Date of Award

Summer 1977

Document Type

Thesis

Degree Name

Master of Arts

First Advisor

V. Minnich

Abstract

This project, Electronic Cell Pattern Recognition Atlas by Agatha Rollins (1977), investigates the feasibility of using electronic blood cell patterns generated by the Technicon Hemalog “D” Differential Cell Analyzer to aid in disease diagnosis. The study aimed to correlate specific peroxidase-stained cell pattern signatures with distinct hematologic conditions, thereby developing a potential diagnostic tool for routine hematology laboratories. The project was structured as an atlas to visually and analytically present patterns found in both normal and pathological blood samples.

The atlas includes comparisons between manual differentials and automated results across a range of conditions such as acute and chronic leukemias, renal disease, eosinophilia, myelofibrosis, infectious mononucleosis, and more. Central to the study is the logic and function of the peroxidase channel in the Hemalog "D", which differentiates cells based on scatter and absorbance characteristics. The research highlights consistent, reproducible x-y oscilloscope display patterns for certain disease states, particularly chronic lymphocytic leukemia (CLL), and discusses the diagnostic implications of these electronic signatures.

Rollins also reviews existing literature supporting the concept of electronic cell recognition and provides detailed terminology and visual data to bridge traditional morphologic evaluation with emerging automated technologies. Limitations noted include the availability of suitable patient samples and minimal existing literature at the time. Nonetheless, the study concludes that recognizable pattern similarities across specific diseases suggest future viability for electronic pattern recognition as a diagnostic adjunct in clinical settings.

Research Highlights

  • The Problem: Manual morphological blood cell differentials are limited by observer subjectivity and the difficulty of identifying primitive cell types (blasts) in acute leukemia. Agatha Rollins addressed the need to determine if reproducible electronic cell patterns generated by the Technicon Hemalog "D" could accurately identify specific disease states.

  • The Method: The study utilized the Technicon Hemalog "D" Differential Cell Analyzer to perform automated cytochemical analysis and sizing of 64 blood samples from patients at the Barnes Clinical Hematology and Urinalysis Laboratory. Analysis focused on the peroxidase (perox) channel x-y oscilloscope displays to correlate electronic signatures—based on light scatter (cell size) and absorbance (peroxidase staining)—with disease-specific morphologies. 

  • Quantitative Finding: The study evaluated patient cases including 14 Acute Myelomonocytic Leukemia, 10 Chronic Lymphocytic Leukemia, 7 Acute Myelocytic Leukemia, 6 Leukemoid Reaction, 5 Myelofibrosis, 5 Myeloperoxidase Deficiency, 4 Chronic Myelocytic Leukemia, 4 Monocytosis, 3 Renal Disease, 2 Infectious Mononucleosis, 1 Acute Lymphocytic Leukemia, and 1 Eosinophilia. Electronic cell pattern recognition correctly identified Chronic Lymphocytic Leukemia patterns in 100% of cases, while Myeloperoxidase Deficiency showed distinctive patterns in 100% of findings. 

  • Qualitative Finding: Chronic Lymphocytic Leukemia (CLL) produced a characteristic "spiked" population terminating in the extreme left middle of the x-y display; Infectious Mononucleosis and Monocytosis displayed unique increases in specific display areas like the "Mono peroxidase" area; Myelofibrosis was often identifiable by increased "noise" caused by nucleated red blood cells. 

  • Finding: While electronic pattern recognition serves as a potent screening or alert function, particularly for identifying enzyme-deficient cells or tracking treatment baselines, it remains a supplement to traditional morphology, which provides critical data on cell inclusion bodies and platelet characteristics.

Download

Share

COinS