When we use radiation to treat cancer patients, the irradiated volume usually encloses the detectable tumor and any metastatic areas that are thought to be at risk. Usually, in three-dimensional (3D) radiation therapy, we need to define the clinical target volume (CTV), which identifies the areas suspected of containing microscopic metastasis. We can use a set of computed tomographic (CT) images to define these suspected areas, which we relate to nodal station distributions. The traditional method uses special anatomic landmarks for such definition. This lymph node image guidance method uses a set of CT images to show the relevant nodal stations and find the correlation between the suspected areas and the lymphatic pathways. First, we used CT images from head and neck cancer patients. Then, we designed a program to delineate the regional lymphatic areas in order to distinguish the position of all the major lymph nodes in the head and neck region. These lymph nodes were then used to assist the oncologist in developing treatment references. Finally, we established a database management system to determine the relationships between different lesions and lymph nodes, as well as correlations, with metastatic pathways. Defining these relationships on the CT images provided oncologists specific information that could be used to define the CTV more precisely for 3D radiation therapy.