scholarly journals Diagnosis prediction of tumours of unknown origin using ImmunoGenius, a machine learning-based expert system for immunohistochemistry profile interpretation

2021 ◽  
Vol 16 (1) ◽  
Author(s):  
Yosep Chong ◽  
Nishant Thakur ◽  
Ji Young Lee ◽  
Gyoyeon Hwang ◽  
Myungjin Choi ◽  
...  
Keyword(s):  
Machine Learning ◽  
Expert System ◽  
Learning Algorithm ◽  
Unknown Origin ◽  
Test Dataset ◽  
Bayesian Theorem ◽  
Hit Rate ◽  
Significant Difference ◽  
Treatment Plans ◽  
Case Data

Abstract Background Immunohistochemistry (IHC) remains the gold standard for the diagnosis of pathological diseases. This technique has been supporting pathologists in making precise decisions regarding differential diagnosis and subtyping, and in creating personalized treatment plans. However, the interpretation of IHC results presents challenges in complicated cases. Furthermore, rapidly increasing amounts of IHC data are making it even harder for pathologists to reach to definitive conclusions. Methods We developed ImmunoGenius, a machine-learning-based expert system for the pathologist, to support the diagnosis of tumors of unknown origin. Based on Bayesian theorem, the most probable diagnoses can be drawn by calculating the probabilities of the IHC results in each disease. We prepared IHC profile data of 584 antibodies in 2009 neoplasms based on the relevant textbooks. We developed the reactive native mobile application for iOS and Android platform that can provide 10 most possible differential diagnoses based on the IHC input. Results We trained the software using 562 real case data, validated it with 382 case data, tested it with 164 case data and compared the precision hit rate. Precision hit rate was 78.5, 78.0 and 89.0% in training, validation and test dataset respectively. Which showed no significant difference. The main reason for discordant precision was lack of disease-specific IHC markers and overlapping IHC profiles observed in similar diseases. Conclusion The results of this study showed a potential that the machine-learning algorithm based expert system can support the pathologic diagnosis by providing second opinion on IHC interpretation based on IHC database. Incorporation with contextual data including the clinical and histological findings might be required to elaborate the system in the future.

scholarly journals Diagnosis Prediction of Tumours of Unknown Origin Using ImmunoGenius, A Machine Learning-based Expert System for Immunohistochemistry Profile Interpretation

2021 ◽  
Author(s):  
Yosep Chong ◽  
Nishant Thakur ◽  
Ji Young Lee ◽  
Gyoyeon Hwang ◽  
Myungjin Choi ◽  
...  
Keyword(s):  
Machine Learning ◽  
Expert System ◽  
Learning Algorithm ◽  
Unknown Origin ◽  
Test Dataset ◽  
Bayesian Theorem ◽  
Hit Rate ◽  
Significant Difference ◽  
Treatment Plans ◽  
Case Data

Abstract BackgroundImmunohistochemistry (IHC) remains the gold standard for the diagnosis of pathological diseases. This technique has been supporting pathologists in making precise decisions regarding differential diagnosis and subtyping, and in creating personalized treatment plans. However, the interpretation of IHC results presents challenges in complicated cases. Furthermore, rapidly increasing amounts of IHC data are making it even harder for pathologists to reach to definitive conclusions. MethodsWe developed ImmunoGenius, a machine-learning-based expert system for the pathologist, to support the diagnosis of tumors of unknown origin. Based on Bayesian theorem, the most probable diagnoses can be drawn by calculating the probabilities of the IHC results in each disease. We prepared IHC profile data of 584 antibodies in 2009 neoplasms based on the relevant textbooks. We developed the reactive native mobile application for iOS and Android platform that can provide 10 most possible differential diagnoses based on the IHC input. ResultsWe trained the software using 562 real case data, validated it with 382 case data, tested it with 164 case data and compared the precision hit rate. Precision hit rate was 78.5%, 78.0% and 89.0% in training, validation and test dataset respectively. which showed no significant difference. The main reason for discordant precision was lack of disease-specific IHC markers and overlapping IHC profiles observed in similar diseases.Conclusion The results of this study showed a potential that the machine-learning algorithm based expert system can support the pathologic diagnosis by providing second opinion on IHC interpretation based on IHC database. Incorporation with contextual data including the clinical and histological findings might be required to elaborate the system in the future.

