Mathematical Models to Enhance the Value of Information from Current Laboratory Platforms

[摘要] Buckminster Fuller once said, “It is one of our most exciting discoveries that local discovery leads to a complex of further discoveries.” Every minute, a clinical laboratory generates local discoveries. Examples are a creatinine value of 3 mg/dL (265 μmol/L), a potassium concentration of 6 mEq/L, or a hematocrit of 24%. Each observation provides the clinician with the potential to discover the cause of an urgent problem. Some researchers believe we can discover disease earlier by using existing laboratory platforms. For example, while performing a complete blood count (CBC),3 we can collect up to 80 000 measurements at a time. “Clinically, the only thing typically used is the average red blood cell volume and maybe the variance,” said Dr. John Higgins, assistant pathologist at Massachusetts General Hospital in Boston. “It was shocking to me.” Higgins wanted to know what could be done with some of what he considers as high-throughput data. We spoke with Higgins about his pioneering work on a mathematical model that he hopes will make use of the wealth of these data to identify various forms of disease and increase opportunities for the early detection of illness.The body is known to create 200 × 109 red blood cells (RBCs) per day. The body also loses the same number of cells through senescence, thereby maintaining an overall stability of the RBC population; however, little is known about the aspect of the 120-day life cycle of RBCs that occurs after they leave the bone marrow. To try to elucidate more about the RBC aging process, researchers began studying the volume and hemoglobin mass of RBCs, only to find that the 2 parameters are linked: New reticulocytes have a high volume and contain a high hemoglobin mass, and both variables decrease as the cell ages (although cell …

[学科分类] 过敏症与临床免疫学