Real time blood testing using quantitative phase imaging

• Published in: PloS one Vol. 8; no. 2; p. e55676
• Main Authors: Pham, Hoa V, Bhaduri, Basanta, Tangella, Krishnarao, Best-Popescu, Catherine, Popescu, Gabriel
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 06.02.2013; Public Library of Science (PLoS)
• Subjects: Algorithms; Anemia; Anemia, Iron-Deficiency - pathology; Anemia, Macrocytic - pathology; Biology; Blood; Blood & organ donations; Blood diseases; Blood tests; Case-Control Studies; Cell morphology; Cell size; Cell surface; Cellular communication; Computer engineering; Computer Science; Cytology; Diagnostic systems; Economic conditions; Engineering; Erythrocytes; Erythrocytes - pathology; Field of view; Hematology; Hemoglobins - analysis; Humans; Image Processing, Computer-Assisted; Imaging; Labeling; Laboratories; Light; Medical imaging; Medical testing products; Medical tests; Medical wastes; Medicine; Memory; Microscopy; Microscopy, Electron, Transmission; Microscopy, Phase-Contrast; Morphology; Pathology; Patients; Physics; Programming languages; Real time; Refractometry; Smear; United States > US; Illinois
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0055676

We demonstrate a real-time blood testing system that can provide remote diagnosis with minimal human intervention in economically challenged areas. Our instrument combines novel advances in label-free optical imaging with parallel computing. Specifically, we use quantitative phase imaging for extracting red blood cell morphology with nanoscale sensitivity and NVIDIA's CUDA programming language to perform real time cellular-level analysis. While the blood smear is translated through focus, our system is able to segment and analyze all the cells in the one megapixel field of view, at a rate of 40 frames/s. The variety of diagnostic parameters measured from each cell (e.g., surface area, sphericity, and minimum cylindrical diameter) are currently not available with current state of the art clinical instruments. In addition, we show that our instrument correctly recovers the red blood cell volume distribution, as evidenced by the excellent agreement with the cell counter results obtained on normal patients and those with microcytic and macrocytic anemia. The final data outputted by our instrument represent arrays of numbers associated with these morphological parameters and not images. Thus, the memory necessary to store these data is of the order of kilobytes, which allows for their remote transmission via, for example, the cellular network. We envision that such a system will dramatically increase access for blood testing and furthermore, may pave the way to digital hematology.