An analytical approach for leukemia diagnosis from light microscopic images of Rbcs (Computational approach for leukemia diagnosis)

Leukemia is routinely diagnosed by light microscopic images. However, pathologists' criteria for a disease diagnosis from images are mostly qualitative and empirical in nature. Reports suggest that though leukemia is a cancer of leukocytes; however, there are morphological alterations of red blood cells (RBCs) under the condition of leukemia. This has been evident by in observation of ultra-structural images of RBCs. Recently computational analysis of those ultra-structural images helps in revealing the quantitative understanding of the changes under the condition of leukemia. Light microscopic image analysis may further propel this approach towards clinical feasibility. Hence, development of computational analytical method for light microscopic images would the most suited way for direct application in clinics as the dragged quantitative information may help in an understanding of grading of the disease. Moreover wider application of the method may provide a hint towards the pre-leukemic state and the residual disease in future.

I. Introduction

Morphological alterations of red blood cells (RBCs) are routinely noted in different pathological conditions like in different hematological diseases and liver cirrhosis. In leukemia morphological alterations of RBCs have been reported, though leukemia is a cancer of white blood cells (WBCs). Main morphological features that have been observed in RBCs isolated from leukemic patients are the loss of biconcave shape due to presence of either thorn or horns structures or flaccid appearance in leukemic cases [1]. It has been claimed that these manifestation could be used as a marker of cancer cachexia and pre-cancer state. Analytical model is developed to predict the effect of cancer cachexia in the therapeutic outcome of leukemia treatment [2]. However, to substantiate this rigorous quantitative information is necessary.