scholarly journals Diagnosis Prediction of Tumours of Unknown Origin Using Immunogenius, A Machine Learning-based Expert Support System for Immunohistochemistry Profile Interpretation

2020 ◽  
Author(s):  
Yosep Chong ◽  
Ji Young Lee ◽  
Nishant Thakur ◽  
Gyoyeon Hwang ◽  
Myungjin Choi ◽  
...  
Keyword(s):  
Machine Learning ◽  
Support System ◽  
Unknown Origin ◽  
Training Dataset ◽  
Validation Dataset ◽  
Hit Rate ◽  
Significant Difference ◽  
Expert Support System ◽  
Case Data ◽  
Expert Support

Abstract Immunohistochemistry (IHC) remains the gold standard for the diagnosis of pathological diseases. This technique has been supporting pathologists in making precise decisions regarding differential diagnosis and subtyping, and in creating personalized treatment plans. However, the interpretation of IHC results presents challenges in complicated cases. Furthermore, rapidly increasing amounts of IHC data are making it even harder for pathologists to reach to definitive conclusions. MethodsWe developed ImmunoGenius, a machine-learning-based expert support system for the pathologist, to support the diagnosis of tumors of unknown origin. Based on the Bayesian theorem, we developed the reactive native mobile application for iOS and Android platform. ImmunoGenius include the IHC profile of 584 antibodies in 2009 neoplasms. ResultsWe trained the software using 634 real case data, validated it with 382 case data, and compared the precision hit rate. Precision hit rate of the training dataset was 78.5 % and the hit rate of the validation dataset was 78.0%, which showed no significant difference. The main reason for discordant precision was lack of disease-specific IHC markers and overlapping IHC profiles observed in similar diseases.ConclusionThe results of this study showed a potential that the machine-learning algorithm based expert system can support the pathologic diagnosis by providing second opinion on IHC interpretation based on IHC database. Incorporation with contextual data including the clinical and histological findings might be required to elaborate the system in the future.

scholarly journals Machine Learning Approaches for Early Prediction of Gestational Diabetes Mellitus Based on Prospective Cohort Study

2021 ◽  
Author(s):  
Jingyuan Wang ◽  
Xiujuan Chen ◽  
Yueshuai Pan ◽  
Kai Chen ◽  
Yan Zhang ◽  
...  
Keyword(s):  
Diabetes Mellitus ◽  
Machine Learning ◽  
Cohort Study ◽  
Gestational Diabetes ◽  
Learning Algorithm ◽  
Training Dataset ◽  
Machine Learning Algorithm ◽  
Early Prediction ◽  
Test Dataset ◽  
Svm Model

Abstract Purpose: To develop and verify an early prediction model of gestational diabetes mellitus (GDM) using machine learning algorithm.Methods: The dataset collected from a pregnant cohort study in eastern China, from 2017 to 2019. It was randomly divided into 75% as the training dataset and 25% as the test dataset using the train_test_split function. Based on Python, four classic machine learning algorithm and a New-Stacking algorithm were first trained by the training dataset, and then verified by the test dataset. The four models were Logical Regression (LR), Random Forest (RT), Artificial Neural Network (ANN) and Support Vector Machine (SVM). The sensitivity, specificity, accuracy, and area under the Receiver Operating Characteristic Curve (AUC) were used to analyse the performance of models.Results: Valid information from a total of 2811 pregnant women were obtained. The accuracies of the models ranged from 80.09% to 86.91% (RF), sensitivities ranged from 63.30% to 81.65% (SVM), specificities ranged from 79.38% to 97.53% (RF), and AUCs ranged from 0.80 to 0.82 (New-Stacking).Conclusion: This paper successfully constructed a New-Stacking model theoretically, for its better performance in specificity, accuracy and AUC. But the SVM model got the highest sensitivity, the SVM model was recommends as the prediction model for clinical.

scholarly journals Evaluation and calibration of a low-cost particle sensor in ambient conditions using machine-learning methods

