scholarly journals A Preliminary Study on Machine Learning-Based Evaluation of Static and Dynamic FET-PET for the Detection of Pseudoprogression in Patients with IDH-Wildtype Glioblastoma

Cancers ◽  
2020 ◽  
Vol 12 (11) ◽  
pp. 3080
Author(s):  
Sied Kebir ◽  
Teresa Schmidt ◽  
Matthias Weber ◽  
Lazaros Lazaridis ◽  
Norbert Galldiks ◽  
...  
Keyword(s):  
Machine Learning ◽  
Roc Curve ◽  
Response Assessment ◽  
Linear Discriminant ◽  
Fet Pet ◽  
Contrast Enhanced ◽  
The Mean ◽  
Preliminary Study ◽  
Validation Set ◽  
Dynamic Fet Pet

Pseudoprogression (PSP) detection in glioblastoma remains challenging and has important clinical implications. We investigated the potential of machine learning (ML) in improving the performance of PET using O-(2-[18F]-fluoroethyl)-L-tyrosine (FET) for differentiation of tumor progression from PSP in IDH-wildtype glioblastoma. We retrospectively evaluated the PET data of patients with newly diagnosed IDH-wildtype glioblastoma following chemoradiation. Contrast-enhanced MRI suspected PSP/TP and all patients underwent subsequently an additional dynamic FET-PET scan. The modified Response Assessment in Neuro-Oncology (RANO) criteria served to diagnose PSP. We trained a Linear Discriminant Analysis (LDA)-based classifier using FET-PET derived features on a hold-out validation set. The results of the ML model were compared with a conventional FET-PET analysis using the receiver-operating-characteristic (ROC) curve. Of the 44 patients included in this preliminary study, 14 patients were diagnosed with PSP. The mean (TBRmean) and maximum tumor-to-brain ratios (TBRmax) were significantly higher in the TP group as compared to the PSP group (p = 0.014 and p = 0.033, respectively). The area under the ROC curve (AUC) for TBRmax and TBRmean was 0.68 and 0.74, respectively. Using the LDA-based algorithm, the AUC (0.93) was significantly higher than the AUC for TBRmax. This preliminary study shows that in IDH-wildtype glioblastoma, ML-based PSP detection leads to better diagnostic performance.

scholarly journals NIMG-14. MACHINE LEARNING-BASED EVALUATION OF STATIC AND DYNAMIC FET-PET FOR THE DETECTION OF PSEUDOPROGRESSION IN PATIENTS WITH IDH-WILDTYPE GLIOBLASTOMA

2020 ◽  
Vol 22 (Supplement_2) ◽  
pp. ii149-ii150
Author(s):  
Sied Kebir ◽  
Teresa Schmidt ◽  
Matthias Weber ◽  
Lazaros Lazaridis ◽  
Norbert Galldiks ◽  
...  
Keyword(s):  
Machine Learning ◽  
Roc Curve ◽  
Roc Analysis ◽  
Operating Characteristic ◽  
Response Assessment ◽  
Linear Discriminant ◽  
Fet Pet ◽  
Contrast Enhanced ◽  
The Mean ◽  
Dynamic Fet Pet

