Radiomics Nomogram Improves the Prediction of Epilepsy in Patients With Gliomas

Morphometric and radiomics analysis toward the prediction of epilepsy associated with supratentorial low-grade glioma in children

OBJECTIVES

Understanding the impact of epilepsy on pediatric brain tumors is crucial to diagnostic precision and optimal treatment selection. This study investigated MRI radiomics features, tumor location, voxel-based morphometry (VBM) for gray matter density, and tumor volumetry to differentiate between children with low grade glioma (LGG)-associated epilepsies and those without, and further identified key radiomics features for predicting of epilepsy risk in children with supratentorial LGG to construct an epilepsy prediction model.

METHODS

A total of 206 radiomics features of tumors and voxel-based morphometric analysis of tumor location features were extracted from T2-FLAIR images in a primary cohort of 48 children with LGG with epilepsy (N = 23) or without epilepsy (N = 25), prior to surgery. Feature selection was performed using the minimum redundancy maximum relevance algorithm, and leave-one-out cross-validation was applied to assess the predictive performance of radiomics and tumor location signatures in differentiating epilepsy-associated LGG from non-epilepsy cases.

RESULTS

Voxel-based morphometric analysis showed significant positive t-scores within bilateral temporal cortex and negative t-scores in basal ganglia between epilepsy and non-epilepsy groups. Eight radiomics features were identified as significant predictors of epilepsy in LGG, encompassing characteristics of 2 locations, 2 shapes, 1 image gray scale intensity, and 3 textures. The most important predictor was temporal lobe involvement, followed by high dependence high grey level emphasis, elongation, area density, information correlation 1, midbrain and intensity range. The Linear Support Vector Machine (SVM) model yielded the best prediction performance, when implemented with a combination of radiomics features and tumor location features, as evidenced by the following metrics: precision (0.955), recall (0.913), specificity (0.960), accuracy (0.938), F-1 score (0.933), and area under curve (AUC) (0.950).

CONCLUSION

Our findings demonstrated the efficacy of machine learning models based on radiomics features and voxel-based anatomical locations in predicting the risk of epilepsy in supratentorial LGG. This model provides a highly accurate tool for distinguishing epilepsy-associated LGG in children, supporting precise treatment planning.

TRIAL REGISTRATION

Not applicable.

Prediction of preoperative tumor-related epilepsy using XGBoost radiomics models with 4 MRI sequences

Introduction. Tumor-related epilepsy is a prevalent condition in patients with gliomas. Accurate prediction of epilepsy is crucial for early treatment. This study aimed to evaluate the novel application of the eXtreme Gradient Boost (XGBoost) machine learning (ML) algorithm into a radiomics model predicting preoperative tumor-related epilepsy (PTRE). Its performance was compared with 4 conventional ML algorithms, including the least absolute shrinkage and selection operator (LASSO), elastic net, random forest, and support vector machine.Methods.This study used four magnetic resonance imaging (MRI) images consisting of four sequences (T1-weighted [T1W], T1-weighted contrast [T1WC], T2-weighted [T2W], and T2-weighted fluid-attenuated inversion recovery [T2W FLAIR]) acquired from 74 glioma patients, 30 with PTRE and 44 without PTRE. 394 radiomics features were extracted from the MRI scans usingPyradiomics, alongside 12 clinical features from the medical records. The ML algorithms were mixed and matched to create 20 radiomics models with two stages for: (1) feature selection and (2) prediction of PTRE. Nested cross-validation was used to tune the algorithms and select the stable features.Results.The XGBoost radiomics model demonstrated the second-highest balanced accuracy and F1-score of 0.81 ± 0.01 and 0.80 ± 0.01 respectively. It also achieved the highest recall of 0.81 ± 0.02. It used mostly textural radiomics features from the T1W, T2W and T2W FLAIR sequences to make the predictions.Conclusion.This study demonstrates that XGBoost is a viable alternative to conventional ML algorithms for developing a radiomics model to predict PTRE, as the model produced from XGBoost had among the highest metrics. XGBoost selected features with a higher predictive value than other models. The features selected by XGBoost were more stable, which is a useful property for radiomics analysis. Features selected from multiple MRI sequences were important in the model's decision.

