-Renal epithelial neoplasms have characteristic chromosomal imbalances, and we have shown previously that virtual karyotypes derived from single-nucleotide polymorphism microarrays can be performed on formalin-fixed, paraffin-embedded tissue.
-To perform a direct comparison of virtual and conventional karyotypes to evaluate concordance of results.
-Twenty archival formalin-fixed, paraffin-embedded tumor samples with preexisting, conventional cytogenetic results were analyzed with Affymetrix 10K 2.0 or 250K Nsp single-nucleotide polymorphism microarrays.
-Nineteen samples yielded adequate virtual karyotypes for interpretation. Eight samples showed complete agreement between the 2 techniques, and 8 samples showed partial agreement. The disease-defining lesions (eg, loss of 3p for clear cell carcinoma) were identified in all 19 cases by virtual karyotypes and in 15 cases by conventional karyotypes. Virtual and conventional karyotypic findings were concordant in the identification of these disease-defining lesions in 86% (13 of 15) of cases. In 3 cases, virtual karyotypes identified lesions consistent with the morphologic diagnosis, whereas the conventional karyotypes were unsuccessful because of insufficient tumor representation or stromal overgrowth. Two cases with acquired uniparental disomy were identified by single-nucleotide polymorphism arrays, and 5 cases with translocations were identified by conventional karyotype.
-Our results show that both techniques are able to identify the characteristic chromosomal abnormality for renal tumor subtypes in most cases. Discrepancies can be explained by inherent limitations of each technique, inadequate tumor sampling, and tumor heterogeneity. We conclude that virtual karyotyping is a robust alternative to conventional cytogenetics for the evaluation of chromosomal anomalies in formalin-fixed, paraffin-embedded tissues from renal epithelial neoplasms.