Abstract BACKGROUND Pseudoprogression (PSP) detection in glioblastoma has important clinical implications and remains a challenging task. With the significant advances provided by machine learning (ML) in health care, we investigated the potential of ML in improving the performance of PET using O-(2-[18F]-fluoroethyl)-L-tyrosine (FET) for differentiation of tumor progression from PSP in IDH-wildtype glioblastoma. METHODS We retrospectively evaluated the PET data of patients with newly diagnosed IDH-wildtype glioblastoma following chemoradiation. All patients presented imaging findings suspected of PSP/TP on contrast-enhanced MRI. For further diagnostic evaluation, patients underwent subsequently an additional dynamic FET-PET scan. The modified Response Assessment in Neuro-Oncology (RANO) criteria served to diagnose PSP. To develop a robust ML model, we trained a Linear Discriminant Analysis (LDA)-based classifier using FET-PET derived features on a training cohort and validated the results on a separate test cohort. The results of the ML model were compared with a conventional FET-PET analysis using the receiver-operating-characteristic (ROC) curve. RESULTS Of the 44 patients included in this study, 14 patients were diagnosed with PSP. The mean (TBRmean) and maximum tumor-to-brain ratios (TBRmax) were significantly higher in the TP group as compared to the PSP group (p=0.010 and p=0.047, respectively). The area under the ROC curve (AUC) for TBRmax and TBRmean was 0.68 and 0.74, respectively. Using the LDA-based algorithm, the AUC (0.93) was significantly higher than the AUC for TBRmax. CONCLUSIONS This study shows that in IDH-wildtype glioblastoma, ML-based PSP detection leads to better diagnostic performance compared to conventional ROC analysis.

scholarly journals Validity of Autism Behavior Checklist (ABC): preliminary study

2005 ◽  
Vol 27 (4) ◽  
pp. 295-301 ◽  
Author(s):  
Márcia Regina Fumagalli Marteleto ◽  
Márcia Regina Marcondes Pedromônico
Keyword(s):  
Roc Curve ◽  
Criterion Validity ◽  
Cutoff Score ◽  
Autistic Children ◽  
Autism Behavior Checklist ◽  
Dsm Iv ◽  
The Mean ◽  
Preliminary Study ◽  
Autistic Disorders ◽  
Autism Behavior

OBJECTIVE: To examine the concurrent and criterion validity of the Autism Behavior Checklist (ABC). METHODS: Three groups, comprising 38 mothers of children previously diagnosed with autism (DSM IV-TR, 2002), 43 mothers of children with language disorders other than autism, and 52 mothers of children who had no linguistic or behavioral complaints, were interviewed. In order to minimize the effect of maternal level of education, the questionnaire was completed by the researcher. To determine the concurrent validation, ANOVA and discriminant analysis were used. The ROC curve was used to establish the cutoff score of the sample and to examine the criterion validity. RESULTS: The mean total score was significantly higher in the group of mothers of autistic children than in the other groups. The ABC correctly identified 81.6% of the autistic children. The ROC curve cutoff score was 49, and the sensitivity was 92.1%, higher than the 57.89% found when a cutoff score of 68 was used. The specificity was 92.6%, similar to the 94.73% obtained with a cutoff score of 68. CONCLUSIONS: The ABC shows promise as an instrument for identifying children with autistic disorders, both in clinical and educational contexts, especially when a cutoff score of 49 is used.

scholarly journals FET PET Radiomics for Differentiating Pseudoprogression from Early Tumor Progression in Glioma Patients Post-Chemoradiation

Cancers ◽  
2020 ◽  
Vol 12 (12) ◽  
pp. 3835
Author(s):  
Philipp Lohmann ◽  
Mai A. Elahmadawy ◽  
Robin Gutsche ◽  
Jan-Michael Werner ◽  
Elena K. Bauer ◽  
...  
Keyword(s):  
Tumor Progression ◽  
Negative Predictive Value ◽  
Predictive Value ◽  
Diagnostic Performance ◽  
Data Augmentation ◽  
Response Assessment ◽  
Lower Sensitivity ◽  
Fet Pet ◽  
Early Tumor ◽  
Dynamic Fet Pet

