Multiparameter MRI-based radiomics nomogram for preoperative prediction of brain invasion in atypical meningioma:a multicentre study.

OBJECTIVE: To develop a nomogram based on tumor and peritumoral edema (PE) radiomics features extracted from preoperative multiparameter MRI for predicting brain invasion (BI) in atypical meningioma (AM). METHODS: In this retrospective study, according to the 2021 WHO classification criteria, a total of 469 patients with pathologically confirmed AM from three medical centres were enrolled and divided into training (n = 273), internal validation (n = 117) and external validation (n = 79) cohorts. BI was diagnosed based on the histopathological examination. Preoperative contrast-enhanced T1-weighted MR images (T1C) and T2-weighted MR images (T2) for extracting meningioma features and T2-fluid attenuated inversion recovery (FLAIR) sequences for extracting meningioma and PE features were obtained. The multiple logistic regression was applied to develop separate multiparameter radiomics models for comparison. A nomogram was developed by combining radiomics features and clinical risk factors, and the clinical usefulness of the nomogram was verified using decision curve analysis. RESULTS: Among the clinical factors, PE volume and PE/tumor volume ratio are the risk of BI in AM. The combined nomogram based on multiparameter MRI radiomics features of meningioma and PE and clinical indicators achieved the best performance in predicting BI in AM, with area under the curve values of 0.862 (95% CI, 0.819-0.905) in the training cohort, 0.834 (95% CI, 0.780-0.908) in the internal validation cohort and 0.867 (95% CI, 0.785-0.950) in the external validation cohort, respectively. CONCLUSIONS: The nomogram based on tumor and PE radiomics features extracted from preoperative multiparameter MRI and clinical factors can predict the risk of BI in patients with AM.

Concomitant Prediction of the Ki67 and PIT-1 Expression in Pituitary Adenoma Using Different Radiomics Models.

OBJECTIVES: To preoperatively predict the high expression of Ki67 and positive pituitary transcription factor 1 (PIT-1) simultaneously in pituitary adenoma (PA) using three different radiomics models. METHODS: A total of 247 patients with PA (training set: n = 198; test set: n = 49) were included in this retrospective study. The imaging features were extracted from preoperative contrast-enhanced T1WI (T1CE), T1-weighted imaging (T1WI), and T2-weighted imaging (T2WI). Feature selection was performed using Spearman's rank correlation coefficient and least absolute shrinkage and selection operator (LASSO). The classic machine learning (CML), deep learning (DL), and deep learning radiomics (DLR) models were constructed using logistic regression (LR), support vector machine (SVM), and multi-layer perceptron (MLP) algorithms. The area under the receiver operating characteristic (ROC) curve (AUC), sensitivity, specificity, accuracy, negative predictive value (NPV) and positive predictive value (PPV) were calculated for the training and test sets. In addition, combined with clinical characteristics, the best CML and the best DL models (SVM classifier), the DL radiomics nomogram (DLRN) was constructed to aid clinical decision-making. RESULTS: Seven CML features, 96 DL features, and 107 DLR features were selected to construct CML, DL and DLR models. Compared to CML and DL model, the DLR model had the best performance. The AUC, sensitivity, specificity, accuracy, NPV and PPV were 0.827, 0.792, 0.800, 0.796, 0.800 and 0.792 in the test set, respectively. CONCLUSIONS: Compared with CML and DL models, the DLR model shows the best performance in predicting the Ki67 and PIT-1 expression in PAs simultaneously.

Imaging manifestations of papillary glioneuronal tumors.

