Imaging Features of Common Pediatric Intracranial Tumours: A Primer for the Radiology Trainee

Mata‐Mbemba, Daddy; Donnellan, John; Krishnan, Pradeep; Shroff, Manohar; Muthusami, Prakash

doi:10.1016/j.carj.2017.10.006

“…Among the 3 classes studied here, PA is the most distinguishable on routine MRI because of characteristically high T2 signal due to its low cellular density and frequent cystic component. 2,4,5 Our first-stage model similarly prioritized this T2 signal by ranking T2 uniformity, a measure of signal homogeneity, as its most important variable (Figure 2; Supplementary Figure 1, Supplemental Digital Content 9). The predictive value of this feature is visibly appreciable on the corresponding density plot by the strong separation between PA and the other 2 tumors.…”

Section: Feature Interpretationmentioning

confidence: 99%

“…These nuances leave room for improvement in diagnostic accuracy and create a potential role for machine learning in assisting clinicians. 2,4,5 Radiomics-based machine learning has shown clinical utility for management of neurosurgical problems. [6][7][8][9] However, previous applications for pediatric PF tumors had limited accuracy and reproducibility, not only because of small cohorts and obscure feature extraction methods [10][11][12] but also because of the failure to adapt machine-learning strategies to the unique situation, eg, applying a single classifier method to a wide range of diagnoses.…”

mentioning

confidence: 99%

Machine Assist for Pediatric Posterior Fossa Tumor Diagnosis: A Multinational Study

Zhang¹,

Wong²,

Wright³

et al. 2021

View full text Add to dashboard Cite

BACKGROUND Clinicians and machine classifiers reliably diagnose pilocytic astrocytoma (PA) on magnetic resonance imaging (MRI) but less accurately distinguish medulloblastoma (MB) from ependymoma (EP). One strategy is to first rule out the most identifiable diagnosis. OBJECTIVE To hypothesize a sequential machine-learning classifier could improve diagnostic performance by mimicking a clinician's strategy of excluding PA before distinguishing MB from EP. METHODS We extracted 1800 total Image Biomarker Standardization Initiative (IBSI)-based features from T2- and gadolinium-enhanced T1-weighted images in a multinational cohort of 274 MB, 156 PA, and 97 EP. We designed a 2-step sequential classifier – first ruling out PA, and next distinguishing MB from EP. For each step, we selected the best performing model from 6-candidate classifier using a reduced feature set, and measured performance on a holdout test set with the microaveraged F1 score. RESULTS Optimal diagnostic performance was achieved using 2 decision steps, each with its own distinct imaging features and classifier method. A 3-way logistic regression classifier first distinguished PA from non-PA, with T2 uniformity and T1 contrast as the most relevant IBSI features (F1 score 0.8809). A 2-way neural net classifier next distinguished MB from EP, with T2 sphericity and T1 flatness as most relevant (F1 score 0.9189). The combined, sequential classifier was with F1 score 0.9179. CONCLUSION An MRI-based sequential machine-learning classifiers offer high-performance prediction of pediatric posterior fossa tumors across a large, multinational cohort. Optimization of this model with demographic, clinical, imaging, and molecular predictors could provide significant advantages for family counseling and surgical planning.

show abstract

“…21,22 While MRI serves as the cornerstone for operative planning, imaging clues of high-risk EP remain sparse and mostly unexplored. [23][24][25][26] Further, no prior study has examined quantitative, high-dimensional MRI features of EP that either relate to underlying molecular subgroups or prognosis. Here, we apply machine-enabled strategies on MRI to identify high-risk profile of posterior fossa EP tumors.…”

Section: Accepted Manuscriptmentioning

confidence: 99%

Radiomic signatures of posterior fossa ependymoma: Molecular subgroups and risk profiles