Currently, a reliable diagnostic test for differentiating pseudoprogression from early tumor progression is lacking. We explored the potential of O-(2-[18F]fluoroethyl)-L-tyrosine (FET) positron emission tomography (PET) radiomics for this clinically important task. Thirty-four patients (isocitrate dehydrogenase (IDH)-wildtype glioblastoma, 94%) with progressive magnetic resonance imaging (MRI) changes according to the Response Assessment in Neuro-Oncology (RANO) criteria within the first 12 weeks after completing temozolomide chemoradiation underwent a dynamic FET PET scan. Static and dynamic FET PET parameters were calculated. For radiomics analysis, the number of datasets was increased to 102 using data augmentation. After randomly assigning patients to a training and test dataset, 944 features were calculated on unfiltered and filtered images. The number of features for model generation was limited to four to avoid data overfitting. Eighteen patients were diagnosed with early tumor progression, and 16 patients had pseudoprogression. The FET PET radiomics model correctly diagnosed pseudoprogression in all test cohort patients (sensitivity, 100%; negative predictive value, 100%). In contrast, the diagnostic performance of the best FET PET parameter (TBRmax) was lower (sensitivity, 81%; negative predictive value, 80%). The results suggest that FET PET radiomics helps diagnose patients with pseudoprogression with a high diagnostic performance. Given the clinical significance, further studies are warranted.

scholarly journals Reproducibility of Ablated Volume Measurement Is Higher with Contrast-Enhanced Ultrasound than with B-Mode Ultrasound after Benign Thyroid Nodule Radiofrequency Ablation—A Preliminary Study

2020 ◽  
Vol 9 (5) ◽  
pp. 1504 ◽  
Author(s):  
Simone Schiaffino ◽  
Francesca Serpi ◽  
Duccio Rossi ◽  
Valerio Ferrara ◽  
Ciriaco Buonomenna ◽  
...  
Keyword(s):  
Radiofrequency Ablation ◽  
Thyroid Nodules ◽  
Observer Agreement ◽  
Contrast Enhanced Ultrasound ◽  
Limits Of Agreement ◽  
Contrast Enhanced ◽  
The Mean ◽  
Preliminary Study ◽  
Benign Thyroid Nodules ◽  
Benign Thyroid

The reproducibility of contrast-enhanced ultrasound (CEUS) and standard B-mode ultrasound in the assessment of radiofrequency-ablated volume of benign thyroid nodules was compared. A preliminary study was conducted on consecutive patients who underwent radiofrequency ablation (RFA) of benign thyroid nodules between 2014 and 2016, with available CEUS and B-mode post-ablation checks. CEUS and B-mode images were retrospectively evaluated by two radiologists to assess inter- and intra-observer agreement in the assessment of ablated volume (Bland–Altman test). For CEUS, the mean inter-observer difference (95% limits of agreement) was 0.219 mL (-0.372–0.809 mL); for B-mode, the mean difference was 0.880 mL (-1.655–3.414 mL). Reproducibility was significantly higher for CEUS (85%) than for B-mode (27%). Mean intra-observer differences (95% limits of agreement) were 0.013 mL (0.803–4.097 mL) for Reader 1 and 0.031 mL (0.763–3.931 mL) for Reader 2 using CEUS, while they were 0.567 mL (-2.180–4.317 mL, Reader 1) and 0.759 mL (-2.584–4.290 mL, Reader 2) for B-mode. Intra-observer reproducibility was significantly higher for CEUS (96% and 95%, for the two readers) than for B-mode (21% and 23%). In conclusion, CEUS had higher reproducibility and inter- and intra-observer agreement compared to conventional B-mode in the assessment of radiofrequency-ablated volume of benign thyroid nodules.

scholarly journals Ensemble Learning Models Based on Noninvasive Features for Type 2 Diabetes Screening: Model Development and Validation (Preprint)

2019 ◽  
Author(s):  
Tianzhou Yang ◽  
Li Zhang ◽  
Liwei Yi ◽  
Huawei Feng ◽  
Shimeng Li ◽  
...  
Keyword(s):  
Machine Learning ◽  
Prediction Models ◽  
Ensemble Methods ◽  
Diabetes Screening ◽  
Linear Discriminant ◽  
Machine Learning Methods ◽  
Diabetes Prediction ◽  
Validation Set ◽  
Fold Cross Validation

