MR imaging based fractal analysis for differentiating primary CNS lymphoma and glioblastoma

Non-invasive differential diagnosis of teratomas from other intracranial germ cell tumours using MRI-based fractal and radiomic analyses

OBJECTIVES

The histologic subtype of intracranial germ cell tumours (IGCTs) is an important factor in deciding the treatment strategy, especially for teratomas. In this study, we aimed to non-invasively diagnose teratomas based on fractal and radiomic features.

MATERIALS AND METHODS

This retrospective study included 330 IGCT patients, including a discovery set (n = 296) and an independent validation set (n = 34). Fractal and radiomic features were extracted from T1-weighted, T2-weighted, and post-contrast T1-weighted images. Five classifiers, including logistic regression, random forests, support vector machines, K-nearest neighbours, and XGBoost, were compared for our task. Based on the optimal classifier, we compared the performance of clinical, fractal, and radiomic models and the model combining these features in predicting teratomas.

RESULTS

Among the diagnostic models, the fractal and radiomic models performed better than the clinical model. The final model that combined all the features showed the best performance, with an area under the curve, precision, sensitivity, and specificity of 0.946 [95% confidence interval (CI): 0.882-0.994], 95.65% (95% CI: 88.64-100%), 88.00% (95% CI: 77.78-96.36%), and 91.67% (95% CI: 78.26-100%), respectively, in the test set of the discovery set, and 0.944 (95% CI: 0.855-1.000), 85.71% (95% CI: 68.18-100%), 94.74% (95% CI: 83.33-100%), and 80.00% (95% CI: 58.33-100%), respectively, in the independent validation set. SHapley Additive exPlanations indicated that two fractal features, two radiomic features, and age were the top five features highly associated with the presence of teratomas.

CONCLUSION

The predictive model including image and clinical features could help guide treatment strategies for IGCTs.

CLINICAL RELEVANCE STATEMENT

Our machine learning model including image and clinical features can non-invasively predict teratoma components, which could help guide treatment strategies for intracranial germ cell tumours (IGCT).

KEY POINTS

• Fractals and radiomics can quantitatively evaluate imaging characteristics of intracranial germ cell tumours. • Model combing imaging and clinical features had the best predictive performance. • The diagnostic model could guide treatment strategies for intracranial germ cell tumours.

Do Habitat MRI and Fractal Analysis Help Distinguish Triple-Negative Breast Cancer From Non-Triple-Negative Breast Carcinoma

Purpose: To determine whether multiparametric MRI-based spatial habitats and fractal analysis can help distinguish triple-negative breast cancer (TNBC) from non-TNBC. Method: Multiparametric DWI and DCE-MRI at 3T were obtained from 142 biopsy- and surgery-proven breast cancer with 148 breast lesions (TNBC = 26 and non-TNBC = 122). The contrast-enhancing lesions were divided into 3 spatial habitats based on perfusion and diffusion patterns using K-means clustering. The fractal dimension (FD) of the tumour subregions was calculated. The accuracy of the habitat segmentation was measured using the Dice index. Inter- and intra-reader reliability were evaluated with the intraclass correlation coefficient (ICC). The ability to predict TNBC status was assessed using the receiver operating characteristic curve. Results: The Dice index for the whole tumour was 0.81 for inter-reader and 0.88 for intra-reader reliability. The inter- and intra-reader reliability were excellent for all 3 tumour habitats and fractal features (ICC > 0.9). TNBC had a lower hypervascular cellular habitat and higher FD 1 compared to non-TNBC (all P < .001). Multivariate analysis confirmed that hypervascular cellular habitat (OR = 0.88) and FD 1 (OR = 1.35) were independently associated with TNBC (all P < .001) after adjusting for rim enhancement, axillary lymph nodes status, and histological grade. The diagnostic model combining hypervascular cellular habitat and FD 1 showed excellent discriminatory ability for TNBC, with an AUC of 0.951 and an accuracy of 91.9%. Conclusions: The fraction of hypervascular cellular habitat and its FD may serve as useful imaging biomarkers for predicting TNBC status.

