Limited utility despite accuracy of the national SEER dataset for the study of craniopharyngioma

Abstract

Craniopharyngioma is histologically benign and associated with high survival rates but poor quality of life. The SEER Program is among the most cited data sources regarding malignancies in the United States. SEER began collecting data regarding craniopharyngiomain 2004. SEER-STAT v7.0.5 was utilized to identify patients (January 1, 2004-December 31, 2008) with ICD-O-3 codes for craniopharyngioma. Age was categorized into 3 groups: ≤19, 20-34, and ≥35 years, as was surgical intervention: none, subtotal resection, and gross total/radical resection. Demographic, initial treatment, and follow-up data were collected. 635/662 (95.9 %) patients had complete data. Incidence per million patient-years by age group was 1.9, 1.1, and 1.9, respectively (p < 0.0001). There was bimodal incidence, with peaks at 5-9 and 60-74 years. Surgery occurred in 528 patients (83.1 %), without association between age group and extent of surgery (p = 0.14). Radiation was delivered in 139 (21.9 %) cases, with no association between treatment and extent of surgery (p = 0.73) or age group (p = 0.14). Median follow-up was 23 months. Overall and Cause-specific Survival were 87.9 and 94.5 %, respectively. Neither was associated with extent of surgery but both were positively associated with radiation (p = 0.0003 and 0.0007, respectively). There was no difference in OS or CSS when comparing STR ± RT versus GTR alone (p = 0.38 and 0.56, respectively). SEER provides reliable demographic and survival data regarding craniopharyngioma. SEER's focus on mortality statistics limits utility for outcomes studies in tumors with high survival rates, such as craniopharyngioma. Initial treatment data from SEER varies somewhat from current literature, meriting further investigation.