ABSTRACT
OBJECTIVE: To describe a genetic progression pathway in breast cancer by a maximum likelihood-based tree model representing the dependencies between chromosomal imbalances. STUDY DESIGN: One hundred six cases were studied by comparative genomic hybridization, followed by maximum likelihood estimation of an oncogenetic tree model. RESULTS: The tree model identified 3 clusters with correlated chromosomal imbalances. The first cluster included losses at 4q, 5q, 6q, 9p, 13q and a gain at 17q; the second cluster included gains at 1q, 8q, 16p and 20q; the third cluster included losses at 8p, 11q, 16q and 18q. The imbalances nearest the root of the tree were the loss at 13q (cluster 1), the gain at 1q (cluster 2) and the loss at 18q (cluster 3), reflecting an early change in breast cancer evolution. Cox regression analysis revealed the tumor stage and the grade as relevant for overall survival (p = 0.001) and the tumor stage, the grade and the loss at 16q as relevant for disease-free survival (p = 0.001). CONCLUSION: Methods like oncogenetic tree analysis provide insights into the genetic progression of breast cancer and may extract relevant markers detected by screening methods like comparative genomic hybridization for further studies.
Subject(s)
Breast Neoplasms/classification , Breast Neoplasms/genetics , Chromosome Aberrations , Chromosomes, Human , Comparative Genomic Hybridization , Models, Statistical , Adult , Aged , Aged, 80 and over , Biomarkers, Tumor/analysis , Breast Neoplasms/pathology , Disease Progression , Female , Humans , Middle Aged , Models, Genetic , Neoplasm Staging , Regression AnalysisABSTRACT
The purpose of this study was to compare the chromosomal genotype between breast cancers with and without secondary manifestations and between primary tumors and their secondary manifestations. Eighty six breast cancers, twenty lymph node metastases, ten distant metastases and ten local recurrences were analyzed by comparative genomic hybridization. Tumors with local recurrences showed significant more frequent losses at 2q32 than the tumors without recurrences. Lymph node positive cases showed significant more frequent losses at 9p21 than node negative cases. Lymph node metastases exhibited significant more frequent losses at 7q11, 14q24.3-q31 and 17q22-q24 than their primary tumors. In cases with distant metastases, losses at 5q23 were more frequent than in those without, but not reaching the significance level. The distant metastases showed significant more frequent losses at 5p15, 12q24 and 17q22-q24 than the primary tumors. These results reveal strong evidence that the potential for progression is determined in the primary tumor and that different ways of the development of local recurrences, lymph node and distant metastases exist. After confirmation of the results by interphase FISH on tissue micro arrays, the detection of these specific chromosomal imbalances may contribute to a more individual prediction of prognosis in breast cancer.
