AI based Differentiation of Beta-Thalassemia and Iron Deficiency Anemia.

Thalassemia is a genetic disorder that leads to variation in the protein structure of the hemoglobin molecule, which causes profound anemia that kills untreated affected children before the age of 3 years. Unfortunately, the traditional RBC indices based methods do not distinguish between β-thalassemia and a normal iron deficiency anemia, causing wrong treatment to be administered. Techniques such as HPLC, serum ferritin analysis and molecular studies are needed to help differentiate them. The main drawbacks of these techniques are that, it takes a long time to issue the results, it is expensive and is not available everywhere. Here, we attempt to use a heuristic algorithm as well as AI based techniques to accurately predict the occurrence of thalassemia from the easily available RBC data itself. The AI technique is based on support vector machine. An accuracy of 60 % and 96% are obtained for the heuristic algorithm and the AI technique respectively. Details of this work are presented in this paper.