Using partially shared radiomics features to simultaneously identify isocitrate dehydrogenase mutation status and epilepsy in glioma patients from MRI images

Prediction of isocitrate dehydrogenase (IDH) mutation status and epilepsy occurrence are important to glioma patients. Although machine learning models have been constructed for both issues, the correlation between them has not been explored. Our study aimed to exploit this correlation to improve the performance of both of the IDH mutation status identification and epilepsy diagnosis models in patients with glioma II-IV. 399 patients were retrospectively enrolled and divided into a training (n = 279) and an independent test (n = 120) cohort. Multi-center dataset (n = 228) from The Cancer Imaging Archive (TCIA) was used for external test for identification of IDH mutation status. Region of interests comprising the entire tumor and peritumoral edema were automatically segmented using a pre-trained deep learning model. Radiomic features were extracted from T1-weighted, T2-weighted, post-Gadolinium T1 weighted, and T2 fluid-attenuated inversion recovery images. We proposed an iterative approach derived from LASSO to select features shared by two tasks and features specific to each task, before using them to construct the final models. Receiver operating characteristic (ROC) analysis was employed to evaluate the model. The IDH mutation identification model achieved area under the ROC curve (AUC) values of 0.948, 0.946 and 0.860 on the training, internal test, and external test cohorts, respectively. The epilepsy diagnosis model achieved AUCs of 0.924 and 0.880 on the training and internal test cohorts, respectively. The proposed models can identify IDH status and epilepsy with fewer features, thus having better interpretability and lower risk of overfitting. This not only improves its chance of application in clinical settings, but also provides a new scheme to study multiple correlated clinical tasks.

Advances in Neuroimaging and Multiple Post-Processing Techniques for Epileptogenic Zone Detection of Drug-Resistant Epilepsy

Among the approximately 20 million patients with drug-resistant epilepsy (DRE) worldwide, the vast majority can benefit from surgery to minimize seizure reduction and neurological impairment. Precise preoperative localization of epileptogenic zone (EZ) and complete resection of the lesions can influence the postoperative prognosis. However, precise localization of EZ is difficult, and the structural and functional alterations in the brain caused by DRE vary by etiology. Neuroimaging has emerged as an approach to identify the seizure-inducing structural and functional changes in the brain, and magnetic resonance imaging (MRI) and positron emission tomography (PET) have become routine noninvasive imaging tools for preoperative evaluation of DRE in many epilepsy treatment centers. Multimodal neuroimaging offers unique advantages in detecting EZ, especially in improving the detection rate of patients with negative MRI or PET findings. This approach can characterize the brain imaging characteristics of patients with DRE caused by different etiologies, serving as a bridge between clinical and pathological findings and providing a basis for individualized clinical treatment plans. In addition to the integration of multimodal imaging modalities and the development of special scanning sequences and image post-processing techniques for early and precise localization of EZ, the application of deep machine learning for extracting image features and deep learning-based artificial intelligence have gradually improved diagnostic efficiency and accuracy. These improvements can provide clinical assistance for precisely outlining the scope of EZ and indicating the relationship between EZ and functional brain areas, thereby enabling standardized and precise surgery and ensuring good prognosis. However, most existing studies have limitations imposed by factors such as their small sample sizes or hypothesis-based study designs. Therefore, we believe that the application of neuroimaging and post-processing techniques in DRE requires further development and that more efficient and accurate imaging techniques are urgently needed in clinical practice. LEVEL OF EVIDENCE: 5 TECHNICAL EFFICACY: Stage 2.

Post-discharge nutritional management for patients with coronary heart disease and frailty: a qualitative study

BACKGROUND

Frail elderly patients experience physiological function and reserve depletion, leading to imbalances in their internal environment, which increases the risk of coronary heart disease recurrence and malnutrition. However, the majority of these patients, who primarily have a low level of education and lack self-management skills, face difficulties actively dealing with obstacles during the transition period after their discharge from hospitalization. Therefore, it is necessary to understand and discuss in depth the nutrition management experience of discharged elderly patients with coronary heart disease and frailty (ages 65-80 years old) and to analyze the promoting and hindering factors that affect scientific diet behavior during the discharge transition period.

