Evaluation of digital images for identification and characterization of monoclonal immunoglobulins by immunofixation

Laura M. Bender, Steven W. Cotten, Yuri Fedoriw, Monte S. Willis, Christopher R. McCudden

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Objectives: High resolution digital imaging systems were recently introduced to capture and visualize serum protein electrophoresis results. In this study, we compared the performance of five, experienced interpreters using digital images and physical gels to identify and characterize monoclonal gammopathies by immunofixation. Design and methods: Immunofixation gels were generated using Sebia's HYDRASYS and digital images were captured with Sebia's Gelscan system. Interpreters blindly reviewed 200 consecutively obtained immunofixation results using physical gels, low resolution (LR) images, and high resolution (HR) images. Results: Interpretations of the physical gels were significantly more sensitive (p ≤ 0.01) than LR and HR images, and significantly more specific (p < 0.001) than the LR images. Interpreters had a sensitivity of 82.0% (45.8-95.7) using the LR images and a specificity of 71.0% (47.8-91.3); using the HR images interpreters had a sensitivity of 80.4% (68.1-86.8) and specificity of 91.8% (80.3-97.8). There was 73.6% agreement between the HR digital images and the physical gel for immunoglobulin isotype characterization. Interpreters using digital images collectively missed 19 patients with monoclonal immunoglobulins that were identified using physical gels. Conclusions: Interpreters using digital images had significantly different performance than when using physical agarose gels. Differences were most pronounced for low concentration monoclonal gammopathies (< 0.3 g/dL) and for complex patterns. Between-interpreter agreement was also lower using digital images. While digital images may serve as a useful resource for retrospective analysis and review of previous results, they are not equivalent to physical gels. Additional studies are warranted to explore the clinical impact of these observed differences.

Original languageEnglish (US)
Pages (from-to)255-258
Number of pages4
JournalClinical Biochemistry
Volume46
Issue number3
DOIs
StatePublished - Feb 1 2013

Fingerprint

Immunoglobulins
Image resolution
Gels
Paraproteinemias
Immunoglobulin Isotypes
Optical resolving power
Electrophoresis
Imaging systems
Sepharose
Blood Proteins

Keywords

  • Digital imaging
  • High resolution
  • Immunofixation
  • Low resolution
  • Monoclonal gammopathy
  • Multiple myeloma
  • Serum protein electrophoresis

ASJC Scopus subject areas

  • Clinical Biochemistry

Cite this

Evaluation of digital images for identification and characterization of monoclonal immunoglobulins by immunofixation. / Bender, Laura M.; Cotten, Steven W.; Fedoriw, Yuri; Willis, Monte S.; McCudden, Christopher R.

In: Clinical Biochemistry, Vol. 46, No. 3, 01.02.2013, p. 255-258.

Research output: Contribution to journalArticle

Bender, Laura M. ; Cotten, Steven W. ; Fedoriw, Yuri ; Willis, Monte S. ; McCudden, Christopher R. / Evaluation of digital images for identification and characterization of monoclonal immunoglobulins by immunofixation. In: Clinical Biochemistry. 2013 ; Vol. 46, No. 3. pp. 255-258.
@article{d15a4c8678d64f4697bf38c7844e7071,
title = "Evaluation of digital images for identification and characterization of monoclonal immunoglobulins by immunofixation",
abstract = "Objectives: High resolution digital imaging systems were recently introduced to capture and visualize serum protein electrophoresis results. In this study, we compared the performance of five, experienced interpreters using digital images and physical gels to identify and characterize monoclonal gammopathies by immunofixation. Design and methods: Immunofixation gels were generated using Sebia's HYDRASYS and digital images were captured with Sebia's Gelscan system. Interpreters blindly reviewed 200 consecutively obtained immunofixation results using physical gels, low resolution (LR) images, and high resolution (HR) images. Results: Interpretations of the physical gels were significantly more sensitive (p ≤ 0.01) than LR and HR images, and significantly more specific (p < 0.001) than the LR images. Interpreters had a sensitivity of 82.0{\%} (45.8-95.7) using the LR images and a specificity of 71.0{\%} (47.8-91.3); using the HR images interpreters had a sensitivity of 80.4{\%} (68.1-86.8) and specificity of 91.8{\%} (80.3-97.8). There was 73.6{\%} agreement between the HR digital images and the physical gel for immunoglobulin isotype characterization. Interpreters using digital images collectively missed 19 patients with monoclonal immunoglobulins that were identified using physical gels. Conclusions: Interpreters using digital images had significantly different performance than when using physical agarose gels. Differences were most pronounced for low concentration monoclonal gammopathies (< 0.3 g/dL) and for complex patterns. Between-interpreter agreement was also lower using digital images. While digital images may serve as a useful resource for retrospective analysis and review of previous results, they are not equivalent to physical gels. Additional studies are warranted to explore the clinical impact of these observed differences.",
keywords = "Digital imaging, High resolution, Immunofixation, Low resolution, Monoclonal gammopathy, Multiple myeloma, Serum protein electrophoresis",
author = "Bender, {Laura M.} and Cotten, {Steven W.} and Yuri Fedoriw and Willis, {Monte S.} and McCudden, {Christopher R.}",
year = "2013",
month = "2",
day = "1",
doi = "10.1016/j.clinbiochem.2012.10.030",
language = "English (US)",
volume = "46",
pages = "255--258",
journal = "Clinical Biochemistry",
issn = "0009-9120",
publisher = "Elsevier Inc.",
number = "3",

