Classifying Pediatric Central Nervous System Tumors through near Optimal Feature Selection and Mutual Information: A Single Center Cohort

Labeling, gathering mutual information, clustering and classification of central nervous system tumors may assist in predicting not only distinct diagnoses based on tumor-specific features but also prognosis. This study evaluates the epidemi- ological features of central nervous system tumors in children who referred to Mahak’s Pediatric Cancer Treatment and Research Center in Tehran, Iran. This cohort (convenience sample) study comprised 198 children (≤15 years old) with central nervous system tumors who referred to Mahak''s Pediatric Cancer Treatment and Research Center from 2007 to 2010. In addition to the descriptive analyses on epidemiological features and mutual information, we used the Least Squares Support Vector Machines method in MATLAB software to propose a preliminary predictive model of pediatric central nervous system tumor feature-label analysis.Of patients, there were 63.1% males and 36.9% females. Patients'' mean±SD age was 6.11±3.65 years. Tumor location was as follows: supra-tentorial (30.3%), infra- tentorial (67.7%) and 2% (spinal). The most frequent tumors registered were: high-grade glioma (supra-tentorial) in 36 (59.99%) patients and medulloblastoma (infra-tentorial) in 65 (48.51%) patients. The most prevalent clinical findings included vomiting, headache and impaired vision. Gender, age, ethnicity, tumor stage and the presence of metastasis were the features predictive of supra-tentorial tumor histology. Our data agreed with previous reports on the epidemiology of central nervous system tumors. Our feature-label analysis has shown how presenting features may partially predict diagnosis. Timely diagnosis and management of central nervous system tumors can lead to decreased disease burden and improved survival. This may be further facilitated through development of partitioning, risk prediction and prognostic models.