METHODS

Fifteen elderly patients with coronary heart disease and frailty who had been discharged from the hospital for 6 months were interviewed using a semistructured method. The directed content analysis approach to descriptive research was used to extract topics from the interview content.

RESULTS

All participants discussed the problems in health nutrition management experience of discharged. Five topics and ten subtopics were extracted, such as ①Weak perceptions and behaviors towards healthy eating (personal habit solidification, negative attitudes towards nutrition management), ②Lack of objective factors for independently adjusting dietary conditions (reliance on subjective feelings, times of appetite change), ③Personal hindrance factors (memory impairment, deficiencies in self-nutrition management), ④Expected external support (assistance care support, ways to obtain nutritional information), ⑤Lack of continuous nutrition management (interruption of professional guidance, avoidance of medical treatment behavior).

CONCLUSIONS

Nutrition management after discharge places a burden on elderly patients with coronary heart disease and frailty. According to the patients' physical conditions, we should develop a diet support system that is coordinated by individuals, families and society.

Automatic detection of mild cognitive impairment based on deep learning and radiomics of MR imaging

PURPOSE

Early and rapid diagnosis of mild cognitive impairment (MCI) has important clinical value in improving the prognosis of Alzheimer's disease (AD). The hippocampus and parahippocampal gyrus play crucial roles in the occurrence of cognitive function decline. In this study, deep learning and radiomics techniques were used to automatically detect MCI from healthy controls (HCs).

METHOD

This study included 115 MCI patients and 133 normal individuals with 3D-T1 weighted MR structural images from the ADNI database. The identification and segmentation of the hippocampus and parahippocampal gyrus were automatically performed with a VB-net, and radiomics features were extracted. Relief, Minimum Redundancy Maximum Correlation, Recursive Feature Elimination and the minimum absolute shrinkage and selection operator (LASSO) were used to reduce the dimensionality and select the optimal features. Five independent machine learning classifiers including Support Vector Machine (SVM), Random forest (RF), Logistic Regression (LR), Bagging Decision Tree (BDT), and Gaussian Process (GP) were trained on the training set, and validated on the testing set to detect the MCI. The Delong test was used to assess the performance of different models.

RESULT

Our VB-net could automatically identify and segment the bilateral hippocampus and parahippocampal gyrus. After four steps of feature dimensionality reduction, the GP models based on combined features (11 features from the hippocampus, and 4 features from the parahippocampal gyrus) showed the best performance for the MCI and normal control subject discrimination. The AUC of the training set and test set were 0.954 (95% CI: 0.929-0.979) and 0.866 (95% CI: 0.757-0.976), respectively. Decision curve analysis showed that the clinical benefit of the line graph model was high.

CONCLUSION

The GP classifier based on 15 radiomics features of bilateral hippocampal and parahippocampal gyrus could detect MCI from normal controls with high accuracy based on conventional MR images. Our fully automatic model could rapidly process the MRI data and give results in 1 minute, which provided important clinical value in assisted diagnosis.

Development and external validation of a clinical-radiomics nomogram for preoperative prediction of LVSI status in patients with endometrial carcinoma

PURPOSE

To develop and validate a model that incorporates radiomics based on MRI scans and clinical characteristics to predict lymphovascular invasion (LVSI) in endometrial cancer (EC) patients.

METHODS

There were 332 patients with EC enrolled retrospectively in this multicenter study. Radiomics score (Radscore) were computed using the valuable radiomics features. The independent predictors of LVSI were identified by univariate logistic analysis. Multivariate logistic regression was used to develop a clinical-radiomics predictive model. Based on the model, a nomogram was developed and validated internally and externally. The nomogram was evaluated with discrimination, calibration, decision curve analysis (DCA), and clinical impact curves (CIC).

RESULTS

Three predictive models were constructed based on clinicopathological features, radiomic factors and a combination of them, and that the clinic-radiomic model performed best among the three models. Four independent factors comprised the clinical-radiomics model: dynamic contrast enhancement rate of late arterial phase (DCE2), deep myometrium invasion (DMI), lymph node metastasis (LNM), and Radscore. Clinical-radiomics model performance was 0.901 (95% CI 0.84-0.96) in the training cohort, 0.80 (95% CI 0.68-0.92) in the internal validation cohort, and 0.81 (95% CI 0.73-0.9) in the external validation cohort for identifying patients with LVSI, respectively. The model is used to develop a nomogram for clinical use.

