One- year mortality in patients with advanced hepatocellular carcinoma on immunotherapy: Prediction using machine learning models (Preprint)

2021 ◽  
Author(s):  
Thomas Ka-Luen Lui ◽  
Ka Shing, Michael Cheung ◽  
Wai Keung Leung
Keyword(s):  
Machine Learning ◽  
Hepatocellular Carcinoma ◽  
Characteristic Curve ◽  
False Negative ◽  
False Negative Rate ◽  
Absolute Error ◽  
Advanced Hepatocellular Carcinoma ◽  
Data Set ◽  
One Year ◽  
Related Mortality

BACKGROUND Immunotherapy is a new promising treatment for patients with advanced hepatocellular carcinoma (HCC), but is costly and potentially associated with considerable side effects. OBJECTIVE This study aimed to evaluate the role of machine learning (ML) models in predicting the one-year cancer-related mortality in advanced HCC patients treated with immunotherapy METHODS 395 HCC patients who had received immunotherapy (including nivolumab, pembrolizumab or ipilimumab) in 2014 - 2019 in Hong Kong were included. The whole data set were randomly divided into training (n=316) and validation (n=79) set. The data set, including 45 clinical variables, was used to construct six different ML models in predicting the risk of one-year mortality. The performances of ML models were measured by the area under receiver operating characteristic curve (AUC) and the mean absolute error (MAE) using calibration analysis. RESULTS The overall one-year cancer-related mortality was 51.1%. Of the six ML models, the random forest (RF) has the highest AUC of 0.93 (95%CI: 0.86-0.98), which was better than logistic regression (0.82, p=0.01) and XGBoost (0.86, p=0.04). RF also had the lowest false positive (6.7%) and false negative rate (2.8%). High baseline AFP, bilirubin and alkaline phosphatase were three common risk factors identified by all ML models. CONCLUSIONS ML models could predict one-year cancer-related mortality of HCC patients treated with immunotherapy, which may help to select patients who would most benefit from this new treatment option.

scholarly journals Registration Status Prediction of Students using Machine Learning in the Context of Private University of Bangladesh

2019 ◽  
Vol 9 (1) ◽  
pp. 2594-2600 ◽  
Keyword(s):  
Higher Education ◽  
Machine Learning ◽  
Random Forest ◽  
Performance Metrics ◽  
False Positive Rate ◽  
False Negative ◽  
Private Universities ◽  
False Negative Rate ◽  
Private University ◽  
Data Set

Bangladesh is a densely populated country where a large portion of citizens is living under poverty. In Bangladesh, a significant portion of higher education is accomplished at private universities. In this twenty-first century, these students of higher education are highly mobile and different from earlier generations. Thus, retaining existing students has become a great challenge for many private universities in Bangladesh. Early prediction of the total number of registered students in a semester can help in this regard. This can have a direct impact on a private university in terms of budget, marketing strategy, and sustainability. In this paper, we have predicted the number of registered students in a semester in the context of a private university by following several machine learning approaches. We have applied seven prominent classifiers, namely SVM, Naive Bayes, Logistic, JRip, J48, Multilayer Perceptron, and Random Forest on a data set of more than a thousand students of a private university in Bangladesh, where each record contains five attributes. First, all data are preprocessed. Then preprocessed data are separated into the training and testing set. Then, all these classifiers are trained and tested. Since a suitable classifier is required to solve the problem, the performances of all seven classifiers need to be thoroughly assessed. So, we have computed six performance metrics, i.e. accuracy, sensitivity, specificity, precision, false positive rate (FPR) and false negative rate (FNR) for each of the seven classifiers and compare them. We have found that SVM outperforms all other classifiers achieving 85.76% accuracy, whereas Random Forest achieved the lowest accuracy which is 79.65%.

scholarly journals Development, evaluation, and validation of machine learning models for COVID-19 detection based on routine blood tests

2020 ◽  
Author(s):  
Cabitza Federico ◽  
Campagner Andrea ◽  
Ferrari Davide ◽  
Di Resta Chiara ◽  
Ceriotti Daniele ◽  
...  
Keyword(s):  
Machine Learning ◽  
Complete Blood Count ◽  
Characteristic Curve ◽  
External Validation ◽  
False Negative ◽  
Turnaround Time ◽  
Training Data ◽  
Data Set ◽  
Blood Tests ◽  
Routine Blood