Balanced transformer: efficient classification of glioblastoma and primary central nervous system lymphoma

Objective.Primary central nervous system lymphoma (PCNSL) and glioblastoma (GBM) are malignant primary brain tumors with different biological characteristics. Great differences exist between the treatment strategies of PCNSL and GBM. Thus, accurately distinguishing between PCNSL and GBM before surgery is very important for guiding neurosurgery. At present, the spinal fluid of patients is commonly extracted to find tumor markers for diagnosis. However, this method not only causes secondary injury to patients, but also easily delays treatment. Although diagnosis using radiology images is non-invasive, the morphological features and texture features of the two in magnetic resonance imaging (MRI) are quite similar, making distinction with human eyes and image diagnosis very difficult. In order to solve the problem of insufficient number of samples and sample imbalance, we used data augmentation and balanced sample sampling methods. Conventional Transformer networks use patch segmentation operations to divide images into small patches, but the lack of communication between patches leads to unbalanced data layers.Approach.To address this problem, we propose a balanced patch embedding approach that extracts high-level semantic information by reducing the feature dimensionality and maintaining the geometric variation invariance of the features. This approach balances the interactions between the information and improves the representativeness of the data. To further address the imbalance problem, the balanced patch partition method is proposed to increase the receptive field by sampling the four corners of the sliding window and introducing a linear encoding component without increasing the computational effort, and designed a new balanced loss function.Main results.Benefiting from the overall balance design, we conducted an experiment using Balanced Transformer and obtained an accuracy of 99.89%, sensitivity of 99.74%, specificity of 99.73% and AUC of 99.19%, which is far higher than the previous results (accuracy of 89.6% ∼ 96.8%, sensitivity of 74.3% ∼ 91.3%, specificity of 88.9% ∼ 96.02% and AUC of 87.8% ∼ 94.9%).Significance.This study can accurately distinguish PCNSL and GBM before surgery. Because GBM is a common type of malignant tumor, the 1% improvement in accuracy has saved many patients and reduced treatment times considerably. Thus, it can provide doctors with a good basis for auxiliary diagnosis.

Demystifying neuroblastoma malignancy through fractal dimension, entropy, and lacunarity

PURPOSE

Neuroblastoma is a pediatric solid tumor with a prognosis associated with histology and age of the patient, which are the parameters of the well-established current classification (Shimada classification). Despite the development of new treatment options, the prognosis of high-risk neuroblastoma patients is still poor. Therefore, there is a continuous need to stratify the children suffering from this tumor. A mathematical and computational approach is proposed to enable automatic and precise cancer diagnosis on the histological slide.

METHODS

We targeted the complexity of neuroblastoma by calculating its image entropy (S), fractal dimension (FD), and lacunarity (λ) in a combined mathematical code. First, we tested the proposed method for patient-derived glioma images. It allowed distinguishing between normal brain tissue, grade II, and grade III glioma, which harbor different outcomes.

RESULTS

In neuroblastoma, our analysis of image's FD, S, and λ combined with a machine learning algorithm automatically predicted tumor malignancy with a receiver operating characteristic curve of 0.82. FD, S, and λ distinguish between neuroblastoma and ganglioneuroma, but they only partially differentiate between the normal samples and the other classes. Ganglioneuroma, the most differentiated form, and poorly-differentiated neuroblastoma display different values of FD, S, and λ.

CONCLUSIONS

FD, S, and λ of imaging recognize groups in neuroblastic tumors. We suggest that future studies including these features may challenge the current Shimada classification of neuroblastoma with categories of favorable and unfavorable histology. It is expected that this methodology could trigger multicenter studies and potentially find practical use in the clinical setting of children's hospitals worldwide.

Three-dimensional fractal dimension and lacunarity features may noninvasively predict TERT promoter mutation status in grade 2 meningiomas

PURPOSE

The 2021 World Health Organization classification includes telomerase reverse transcriptase promoter (TERTp) mutation status as a factor for differentiating meningioma grades. Therefore, preoperative prediction of TERTp mutation may assist in clinical decision making. However, no previous study has applied fractal analysis for TERTp mutation status prediction in meningiomas. The purpose of this study was to assess the utility of three-dimensional (3D) fractal analysis for predicting the TERTp mutation status in grade 2 meningiomas.