}

TY - JOUR

T1 - Evaluation of digital images for identification and characterization of monoclonal immunoglobulins by immunofixation

AU - Bender, Laura M.

AU - Cotten, Steven W.

AU - Fedoriw, Yuri

AU - Willis, Monte S.

AU - McCudden, Christopher R.

PY - 2013/2/1

Y1 - 2013/2/1

N2 - Objectives: High resolution digital imaging systems were recently introduced to capture and visualize serum protein electrophoresis results. In this study, we compared the performance of five, experienced interpreters using digital images and physical gels to identify and characterize monoclonal gammopathies by immunofixation. Design and methods: Immunofixation gels were generated using Sebia's HYDRASYS and digital images were captured with Sebia's Gelscan system. Interpreters blindly reviewed 200 consecutively obtained immunofixation results using physical gels, low resolution (LR) images, and high resolution (HR) images. Results: Interpretations of the physical gels were significantly more sensitive (p ≤ 0.01) than LR and HR images, and significantly more specific (p < 0.001) than the LR images. Interpreters had a sensitivity of 82.0% (45.8-95.7) using the LR images and a specificity of 71.0% (47.8-91.3); using the HR images interpreters had a sensitivity of 80.4% (68.1-86.8) and specificity of 91.8% (80.3-97.8). There was 73.6% agreement between the HR digital images and the physical gel for immunoglobulin isotype characterization. Interpreters using digital images collectively missed 19 patients with monoclonal immunoglobulins that were identified using physical gels. Conclusions: Interpreters using digital images had significantly different performance than when using physical agarose gels. Differences were most pronounced for low concentration monoclonal gammopathies (< 0.3 g/dL) and for complex patterns. Between-interpreter agreement was also lower using digital images. While digital images may serve as a useful resource for retrospective analysis and review of previous results, they are not equivalent to physical gels. Additional studies are warranted to explore the clinical impact of these observed differences.

AB - Objectives: High resolution digital imaging systems were recently introduced to capture and visualize serum protein electrophoresis results. In this study, we compared the performance of five, experienced interpreters using digital images and physical gels to identify and characterize monoclonal gammopathies by immunofixation. Design and methods: Immunofixation gels were generated using Sebia's HYDRASYS and digital images were captured with Sebia's Gelscan system. Interpreters blindly reviewed 200 consecutively obtained immunofixation results using physical gels, low resolution (LR) images, and high resolution (HR) images. Results: Interpretations of the physical gels were significantly more sensitive (p ≤ 0.01) than LR and HR images, and significantly more specific (p < 0.001) than the LR images. Interpreters had a sensitivity of 82.0% (45.8-95.7) using the LR images and a specificity of 71.0% (47.8-91.3); using the HR images interpreters had a sensitivity of 80.4% (68.1-86.8) and specificity of 91.8% (80.3-97.8). There was 73.6% agreement between the HR digital images and the physical gel for immunoglobulin isotype characterization. Interpreters using digital images collectively missed 19 patients with monoclonal immunoglobulins that were identified using physical gels. Conclusions: Interpreters using digital images had significantly different performance than when using physical agarose gels. Differences were most pronounced for low concentration monoclonal gammopathies (< 0.3 g/dL) and for complex patterns. Between-interpreter agreement was also lower using digital images. While digital images may serve as a useful resource for retrospective analysis and review of previous results, they are not equivalent to physical gels. Additional studies are warranted to explore the clinical impact of these observed differences.

KW - Digital imaging

KW - High resolution

KW - Immunofixation

KW - Low resolution

KW - Monoclonal gammopathy

KW - Multiple myeloma

KW - Serum protein electrophoresis

UR - http://www.scopus.com/inward/record.url?scp=84871998788&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84871998788&partnerID=8YFLogxK

U2 - 10.1016/j.clinbiochem.2012.10.030

DO - 10.1016/j.clinbiochem.2012.10.030

M3 - Article

C2 - 23127385

AN - SCOPUS:84871998788

VL - 46

SP - 255

EP - 258

JO - Clinical Biochemistry

JF - Clinical Biochemistry

SN - 0009-9120

IS - 3

ER -