BACKGROUND: Infiltrating WHO grade II and III gliomas are devastating and difficult to treat neoplasms classified as astrocytoma (A2), oligodendroglioma (O2), oligoastrocytoma (OA2), or anaplastic astrocytoma (A3), oligodendroglioma (O3), and oligoastrocytoma (OA3). Anaplastic grade is critically dependent on the presence of mitoses (astrocytomas) or high mitotic activity (oligodendroglial tumors), with CDKN2A deletion considered the most common mechanism for associated cell cycle dysregulation. We therefore sought to determine if p16 loss by immunohistochemistry (IHC) or CDKN2A gene by fluorescence in situ hybridization (FISH) are associated with patient survival across histologic groupings after controlling for age, grade, and other molecular markers. METHODS: Adults (18 years of age) with infiltrating gliomas (grades II or III) were selected from the UCSF Adult Glioma Study (N = 157). The population consisted of 90 males (average age = 43.2) and 67 females (average age = 40.7). Isocitrate dehydrogenase (IDH) status (IDH1 and IDH2) was known for all cases. Histology and grade were as follows: 53 O2, 33 A2, 29 OA2, 20 O3, 15 A3, 7 OA3. CDKN2A FISH was performed using commercial probes and immunostains for p16 and MIB1 (Ki-67) were similarly performed using commercial antibodies. The end point for the analysis was overall survival, analyzed using Cox proportional hazards, stratified by tumor histology and with adjustment for sex, age, grade, and various tumor markers. A total of 61 events (deaths) were observed, and median follow-up time across all subjects was 6.4 years. RESULTS: After controlling for age, sex, and grade, CDKN2A deletion was associated with decreased survival in astrocytoma patients (P = 0.045; HR = 3.1; 95%CI = 1.03-9.2) but not in oligodendroglioma or oligoastrocytoma patients (P = 0.57 and P = 0.67, respectively). A strong inverse association between IDH mutation and CDKN2A deletion was observed in astrocytomas, and the association of CDKN2A deletion with survival was attenuated after controlling for IDH status (P = 0.34). Interestingly, CDKN2A loss was observed in 83% of IDH wild-type versus 33% of IDH-mutant astrocytomas (P = 0.029). CDKN2A loss was only weakly associated with decreased p16 expression (p = 0.09), and the p16 labeling index (LI) was not associated with outcome. There was also no clear association between CDKN2A deletion and MIB1 LI (P = 0.92). CONCLUSIONS: CDKN2A deletion by FISH is associated with shortened survival in astrocytoma patients, but not in patients with oligodendroglial tumors. The strong inverse association between IDH mutation and CDKN2A deletion suggests that CDKN2A status may have prognostic value in the clinical work-up of grade II and III astrocytoma patients. SECONDARY CATEGORY: n/a.