MIXTURES OF RED BLOOD CELL VOLUME DISTRIBUTIONS

点击翻译按钮获取中文摘要

This research is designed to develop a new, diagnostically useful statistical method for the study of red cell volume distributions in anemia. Over the past two decades the diagnosis of anemia has been facilitated by reliable estimates of the mean volume of red cells made by electronic particle counters. current automated hematology-analyzers measure the volume of red cells made by electronic particle counters. Current automated hematology-analyzers measure the volume of each red cell, providing the red cell volume distribution. Procedures developed by the Principal Investigator (McLaren el al., 1986a, 1986b, 1987) for the analysis of single populations of cells have now been adopted by the Expert Panel on Cytometry of the International Committee for Standardization in Hematology (ICSH) for inclusion as an ICSH Reference Method in their recommendations for the standardization of cell size distribution analysis (ICSH, 1988) The major goal of this research is to extend these procedures to permit the detection and characterization of component subpopulations contained within red blood cell volume distributions. Different causes of anemia may result in characteristic alterations in the red cell volume distribution. Recognizing changes in the form of the red cell volume distribution may be diagnostically valuable because some causes of anemia leave the shape unaltered (e.g., many hypoproliferative states), while in other disorders the form is changed sometimes by the presence of subpopulations of red cells. Although mixtures of populations of cells may occur in some blood disorders (early ion deficiency or other types of nutritional anemias, myelodysplastic syndromes, and sideroblastic anemia) separation of the entire distribution into subpopulations has been hampered by lack of statistical methods for characterization of doubly-truncated distributions. To provide valid statistical methods for analysis of distributions with multiple populations of cells, this research project has for specific aims: (1) devise iterative algorithms for distribution parameter estimation and goodness of fit testing suitable for use with red blood cell volume data; (2) determine the statistical conditions under which mixtures of doubly-truncated lognormal distributions can be detected by analysis of computer simulated mixtures; (3) verify the accuracy of the method for analysis of mixed distributions using in vitro mixtures of blood samples with different mean cell volumes and in vivo mixtures of blood samples from patients with treated iron deficiency anemia, vitamin B12 deficiency, and folate deficiency; and (4) apply these statistical procedures in studies of blood samples from patients with disorders in which subpopulations of red cells may be present, e.g. myelodysplastic anemia and sideroblastic anemia. These new statistical techniques for the analysis of red cell volume distributions containing multiple populations of cells will help make generally available the new diagnostic information derived from current electronic particle counters and provide a clinically useful approach to improving the evaluation and management of anemia.