) were obtained from an experiment conducted by the Central Institute of Cotton Research (CICR), Sirsa, Haryana (India) during the Kharif season of the year 2018–2019. The machine learning (ML) classifiers/models, namely k-nearest neighbor (KNN), Classification and Regression Tree (CART), C4.5, Naïve Bayes, random forest (RF), bagging, and boosting were considered for cotton genotypes classification. The performance of these ML classifiers was compared to each other along with the linear discriminant analysis (LDA) and logistic regression. The holdout method was used for cross-validation with an 80:20 ratio of training and testing data. The results of the appraisal based on hold-out cross-validation showed that the RF and AdaBoost performed very well, having only two misclassifications with the same accuracy of 97.26% and the error rate of 2.74%. The LDA classifier performed the worst in terms of accuracy, with nine misclassifications. The other performance measures, namely sensitivity, specificity, precision, F1 score, and G-mean, were all together used to find out the best ML classifier among all those considered. Moreover, the RF and AdaBoost algorithms had the highest value of all the performance measures, with 96.97% sensitivity and 97.50% specificity. Thus, these models were found to be the best in classifying the low- and high-yielding cotton genotypes.

Identification of Barangan Banana Ripeness Treatment Types using k-Nearest Neighbor

Bananas are favored by the public because bananas are rich in nutrients that our bodies need. One of the bananas that people are interested in is the Barangan. Bananas sold in the market have various types of ripeness that vary based on their treatment. This study aimed to identify the type of treatment for Barangan. Identification is carried out based on an analysis of the image of Barangan using color and texture features. The k-Nearest Neighbor (k-NN) method is used in the identification. The k-NN compares the similarity between the unknown data and the sample data. The k values used in this study are k=1, k=3, and k=5. The Euclidean Distance is used to measure the distance between 2 feature vectors. The classification test uses the holdout method, where the percentage of the amount of sample data and test data is 66.67% of training data and 33.33% of test data. The accuracy obtained at k=1 is 86.67%, at k=3 is 76.67%, and at k=5 is 80%. The best accuracy for identifying banana ripeness treatment types using the k-Nearest Neighbor method is obtained at k = 1, with accuracy reaching 86.67%.

Staging of Skin Cancer Based on Hyperspectral Microscopic Imaging and Machine Learning.

Skin cancer, a common type of cancer, is generally divided into basal cell carcinoma (BCC), squamous cell carcinoma (SCC) and malignant melanoma (MM). The incidence of skin cancer has continued to increase worldwide in recent years. Early detection can greatly reduce its morbidity and mortality. Hyperspectral microscopic imaging (HMI) technology can be used as a powerful tool for skin cancer diagnosis by reflecting the changes in the physical structure and microenvironment of the sample through the differences in the HMI data cube. Based on spectral data, this work studied the staging identification of SCC and the influence of the selected region of interest (ROI) on the staging results. In the SCC staging identification process, the optimal result corresponded to the standard normal variate transformation (SNV) for spectra preprocessing, the partial least squares (PLS) for dimensionality reduction, the hold-out method for dataset partition and the random forest (RF) model for staging identification, with the highest staging accuracy of 0.952 ± 0.014, and a kappa value of 0.928 ± 0.022. By comparing the staging results based on spectral characteristics from the nuclear compartments and peripheral regions, the spectral data of the nuclear compartments were found to contribute more to the accurate staging of SCC.

1570P A deep learning model for prediction of chemotherapy infusion related symptoms using continuous heart rate variability

We hypothesized chemotherapy infusion related reaction might be predictable according to continuous monitoring of heart rate recorded by a wearable device. In this study we propose a deep learning model that predicts patient level symptoms per minute during chemotherapy infusion. We prospectively enrolled patients with cancer who will receive chemotherapy and at high risk for infusion related reaction. During chemotherapy infusion, 1 minute interval heart rate was recorded from a smart band and self-recorded symptom log was obtained. Patient's status at every minute was classified into 4 categories; normal, mild symptom without intervention, serious symptom which needed intervention, and patient's voluntary movement. The prediction model used the time windows of 120 minutes to predict infusion related symptoms over the current time. The model was trained 1500 epochs and we used the Adam optimizer with initial learning rate of 1e-3. We trained a multi-percentron layer based deep learning model. The losses were calculated by the cross-entropy. The model was validated using the holdout method with 20% of the datasts. The performance of the model was assessed with the receiver operating characteristic curves (AUROC) of 4 different events. We collected 1-minute interval heart rate record and patient's symptom logs during 4,716 minutes of 35 cycles of chemotherapy administration, from December 2021 to February 2022. The median age of enrolled patients was 60(range 35-85), and breast cancer(68.6%) was the most common diagnosis. Among 4,716 minutes, 277 (5.87%) minutes were labeled as non-normal status. Our model showed AUROC of 0.92 for normal status, 1.00 for mild symptoms, 0.82 for serious symptoms, and 0.97 for patient's voluntary movement. The proposed prediction model shows the 97.38% accuracy for 20% validation set. This study showed that a deep learning model could recognize infusion related symptoms associated heart rate variability obtained by wearable device in real world patients receiving chemotherapy. Further studies for validation of predictive capability in a real time setting are warranted.