To analyse the imaging findings of papillary glioneuronal tumors (PGNTs), in order to improve the accuracy of preoperative diagnosis of this tumor. The clinical and imaging manifestations of 36 cases of PGNT confirmed by pathology were analyzed retrospectively. A total of 17 males and 19 females, averaging 22.47 (± 11.23) years. Initial symptoms included epilepsy in ten, headache in seven, and others in 19 cases. 97.2% (35/36) of the lesions were located in the supratentorial area, and 80.5% (29/36) in the intraventricular or deep white matter adjacent to the lateral ventricles. Twenty-four of the lesions (66.7%) were mixed cystic and solid, four (11.1%) were cystic with mural nodules, four (11.1%) were cystic, and four (11.1%) were solid. Four cases of PGNT of cystic imaging showed a "T2-FLAIR mismatch" sign. 69.4% (25/36) had septations. Nine lesions (25%) were accompanied by edema, and 9 (25%) of the mixed cystic and solid lesions were accompanied by hemorrhage. Among the 18 patients who underwent computed tomography (CT) or susceptibility-weighted imaging (SWI), nine had lesions with calcification. PGNTs mostly manifest as cystic mass with mural nodules or mixed cystic and solid mass in the white matter around the supratentorial ventricle, and the cystic part of the lesion is mostly accompanied by septations. Pure cystic lesions may exhibit the sign of "T2-FLAIR mismatch". PGNT is rarely accompanied by edema but sometimes by calcification and hemorrhage. Patients often present with seizures, headaches, and mass effect symptoms.

Radiomics for predicting MGMT status in cerebral glioblastoma: comparison of different MRI sequences.

Using radiomics to predict O6-methylguanine-DNA methyltransferase promoter methylation status in patients with newly diagnosed glioblastoma and compare the performances of different MRI sequences. Preoperative MRI scans from 215 patients were included in this retrospective study. After image preprocessing and feature extraction, two kinds of machine-learning models were established and compared for their performances. One kind was established using all MRI sequences (T1-weighted image, T2-weighted image, contrast enhancement, fluid-attenuated inversion recovery, DWI_b_high, DWI_b_low and apparent diffusion coefficient), and the other kind was based on single MRI sequence as listed above. For the machine-learning model based on all sequences, a total of seven radiomic features were selected with the Maximum Relevance and Minimum Redundancy algorithm. The predictive accuracy was 0.993 and 0.750 in the training and validation sets, respectively, and the area under curves were 1.000 and 0.754 in the two sets, respectively. For the machine-learning model based on single sequence, the numbers of selected features were 8, 10, 10, 13, 9, 7 and 6 for T1-weighted image, T2-weighted image, contrast enhancement, fluid-attenuated inversion recovery, DWI_b_high, DWI_b_low and apparent diffusion coefficient, respectively, with predictive accuracies of 0.797-1.000 and 0.583-0.694 in the training and validation sets, respectively, and the area under curves of 0.874-1.000 and 0.538-0.697 in the two sets, respectively. Specifically, T1-weighted image-based model performed best, while contrast enhancement-based model performed worst in the independent validation set. The machine-learning models based on seven different single MRI sequences performed differently in predicting O6-methylguanine-DNA methyltransferase status in glioblastoma, while the machine-learning model based on the combination of all sequences performed best.

MRI-Based Machine Learning Fusion Models to Distinguish Encephalitis and Gliomas.

This paper aims to compare the performance of the classical machine learning (CML) model and the deep learning (DL) model, and to assess the effectiveness of utilizing fusion radiomics from both CML and DL in distinguishing encephalitis from glioma in atypical cases. We analysed the axial FLAIR images of preoperative MRI in 116 patients pathologically confirmed as gliomas and clinically diagnosed with encephalitis. The 3 CML models (logistic regression (LR), support vector machine (SVM) and multi-layer perceptron (MLP)), 3 DL models (DenseNet 121, ResNet 50 and ResNet 18) and a deep learning radiomic (DLR) model were established, respectively. The area under the receiver operating curve (AUC) and sensitivity, specificity, accuracy, negative predictive value (NPV) and positive predictive value (PPV) were calculated for the training and validation sets. In addition, a deep learning radiomic nomogram (DLRN) and a web calculator were designed as a tool to aid clinical decision-making. The best DL model (ResNet50) consistently outperformed the best CML model (LR). The DLR model had the best predictive performance, with AUC, sensitivity, specificity, accuracy, NPV and PPV of 0.879, 0.929, 0.800, 0.875, 0.867 and 0.889 in the validation sets, respectively. Calibration curve of DLR model shows good agreement between prediction and observation, and the decision curve analysis (DCA) indicated that the DLR model had higher overall net benefit than the other two models (ResNet50 and LR). Meanwhile, the DLRN and web calculator can provide dynamic assessments. Machine learning (ML) models have the potential to non-invasively differentiate between encephalitis and glioma in atypical cases. Furthermore, combining DL and CML techniques could enhance the performance of the ML models.

