Title: SPATIO-TEMPORAL DATA MINING IN MEDICAL IMAGES

Year of Publication: Jul - 2013
Page Numbers: 197-199
Authors: Anahid Basiri, Pouria Amirian, Adam Winstanley
Conference Name: The Third International Conference on Digital Information and Communication Technology and its Applications (DICTAP2013)
- Czech Republic

Abstract:


Magnetic Resonance Imaging (MRI) is frequently used as follow-up examination. Since there is always possibility of multiplication of cancer cells and recurrence of cancer tumors after first treatments taken, such as surgery, patients are highly recommended to get follow-ups examinations, such as MRIs. Number of MRIs per year depends on the type and location of the firstly diagnosed tumor but based on patients health status specialist may cut down on number of annual MRIs year by year. However this may be dangerous because it is possible for the patients to have a recurrence of tumor after some years and since there is a few MRIs taken every year, tumor may grow fast or become in an inoperable position. In this regard, we suggest doing spatio-temporal data mining, using medical information of similar patients as input data, to determine when the most high-risk time to get another tumor is. In order to do such analysis effective criteria, such as patient age, type of tumor, location of tumor, amount of residual tumor, etc. should be considered. This paper suggested mining the information of patients with brain tumor as one of the most dangerous types of cancer.