Integration of machine learning algorithms and GIS-based approaches to cutaneous leishmaniasis prevalence risk mapping

Cutaneous leishmaniasis is a complex infection that is caused by different species of Leishmania and affects more than 2 million people in 88 countries. Identifying the environmental factors affecting the occurrence of cutaneous leishmaniasis and preparing a risk map is one of the basic tools to control and manage this disease. The aim of this study was a spatial prediction of cutaneous leishmaniasis in Isfahan province, Iran using three machine learning algorithms (decision tree (DT), support vector regression (SVR), and linear regression (LR)). The spatial database was created using data collected on the number of diseases in Isfahan province from 2011 to 2018, as well as ten environmental parameters (temperature, humidity, rainfall, altitude, slope, wind speed, normalized difference vegetation index (NDVI), number of sunny days, number of frosty days, and distance to stream) that affect the incidence of leishmaniasis. Furthermore, the fuzzy method was employed in this study to reduce uncertainty and evaluate the effect of environmental factors on disease prevalence. Using the holdout method and 70:30 ratios, the data were used to model and prepare a leishmaniasis prediction map and evaluate the results, respectively. The accuracy of the maps satisfied with the DT, SVR, and LR algorithms was 0.951, 0.934, and 0.914, respectively, according to the receiver operating characteristic (ROC) curve and area under the curve (AUC). Furthermore, the eastern and southern parts of the province have the lowest risk of leishmaniasis. The result of this issue is the identification of high-risk areas of the disease and increase life and peace for people in the community.

Comparison of Deep Learning Algorithms in Predicting Expert Assessments of Pain Scores during Surgical Operations Using Analgesia Nociception Index.

There are many surgical operations performed daily in operation rooms worldwide. Adequate anesthesia is needed during an operation. Besides hypnosis, adequate analgesia is critical to prevent autonomic reactions. Clinical experience and vital signs are usually used to adjust the dosage of analgesics. Analgesia nociception index (ANI), which ranges from 0 to 100, is derived from heart rate variability (HRV) via electrocardiogram (ECG) signals, for pain evaluation in a non-invasive manner. It represents parasympathetic activity. In this study, we compared the performance of multilayer perceptron (MLP) and long short-term memory (LSTM) algorithms in predicting expert assessment of pain score (EAPS) based on patient′s HRV during surgery. The objective of this study was to analyze how deep learning models differed from the medical doctors′ predictions of EAPS. As the input and output features of the deep learning models, the opposites of ANI and EAPS were used. This study included 80 patients who underwent operations at National Taiwan University Hospital. Using MLP and LSTM, a holdout method was first applied to 60 training patients, 10 validation patients, and 10 testing patients. As compared to the LSTM model, which had a testing mean absolute error (MAE) of 2.633 ± 0.542, the MLP model had a testing MAE of 2.490 ± 0.522, with a more appropriate shape of its prediction curves. The model based on MLP was selected as the best. Using MLP, a seven-fold cross validation method was then applied. The first fold had the lowest testing MAE of 2.460 ± 0.634, while the overall MAE for the seven-fold cross validation method was 2.848 ± 0.308. In conclusion, HRV analysis using MLP algorithm had a good correlation with EAPS; therefore, it can play role as a continuous monitor to predict intraoperative pain levels, to assist physicians in adjusting analgesic agent dosage. Further studies may consider obtaining more input features, such as photoplethysmography (PPG) and other kinds of continuous variable, to improve the prediction performance.

Prediction of Nodal Metastasis in Lung Cancer Using Deep Learning of Endobronchial Ultrasound Images.

Simple SummaryEndobronchial ultrasound-guided transbronchial aspiration is a minimally invasive and highly accurate modality for the diagnosis of lymph node metastasis and is useful for pre-treatment biomarker test sampling in patients with lung cancer. Endobronchial ultrasound image analysis is useful for predicting nodal metastasis; however, it can only be used as a supplemental method to tissue sampling. In recent years, deep learning-based computer-aided diagnosis using artificial intelligence technology has been introduced in research and clinical medicine. This study investigated the feasibility of computer-aided diagnosis for the prediction of nodal metastasis in lung cancer using endobronchial ultrasound images. The outcome of this study may help improve diagnostic efficiency and reduce invasiveness of the procedure.Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is a valid modality for nodal lung cancer staging. The sonographic features of EBUS helps determine suspicious lymph nodes (LNs). To facilitate this use of this method, machine-learning-based computer-aided diagnosis (CAD) of medical imaging has been introduced in clinical practice. This study investigated the feasibility of CAD for the prediction of nodal metastasis in lung cancer using endobronchial ultrasound images. Image data of patients who underwent EBUS-TBNA were collected from a video clip. Xception was used as a convolutional neural network to predict the nodal metastasis of lung cancer. The prediction accuracy of nodal metastasis through deep learning (DL) was evaluated using both the five-fold cross-validation and hold-out methods. Eighty percent of the collected images were used in five-fold cross-validation, and all the images were used for the hold-out method. Ninety-one patients (166 LNs) were enrolled in this study. A total of 5255 and 6444 extracted images from the video clip were analyzed using the five-fold cross-validation and hold-out methods, respectively. The prediction of LN metastasis by CAD using EBUS images showed high diagnostic accuracy with high specificity. CAD during EBUS-TBNA may help improve the diagnostic efficiency and reduce invasiveness of the procedure.