Search

1005 Background: The prevalence of brain metastasis from breast cancer (BCA) is increasing. We have identified several potential therapeutic targets through gene expression profiling of human BCA brain metastases. One potential target, hexokinase II (HK-2), is up-regulated in brain metastases as compared to unlinked primary tumors. HK-2 plays a vital role in the generation of many components required for cell growth and inhibits apoptosis through mitochondrial binding. We hypothesized that HK-2 promotes metastatic cell proliferation in the brain and that its increased expression would negatively impact patient survival post-resection. Methods: The gene expression profile of surgically resected human BCA brain metastases was compared to that of non-related unlinked primary invasive ductal carcinomas using cDNA microarray analysis. In a separate retrospective study, we generated survival curves dividing a cohort of 113 patients based on HK-2 immunoreactivity in resected brain metastases. To evaluate the fun...

3678 Background: Metastatic disease is the hallmark of human malignancy that continues to challenge even the most recent advances in cancer treatment. In colorectal cancer, the major site of metastatic spread is the liver, but other sites include lung, and peritoneum, among others. In addition to elucidating mechanisms of tumor spread, predicting the pattern of metastatic spread could allow a more focused surveillance and adjuvant therapy. The aim of this study was to study the pattern of metastatic spread in colorectal cancer using gene expression micorarrays. Methods: Using microarrays consisting of ∼35,000 spotted oligonucleotides, we profiled 40 primary colorectal cancer samples from patients with resected primary tumors who developed metastatic disease. Patients were grouped into three patterns of isolated or dominant metastatic spread: liver, lung, and peritoneum/mesentery. Class comparison analysis and hierarchical clustering was performed on the array data. Results: An evident distinction could be...