AbstractBackgroundThe rRT-PCR test, the current gold standard for the detection of coronavirus disease (COVID-19), presents with known shortcomings, such as long turnaround time, potential shortage of reagents, false-negative rates around 15–20%, and expensive equipment. The hematochemical values of routine blood exams could represent a faster and less expensive alternative.MethodsThree different training data set of hematochemical values from 1,624 patients (52% COVID-19 positive), admitted at San Raphael Hospital (OSR) from February to May 2020, were used for developing machine learning (ML) models: the complete OSR dataset (72 features: complete blood count (CBC), biochemical, coagulation, hemogasanalysis and CO-Oxymetry values, age, sex and specific symptoms at triage) and two sub-datasets (COVID-specific and CBC dataset, 32 and 21 features respectively). 58 cases (50% COVID-19 positive) from another hospital, and 54 negative patients collected in 2018 at OSR, were used for internal-external and external validation.ResultsWe developed five ML models: for the complete OSR dataset, the area under the receiver operating characteristic curve (AUC) for the algorithms ranged from 0.83 to 0.90; for the COVID-specific dataset from 0.83 to 0.87; and for the CBC dataset from 0.74 to 0.86. The validations also achieved good results: respectively, AUC from 0.75 to 0.78; and specificity from 0.92 to 0.96.ConclusionsML can be applied to blood tests as both an adjunct and alternative method to rRT-PCR for the fast and cost-effective identification of COVID-19-positive patients. This is especially useful in developing countries, or in countries facing an increase in contagions.

scholarly journals Development, evaluation, and validation of machine learning models for COVID-19 detection based on routine blood tests

2020 ◽  
Vol 0 (0) ◽  
Author(s):  
Federico Cabitza ◽  
Andrea Campagner ◽  
Davide Ferrari ◽  
Chiara Di Resta ◽  
Daniele Ceriotti ◽  
...  
Keyword(s):  
Machine Learning ◽  
Complete Blood Count ◽  
Characteristic Curve ◽  
External Validation ◽  
False Negative ◽  
Turnaround Time ◽  
Training Data ◽  
Data Set ◽  
Blood Tests ◽  
Routine Blood

AbstractObjectivesThe rRT-PCR test, the current gold standard for the detection of coronavirus disease (COVID-19), presents with known shortcomings, such as long turnaround time, potential shortage of reagents, false-negative rates around 15–20%, and expensive equipment. The hematochemical values of routine blood exams could represent a faster and less expensive alternative.MethodsThree different training data set of hematochemical values from 1,624 patients (52% COVID-19 positive), admitted at San Raphael Hospital (OSR) from February to May 2020, were used for developing machine learning (ML) models: the complete OSR dataset (72 features: complete blood count (CBC), biochemical, coagulation, hemogasanalysis and CO-Oxymetry values, age, sex and specific symptoms at triage) and two sub-datasets (COVID-specific and CBC dataset, 32 and 21 features respectively). 58 cases (50% COVID-19 positive) from another hospital, and 54 negative patients collected in 2018 at OSR, were used for internal-external and external validation.ResultsWe developed five ML models: for the complete OSR dataset, the area under the receiver operating characteristic curve (AUC) for the algorithms ranged from 0.83 to 0.90; for the COVID-specific dataset from 0.83 to 0.87; and for the CBC dataset from 0.74 to 0.86. The validations also achieved good results: respectively, AUC from 0.75 to 0.78; and specificity from 0.92 to 0.96.ConclusionsML can be applied to blood tests as both an adjunct and alternative method to rRT-PCR for the fast and cost-effective identification of COVID-19-positive patients. This is especially useful in developing countries, or in countries facing an increase in contagions.

scholarly journals Estimation of Soil Cohesion Using Machine Learning Method: A Random Forest Approach

2021 ◽  
Vol 2021 ◽  
pp. 1-14
Author(s):  
Hai-Bang Ly ◽  
Thuy-Anh Nguyen ◽  
Binh Thai Pham
Keyword(s):  
Machine Learning ◽  
Random Forest ◽  
Soil Properties ◽  
Clay Content ◽  
Absolute Error ◽  
Experimental Methods ◽  
Liquid Limit ◽  
Support Vector ◽  
Data Set ◽  
Soil Cohesion