The diagnostic efficiency of integration of 2HG MRS and IVIM versus individual parameters for predicting IDH mutation status in gliomas in clinical scenarios: A retrospective study.

PURPOSE: Currently, there remains a scarcity of established preoperative tests to accurately predict the isocitrate dehydrogenase (IDH) mutation status in clinical scenarios, with limited research has explored the potential synergistic diagnostic performance among metabolite, perfusion, and diffusion parameters. To address this issue, we aimed to develop an imaging protocol that integrated 2-hydroxyglutarate (2HG) magnetic resonance spectroscopy (MRS) and intravoxel incoherent motion (IVIM) by comprehensively assessing metabolic, cellular, and angiogenic changes caused by IDH mutations, and explored the diagnostic efficiency of this imaging protocol for predicting IDH mutation status in clinical scenarios. METHODS: Patients who met the inclusion criteria were categorized into two groups: IDH-wild type (IDH-WT) group and IDH-mutant (IDH-MT) group. Subsequently, we quantified the 2HG concentration, the relative apparent diffusion coefficient (rADC), the relative true diffusion coefficient value (rD), the relative pseudo-diffusion coefficient (rD*) and the relative perfusion fraction value (rf). Intergroup differences were estimated using t-test and Mann-Whitney U test. Finally, we performed receiver operating characteristic (ROC) curve and DeLong's test to evaluate and compare the diagnostic performance of individual parameters and their combinations. RESULTS: 64 patients (female, 21; male, 43; age, 47.0 ± 13.7 years) were enrolled. Compared with IDH-WT gliomas, IDH-MT gliomas had higher 2HG concentration, rADC and rD (P < 0.001), and lower rD* (P = 0.013). The ROC curve demonstrated that 2HG + rD + rD* exhibited the highest areas under curve (AUC) value (0.967, 95%CI 0.889-0.996) for discriminating IDH mutation status. Compared with each individual parameter, the predictive efficiency of 2HG + rADC + rD* and 2HG + rD + rD* shows a statistically significant enhancement (DeLong's test: P < 0.05). CONCLUSIONS: The integration of 2HG MRS and IVIM significantly improves the diagnostic efficiency for predicting IDH mutation status in clinical scenarios.

Malignant transformation of desmoplastic infantile tumors in a non-infant: a case report.

Desmoplastic infantile tumors (DITs) are rare benign intracranial tumors in infants with benign biological behavior and rare malignant transformation characteristics. We present a DIT case that underwent malignant transformation and metastasis. A 6-year-old girl presented with DITs and underwent surgical resection. 12 years later, the tumor recurred and underwent surgical resection again. The pathology report confirmed the lesion to be a glioblastoma multiforme. She received adjuvant treatment. A year after the surgical operation of the lesions, she had intraspinal metastasis and underwent surgical resection again. Multiple spinal cord metastases were subsequently identified in the patient. The patient's condition exhibited severe deterioration during the follow-up period. This case report focuses on the occurrence of DITs and their potential malignant transformation, as assessed through computed tomography and magnetic resonance imaging.

The clinical, radiological, and surgical characteristics of anterior perforated substance glioma: a retrospective study.

BACKGROUND: To explore the clinical, radiological, and surgical characteristics of anterior perforated substance (APS) gliomas. METHODS: Twenty patients with APS glioma who were treated with surgery between March 2019 and January 2022 from Tiantan hospital were retrospectively reviewed. The clinical, histological and radiological data were collected. RESULTS: Twenty patients, including 7 males (55%) and 13 females (45%), with a mean age at diagnosis of 37.9 years (range, 28-53 years) underwent operative intervention for APS. Headaches and dizziness were the most common preoperative symptoms in the majority patients (14, 70%). Based on radiological features of MRI, the APS was classified into two subtypes, type A and type B. Seven patients (40%) in type A indicated a clear tumor margin, while 13 patients (60%) in type B showed an ill-defined margin. The surgical approach including frontal, temporal, and coronal frontal incisions for type A and type B tumors, respectively. Three patients in type A received total resection, while one patient in type B were total resected. Pathologically, 12 cases (60%, 12/20) were diagnosed as astrocytoma and 8 cases (20%, 8/20) were oligodendroglioma. Meanwhile, 17 cases (85%, 17/20) had MGMT promotor methylation. CONCLUSION: In this study, we performed the first systematic research of patients with APS glioma. Most of patients with APS presented headaches and dizziness symptoms. The APS glioma was further divided into two major radiological subtypes with relevant different surgical approaches. The APS glioma in type A were more likely to receive total resection.

