General Morphological Features, Ki-67 and Telomerase/Telomere Reliably Predict Common Malignancies and Subclassify Carcinomas
J Moorhead, J Gonzalez, A Blanes, SJ Diaz-Cano. King's College Hospital, London, United Kingdom; University of Malaga School of Medicine, Malaga, Spain
Background: Tumor subclassification relies heavily on immunohistochemical markers and is site specific. Biological definition based on general morphology and kinetic regardless of location is not available.
Design: We analyzed primary and secondary growth patterns (tubulo-papillary, nested-trabecular, nodular-solid, diffuse), nuclear grade (including chromatin, nucleolus, pleomorphism and anisokaryosis), desmoplasia, and confluent necrosis in common malignancies: carcinomas (45 basal cell, 75 squamous cell, 104 adenocarcinomas, 30 urothelial, and 15 neuroendocrine), sarcomas (75), lymphomas (75) and melanomas (45). Representative samples were evaluated by standard immunohistochemistry for Ki67, telomerase, mlh1, msh2, TUNEL assay for apoptosis, telomere PNA-FISH, and selective cDNA miniarray (telomerase, p53, mdm2, p21, cdk2, cyclin E, pRB, Egr2, JunB, and FosB). Total RNA was extracted, cleaned from normal and neoplastic tissues (RNeasy columns), first-strand cDNA synthesized using T7-(dT24)-oligomer and used as template for cRNA synthesis. The cRNA was fragmented, Cy3-/Cy5-labeled, and hybridized to miniarray noncompetitively, cross-validating the results (expression factor>2, significance<0.01). Appropriate controls were run. Fisher's exact tests and analysis of variance (significant if P<0.05) were used for comparison; significant variables were used for cross-validated discriminant analyses for a biological definition by diagnostic groups.
Results: Discriminant analysis for the main diagnostic groups selected as independent variables growth patterns, nuclear pseudoinclusions, necrosis distribution, Ki-67 index, desmoplastic reaction and nuclear grade (nucleolus number, chromatin distribution and pleomorphism). Telomerase expression and telomere positive cells (%) were the added variables for carcinoma subclassification. The percentage of cases correctly classified with these variables were always >97% in all cross-validated models. Telomerase/telomere indices directly correlated with the kinetic index, being significantly higher in high-grade malignancies with upregulation of p53, cyclin E, Egr2, JunB and FosB.
Conclusions: A careful selection of a small number of morphological (growth patterns, desmoplastic reaction, nuclear grading) and kinetic features provides a reliable and reproducible classification of common malignancies. Proliferation and telomerase/telomere indices improve the prediction for carcinoma subclassification only.
Wednesday, March 24, 2010 9:30 AM
Poster Session V # 224, Wednesday Morning