Soil cohesion (C) is one of the critical soil properties and is closely related to basic soil properties such as particle size distribution, pore size, and shear strength. Hence, it is mainly determined by experimental methods. However, the experimental methods are often time-consuming and costly. Therefore, developing an alternative approach based on machine learning (ML) techniques to solve this problem is highly recommended. In this study, machine learning models, namely, support vector machine (SVM), Gaussian regression process (GPR), and random forest (RF), were built based on a data set of 145 soil samples collected from the Da Nang-Quang Ngai expressway project, Vietnam. The database also includes six input parameters, that is, clay content, moisture content, liquid limit, plastic limit, specific gravity, and void ratio. The performance of the model was assessed by three statistical criteria, namely, the correlation coefficient (R), mean absolute error (MAE), and root mean square error (RMSE). The results demonstrated that the proposed RF model could accurately predict soil cohesion with high accuracy (R = 0.891) and low error (RMSE = 3.323 and MAE = 2.511), and its predictive capability is better than SVM and GPR. Therefore, the RF model can be used as a cost-effective approach in predicting soil cohesion forces used in the design and inspection of constructions.

Adoption of Machine Learning With Adaptive Approach for Securing CPS

2020 ◽  
pp. 388-415
Author(s):  
Rama Mercy Sam Sigamani
Keyword(s):  
Machine Learning ◽  
Cyber Security ◽  
Smart Grids ◽  
False Negative ◽  
False Negative Rate ◽  
Cyber Physical Systems ◽  
Detection Accuracy ◽  
Physical Systems ◽  
Positive Rate ◽  
Iot Devices

The cyber physical system safety and security is the major concern on the incorporated components with interface standards, communication protocols, physical operational characteristics, and real-time sensing. The seamless integration of computational and distributed physical components with intelligent mechanisms increases the adaptability, autonomy, efficiency, functionality, reliability, safety, and usability of cyber-physical systems. In IoT-enabled cyber physical systems, cyber security is an essential challenge due to IoT devices in industrial control systems. Computational intelligence algorithms have been proposed to detect and mitigate the cyber-attacks in cyber physical systems, smart grids, power systems. The various machine learning approaches towards securing CPS is observed based on the performance metrics like detection accuracy, average classification rate, false negative rate, false positive rate, processing time per packet. A unique feature of CPS is considered through structural adaptation which facilitates a self-healing CPS.

Machine-Learning-Based External Plagiarism Detecting Methodology From Monolingual Documents

2018 ◽  
pp. 122-139
Author(s):  
Saugata Bose ◽  
Ritambhra Korpal
Keyword(s):  
Machine Learning ◽  
Language Processing ◽  
Confusion Matrix ◽  
False Negative ◽  
False Negative Rate ◽  
Search Space ◽  
Machine Learning Algorithms ◽  
C4.5 Decision Tree ◽  
N Gram ◽  
Four Levels

In this chapter, an initiative is proposed where natural language processing (NLP) techniques and supervised machine learning algorithms have been combined to detect external plagiarism. The major emphasis is on to construct a framework to detect plagiarism from monolingual texts by implementing n-gram frequency comparison approach. The framework is based on 120 characteristics which have been extracted during pre-processing steps using simple NLP approach. Afterward, filter metrics has been applied to select most relevant features and supervised classification learning algorithm has been used later to classify the documents in four levels of plagiarism. Then, confusion matrix was built to estimate the false positives and false negatives. Finally, the authors have shown C4.5 decision tree-based classifier's suitability on calculating accuracy over naive Bayes. The framework achieved 89% accuracy with low false positive and false negative rate and it shows higher precision and recall value comparing to passage similarities method, sentence similarity method, and search space reduction method.

scholarly journals The Application of Machine Learning to a General Risk–Need Assessment Instrument in the Prediction of Criminal Recidivism

2020 ◽  
pp. 009385482096975
Author(s):  
Mehdi Ghasemi ◽  
Daniel Anvari ◽  
Mahshid Atapour ◽  
J. Stephen wormith ◽  
Keira C. Stockdale ◽  
...  
Keyword(s):  
Machine Learning ◽  
Predictive Accuracy ◽  
Characteristic Curve ◽  
Assessment Instrument ◽  
Support Vector ◽  
Data Set ◽  
Applied Machine Learning ◽  
Vector Machines ◽  
Individual Scores