Radiogenomic Analysis of Vascular Endothelial Growth Factor in Patients With Glioblastoma.

OBJECTIVES: This article aims to predict the presence of vascular endothelial growth factor (VEGF) expression and to predict the expression level of VEGF by machine learning based on preoperative magnetic resonance imaging (MRI) of glioblastoma (GBM). METHODS: We analyzed the axial T2-weighted images (T2WI) and T1-weighted contrast-enhancement images of preoperative MRI in 217 patients with pathologically diagnosed GBM. Patients were divided into negative and positive VEGF groups, with the latter group further subdivided into low and high expression. The machine learning models were established with the maximum relevance and minimum redundancy algorithm and the extreme gradient boosting classifier. The area under the receiver operating curve (AUC) and accuracy were calculated for the training and validation sets. RESULTS: Positive VEGF in GBM was 63.1% (137/217), with a high expression ratio of 53.3% (73/137). To predict the positive and negative VEGF expression, 7 radiomic features were selected, with 3 features from T1CE and 4 from T2WI. The accuracy and AUC were 0.83 and 0.81, respectively, in the training set and were 0.73 and 0.74, respectively, in the validation set. To predict high and low levels, 7 radiomic features were selected, with 2 from T1CE, 1 from T2WI, and 4 from the data combinations of T1CE and T2WI. The accuracy and AUC were 0.88 and 0.88, respectively, in the training set and were 0.72 and 0.72, respectively, in the validation set. CONCLUSION: The VEGF expression status in GBM can be predicted using a machine learning model. Radiomic features resulting from data combinations of different MRI sequences could be helpful.

The sinuous, wave-like intratumoral-wall sign is a sensitive and specific radiological biomarker for oligodendrogliomas.

OBJECTIVES: The purpose of this study was to investigate the clinical utility of the sinuous, wave-like intratumoral-wall (SWITW) sign on T2WI in diagnosing isocitrate dehydrogenase (IDH) mutant and 1p/19q codeleted (IDHmut-Codel) oligodendrogliomas, for which a relatively conservative resection strategy might be sufficient due to a better response to chemoradiotherapy and favorable prognosis. METHODS: Imaging data from consecutive adult patients with diffuse lower-grade gliomas (LGGs, histological grades 2-3) in Beijing Tiantan Hospital (December 1, 2013, to October 31, 2021, BTH set, n = 711) and the Cancer Imaging Archive (TCIA) LGGs set (n = 117) were used to develop and validate our findings. Two independent observers assessed the SWITW sign and some well-reported discriminative radiological features to establish a practical diagnostic strategy. RESULTS: The SWITW sign showed satisfying sensitivity (0.684 and 0.722 for BTH and TCIA sets) and specificity (0.938 and 0.914 for BTH and TCIA sets) in defining IDHmut-Codels, and the interobserver agreement was substantial (κ 0.718 and 0.756 for BTH and TCIA sets). Compared to calcification, the SWITW sign improved the sensitivity by 0.28 (0.404 to 0.684) in the BTH set, and 81.0% (277/342) of IDHmut-Codel cases demonstrated SWITW and/ or calcification positivity. Combining the SWITW sign, calcification, low ADC values, and other discriminative features, we established a concise and reliable diagnostic protocol for IDHmut-Codels. CONCLUSIONS: The SWITW sign was a sensitive and specific imaging biomarker for IDHmut-Codels. The integrated protocol provided an explicable, efficient, and reproducible method for precise preoperative diagnosis, which was essential to guide individualized surgical plan-making. KEY POINTS: • The SWITW sign was a sensitive and specific imaging biomarker for IDHmut-Codel oligodendrogliomas. • The SWITW sign was more sensitive than calcification and an integrated strategy could improve diagnostic sensitivity for IDHmut-Codel oligodendrogliomas. • Combining SWITW, calcification, low ADC values, and other discriminative features could make a precise preoperative diagnosis for IDHmut-Codel oligodendrogliomas.