BACKGROUND Early diabetes screening can effectively reduce the burden of disease. However, natural population–based screening projects require a large number of resources. With the emergence and development of machine learning, researchers have started to pursue more flexible and efficient methods to screen or predict type 2 diabetes. OBJECTIVE The aim of this study was to build prediction models based on the ensemble learning method for diabetes screening to further improve the health status of the population in a noninvasive and inexpensive manner. METHODS The dataset for building and evaluating the diabetes prediction model was extracted from the National Health and Nutrition Examination Survey from 2011-2016. After data cleaning and feature selection, the dataset was split into a training set (80%, 2011-2014), test set (20%, 2011-2014) and validation set (2015-2016). Three simple machine learning methods (linear discriminant analysis, support vector machine, and random forest) and easy ensemble methods were used to build diabetes prediction models. The performance of the models was evaluated through 5-fold cross-validation and external validation. The Delong test (2-sided) was used to test the performance differences between the models. RESULTS We selected 8057 observations and 12 attributes from the database. In the 5-fold cross-validation, the three simple methods yielded highly predictive performance models with areas under the curve (AUCs) over 0.800, wherein the ensemble methods significantly outperformed the simple methods. When we evaluated the models in the test set and validation set, the same trends were observed. The ensemble model of linear discriminant analysis yielded the best performance, with an AUC of 0.849, an accuracy of 0.730, a sensitivity of 0.819, and a specificity of 0.709 in the validation set. CONCLUSIONS This study indicates that efficient screening using machine learning methods with noninvasive tests can be applied to a large population and achieve the objective of secondary prevention.

scholarly journals Evaluation of CSTB and DMBT1 Expression in Saliva of Gastric Cancer Patients and Controls

2021 ◽  
Author(s):  
Maryam Koopaie ◽  
Marjan Ghafourian ◽  
Soheila Manifar ◽  
Shima Younespour ◽  
Mansour Davoudi ◽  
...  
Keyword(s):  
Machine Learning ◽  
Gastric Cancer ◽  
Statistical Analysis ◽  
Food Intake ◽  
Roc Curve ◽  
Clinical Characteristics ◽  
Demographic Data ◽  
Healthy Individuals ◽  
Diagnostic Algorithms ◽  
The Mean

Abstract Background: Gastric cancer (GC) is the fifth most common cancer and the third cause of cancer deaths globally with late diagnosis, low survival rate and poor prognosis. This case-control study aimed to evaluate the expression of cystatin B (CSTB) and deleted in malignant brain tumor 1 (DMBT1) in the saliva of GC patients with healthy individuals to construct diagnostic algorithms using statistical analysis and machine learning methods.Methods: Demographic data, clinical characteristics and food intake habits of the case and control group were gathered through a standard checklist. Unstimulated whole saliva samples were taken from 31 healthy individuals and 31 GC patients. Through ELISA test and statistical analysis, the expression of salivary CSTB and DMBT1 proteins were evaluated. To construct diagnostic algorithms, we used the machine learning method.Results: The mean salivary expression of CSTB in GC patients was significantly lower (115.55±7.06, p=0.001) and the mean salivary expression of DMBT1 in GC patients was significantly higher (171.88±39.67, p=0.002) than the control. Multiple linear regression analysis demonstrated that GC was significantly correlated with high levels of DMBT1 after controlling the effects of age of participants (R2=0.20, p<0.001). Considering salivary CSTB greater than 119.06 ng/mL as an optimal cut-off value, the sensitivity and specificity of CSTB in the diagnosis of GC was 83.87% and 70.97%, respectively The area under the ROC curve was calculated as 0.728. The optimal cut-off value of DMBT1 for differentiating GC patients from controls was greater than 146.33 ng/mL (sensitivity=80.65% and specificity=64.52%). The area under the ROC curve was up to 0.741. As a result of the machine learning method, the area under the receiver-operating characteristic curve for the diagnostic ability of CSTB, DMBT1, demographic data, clinical characteristics and food intake habits was 0.95. The machine learning model's sensitivity, specificity, and accuracy were 100%, 70.8%, and 80.5%, respectively. Conclusion: Salivary levels of DMBT1 and CSTB may be accurate in diagnosing GCs. Machine learning analyses using salivary biomarkers, demographic, clinical and nutrition habits data simultaneously could provide affordability models with acceptable accuracy for differentiation of GC by a cost-effective and non-invasive method.