The Level of Service/Case Management Inventory (LS/CMI) is one of the most frequently used tools to assess criminogenic risk–need in justice-involved individuals. Meta-analytic research demonstrates strong predictive accuracy for various recidivism outcomes. In this exploratory study, we applied machine learning (ML) algorithms (decision trees, random forests, and support vector machines) to a data set with nearly 100,000 LS/CMI administrations to provincial corrections clientele in Ontario, Canada, and approximately 3 years follow-up. The overall accuracies and areas under the receiver operating characteristic curve (AUCs) were comparable, although ML outperformed LS/CMI in terms of predictive accuracy for the middle scores where it is hardest to predict the recidivism outcome. Moreover, ML improved the AUCs for individual scores to near 0.60, from 0.50 for the LS/CMI, indicating that ML also improves the ability to rank individuals according to their probability of recidivating. Potential considerations, applications, and future directions are discussed.

scholarly journals Multi-Hazard Exposure Mapping Using Machine Learning Techniques: A Case Study from Iran

2019 ◽  
Vol 11 (16) ◽  
pp. 1943 ◽  
Author(s):  
Omid Rahmati ◽  
Saleh Yousefi ◽  
Zahra Kalantari ◽  
Evelyn Uuemaa ◽  
Teimur Teimurian ◽  
...  
Keyword(s):  
Machine Learning ◽  
State Of The Art ◽  
Characteristic Curve ◽  
Machine Learning Techniques ◽  
Support Vector ◽  
Mountainous Area ◽  
Data Set ◽  
Boosted Regression Tree ◽  
Hazard Exposure ◽  
Learning Techniques

Mountainous areas are highly prone to a variety of nature-triggered disasters, which often cause disabling harm, death, destruction, and damage. In this work, an attempt was made to develop an accurate multi-hazard exposure map for a mountainous area (Asara watershed, Iran), based on state-of-the art machine learning techniques. Hazard modeling for avalanches, rockfalls, and floods was performed using three state-of-the-art models—support vector machine (SVM), boosted regression tree (BRT), and generalized additive model (GAM). Topo-hydrological and geo-environmental factors were used as predictors in the models. A flood dataset (n = 133 flood events) was applied, which had been prepared using Sentinel-1-based processing and ground-based information. In addition, snow avalanche (n = 58) and rockfall (n = 101) data sets were used. The data set of each hazard type was randomly divided to two groups: Training (70%) and validation (30%). Model performance was evaluated by the true skill score (TSS) and the area under receiver operating characteristic curve (AUC) criteria. Using an exposure map, the multi-hazard map was converted into a multi-hazard exposure map. According to both validation methods, the SVM model showed the highest accuracy for avalanches (AUC = 92.4%, TSS = 0.72) and rockfalls (AUC = 93.7%, TSS = 0.81), while BRT demonstrated the best performance for flood hazards (AUC = 94.2%, TSS = 0.80). Overall, multi-hazard exposure modeling revealed that valleys and areas close to the Chalous Road, one of the most important roads in Iran, were associated with high and very high levels of risk. The proposed multi-hazard exposure framework can be helpful in supporting decision making on mountain social-ecological systems facing multiple hazards.

First-line avelumab + axitinib in patients with advanced hepatocellular carcinoma: Results from a phase 1b trial (VEGF Liver 100).

2019 ◽  
Vol 37 (15_suppl) ◽  
pp. 4072-4072 ◽  
Author(s):  
Masatoshi Kudo ◽  
Kenta Motomura ◽  
Yoshiyuki Wada ◽  
Yoshitaka Inaba ◽  
Yasunari Sakamoto ◽  
...  
Keyword(s):  
Hepatocellular Carcinoma ◽  
Kinase Inhibitor ◽  
Objective Response ◽  
Study Data ◽  
Clinical Activity ◽  
Advanced Hepatocellular Carcinoma ◽  
Data Set ◽  
Phase 1B ◽  
Grade 3