A multi-task two-path deep learning system for predicting the invasiveness of craniopharyngioma.

BACKGROUND AND OBJECTIVE: Craniopharyngioma is a kind of benign brain tumor in histography. However, it might be clinically aggressive and have severe manifestations, such as increased intracranial pressure, hypothalamic-pituitary dysfunction, and visual impairment. It is considered challenging for radiologists to predict the invasiveness of craniopharyngioma through MRI images. Therefore, developing a non-invasive method that can predict the invasiveness and boundary of CP as a reference before surgery is of clinical value for making more appropriate and individualized treatment decisions and reducing the occurrence of inappropriate surgical plan choices. METHODS: The MT-Brain system has consisted of two pathways, a sub-path based on 2D CNN for capturing the features from each slice of MRI images, and a 3D sub-network for capturing additional context information between slices. By introducing the two-path architecture, our system can make full use of the fusion of the above 2D and 3D features for classification. Furthermore, position encoding and mask-guided attention also have been introduced to improve the segmentation and diagnosis performance. To verify the performance of the MT-Brain system, we have enrolled 1032 patients with craniopharyngioma (302 invasion and 730 non-invasion patients), segmented the tumors on postcontrast coronal T1WI and randomized them into a training dataset and a testing dataset at a ratio of 8:2. RESULTS: The MT-Brain system achieved a remarkable performance in diagnosing the invasiveness of craniopharyngioma with the AUC of 83.84%, the accuracy of 77.94%, the sensitivity of 70.97%, and the specificity of 80.99%. In the lesion segmentation task, the predicted boundaries of lesions were similar to those labeled by radiologists with the dice of 66.36%. In addition, some explorations also have been made on the interpretability of deep learning models, illustrating the reliability of the model. CONCLUSIONS: To the best of our knowledge, this study is the first to develop an integrated deep learning model to predict the invasiveness of craniopharyngioma preoperatively and locate the lesion boundary synchronously on MRI. The excellent performances indicate that the MT-Brain system has great potential in real-world clinical applications.

Combining hyperintense FLAIR rim and radiological features in identifying IDH mutant 1p/19q non-codeleted lower-grade glioma.

OBJECTIVES: Even very small residual tumors of IDH mutant 1p/19q non-codeleted (IDHmut-Noncodel) astrocytoma could have a significantly negative impact on survival; thus, accurate preoperative diagnosis is of utmost importance to guide aggressive tumor resection strategy for this subtype. This study aimed to diagnose IDHmut-Noncodel from IDH mutant 1p/19q codeleted (IDHmut-Codel) and IDH wild-type gliomas by preoperative MRI and CT to guide surgical plan-making. METHODS: Consecutive adult patients diagnosed with diffuse lower-grade glioma (LGG, histological grade 2-3) from December 1, 2013 to December 31, 2020, were retrospectively included in this study. Clinical and radiological features were recorded and analyzed. Patients were divided into cohort A and cohort B for training and validation based on the operation date (2:1). RESULTS: A total of 585 patients were included in this study (cohort A, 390; cohort B, 195). The hyperintense FLAIR rim with hypointense core (hyperFLAIRrim) was a more sensitive sign than T2-FLAIR mismatch (T2FM) in defining IDHmut-Noncodel astrocytoma (sensitivity in cohort A: 0.713, 0.539, respectively; in cohort B: 0.713, 0.489, respectively) without compromised specificity (all 1.00). The hyperFLAIRrim, higher rADC, homogenous pattern on T2WI, non-calcification, and younger age were the most important factors associated with IDHmut-Noncodel astrocytoma. Combining these factors, the random forest model showed the best predictive ability. CONCLUSION: The hyperFLAIRrim sign was a specific and more sensitive sign in diagnosing IDHmut-Noncodel astrocytoma. Combining hyperFLAIRrim, higher rADC, homogenous pattern, non-calcification, and younger age could precisely predict glioma subtype for subsequent surgical plan-making. KEY POINTS: • A single hyperintense FLAIR rim (hyperFLAIRrim) sign with a hypointense core, regardless of T2 appearance, was more sensitive than T2FM in diagnosing IDHmut-Noncodel astrocytoma with high specificity. • The higher rADC value, homogenous pattern on T2WI, non-calcification, and younger age have a close relationship with IDHmut-Noncodel astrocytoma. • Neurosurgeons should perform a more aggressive resection strategy to prolong survival for radiologically indicated IDHmut-Noncodel astrocytoma. Our study provided a usable, practicable, and reliable protocol for neurosurgeons to make an individualized surgical strategy.