2020 ◽  
Vol 13 (4) ◽  
pp. 1693-1707 ◽  
Author(s):  
Minxing Si ◽  
Ying Xiong ◽  
Shan Du ◽  
Ke Du
Keyword(s):  
Machine Learning ◽  
Linear Regression ◽  
Random Search ◽  
Low Cost ◽  
Learning Algorithms ◽  
Machine Learning Algorithms ◽  
Ambient Conditions ◽  
Test Dataset ◽  
Compact Size ◽  
Significant Difference

Abstract. Particle sensing technology has shown great potential for monitoring particulate matter (PM) with very few temporal and spatial restrictions because of its low cost, compact size, and easy operation. However, the performance of low-cost sensors for PM monitoring in ambient conditions has not been thoroughly evaluated. Monitoring results by low-cost sensors are often questionable. In this study, a low-cost fine particle monitor (Plantower PMS 5003) was colocated with a reference instrument, the Synchronized Hybrid Ambient Real-time Particulate (SHARP) monitor, at the Calgary Varsity air monitoring station from December 2018 to April 2019. The study evaluated the performance of this low-cost PM sensor in ambient conditions and calibrated its readings using simple linear regression (SLR), multiple linear regression (MLR), and two more powerful machine-learning algorithms using random search techniques for the best model architectures. The two machine-learning algorithms are XGBoost and a feedforward neural network (NN). Field evaluation showed that the Pearson correlation (r) between the low-cost sensor and the SHARP instrument was 0.78. The Fligner and Killeen (F–K) test indicated a statistically significant difference between the variances of the PM2.5 values by the low-cost sensor and the SHARP instrument. Large overestimations by the low-cost sensor before calibration were observed in the field and were believed to be caused by the variation of ambient relative humidity. The root mean square error (RMSE) was 9.93 when comparing the low-cost sensor with the SHARP instrument. The calibration by the feedforward NN had the smallest RMSE of 3.91 in the test dataset compared to the calibrations by SLR (4.91), MLR (4.65), and XGBoost (4.19). After calibrations, the F–K test using the test dataset showed that the variances of the PM2.5 values by the NN, XGBoost, and the reference method were not statistically significantly different. From this study, we conclude that a feedforward NN is a promising method to address the poor performance of low-cost sensors for PM2.5 monitoring. In addition, the random search method for hyperparameters was demonstrated to be an efficient approach for selecting the best model structure.

scholarly journals Analysis of the mandibular canal course using unsupervised machine learning algorithm

PLoS ONE ◽  
2021 ◽  
Vol 16 (11) ◽  
pp. e0260194
Author(s):  
Young Hyun Kim ◽  
Kug Jin Jeon ◽  
Chena Lee ◽  
Yoon Joo Choi ◽  
Hoi-In Jung ◽  
...  
Keyword(s):  
Machine Learning ◽  
Cluster Analysis ◽  
Learning Algorithm ◽  
Mandibular Canal ◽  
Unsupervised Machine Learning ◽  
Computed Tomography Images ◽  
Significant Difference ◽  
Cluster 2 ◽  
Axial View

Objectives Anatomical structure classification is necessary task in medical field, but the inevitable variability of interpretation among experts makes reliable classification difficult. This study aims to introduce cluster analysis, unsupervised machine learning method, for classification of three-dimensional (3D) mandibular canal (MC) courses, and to visualize standard MC courses derived from cluster analysis in the Korean population. Materials and methods A total of 429 cone-beam computed tomography images were used. Four sites in the mandible were selected for the measurement of the MC course and four parameters, two vertical and two horizontal parameters were measured per site. Cluster analysis was carried out as follows: parameter measurement, parameter normalization, cluster tendency evaluation, optimal number of clusters determination, and k-means cluster analysis. The 3D MC courses were classified into three types with statistically significant mean differences by cluster analysis. Results Cluster 1 showed a smooth line running towards the lingual side in the axial view and a steep slope in the sagittal view. Cluster 2 ran in an almost straight line closest to the lingual and inferior border of mandible. Cluster 3 showed the pathway with a bent buccally in the axial view and an increasing slope in the sagittal view in the posterior area. Cluster 2 showed the highest distribution (42.1%), and males were more widely distributed (57.1%) than the females (42.9%). Cluster 3 comprised similar ratio of male and female cases and accounted for 31.9% of the total distribution. Cluster 1 had the least distribution (26.0%) Distributions of the right and left sides did not show a statistically significant difference. Conclusion The MC courses were automatically classified as three types through cluster analysis. Cluster analysis enables the unbiased classification of the anatomical structures by reducing observer variability and can present representative standard information for each classified group.