scholarly journals Value of contrast-enhanced CT based radiomic machine learning algorithm in differentiating gastrointestinal stromal tumors with KIT exon 11 mutation: a two-center study

2020 ◽  
Author(s):  
Bo Liu ◽  
Hexiang Wang ◽  
Shunli Liu ◽  
Shifeng Yang ◽  
Yancheng Song ◽  
...  
Keyword(s):  
Machine Learning ◽  
Gastrointestinal Stromal Tumors ◽  
Ct Images ◽  
Arterial Phase ◽  
Training Set ◽  
Clinical Model ◽  
Stromal Tumors ◽  
Contrast Enhanced ◽  
Validation Set ◽  
Exon 11

Abstract Background Knowing the genetic phenotype of gastrointestinal stromal tumors (GISTs) is essential for patients who receive therapy with tyrosine kinase inhibitors.Methods We enrolled 106 patients (80 in the training set, 26 in the validation set) with clinicopathologically confirmed GISTs from two centers. Preoperative and postoperative clinical characteristics were selected and analyzed to construct the clinical model. Arterial phase (A-phase), venous phase (V-phase), delayed phase (D-phase), and combined radiomics algorithms were generated from the training set based on contrast-enhanced computed tomography (CE-CT) images. Various radiomics feature selection methods were used, namely least absolute shrinkage and selection operator (LASSO); minimum redundancy maximum relevance (mRMR); and generalized linear model (GLM) as a machine-learning classifier. Independent predictive factors were determined to construct preoperative and postoperative radiomics nomograms by multivariate logistic regression analysis. The performances of the clinical model, radiomics algorithm, and radiomics nomogram in distinguishing GISTs with the KIT exon 11 mutation were evaluated by area under the curve (AUC) of the receiver operating characteristic (ROC).Results The combined radiomics algorithm was found to be the best prediction model for differentiating the expression status of the KIT exon 11 mutation (AUC = 0.836; 95% confidence interval (CI), 0.640–0.951) in the validation set. The clinical model, and preoperative and postoperative radiomics nomograms had AUCs of 0.606 (95% CI, 0.397–0.790), 0.715 (95% CI, 0.506–0.873), and 0.679 (95% CI, 0.468–0.847), respectively, with the validation set.Conclusion The radiomics algorithm could distinguish GISTs with the KIT exon 11 mutation based on CE-CT images and could potentially be used for selective genetic analysis to support the precision medicine of GISTs.

scholarly journals Ensemble Learning Models Based on Noninvasive Features for Type 2 Diabetes Screening: Model Development and Validation

10.2196/15431 ◽  
2020 ◽  
Vol 8 (6) ◽  
pp. e15431
Author(s):  
Tianzhou Yang ◽  
Li Zhang ◽  
Liwei Yi ◽  
Huawei Feng ◽  
Shimeng Li ◽  
...  
Keyword(s):  
Machine Learning ◽  
Prediction Models ◽  
Ensemble Methods ◽  
Diabetes Screening ◽  
Linear Discriminant ◽  
Machine Learning Methods ◽  
Diabetes Prediction ◽  
Validation Set ◽  
Fold Cross Validation

