* Faculty       * Staff       * Contact       * Institute Directory
* Research Groups      
* Undergraduate       * Graduate       * Institute Admissions: Undergraduate | Graduate      
* Events       * Institute Events      
* Lab Manual       * Institute Computing      
No Menu Selected

* Research

Ph.D. Theses

The Cell-Graphs of Brain Cancer

By Cigdem Demir
Advisor: Bulent Yener
December 6, 2005

In the current practice of medicine, pathologists traditionally diagnose cancer from tissue samples. Examining such biopsy samples under a microscope, a pathologist typically makes assessments based on visual interpretation of cell morphology and tissue distribution. This, however, leads to a certain level of subjectivity. To circumvent this problem, it is important to develop computational diagnostic tools that operate on quantitative measures. Such automated diagnostic tools facilitate fast, objective, mathematical judgment complementary to that of a pathologist, reducing the subjectivity.

For the purpose of automated cancer diagnosis, we introduce a novel graph-theoretical tissue representation (cell-graphs) that enables us to quantify the topological properties of cluster formation in a tissue sample. By making use of these distinctive properties, we automatically distinguish cancerous tissues from their counterparts with high accuracy.

In this cell-graph representation, nodes represent a group of cells (cell-clusters) and edges represent the spatial interrelations of these cell-clusters. Thus, this representation quantifies the spatial distribution of the cell-clusters across a tissue. In this thesis, we present the methodology of cell-graph generation along with a theoretical framework and experimental demonstrations. We also introduce the definitions of different sets of distinctive cell-graph features.

In this thesis, we report on the experimental demonstrations obtained on clinical data for the diagnosis of brain cancer (malignant glioma). Our experiments show that the cell-graph approach is able to differentiate a cancerous tissue from non-cancerous tissues (from a healthy tissue and a benign inflammatory process) with high accuracies. Moreover, we extend our experiments to the breast tissues for the purpose of automated breast cancer diagnosis, which demonstrates that the cell-graph approach is not limited to the diagnosis of brain cancer.

* Return to main PhD Theses page