scholarly journals Multifactor Prediction of Embryo Transfer Outcomes Based on a Machine Learning Algorithm

2021 ◽  
Vol 12 ◽  
Author(s):  
Ran Liu ◽  
Shun Bai ◽  
Xiaohua Jiang ◽  
Lihua Luo ◽  
Xianhong Tong ◽  
...  
Keyword(s):  
Machine Learning ◽  
Embryo Transfer ◽  
Learning Algorithm ◽  
Conditional Inference ◽  
Support Vector ◽  
Machine Learning Algorithm ◽  
Significant Difference ◽  
Predictive Rate ◽  
Ivf Et ◽  
Day Of Embryo Transfer

In vitro fertilization-embryo transfer (IVF-ET) technology make it possible for infertile couples to conceive a baby successfully. Nevertheless, IVF-ET does not guarantee success. Frozen embryo transfer (FET) is an important supplement to IVF-ET. Many factors are correlated with the outcome of FET which is unpredictable. Machine learning is a field of study that predict various outcomes by defining data attributes and using relevant data and calculation algorithms. Machine learning algorithm has been widely used in clinical research. The present study focuses on making predictions of early pregnancy outcomes in FET through clinical characters, including age, body mass index (BMI), endometrial thickness (EMT) on the day of progesterone treatment, good-quality embryo rate (GQR), and type of infertility (primary or secondary), serum estradiol level (E2) on the day of embryo transfer, and serum progesterone level (P) on the day of embryo transfer. We applied four representative machine learning algorithms, including logistic regression (LR), conditional inference tree, random forest (RF) and support vector machine (SVM) to build prediction models and identify the predictive factors. We found no significant difference among the models in the sensitivity, specificity, positive predictive rate, negative predictive rate or accuracy in predicting the pregnancy outcome of FET. For example, the positive/negative predictive rate of the SVM (gamma = 1, cost = 100, 10-fold cross validation) is 0.56 and 0.55. This approach could provide a reference for couples considering FET. The prediction accuracy of the present study is limited, which suggests that there may be some other more effective predictors to be developed in future work.

scholarly journals Machine Learning Algorithm-Based Analysis of Efficacy of Pulmonary Surfactant Combined with Mucosolvan in Meconium Aspiration Syndrome of Newborns through Ultrasonic Images

2021 ◽  
Vol 2021 ◽  
pp. 1-7
Author(s):  
Yanni Ji ◽  
Wenqian Lou ◽  
Jianwei Ji
Keyword(s):  
Machine Learning ◽  
Mechanical Ventilation ◽  
Learning Algorithm ◽  
Combination Group ◽  
Machine Learning Algorithm ◽  
Ventilation Time ◽  
Hospitalization Time ◽  
Significant Difference ◽  
Mechanical Ventilation Time ◽  
Ultrasonic Images

