vinamrak / machine-leaning-for-leukemia-classification

Machine-Leaning-for-Leukemia-Classification

Cancer treatment has been one of the most active areas of medical research for many decades now. One of the main challenges that cancer treatment poses today is targeting tumor specific therapy. This is essential to maximize the efficiency of treatment and to reduce the toxicity of treatment at the same time. Accurate cancer classification is thus central to advancing treatment today. Till date, classification has mostly been done by observing the morphological appearance of tumors, but this approach is quite naive as different classes can often look similar, but react very differently to therapy. This calls for the need of a new approach for classifying cancer. This is where Genomics and Machine Learning come into the scenario. Gene expression data using DNA microarrays has been suggested to be able to provide a tool for classifying cancer. This is what we have set out to do in this chapter. We will be using gene expression data to build a class predictor taking acute leukemias as our test case. Leukemia has mainly two classes - acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML). Leukemia classification is a lengthy process with steps involved such as – interpreting the tumor’s morphology, histochemistry, immunophenotyping, and cytogenetic analysis. All of these steps have to be carried out in separate, highly specialized laboratories. And although the classification is mostly accurate, errors still happen. In this chapter, we shall be using Machine Learning on gene expression data to try to build a classifier that correctly classifies ALL from AML.