Background Early diabetes screening can effectively reduce the burden of disease. However, natural population–based screening projects require a large number of resources. With the emergence and development of machine learning, researchers have started to pursue more flexible and efficient methods to screen or predict type 2 diabetes. Objective The aim of this study was to build prediction models based on the ensemble learning method for diabetes screening to further improve the health status of the population in a noninvasive and inexpensive manner. Methods The dataset for building and evaluating the diabetes prediction model was extracted from the National Health and Nutrition Examination Survey from 2011-2016. After data cleaning and feature selection, the dataset was split into a training set (80%, 2011-2014), test set (20%, 2011-2014) and validation set (2015-2016). Three simple machine learning methods (linear discriminant analysis, support vector machine, and random forest) and easy ensemble methods were used to build diabetes prediction models. The performance of the models was evaluated through 5-fold cross-validation and external validation. The Delong test (2-sided) was used to test the performance differences between the models. Results We selected 8057 observations and 12 attributes from the database. In the 5-fold cross-validation, the three simple methods yielded highly predictive performance models with areas under the curve (AUCs) over 0.800, wherein the ensemble methods significantly outperformed the simple methods. When we evaluated the models in the test set and validation set, the same trends were observed. The ensemble model of linear discriminant analysis yielded the best performance, with an AUC of 0.849, an accuracy of 0.730, a sensitivity of 0.819, and a specificity of 0.709 in the validation set. Conclusions This study indicates that efficient screening using machine learning methods with noninvasive tests can be applied to a large population and achieve the objective of secondary prevention.

P09.01 Radiomics-based machine learning approach in differentiation between vestibular schwannoma and meningioma in the cerebellopontine angle

2021 ◽  
Vol 23 (Supplement_2) ◽  
pp. ii26-ii26
Author(s):  
Y Zhang ◽  
C Chen ◽  
J Xu
Keyword(s):  
Machine Learning ◽  
Vestibular Schwannoma ◽  
Cerebellopontine Angle ◽  
Gradient Boosting ◽  
Surgical Approaches ◽  
Support Vector ◽  
Selection Methods ◽  
Linear Discriminant ◽  
Machine Learning Classification ◽  
Validation Set

Abstract BACKGROUND Vestibular schwannoma (VS) and meningioma are the most two common tumors in the cerebellopontine angle (CPA). Accurate preoperative differentiation of the two lesions is important due to their different surgical approaches and outcomes for the preservation of hearing and facial nerve function. Magnetic resonance (MR) scan is commonly performed to preoperatively evaluate CPA tumors and to differentiate VS from meningioma in clinical routine. However, in some cases, overlaps of conventional MR imaging patterns between the two lesions could make preoperative diagnosis challenging. The purpose of this study is to investigate the ability of radiomics, a novel method providing objective and quantitative information beyond visual assessment, in differentiation between VS and meningioma located at CPA using machine learning technology. MATERIAL AND METHODS This retrospective study enrolled eligible patients who were diagnosed with VS (N = 50) or meningioma (N = 41) in the CPA. A set of mineable three-dimensional radiomic parameters were extracted from preoperative contrast-enhanced T1-weighted images. Optimal features were selected first with three selection methods including distance correlation, least absolute shrinkage and selection operator (LASSO) and gradient boosting decision tree (GBDT). Then three machine learning classification algorithms, namely linear discriminant analysis (LDA), support vector machine (SVM) and random forest were employed to build discriminative models. Area under the curve (AUC), accuracy, sensitivity and specificity were calculated to assess the performance of each model. RESULTS Nine models were established with different combinations of selection methods and machine learning classifiers. Three classifiers with the suitable selection method all represented feasible ability in differentiation with AUC more than 0.86 in the validation set, and LDA-based models seemed to show better diagnostic performance than those based on the other two classifiers. The combination of LASSO and LDA classifier was found to show the highest AUC of 0.942 in the validation set. CONCLUSION Radiomics-based models via machine learning approaches could potentially be utilized to assist in preoperative differentiation between VS and meningioma in the CPA.

Sign in / Sign up

Export Citation Format

Share Document