METHODS

Forty-eight patients with surgically confirmed grade 2 meningiomas (41 TERTp-wildtype and 7 TERTp-mutant) were included. 3D fractal dimension (FD) and lacunarity values were extracted from the fractal analysis. A predictive model combining clinical, conventional, and fractal parameters was built using logistic regression analysis. Receiver operating characteristic curve analysis was used to assess the ability of the model to predict TERTp mutation status.

RESULTS

Patients with TERTp-mutant grade 2 meningiomas were older (P = 0.029) and had higher 3D FD (P = 0.026) and lacunarity (P = 0.004) values than patients with TERTp-wildtype grade 2 meningiomas. On multivariable logistic analysis, higher 3D FD values (odds ratio = 32.50, P = 0.039) and higher 3D lacunarity values (odds ratio = 20.54, P = 0.014) were significant predictors of TERTp mutation status. The area under the curve, accuracy, sensitivity, and specificity of the multivariable model were 0.84 (95% confidence interval 0.71-0.93), 83.3%, 71.4%, and 85.4%, respectively.

CONCLUSION

3D FD and lacunarity may be useful imaging biomarkers for predicting TERTp mutation status in grade 2 meningiomas.

The Association of Fractal Dimension with Vascularity and Clinical Outcomes in Glioblastoma

BACKGROUND

Growing evidence indicates fractal analysis (FA) has potential as a computational tool to assess tumor microvasculature in glioblastoma (GBM). As fractal parameters of microvasculature have shown to be reliable quantitative biomarkers in brain tumors, there has been similar success in measuring the architecture of tumor tissue using FA in other tumor types. However, evaluating fractal parameters of tissue structure in relation to the microvasculature has not yet been implemented in GBM. We aimed to assess the utility of this methodology in quantifying structural characteristics of GBM cytoarchitecture and vascularity by correlating fractal parameters with gene expression.

METHODS

Formalin-fixed paraffin-embedded specimens were retrospectively collected from 43 patients following resection of a newly diagnosed GBM; 4 normal brain specimens were obtained from epilepsy surgeries as controls. Tumor samples were processed using FA employing a software-based box-counting method algorithm and custom messenger RNA expression assays. Fractal parameters were then correlated with clinical features, outcomes, and a panel of 92 genes associated with vascularity and angiogenesis.

RESULTS

Statistical analysis demonstrated that fractal-based indices were not adequate parameters for distinction of GBM cytoarchitecture compared with normal brain specimens. Correlation analysis of our gene expression findings suggested that hematoxylin and eosin-based FA may have adequate sensitivity to detect associations with vascular gene expression.

CONCLUSIONS

The combination of neuropathological assessment and histology does not provide optimized data for FA in GBM. However, an association between FA and gene expression in GBM of genes pertaining to cytoarchitecture and angiogenesis warrants further investigation.

Fractal dimension of the brain in neurodegenerative disease and dementia: A systematic review

Sensitive and specific antemortem biomarkers of neurodegenerative disease and dementia are crucial to the pursuit of effective treatments, required both to reliably identify disease and to track its progression. Atrophy is the structural magnetic resonance imaging (MRI) hallmark of neurodegeneration. However in most cases it likely indicates a relatively advanced stage of disease less susceptible to treatment as some disease processes begin decades prior to clinical onset. Among emerging metrics that characterise brain shape rather than volume, fractal dimension (FD) quantifies shape complexity. FD has been applied in diverse fields of science to measure subtle changes in elaborate structures. We review its application thus far to structural MRI of the brain in neurodegenerative disease and dementia. We identified studies involving subjects who met criteria for mild cognitive impairment, Alzheimer's Disease, Vascular Dementia, Lewy Body Dementia, Frontotemporal Dementia, Amyotrophic Lateral Sclerosis, Parkinson's Disease, Huntington's Disease, Multiple Systems Atrophy, Spinocerebellar Ataxia and Multiple Sclerosis. The early literature suggests that neurodegenerative disease processes are usually associated with a decline in FD of the brain. The literature includes examples of disease-related change in FD occurring independently of atrophy, which if substantiated would represent a valuable advantage over other structural imaging metrics. However, it is likely to be non-specific and to exhibit complex spatial and temporal patterns. A more harmonious methodological approach across a larger number of studies as well as careful attention to technical factors associated with image processing and FD measurement will help to better elucidate the metric's utility.