Objective. The study aimed to explore the efficacy of pulmonary surfactant (PS) combined with Mucosolvan in the diagnosis of meconium aspiration syndrome (MAS) in newborns through ultrasonic images of lung based on machine learning. Methods. 138 cases of infants with MAS were selected as the research subjects and randomly divided into PS group (n = 46), Mucosolvan group (n = 46), and combination group (n = 46). Then, ultrasonic images based on machine learning algorithm were used for examination. On the basis of conventional treatment, the PS group accepted intratracheal PS drip treatment with 100 mg/kg. For the Mucosolvan group, 7.5 mg/kg of Mucosolvan was added with 50 g/L glucose, which was diluted to 3 mL. Then, the mixture was injected intravenously with a micropump for more than 5 min. The combination group received combined treatment of PS and Mucosolvan. If there was no relief or the symptoms aggravated after 12 h of PS treatment, the patient should be treated again. 7.5 mg/kg/d of Mucosolvan was given for 7 days. Mechanical ventilation time, hospitalization time, oxygenation index (OI) before treatment, at 3 d and at 7 d after treatment, and arterial/alveolar oxygen ratio (a/APO2) of the three groups were detected and compared. Besides, in-hospital mortality and complication rate of the three groups were statistically compared. Results. Ultrasonic image edge detection based on machine learning algorithm was more condensed and better than Sobel operator. Compared with the PS group and the Mucosolvan group, treatment efficiency, OI at 3 d and at 7 d after treatment, and a/APO2 of combination group were increased. Mechanical ventilation time and hospitalization time of the combination group were shortened, and mortality rate of the combination group was reduced ( P  < 0.05). Compared with the situation before treatment, OI at 3 d and at 7 d after treatment and a/APO2 of the combination group were increased, and OI at 7 d after treatment and a/APO2 of the PS group and the Mucosolvan group were increased ( P  < 0.05). Curative effect, mechanical ventilation time, hospitalization time, OI before and after treatment, a/APO2, and mortality rate during hospitalization of the PS group and the Mucosolvan group had no significant difference ( P  > 0.05). There was no significant difference in the complications rates in the three groups ( P  > 0.05). Conclusion. Pulmonary ultrasound based on machine learning algorithm can be used in the diagnosis of MAS in neonates. PS combined with Mucosolvan is feasible and safe in treating neonatal MAS and effectively improves the pulmonary oxygenation function. Therefore, it is worthy of clinical application.

scholarly journals Evaluation and Calibration of a Low-cost Particle Sensor in Ambient Conditions Using Machine Learning Technologies

2019 ◽  
Author(s):  
Minxing Si ◽  
Ying Xiong ◽  
Shan Du ◽  
Ke Du
Keyword(s):  
Machine Learning ◽  
Linear Regression ◽  
Random Search ◽  
Low Cost ◽  
Learning Algorithms ◽  
Machine Learning Algorithms ◽  
Ambient Conditions ◽  
Test Dataset ◽  
Compact Size ◽  
Significant Difference

Abstract. Particle sensing technology has shown great potential for monitoring particulate matter (PM) with very few temporal and spatial restrictions because of low-cost, compact size, and easy operation. However, the performance of low-cost sensors for PM monitoring in ambient conditions has not been thoroughly evaluated. Monitoring results by low-cost sensors are often questionable. In this study, a low-cost fine particle monitor (Plantower PMS 5003) was co-located with a reference instrument, named Synchronized Hybrid Ambient Real-time Particulate (SHARP) monitor, in Calgary Varsity air monitoring station from December 2018 to April 2019. The study evaluated the performance of this low-cost PM sensor in ambient conditions and calibrated its readings using simple linear regression (SLR), multiple linear regression (MLR), and two more powerful machine learning algorithms using random search techniques for the best model architectures. The two machine learning algorithms are XGBoost and feedforward neural network (NN). Field evaluation showed that the Pearson r between the low-cost sensor and the SHAPR instrument was 0.78. Fligner and Killeen (F-K) test indicated a statistically significant difference between the variances of the PM2.5 values by the low-cost sensor and by the SHARP instrument. Large overestimations by the low-cost sensor before calibration were observed in the field and were believed to be caused by the variation of ambient relative humidity. The root mean square error (RMSE) was 9.93 when comparing the low-cost sensor with the SHARP instrument. The calibration by the feedforward NN had the smallest RMSE of 3.91 in the test dataset, compared to the calibrations by SLR (4.91), MLR (4.65), and XGBoost (4.19). After calibrations, the F-K test using the test dataset showed that the variances of the PM2.5 values by the NN and the XGBoost and by the reference method were not statistically significantly different. From this study, we conclude that feedforward NN is a promising method to address the poor performance of the low-cost sensors for PM2.5 monitoring. In addition, the random search method for hyperparameters was demonstrated to be an efficient approach for selecting the best model structure.

Sign in / Sign up

Export Citation Format

Share Document