Corrigendum: Machine Learning-Based Analysis of Magnetic Resonance Radiomics for the Classification of Gliosarcoma and Glioblastoma.

[This corrects the article DOI: 10.3389/fonc.2021.699789.].

Deep Learning With Data Enhancement for the Differentiation of Solitary and Multiple Cerebral Glioblastoma, Lymphoma, and Tumefactive Demyelinating Lesion.

OBJECTIVES: To explore the MRI-based differential diagnosis of deep learning with data enhancement for cerebral glioblastoma (GBM), primary central nervous system lymphoma (PCNSL), and tumefactive demyelinating lesion (TDL). MATERIALS AND METHODS: This retrospective study analyzed the MRI data of 261 patients with pathologically diagnosed solitary and multiple cerebral GBM (n = 97), PCNSL (n = 92), and TDL (n = 72). The 3D segmentation model was trained to capture the lesion. Different enhancement data were generated by changing the pixel ratio of the lesion and non-lesion areas. The 3D classification network was trained by using the enhancement data. The accuracy, sensitivity, specificity, and area under the curve (AUC) were used to assess the value of different enhancement data on the discrimination performance. These results were then compared with the neuroradiologists' diagnoses. RESULTS: The diagnostic performance fluctuated with the ratio of lesion to non-lesion area changed. The diagnostic performance was best when the ratio was 1.5. The AUCs of GBM, PCNSL, and TDL were 1.00 (95% confidence interval [CI]: 1.000-1.000), 0.96 (95% CI: 0.923-1.000), and 0.954 (95% CI: 0.904-1.000), respectively. CONCLUSIONS: Deep learning with data enhancement is useful for the accurate identification of GBM, PCNSL, and TDL, and its diagnostic performance is better than that of the neuroradiologists.

Machine Learning-Based Analysis of Magnetic Resonance Radiomics for the Classification of Gliosarcoma and Glioblastoma.

OBJECTIVE: To identify optimal machine-learning methods for the radiomics-based differentiation of gliosarcoma (GSM) from glioblastoma (GBM). MATERIALS AND METHODS: This retrospective study analyzed cerebral magnetic resonance imaging (MRI) data of 83 patients with pathologically diagnosed GSM (58 men, 25 women; mean age, 50.5 ± 12.9 years; range, 16-77 years) and 100 patients with GBM (58 men, 42 women; mean age, 53.4 ± 14.1 years; range, 12-77 years) and divided them into a training and validation set randomly. Radiomics features were extracted from the tumor mass and peritumoral edema. Three feature selection and classification methods were evaluated in terms of their performance in distinguishing GSM and GBM: the least absolute shrinkage and selection operator (LASSO), Relief, and Random Forest (RF); and adaboost classifier (Ada), support vector machine (SVM), and RF; respectively. The area under the receiver operating characteristic curve (AUC) and accuracy (ACC) of each method were analyzed. RESULTS: Based on tumor mass features, the selection method LASSO + classifier SVM was found to feature the highest AUC (0.85) and ACC (0.77) in the validation set, followed by Relief + RF (AUC = 0.84, ACC = 0.72) and LASSO + RF (AUC = 0.82, ACC = 0.75). Based on peritumoral edema features, Relief + SVM was found to have the highest AUC (0.78) and ACC (0.73) in the validation set. Regardless of the method, tumor mass features significantly outperformed peritumoral edema features in the differentiation of GSM from GBM (P < 0.05). Furthermore, the sensitivity, specificity, and accuracy of the best radiomics model were superior to those obtained by the neuroradiologists. CONCLUSION: Our radiomics study identified the selection method LASSO combined with the classifier SVM as the optimal method for differentiating GSM from GBM based on tumor mass features.

Differentiation Between Glioblastoma Multiforme and Metastasis From the Lungs and Other Sites Using Combined Clinical/Routine MRI Radiomics.