Glioblastoma radiomics to predict survival: Diffusion characteristics of surrounding nonenhancing tissue to select patients for extensive resection

BACKGROUND AND PURPOSE

Glioblastoma (GBM) is an aggressive primary CNS neoplasm with poor overall survival (OS) despite standard of care. On MRI, GBM is usually characterized by an enhancing portion (CET) (surgery target) and a nonenhancing surrounding (NET). Extent of resection is a long debated issue in GBM, with recent evidence suggesting that both CET and NET should be resected in <65 years old patients, regardless of other risk factors (i.e., molecular biomarkers). Our aim was to test a radiomic model for patient survival stratification in <65 years old patients, by analyzing MRI features of NET, to aid tumor resection.

METHODS

Sixty-eight <65 years old GBM patients, with extensive CET resection, were selected. Resection was evaluated by manually segmenting CET on volumetric T1-weighted MRI pre and postsurgery (within 72 h). All patients underwent the same treatment protocol including chemoradiation. NET radiomic features were extracted with a custom version of Pyradiomics. Feature selection was performed with principal component analysis (PCA) and its effect on survival tested with Cox regression model. Twelve months OS discrimination was tested by t-test followed by logistic regression. Statistical significance was set at p<0.05. The most relevant features were identified from the component matrix.

RESULTS

Five PCA components (PC1-5) explained 90% of the variance. PC5 resulted significant in the Cox model (p = 0.002; exp(B) = 0.686), at t-test (p = 0.002) and logistic regression analysis (p = 0.006). Apparent diffusion coefficient (ADC)-based features were the most significant for patient survival stratification.

CONCLUSIONS

ADC radiomic features on NET predict survival after standard therapy and could be used to improve patient selection for more extensive surgery.

Radiomics-based differentiation between glioblastoma and primary central nervous system lymphoma: a comparison of diagnostic performance across different MRI sequences and machine learning techniques

OBJECTIVES

Despite the robust diagnostic performance of MRI-based radiomic features for differentiating between glioblastoma (GBM) and primary central nervous system lymphoma (PCNSL) reported on prior studies, the best sequence or a combination of sequences and model performance across various machine learning pipelines remain undefined. Herein, we compare the diagnostic performance of multiple radiomics-based models to differentiate GBM from PCNSL.

METHODS

Our retrospective study included 94 patients (34 with PCNSL and 60 with GBM). Model performance was assessed using various MRI sequences across 45 possible model and feature selection combinations for nine different sequence permutations. Predictive performance was assessed using fivefold repeated cross-validation with five repeats. The best and worst performing models were compared to assess differences in performance.

RESULTS

The predictive performance, both using individual and a combination of sequences, was fairly robust across multiple top performing models (AUC: 0.961-0.977) but did show considerable variation between the best and worst performing models. The top performing individual sequences had comparable performance to multiparametric models. The best prediction model in our study used a combination of ADC, FLAIR, and T1-CE achieving the highest AUC of 0.977, while the second ranked model used T1-CE and ADC, achieving a cross-validated AUC of 0.975.

CONCLUSION

Radiomics-based predictive accuracy can vary considerably, based on the model and feature selection methods as well as the combination of sequences used. Also, models derived from limited sequences show performance comparable to those derived from all five sequences.

KEY POINTS

• Radiomics-based diagnostic performance of various machine learning models for differentiating glioblastoma and PCNSL varies considerably. • ML models using limited or multiple MRI sequences can provide comparable performance, based on the chosen model. • Embedded feature selection models perform better than models using a priori feature reduction.

