Prognostic Importance of Age, Tumor Location, and Tumor Grade in Grade II Astrocytomas: An Integrated Analysis of the Cancer Genome Atlas and the Surveillance, Epidemiology, and End Results Database

Abstract

Previous work in anaplastic astrocytoma (AA) demonstrated that the survival benefit from gross total resection (GTR) is modified by age and tumor location. Here, we determined the influence of age and tumor location on survival benefit from GTR in diffuse astrocytoma (DA). We used the Surveillance, Epidemiology, and End Results (SEER) database (1999-2010). We used Kaplan-Meier curves and Cox survival models to determine the survival benefit from GTR in populations stratified by age and tumor location. We determined the prevalence of the mutated isocitrate dehydrogenase (mIDH) using The Cancer Genome Atlas (TCGA). We identified 1980 patients with DA. For frontal DAs, GTR resulted in improved survival relative to subtotal resection in all ages (age ≤50 years hazard ratio [HR], 0.56; P= 0.002; age >50 years HR, 0.41; P < 0.001). For nonfrontal DAs, only patients ≤50 years experienced improved survival with GTR (age ≤50 years HR, 0.55; P=0.002; age >50 years HR, 0.78; P= 0.114). For patients ≤50 years with frontal tumors, survival was comparable between DA and AA after GTR (75% survival DA: 80 months, AA: 89 months, P= 0.973). In TCGA, these tumors were nearly uniformly mIDH (DA: 98%; AA: 90%, P= 0.11). However, for patients ≤50 years with nonfrontal tumors, there was a survival difference after GTR (75% survival DA: 80 months, AA: 30 months, P= 0.001) despite comparable mIDH prevalence (DA: 82%, AA: 75%, P= 0.49). Age and tumor location modify the survival benefit derived from GTR in DA. Survival patterns in SEER imperfectly correlated with mIDH prevalence in TCGA, suggesting that tumor grade and mIDH status convey nonredundant prognostic information in select clinical contexts.