1p/19q Deletion in Brain Tumors: An Innovative Approach for Automated FISH Analysis
G DeMaglio, G Falconieri, F Curcio, S Pizzolitto. General Hospital, Udine, Italy; University Hospital, Udine, Italy
Background: The prognostic and predictive value of 1p/19q deletion in brain tumors has been addressed during the last few years. In particular, the 1p/19q combined loss has been described in more than 70% of oligodendrogliomas, and this genetic alteration is reportedly associated with a better chemo- and radiotherapy response and overall survival rates. Several technical approaches have been devised for the 1p/19q status assessment, including fluorescent in situ hybridization (FISH), comparative genomic hybridization, and loss of heterozygosity. Although FISH is most commonly used, the resulting analysis is labor intensive and time consuming. In addition, the cutoff deletion limits are often not standardized. Although an automated system for FISH analysis of hybridization patterns may represent a valid alternative to the shortcomings of visual counting techniques, published data comparing the systems are currently limited.
Design: FISH analysis for 1p/19q loss was performed on paraffin sections obtained from 83 primary brain tumors. Tumors included oligoastrocytomas (n=32, 39%), oligodendrogliomas (n=24, 29%), astrocytomas (n=15, 18%), high-grade gliomas (n=9, 11%), and other brain tumors (n=3, 3%). Commercial probes against 1p36/1q25 and 19q13/19p13 (Abbott) were used. An automatic system for FISH analysis (Metafer 4, Metasystems, Altlussheim, Germany) was applied for reevaluation of visually screened samples. We compared manual and automatic scoring systems in both pathologic and non-neoplastic perilesional tissues.
Results: The concordance rate of manual versus automatic score was near total (97%) for non-neoplastic areas and virtually total (>99%) for tumoral areas. In about 10% of scannings (17/166), the results were deemed not reliable because of microscopic features that were capable of preventing adequate signaling interpretation, such as tissue necrosis, suboptimal sample preparation, fluorochrome bleaching, and excessive tissue autofluorescence.
Conclusions: An automatic system for FISH analysis proved superior to visual counting because it allowed a faster and reliable evaluation of a greater number of tumor cell nuclei and prevented potential observer bias in the selection of both normal and tumor nuclei. Manual review of the image gallery after automatic analysis is effective in obtaining more reliable results. This system may improve the overall FISH testing efficiency, especially in high-volume laboratories.
Tuesday, March 10, 2009 9:30 AM
Poster Session III # 188, Tuesday Morning