Assessment of EGFRvIII and PTEN profiles on routine histologic material from astrocytic tumors for patient selection for anti-EGFR treatment.

B32

Based on the reported predictive value of the EGFRvIII+/PTEN+ profile for targeted therapeutic approaches with EGFR kinase inhibitors in patients with glioblastoma, we started assessing this profile prospectively, when patient re-evaluation was mandatory in routine clinical practice, in order to select those who would benefit from such targeted treatment. Herein, we present our experience with this testing during the last 2 years. In order to evaluate the methods to be used, we initially assessed available archived tissue material (paraffin blocks) from 106 previously diagnosed astrocytic tumors (66 glioblastomas [GBM], 19 anaplastic astrocytomas [AA], 21 diffuse fibrillary astrocytomas [DFA]) for EGFR gene amplification (real time PCR), EGFRwt and EGFRvIII expression (semiquantitative RT-PCR, strong and low expression), as well as PTEN protein presence (immunohistochemistry [IHC]). Semiquantitative EGFRvIII determination was used in order to exclude cases where vIII would be expressed by only a minority of neoplastic cells. EGFR gene dosage assessment was used as a control for EGFRvIII expression in GBMs, where vIII expression has been repeatedly observed exclusively in cases with EGFR gene amplification. The cutoff for PTEN protein positivity was set at 75% with 2+ intensity. By using the same methods, diagnostic tissue material from 17 patients with GBM who suffered relapse upon initial treatment failure (mostly temozolomide and radiotherapy) was prospectively evaluated for new treatment assessment. EGFRvIII expression was observed in 34% of all tumors tested, including AAs and DFAs. Low EGFR vIII expression was also found in low grade astrocytomas (50% of AAs), where it was unrelated to EGFR gene amplification. Strong EGFRvIII expression was observed in only 15% of all tumors, almost exclusively GBMs (p < .001) with EGFR gene amplification (p < .001), although it was also encountered in 2 AAs with normal gene dosage. Among tumors with strong EGFRvIII expression, 80% exhibited PTEN immunopositivity. Thus, strong EGFRvIII/PTEN positivity was observed in approximately 10% of all astrocytic tumors and in 16% of GBM patients. Two such patients were identified in the prospectively tested series. Both received erlotinib, initially as monotherapy and in the next in combination with low dose bevacizumab and irinotecan. Both patients responded well (16 and 12 months as of current follow up). Thus, although the percentage of patients who may benefit from treatments with small molecular anti-EGFR inhibitors is small in routine practice, it seems worthy selecting these patients by testing for strong EGFRvIII/PTEN expression, especially if no other treatment option is available.