BACKGROUND: Differentiation between cerebral glioblastoma multiforme (GBM) and solitary brain metastasis (MET) is important. The existing radiomic differentiation method ignores the clinical and routine magnetic resonance imaging (MRI) features. PURPOSE: To differentiate between GBM and MET and between METs from the lungs (MET-lung) and other sites (MET-other) through clinical and routine MRI, and radiomics analyses. METHODS AND MATERIALS: A total of 350 patients were collected from two institutions, including 182 patients with GBM and 168 patients with MET, which were all proven by pathology. The ROI of the tumor was obtained on axial postcontrast MRI which was performed before operation. Seven radiomic feature selection methods and four classification algorithms constituted 28 classifiers in two classification strategies, with the best classifier serving as the final radiomics model. The clinical and combination models were constructed using the nomograms developed. The performance of the nomograms was evaluated in terms of calibration, discrimination, and clinical usefulness. Student's t-test or the chi-square test was used to assess the differences in the clinical and radiological characteristics between the training and internal validation cohorts. Receiver operating characteristic curve analysis was performed to assess the performance of developed models with the area under the curve (AUC). RESULTS: The classifier fisher_decision tree (fisher_DT) showed the best performance (AUC: 0.696, 95% CI:0.608-0.783) for distinguishing between GBM and MET in internal validation cohorts; the classifier reliefF_random forest (reliefF_RF) showed the best performance (AUC: 0.759, 95% CI: 0.613-0.904) for distinguishing between MET-lung and MET-other in internal validation cohorts. The combination models incorporating the radiomics signature and clinical-radiological characteristics were superior to the clinical-radiological models in the two classification strategies (AUC: 0.764 for differentiation between GBM in internal validation cohorts and MET and 0.759 or differentiation between MET-lung and MET-other in internal validation cohorts). The nomograms showed satisfactory performance and calibration and were considered clinically useful, as revealed in the decision curve analysis. DATA CONCLUSION: The combination of radiomic and non-radiomic features is helpful for the differentiation among GBM, MET-lung, and MET-other.

A novel extended Kalman filter with support vector machine based method for the automatic diagnosis and segmentation of brain tumors.

BACKGROUND: Brain tumors are life-threatening, and their early detection is crucial for improving survival rates. Conventionally, brain tumors are detected by radiologists based on their clinical experience. However, this process is inefficient. This paper proposes a machine learning-based method to 1) determine the presence of a tumor, 2) automatically segment the tumor, and 3) classify it as benign or malignant. METHODS: We implemented an Extended Kalman Filter with Support Vector Machine (EKF-SVM), an image analysis platform based on an SVM for automated brain tumor detection. A development dataset of 120 patients which supported by Tiantan Hospital was used for algorithm training. Our machine learning algorithm has 5 components as follows. Firstly, image standardization is applied to all the images. This is followed by noise removal with a non-local means filter, and contrast enhancement with improved dynamic histogram equalization. Secondly, a gray-level co-occurrence matrix is utilized for feature extraction to get the image features. Thirdly, the extracted features are fed into a SVM for classify the MRI initially, and an EKF is used to classify brain tumors in the brain MRIs. Fourthly, cross-validation is used to verify the accuracy of the classifier. Finally, an automatic segmentation method based on the combination of k-means clustering and region growth is used for detecting brain tumors. RESULTS: With regard to the diagnostic performance, the EKF-SVM had a 96.05% accuracy for automatically classifying brain tumors. Segmentation based on k-means clustering was capable of identifying the tumor boundaries and extracting the whole tumor. CONCLUSION: The proposed EKF-SVM based method has better classification performance for positive brain tumor images, which was mainly due to the dearth of negative examples in our dataset. Therefore, future work should obtain more negative examples and investigate the performance of deep learning algorithms such as the convolutional neural networks for automatic diagnosis and segmentation of brain tumors.

Non-Invasive Radiomics Approach Predict Invasiveness of Adamantinomatous Craniopharyngioma Before Surgery.

