Research For Anemia Screening Based On Elastic Light Scattering

Anemia is one of the most serious public health problems with an estimated from one forth to one third of the world's population being affected,with most of those cases concentrated in the developing world.Iron deficiency anemia(IDA)and thalassemia trait(TT)are the two most common types of anemia in China.Among them,iron deficiency anemia is due to the inability of the body to absorb enough iron from the diet,which causes the hemoglobin synthesis in red cells to be blocked,and then causes the body's anemia.Such patients can be effectively treated by supplementing their diet with oral iron.Thalassemia is a hereditary disease.It is characterized by the absence or decreased production of one of the globin subunits.The most common forms are ?-thalassemia and ?-thalassemia,which affect the synthesis of ?-and ?-globin subunits respectively.Studies have shown that thalassemia patients have a reduced capacity to process iron.Therefore,treating thalassemia patients with iron supplements will lead to iron poisoning,thus the correct diagnosis of different types of anemia is the key to effective treatment for anemia patients.Prior research has shown a critical distinction between IDA and TT is found in the red cell morphology,where IDA presents with a small mean cell volume and a broad distribution of cell sizes,and TT presents with an even smaller cell volume and narrow distribution of cell sizes.Several prior groups have even put forward a series of diagnostic indexes based on the morphological parameters of red blood cells to distinguish these two types of anemia,calculated from red cell morphology as measured by conventional flow cytometers and other laboratory instruments.But most of these indices are just "simple" metrics that linearly combine parameters,and thus lack statistical robustness.Based on multivariate statistical analysis,our team introduced quadratic discriminant method into the diagnosis of anemia types,and proposed a diagnostic index based on the mean cell volume(MCV),mean cell hemoglobin concentration(MCHC),and red cell distribution width(RDW).Using our method we achieve 94%sensitivity and 90%specificity by using only these three parameters,outperforming all prior indices.In order to measure these parameters easily in the field,without relying on bulky standard instruments,we developed a low cost portable light scattering anemia diagnostic instrument.Shining laser light of two colors onto the sample and collecting the scattered light,resolved versus scattering angle,we can measure scattering patterns from suspensions of blood cells from which we can determine MCV,MCHC,and RDW,exactly those parameters used in the diagnostic index.And the instrument combined with automatic analysis program can automatically give the final diagnosis result without the need of users'intervention.We explored two separate analysis models,a physics-based model that fits the scattering curves to Mie theory,and a machine learning method that correlates the raw blood scattering curves with the clinical indices.In our testing we discovered that the physics-based model was highly intolerant of slight deviations of our pattern from theory,caused by variations in optical throughput at different angles,and slight sample-dependent backgrounds.It also could not accurately extract information from highly anemic samples due to the low contrast of the scattering fringes.The machine learning model,meanwhile,had excellent performance approaching the clinical instrument's accuracy.In view of the high incidence of anemia in remote areas with relatively backward economy,we use cheap parts and optimized structural design to make the whole system have the advantages of low cost,portable,high stability and so on.The instrument was validated on a large-scale pilot study of more than 250 Chinese children.Results showed high sensitivity and specificity.Our results demonstrate that this technology has a broad prospect to solve the key problems in the field of anemia screening.