CONCLUSIONS

The MRI-based radiomics nomogram could serve as a noninvasive tool to predict LVSI in EC patients.

Influencing Factors and Nomogram for the Development of Epilepsy in Advanced Lung Cancer Patients With Brain Metastases

BACKGROUND

Epilepsy is a prevalent comorbidity in patients with brain metastases (BM) and could result in sudden and accidental damage, as well as increased disease burden due to its rapid onset. Foreseeing the potential for the development of epilepsy may permit timely and efficient measures. This study aimed to analyze the influencing factors of epilepsy in advanced lung cancer (ALC) patients with BM and construct a nomogram model to predict the likelihood of developing epilepsy.

METHODS

Socio-demographic and clinical data of ALC patients with BM were retrospectively collected from the First Affiliated Hospital of Zhejiang University School of Medicine between September 2019 and June 2021. Univariate and multivariate logistic regression analyses were applied to determine the influencing factors for epilepsy in ALC patients with BM. Based on the results of the logistic regression analysis, a nomogram was built to represent the contribution of each influencing factor in predicting the probability of epilepsy development in ALC patients with BM. The Hosmer-Lemeshow test and receiver operating characteristic (ROC) curve were utilized to evaluate the goodness of fit and prediction performance of the model.

RESULTS

The incidence of epilepsy among 138 ALC patients with BM was 29.7%. On the multivariate analysis, having a higher number of supratentorial lesions (odds ratio [OR] = 1.727; p = 0.022), hemorrhagic foci (OR = 4.922; p = .021), and a high-grade of peritumoral edema (OR = 2.524; p < .001) were independent risk factors for developing epilepsy, while undergoing gamma knife radiosurgery (OR = .327; p = .019) was an independent protective factor. The p-value of the Hosmer-Lemeshow test was .535 and the area under the ROC curve (AUC) was .852 (95% CI: .807-.897), suggesting the model had a good fit and exhibited strong predictive accuracy.

CONCLUSION

The nomogram was constructed that can predict the probability of epilepsy development for ALC patients with BM, which is helpful for healthcare professionals to identify high-risk groups early and allows for individualized interventions.

Early acquired resistance to EGFR-TKIs in lung adenocarcinomas before radiographic advanced identified by CT radiomic delta model based on two central studies

Early acquired resistance (EAR) to epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) in lung adenocarcinomas before radiographic advance cannot be perceived by the naked eye. This study aimed to discover and validate a CT radiomic model to precisely identify the EAR. Training cohort (n = 67) and internal test cohort (n = 29) were from the First Affiliated Hospital of Fujian Medical University, and external test cohort (n = 29) was from the Second Affiliated Hospital of Xiamen Medical College. Follow-up CT images at three different times of each patient were collected: (1) baseline images before EGFR-TKIs therapy; (2) first follow-up images after EGFR-TKIs therapy (FFT); (3) EAR images, which were the last follow-up images before radiographic advance. The features extracted from FFT and EAR were used to construct the classic radiomic model. The delta features which were calculated by subtracting the baseline from either FFT or EAR were used to construct the delta radiomic model. The classic radiomic model achieved AUC 0.682 and 0.641 in training and internal test cohorts, respectively. The delta radiomic model achieved AUC 0.730 and 0.704 in training and internal test cohorts, respectively. Over the external test cohort, the delta radiomic model achieved AUC 0.661. The decision curve analysis showed that when threshold of the probability of the EAR to the EGFR-TKIs was between 0.3 and 0.82, the proposed model was more benefit than treating all patients. Based on two central studies, the delta radiomic model derived from the follow-up non-enhanced CT images can help clinicians to identify the EAR to EGFR-TKIs in lung adenocarcinomas before radiographic advance and optimize clinical outcomes.