Development of a predictive model of growth hormone deficiency and idiopathic short stature in children

The aim of the present study was to develop predictive models using clinical features and MRI texture features for distinguishing between growth hormone deficiency (GHD) and idiopathic short stature (ISS) in children with short stature. This retrospective study included 362 children with short stature from Children's Hospital of Hebei Province. GHD and ISS were identified via the GH stimulation test using arginine. Overall, there were 190 children with GHD and 172 with ISS. A total of 57 MRI texture features were extracted from the pituitary gland region of interest using C++ language and Matlab software. In addition, the laboratory examination data were collected. Receiver operating characteristic (ROC) regression curves were generated for the predictive performance of clinical features and MRI texture features. Logistic regression models based on clinical and texture features were established for discriminating children with GHD and ISS. Two clinical features [IGF-1 (insulin growth factor-1) and IGFBP-3 (IGF binding protein-3) levels] were used to build the clinical predictive model, whereas the three best MRI textures were used to establish the MRI texture predictive model. The ROC analysis of the two models revealed predictive performance for distinguishing GHD from ISS. The accuracy of predicting ISS from GHD was 64.5% in ROC analysis [area under the curve (AUC), 0.607; sensitivity, 57.6%; specificity, 72.1%] of the clinical model. The accuracy of predicting ISS from GHD was 80.4% in ROC analysis (AUC, 0.852; sensitivity, 93.6%; specificity, 65.8%) of the MRI texture predictive model. In conclusion, these findings indicated that a texture predictive model using MRI texture features was superior for distinguishing children with GHD from those with ISS compared with the model developed using clinical features.

Glioblastoma and primary central nervous system lymphoma: differentiation using MRI derived first-order texture analysis - a machine learning study

OBJECTIVES

To evaluate the diagnostic performance of multiple machine learning classifier models derived from first-order histogram texture parameters extracted from T1-weighted contrast-enhanced images in differentiating glioblastoma and primary central nervous system lymphoma.

METHODS

Retrospective study with 97 glioblastoma and 46 primary central nervous system lymphoma patients. Thirty-six different combinations of classifier models and feature selection techniques were evaluated. Five-fold nested cross-validation was performed. Model performance was assessed for whole tumour and largest single slice using receiver operating characteristic curve.

RESULTS

The cross-validated model performance was relatively similar for the top performing models for both whole tumour and largest single slice (area under the curve 0.909-0.924). However, there was a considerable difference between the worst performing model (logistic regression with full feature set, area under the curve 0.737) and the highest performing model for whole tumour (least absolute shrinkage and selection operator model with correlation filter, area under the curve 0.924). For single slice, the multilayer perceptron model with correlation filter had the highest performance (area under the curve 0.914). No significant difference was seen between the diagnostic performance of the top performing model for both whole tumour and largest single slice.

CONCLUSIONS

T1 contrast-enhanced derived first-order texture analysis can differentiate between glioblastoma and primary central nervous system lymphoma with good diagnostic performance. The machine learning performance can vary significantly depending on the model and feature selection methods. Largest single slice and whole tumour analysis show comparable diagnostic performance.

LCK expression is a potential biomarker for distinguishing primary central nervous system lymphoma from glioblastoma multiforme

Glioblastoma multiforme (GBM) and primary central nervous system lymphoma (PCNSL) are both malignant cerebral tumors; however, their treatments are vastly different. Early and precise diagnosis is vital for subsequent adequate treatment to improve prognosis. Reliable biomarkers that can easily distinguish GBM and PCNSL are urgently needed. We evaluated the diagnostic potential of lymphocyte‐specific protein tyrosine kinase (LCK) as a biomarker in differentiating PCNSL from GBM using established computational approaches (Gene Expression Profiling Interactive Analysis, The Cancer Proteome Atlas, Tumor Immune Estimation Resource, GEO, Oncomine) and immunohistochemistry. The results showed that LCK was expressed at a high level in PCNSL patients but at a low level in GBM patients. Moreover, LCK expression positively correlated with the levels of infiltrating B cells in diffuse large B‐cell lymphoma (DLBCL) and GBM. Overall, bioinformatics analysis and immunohistochemistry revealed that LCK expression is a potential biomarker for distinguishing PCNSL from GBM.

