[1809] Immunohistochemical Analysis of IMP3 Expression in Pilocytic Astrocytomas, Glioblastomas, and Recurrent Glioblastomas

Devki M Patel, Xuemo Fan. Cedars-Sinai Medical Center, Los Angeles, CA

Background: Insulin-like growth factor II mRNA-binding protein 3 (IMP3) is an oncofetal protein linked to tumor aggressiveness and has been observed in a variety of malignant neoplasms. This study aims to characterize IMP3 expression in pilocytic astrocytoma (PA), glioblastoma multiforme (GBM), and recurrent GBM and to determine if IMP3 immunostain will have diagnostic value in differentiating PA from GBM in small specimens.
Design: 21 cases of PA and 38 cases of GBM were studied. Of the GBM group, 18 cases were evaluated for both initial GBM and then recurrence. IMP3 immunostaining was performed on one representative tumor block from each case. The staining intensity (1-3+), percent of positive tumor cells, and staining pattern, along with endothelial positivity, were recorded. 5% or greater staining of tumor cells was considered positive.
Results: IMP3 positivity was found in 2 of 21 cases (9.5%) of PA, 22 of 38 cases (58%) of GBM, and 9 of 18 cases (50%) of recurrent GBM, with the difference between PA and GBM being statistically significant (p=0.0003). Patterns of IMP3 staining in tumor included: perivascular accentuation, subarachnoid and subpial accentuation, and scattered single cells. 2 cases of normal brain used as controls showed focal perineuronal dot-like pattern and 3 cases of GBM showed focal axonal and nuclear staining in adjacent normal brain tissue. Endothelial cell staining for IMP3 was found in 1 of 21 cases (4.8%) of PA, 15 of 38 cases (39.5%) of GBM, and 6 of 18 cases (33.3%) of recurrent GBM, and the difference between PA and GBM was statistically significant (p=0.005). In 1 case of GBM, IMP3 highlighted an arcade vasculature pattern but the tumor cells were completely negative. Surprisingly, of the 18 recurrent GBM cases, 5 showed IMP3 negativity in the initial tumor and positivity in the recurrence. In contrast, 4 cases showed IMP3 positivity in initial tumors but negativity in recurrences. The mean IMP3 staining intensity in positive cases was 2.5+ in PA, 2.9+ in GBM, and 2.4+ in recurrent GBM, and the difference between GBM and recurrent GBM was statistically significant (p=0.0336).
Conclusions: The presence of a higher level of IMP3 in tumor cells and associated endothelial cells in GBM compared to PA supports that IMP3 may be useful in differentiating PA from GBM in small biopsy specimens. Recurrent GBM shows similar IMP3 expression to GBM but with lower staining intensity. The significance of IMP3 expression in endothelial cells in GBM is unclear.
Category: Neuropathology

Tuesday, March 20, 2012 1:00 PM

Poster Session IV # 258, Tuesday Afternoon


Close Window