Zhang

¹

,

Wang

²

,

Tam

³

et al. 2021

View full text Add to dashboard Cite

Background The risk profile for posterior fossa ependymoma (EP) depends on surgical and molecular status [Group A (PFA) versus Group B (PFB)]. While subtotal tumor resection is known to confer worse prognosis, MRI-based EP risk-profiling is unexplored. We aimed to apply machine learning strategies to link MRI-based biomarkers of high-risk EP and also to distinguish PFA from PFB. Methods We extracted 1800 quantitative features from presurgical T2-weighted (T2-MRI) and gadolinium-enhanced T1-weighted (T1-MRI) imaging of 157 EP patients. We implemented nested cross-validation to identify features for risk score calculations and apply a Cox model for survival analysis. We conducted additional feature selection for PFA versus PFB and examined performance across three candidate classifiers. Results For all EP patients with GTR, we identified four T2-MRI-based features and stratified patients into high- and low-risk groups, with 5-year overall survival rates of 62% and 100%, respectively (p < 0.0001). Among presumed PFA patients with GTR, four T1-MRI and five T2-MRI features predicted divergence of high- and low-risk groups, with 5-year overall survival rates of 62.7% and 96.7%, respectively (p = 0.002). T1-MRI-based features showed the best performance distinguishing PFA from PFB with an AUC of 0.86. Conclusions We present machine learning strategies to identify MRI phenotypes that distinguish PFA from PFB, as well as high- and low-risk PFA. We also describe quantitative image predictors of aggressive EP tumors that might assist risk-profiling after surgery. Future studies could examine translating radiomics as an adjunct to EP risk assessment when considering therapy strategies or trial candidacy.

show abstract

“…Glioma, meningioma, and metastatic tumor are all common intracranial tumors. Due to the difference in tumor morphology, size, and lesion location, different tumors have different treatment methods [3,4]. erefore, preoperative precise qualitative is important in the selection of tumor treatment methods, guiding the formulation of treatment plans and evaluating the prognosis of patients.…”

Section: Introductionmentioning

confidence: 99%

Adoption of Diffusion Tensor Imaging under Optimized Fuzzy C-Means Cluster Algorithm in Intracerebral Benign and Malignant Tumor Resection

Liu

¹

,

Xu

²

2021

Scientific Programming

0

View full text Add to dashboard Cite

The study aimed to analyze the application of diffusion tensor imaging (DTI) in the surgery of benign and malignant intracranial tumors through improved fuzzy C-means (FCM). First, a method of combining the maximum and minimum distances was proposed to improve the FCM algorithm. Then, the optimized FCM was applied to the diffusion tensor imaging (DTI) diagnosis. The patients were rolled into the benign tumor group and the malignant tumor group, and relevant parameters were compared. Finally, the postoperative total resection rate and disability rate of the DTI experimental group and the traditional control (Ctrl) group were evaluated. It was found that the segmentation accuracy of the optimized FCM algorithm was higher than traditional one and the obtained corpus callosum edge contour was clearer. In 63 patients with intracranial space, there were obvious differences in pairwise comparison of meningioma and glioma, metastatic tumor’s apparent diffusion coefficient (ADC) value, relative apparent diffusion coefficient (r ADC) value, and relative anisotropy fraction (r FA) P < 0.05 . In terms of the ADC, r ADC, and r FA values of tumor parenchymal area, those of benign tumors were larger than malignant tumors P < 0.05 . The ADC value (8.21 ± 1.87) and r FA value (1.36 ± 0.41) of the contralateral normal white matter area of malignant tumor were greater than the ADC value (7.23 ± 2.31) and r FA value (0.61 ± 0.24) of the peritumor white matter area, with statistically significant differences P < 0.05 . The total cut rates of the experimental group and the Ctrl were 87.5% and 54.84%, and the disability rates were 6.25% and 34.38%. In conclusion, the optimized FCM has high accuracy. The ADC, r ADC, and r FA values of DTI are important in the diagnosis of intracranial tumors. Besides, DTI can improve the survival rate in guiding intracranial tumor resection.

show abstract

Imaging Features of Common Pediatric Intracranial Tumours: A Primer for the Radiology Trainee

Cited by 7 publications

References 77 publications

Machine Assist for Pediatric Posterior Fossa Tumor Diagnosis: A Multinational Study

Machine Assist for Pediatric Posterior Fossa Tumor Diagnosis: A Multinational Study

Radiomic signatures of posterior fossa ependymoma: Molecular subgroups and risk profiles

Adoption of Diffusion Tensor Imaging under Optimized Fuzzy C-Means Cluster Algorithm in Intracerebral Benign and Malignant Tumor Resection

Contact Info

Product

Resources

About