Magnetic resonance imaging-based 3-dimensional fractal dimension and lacunarity analyses may predict the meningioma grade

OBJECTIVE

To assess whether 3-dimensional (3D) fractal dimension (FD) and lacunarity features from MRI can predict the meningioma grade.

METHODS

This retrospective study included 131 patients with meningiomas (98 low-grade, 33 high-grade) who underwent preoperative MRI with post-contrast T1-weighted imaging. The 3D FD and lacunarity parameters from the enhancing portion of the tumor were extracted by box-counting algorithms. Inter-rater reliability was assessed with the intraclass correlation coefficient (ICC). Additionally, conventional imaging features such as location, heterogeneous enhancement, capsular enhancement, and necrosis were assessed. Independent clinical and imaging risk factors for meningioma grade were investigated using multivariable logistic regression. The discriminative value of the prediction model with and without fractal features was evaluated. The relationship of fractal parameters with the mitosis count and Ki-67 labeling index was also assessed.

RESULTS

The inter-reader reliability was excellent, with ICCs of 0.99 for FD and 0.97 for lacunarity. High-grade meningiomas had higher FD (p < 0.001) and higher lacunarity (p = 0.007) than low-grade meningiomas. In the multivariable logistic regression, the diagnostic performance of the model with clinical and conventional imaging features increased with 3D fractal features for predicting the meningioma grade, with AUCs of 0.78 and 0.84, respectively. The 3D FD showed significant correlations with both mitosis count and Ki-67 labeling index, and lacunarity showed a significant correlation with the Ki-67 labeling index (all p values < 0.05).

CONCLUSION

The 3D FD and lacunarity are higher in high-grade meningiomas and fractal analysis may be a useful imaging biomarker for predicting the meningioma grade.

KEY POINTS

• Fractal dimension (FD) and lacunarity are the two parameters used in fractal analysis to describe the complexity of a subject and may aid in predicting meningioma grade. • High-grade meningiomas had a higher fractal dimension and higher lacunarity than low-grade meningiomas, suggesting higher complexity and higher rotational variance. • The discriminative value of the predictive model using clinical and conventional imaging features improved when combined with 3D fractal features for predicting the meningioma grade.

Comparison of Diagnostic Performance of Two-Dimensional and Three-Dimensional Fractal Dimension and Lacunarity Analyses for Predicting the Meningioma Grade

BACKGROUND

To compare the diagnostic performance of two-dimensional (2D) and three-dimensional (3D) fractal dimension (FD) and lacunarity features from MRI for predicting the meningioma grade.

METHODS

This retrospective study included 123 meningioma patients [90 World Health Organization (WHO) grade I, 33 WHO grade II/III] with preoperative MRI including post-contrast T1-weighted imaging. The 2D and 3D FD and lacunarity parameters from the contrast-enhancing portion of the tumor were calculated. Reproducibility was assessed with the intraclass correlation coefficient. Multivariable logistic regression analysis using 2D or 3D fractal features was performed to predict the meningioma grade. The diagnostic ability of the 2D and 3D fractal models were compared.

RESULTS

The reproducibility between observers was excellent, with intraclass correlation coefficients of 0.97, 0.95, 0.98, and 0.96 for 2D FD, 2D lacunarity, 3D FD, and 3D lacunarity, respectively. WHO grade II/III meningiomas had a higher 2D and 3D FD (p=0.003 and p<0.001, respectively) and higher 2D and 3D lacunarity (p=0.002 and p=0.006, respectively) than WHO grade I meningiomas. The 2D fractal model showed an area under the curve (AUC), accuracy, sensitivity, and specificity of 0.690 [95% confidence interval (CI) 0.581-0.799], 72.4%, 75.8%, and 64.4%, respectively. The 3D fractal model showed an AUC, accuracy, sensitivity, and specificity of 0.813 (95% CI 0.733-0.878), 82.9%, 81.8%, and 70.0%, respectively. The 3D fractal model exhibited significantly better diagnostic performance than the 2D fractal model (p<0.001).

CONCLUSION

The 3D fractal analysis proved superiority in diagnostic performance to 2D fractal analysis in grading meningioma.