4072 Background: Combining an immune checkpoint inhibitor with a targeted antiangiogenic agent may leverage complementary mechanisms of action for treatment of advanced/metastatic (a/m) hepatocellular carcinoma (HCC). Avelumab is a human anti–PD-L1 IgG1 antibody with clinical activity in various tumor types; axitinib is a tyrosine kinase inhibitor selective for VEGF receptors 1/2/3. VEGF Liver 100 (NCT03289533) is a phase 1b study evaluating safety and efficacy of avelumab + axitinib in treatment-naive patients (pts) with HCC; interim results are reported here. Methods: Eligible pts had confirmed a/m HCC, ≥1 measurable lesion, a fresh or archival tumor specimen, ECOG PS ≤1, and Child-Pugh class A. Pts received avelumab 10 mg/kg IV Q2W + axitinib 5 mg orally BID until progression, unacceptable toxicity, or withdrawal. Endpoints included safety and objective response (RECIST v1.1; modified [m] RECIST for HCC). Results: Interim assessment was performed after a minimum follow up of 6 months based on the released study data set (clinical cut-off date: Aug 1, 2018). As of the cut-off date, 22 pts (median age: 68.5 y) were treated with avelumab (median: 20.0 wk) and axitinib (median: 19.9 wk). The most common grade 3 treatment-related adverse events (TRAEs) (≥10% of patients) were hypertension (50.0%) and hand-foot syndrome (22.7%); no grade 4/5 TRAEs were reported. Immune-related AEs (irAEs) (≥10% of pts) were hypothyroidism (31.8%) and hyperthyroidism (13.6%). No grade ≥3 irAEs were reported; no pts discontinued treatment due to TRAEs or irAEs. Based on Waterfall plot calculations, tumor shrinkage was observed in 15 (68.2%) and 16 (72.7%) pts by RECIST and mRECIST, respectively. ORR was 13.6% (95% CI, 2.9%-34.9%) and 31.8% (95% CI, 13.9%-54.9%) by RECIST and mRECIST, respectively. OS data were immature at data cutoff. Conclusions: The preliminary safety of avelumab + axitinib in HCC is manageable and consistent with the known safety profiles of avelumab and axitinib when administered as monotherapies. This study demonstrates antitumor activity of the combination in HCC. Follow-up is ongoing. Clinical trial information: NCT03289533. [Table: see text]

Correlation of decreased expression of PD-L1 on circulating tumor cells and clinical benefit in SCLC Patients treated with RRx-001, a CD47 downregulator, in a phase II trial.

2020 ◽  
Vol 38 (15_suppl) ◽  
pp. 9062-9062
Author(s):  
Corey Carter ◽  
Yusuke Tomita ◽  
Akira Yuno ◽  
Jonathan Baker ◽  
Min-Jung Lee ◽  
...  
Keyword(s):  
Tumor Cells ◽  
Circulating Tumor Cells ◽  
Clinical Benefit ◽  
Logistic Model ◽  
Goodness Of Fit ◽  
False Negative ◽  
False Negative Rate ◽  
Cycle Duration ◽  
Data Set ◽  
Platinum Doublet

9062 Background: In a Phase 2 trial called QUADRUPLE THREAT (QT) (NCT02489903), where 2nd line+ small cell lung cancer (SCLC) patients were treated with RRx-001 and a platinum doublet, the programmed death-ligand 1 (PD-L1) status of circulating tumor cells (CTCs) in 14 patient samples was evaluated. Methods: 26 consented patients received weekly RRx-001 4 mg followed by a reintroduced platinum doublet; epithelial cell adhesion molecule (EPCAM+) CTCs from 10 ml of blood on two consecutive timepoints cycle 1 day 1 and cycle 3 day 8 (cycle duration = 1 week) were detected by EpCAM-based immunomagnetic capture and flow cytometric analysis. CTCs were further characterized for protein expression of PD-L1. Tumor response was classified as partial or complete response based on the response evaluation criteria in solid tumors (RECISTv1.1) measured every 6 weeks. Results: The analyzed clinical data set comprised 14 RECIST-evaluable patients. 50% were females (7/14) and the median age (years) at baseline was 64.5 (Min = 48.5, Max = 84.2, SD = 10.3). The logistic model McFadden goodness of fit score (0 to 100) is 0.477, which is a strong correlation value. The logistic model analyzing the association of CTC PD-L1 expression at two timepoints and response had an approximate 92.8% accuracy in its prediction of clinical benefit (SD/PR/CR). Accuracy is defined in the standard way as 1- (False positive rate + False negative rate). The estimated ROC displayed in Figure 1 suggests a ROC AUC of 0.93 (95% CI: 0.78, 0.99), an excellent measure of performance. Conclusions: Reduction of PD-L1 expression was correlated with good clinical outcome after RRx-001 + platinum doublet treatment. PD-L1 expression reduction in favor of RRx-001 RECIST clinical benefit was clinically significant as compared to non-responders with progressive disease (PD). In the ongoing SCLC Phase 3 study called REPLATINUM (NCT03699956), analyses are planned to correlate response and survival with expression of CD47 and PD-L1 on CTCs. Clinical trial information: NCT02489903.

Sign in / Sign up

Export Citation Format

Share Document