Multiparametric MRI: From Simultaneous Rapid Acquisition Methods and Analysis Techniques Using Scoring, Machine Learning, Radiomics, and Deep Learning to the Generation of Novel Metrics

ABSTRACT

With the recent advancements in rapid imaging methods, higher numbers of contrasts and quantitative parameters can be acquired in less and less time. Some acquisition models simultaneously obtain multiparametric images and quantitative maps to reduce scan times and avoid potential issues associated with the registration of different images. Multiparametric magnetic resonance imaging (MRI) has the potential to provide complementary information on a target lesion and thus overcome the limitations of individual techniques. In this review, we introduce methods to acquire multiparametric MRI data in a clinically feasible scan time with a particular focus on simultaneous acquisition techniques, and we discuss how multiparametric MRI data can be analyzed as a whole rather than each parameter separately. Such data analysis approaches include clinical scoring systems, machine learning, radiomics, and deep learning. Other techniques combine multiple images to create new quantitative maps associated with meaningful aspects of human biology. They include the magnetic resonance g-ratio, the inner to the outer diameter of a nerve fiber, and the aerobic glycolytic index, which captures the metabolic status of tumor tissues.

MRI-based radiomics models for the early prediction of radiation-induced temporal lobe injury in nasopharyngeal carcinoma

Objective

This study was conducted to develop and validate a radiomics-clinics combined model-based magnetic resonance imaging (MRI) radiomics and clinical features for the early prediction of radiation-induced temporal lobe injury (RTLI) in patients with nasopharyngeal carcinoma (NPC).

Methods

This retrospective study was conducted using data from 130 patients with NPC (80 patients with and 50 patients without RTLI) who received radiotherapy. Cases were assigned randomly to training (n = 91) and testing (n = 39) datasets. Data on 168 medial temporal lobe texture features were extracted from T1WI, T2WI, and T1WI-CE MRI sequences obtained at the end of radiotherapy courses. Clinics, radiomics, and radiomics–clinics combined models (based on selected radiomics signatures and clinical factors) were constructed using machine learning software. Univariate logistic regression analysis was performed to identify independent clinical factors. The area under the ROC curve (AUC) was performed to evaluate the performance of three models. A nomogram, decision curves, and calibration curves were used to assess the performance of the combined model.

Results

Six texture features and three independent clinical factors associated significantly with RTLI were used to build the combined model. The AUCs for the combined and radiomics models were 0.962 [95% confidence interval (CI), 0.9306–0.9939] and 0.904 (95% CI, 0.8431–0.9651), respectively, for the training cohort and 0.947 (95% CI, 0.8841–1.0000) and 0.891 (95% CI, 0.7903–0.9930), respectively, for the testing cohort. All of these values exceeded those for the clinics model (AUC = 0.809 and 0.713 for the training and testing cohorts, respectively). Decision curve analysis showed that the combined model had a good corrective effect.

Conclusion

The radiomics–clinics combined model developed in this study showed good performance for predicting RTLI in patients with NPC.

18F-FDG PET/CT-based radiomics nomogram could predict bone marrow involvement in pediatric neuroblastoma

Objective

To develop and validate an ¹⁸F-fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT)-based radiomics nomogram for non-invasively prediction of bone marrow involvement (BMI) in pediatric neuroblastoma.

Methods

A total of 133 patients with neuroblastoma were retrospectively included and randomized into the training set (n = 93) and test set (n = 40). Radiomics features were extracted from both CT and PET images. The radiomics signature was developed. Independent clinical risk factors were identified using the univariate and multivariate logistic regression analyses to construct the clinical model. The clinical-radiomics model, which integrated the radiomics signature and the independent clinical risk factors, was constructed using multivariate logistic regression analysis and finally presented as a radiomics nomogram. The predictive performance of the clinical-radiomics model was evaluated by receiver operating characteristic curves, calibration curves and decision curve analysis (DCA).

Results

Twenty-five radiomics features were selected to construct the radiomics signature. Age at diagnosis, neuron-specific enolase and vanillylmandelic acid were identified as independent predictors to establish the clinical model. In the training set, the clinical-radiomics model outperformed the radiomics model or clinical model (AUC: 0.924 vs. 0.900, 0.875) in predicting the BMI, which was then confirmed in the test set (AUC: 0.925 vs. 0.893, 0.910). The calibration curve and DCA demonstrated that the radiomics nomogram had a good consistency and clinical utility.

Conclusion

The ¹⁸F-FDG PET/CT-based radiomics nomogram which incorporates radiomics signature and independent clinical risk factors could non-invasively predict BMI in pediatric neuroblastoma.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13244-022-01283-8.