Clinical-MRI radiomics enables the prediction of preoperative cerebral spinal fluid dissemination in children with medulloblastoma

Abstract

Background Medulloblastoma (MB) is the most common pediatric embryonal tumor. Accurate identification of cerebral spinal fluid (CSF) dissemination is important in prognosis prediction. Both MRI of the central nervous system (CNS) and CSF cytology will appear false positive and negative. Our objective was to investigate the added value of preoperative-enhanced T1-weighted image-based radiomic features to clinical characteristics in predicting preoperative CSF dissemination for children with MB. Materials and methods This retrospective study included 84 children with histopathologically confirmed MB between November 2006 and November 2018 (training cohort, n=60; internal validation cohort, n=24). A set of cases between December 2018 and February 2020 were used for external validation (n=40). The children with normal head and spine magnetic resonance images (MRI) and no subsequent dissemination in 1 year were diagnosed as non-CSF dissemination. The CSF dissemination was manifested as intracranial or intraspinal nodular-enhanced lesions. Clinical features were collected, and conventional MRI features of preoperative head MRI examinations were evaluated. A total of 385 radiomic features were extracted from preoperative-enhanced T1-weighted images. Minimum redundancy, maximum correlation, and least absolute shrinkage and selection operator were performed to select the features with the best performance in predicting preoperative CSF dissemination. A combined clinical-MRI radiomic prediction model was developed using multivariable logistic regression. Receiver operating curve analysis (ROC) was used to validate the predictive performance. Nomogram and decision curve analysis (DCA) were developed to evaluate the clinical utility of the combined model. Results One clinical and nine radiomic features were selected for predicting preoperative CSF dissemination. The combined model incorporating clinical and radiomic features had the best predictive performance in the training cohort with an AUC of 0.89. This was validated in the internal and external cohorts with AUCs of 0.87 and 0.73. The clinical utility of the model was confirmed by a clinical-MRI radiomic nomogram and DCA. Conclusions The combined model incorporating clinical, conventional MRI, and radiomic features could be applied to predict preoperative CSF dissemination for children with MB as a noninvasive biomarker, which could aid in risk evaluation.

Current role of machine learning and radiogenomics in precision neuro-oncology

“…Pre-operative identification of the molecular subgroup of medulloblastoma is extremely important as the extent of resection depends on the molecular subtype [56]. Therefore, Thompson et al [57] analyzed the prognostic value of the extent of resection in a retrospective multi-institutional cohort involving 787 patients with medulloblastoma. Only in group 4 tumors, wide neurosurgical resection was associated with a significant increase in progression-free survival, but not in overall survival.…”

Clinical, Histological, and Molecular Prognostic Factors in Childhood Medulloblastoma: Where Do We Stand?

“…The latter is of paramount importance, as it would assist with further understanding of the diverse nature of MBs and the designation of appropriate treatment strategies. Until recently, radiomic analysis had managed to indirectly predict the prognosis of patients, as MRI findings were associated with the molecular subgroup of the tumor [58,59] or the dissemination to the CSF [60]. However, the co-estimation of imaging findings, clinical characteristics, and molecular subgroup led to the development of nomograms, predicting both PFS [61] and OS [62].…”

Radiomics and radiogenomics in pediatric neuro-oncology: A review

“… 19 In MB patients, Zheng et al found a clinical-radiomic model to be predictive of preoperative Cerebrospinal Fluid (CSF) dissemination (as determined by head and spine MRI, as well as no subsequent dissemination at the 1-year follow-up for the non-CSF dissemination group). 33 …”

“…Some authors noted that moving forward, multicenter data could be used to increase sample size and evaluate model generalizability. 32 , 33 There exist a few consortiums to facilitate data collection and sharing, 41 , 42 such as the Children’s Brain Tumor Network , which has several cloud-based platforms, outlined in Figure 2 , where researchers can access the data. Other databases relevant for radiomic studies include the National Cancer Institute’s Clinical Proteomic Tumor Analysis Consortium (CPTAC) and the Pediatric Brain Tumor Consortium.…”