PURPOSE: Craniopharyngiomas (CPs) are benign tumors, complete tumor resection is considered to be the optimal treatment. However, although histologically benign, the local invasiveness of CPs commonly contributes to incomplete resection and a poor prognosis. At present, some advocate less aggressive surgery combined with radiotherapy as a more reasonable and effective means of protecting hypothalamus function and preventing recurrence in patients with tight tumor adhesion to the hypothalamus. Hence, if a method can be developed to predict the invasiveness of CP preoperatively, it will help in the development of a more personalized surgical strategy. The aim of the study was to report a radiomics-clinical nomogram for the individualized preoperative prediction of the invasiveness of adamantinomatous CP (ACPs) before surgery. METHODS: In total, 1,874 radiomics features were extracted from whole tumors on contrast-enhanced T1-weighted images. A support vector machine trained a predictive model that was validated using receiver operating characteristic (ROC) analysis on an independent test set. Moreover, a nomogram was constructed incorporating clinical characteristics and the radiomics signature for individual prediction. RESULTS: Eleven features associated with the invasiveness of ACPs were selected by using the least absolute shrinkage and selection operator (LASSO) method. These features yielded area under the curve (AUC) values of 79.09 and 73.5% for the training and test sets, respectively. The nomogram incorporating peritumoral edema and the radiomics signature yielded good calibration in the training and test sets with the AUCs of 84.79 and 76.48%, respectively. CONCLUSION: The developed model yields good performance, indicating that the invasiveness of APCs can be predicted using noninvasive radiological data. This reliable, noninvasive tool can help clinical decision making and improve patient prognosis.

Abnormal Interactions of the Salience Network, Central Executive Network, and Default-Mode Network in Patients With Different Cognitive Impairment Loads Caused by Leukoaraiosis.

Leukoaraiosis (LA) is associated with cognitive impairment in the older people which can be demonstrated in functional connectivity (FC) based on resting-state functional magnetic resonance imaging (rs-fMRI). This study is to explore the FC changes in LA patients with different cognitive status by three network models. Fifty-three patients with LA were divided into three groups: the normal cognition (LA-NC; n = 14, six males), mild cognitive impairment (LA-MCI; n = 27, 13 males), and vascular dementia (LA-VD; n = 12, six males), according to the Mini Mental State Exam (MMSE) and Clinical Dementia Rating (CDR). The three groups and 30 matched healthy controls (HCs; 11 males) underwent rs-fMRI. The data of rs-fMRI were analyzed by independent components analysis (ICA) and region of interest (ROI) analysis by the REST toolbox. Then the FC was respectively analyzed by the default-mode network (DMN), salience networks (SNs) and the central executive network (CEN) with their results compared among the different groups. For inter-brain network analysis, there were negative FC between the SN and DMN in LA groups, and the FC decreased when compared with HC group. While there were enhanced inter-brain network FC between the SN and CEN as well as within the SN. The FC in patients with LA can be detected by different network models of rs-fMRI. The multi-model analysis is helpful for the further understanding of the cognitive changes in those patients.

A deep learning radiomics model for preoperative grading in meningioma.

OBJECTIVES: To noninvasively differentiate meningioma grades by deep learning radiomics (DLR) model based on routine post-contrast MRI. METHODS: We enrolled 181 patients with histopathologic diagnosis of meningioma who received post-contrast MRI preoperative examinations from 2 hospitals (99 in the primary cohort and 82 in the validation cohort). All the tumors were segmented based on post-contrast axial T1 weighted images (T1WI), from which 2048 deep learning features were extracted by the convolutional neural network. The random forest algorithm was used to select features with importance values over 0.001, upon which a deep learning signature was built by a linear discriminant analysis classifier. The performance of our DLR model was assessed by discrimination and calibration in the independent validation cohort. For comparison, a radiomic model based on hand-crafted features and a fusion model were built. RESULTS: The DLR signature comprised 39 deep learning features and showed good discrimination performance in both the primary and validation cohorts. The area under curve (AUC), sensitivity, and specificity for predicting meningioma grades were 0.811(95% CI, 0.635-0.986), 0.769, and 0.898 respectively in the validation cohort. DLR performance was superior over the hand-crafted features. Calibration curves of DLR model showed good agreements between the prediction probability and the observed outcome of high-grade meningioma. CONCLUSIONS: Using routine MRI data, we developed a DLR model with good performance for noninvasively individualized prediction of meningioma grades, which achieved a quantization capability superior over the hand-crafted features. This model has potential to guide and facilitate the clinical decision-making of whether to observe or to treat patients by providing prognostic information.